a JY Lge pf » } £7 R BL cfr pe, gg fy or Seesee e 3 aH 1 rR Ey EC 2 < i =<] : 4 a k i : Ee & % { uF I&/3 i 5 E 4 B : Bl 8 8 r "A TIA STUDIES IN STATISTICS. “He apprehends a world of figures here,” SHARSPEARE. STUDIES IN STATISTICS, Social, Political, and Fedical, GEORGE BLUNDELL {LOBCEPAPY, M.A., M.B., CERT. PREVENTIVE MEDICINE, OXON., F.R.C.P., ETC WITH MAPS AND DIAGRAMS. LONDON: EDWARD STANFORD, 96 & 27 COCKSPUR STREET, CHARING CROSS, S.W. R071 All rights rescrved.] . rE "ty : SiverPitir ; Tr 3 w at Biel ot : ils pe TO SIR JOHN SIMON, K.C.B. D.C.L., LL.D., F.R.8., ETC., ETC.. THIS SMALL WORK IS DEDICATED RY AN ADMIRING PUPIL. o om oo oO [=] = on < g wv 0 £ Qn + Q QO [a < ~~ mm PREFACE. A tpi oO ON ™ TAE numerous and voluminous reports on the very dusty shelves of the library of the Statistical Department of the General Register Office form a vast reservoir, into which a ceaseless stream of facts has been flowing for more than half a century. To disturb the dust upon those shelves; to fish out facts; to group, arrange, compare, and ponder over them, has long been my holby. In spite of the laborious investigations into the Geographical Distribution of Disease by Dr. Haviland and the very valuable papers by Mr. Alex. Buchan and Dr. Arthur Mitchell on the relations of disease to meteorology, it must be admitted that the medical profession has hitherto sadly neglected the materials so long and so patiently gathered together under the superintendence of Dr. Farr and Dr. Ogle. The paucity of workers in this direction has induced me to republish (after careful revision) some essays which were not unfavourably received by the respective scientific societies to which they were originally addressed. This, the medical portion of the work, is comprised in Chapters XIII. —XXT., and will perhaps prove too technical for many readers. on = QO; ££ < . i on 3 ~— © I~ 0 QD ~ 0 a 0 54) % ON Mm ON 6) 75 g x - « PREFACE The first three chapters are of an introductory and elementary character, and will be wisely skipped by the expert. The remainder of the work (Chapters IV.—XIL) consists of an exposition, as clear as I could make it, of some of the more striking facts and more pregnant results of the vast growth of population in Europe, America, and the British Colonies within the last half century. These chapters contain matter which I hope may be of {nterest alike to the historian, the politician, the economist, and even the general public. By a free use of the graphic method, by maps and diagrams, I have endeavoured to present dry matter in as attractive a way as the subject admits of, but at the same time I claim to have in no respect sacrificed scientific accuracy to clearness of expression. Systematic statisticians will in many cases look in vain for the figures upon which the diagrams have been constructed, but after much consideration I decided not to overload the work with long tables that few readers would ever consult. Where the authority for the figures made use of is not specially mentioned, it may be taken for granted that they were extracted either from the Census Reports or from the Registrar-General’s Returns. These last are of five descriptions :— (1) Annual Reports. (2) Decennial Supplements to Annual Reports. (8) Quarterly Returns of Marriages. Births, and Deaths in Enoland. PREFACE (4) Weekly Returns of Births, Deaths, and causes of Death in London and other great towns. (5) Annual Summaries of Births, Deaths, and Causes of Death in London and other great towns. In many cases I have not thought it necessary, or desirable, to quote the precise reports made use of, but the reader who is versed in Blue-hooks will not find much difficulty in verifying any facts in which he may be especially interested. The statistics of migration are not very satisfactory, and it appears to be impossible to get at the absolute facts. As a rule, the emigrants from Europe are recorded, more or less imperfectly, at this end, while the immigrants are also recorded in the new lands beyond the seas. It is difficult, no doubt, to define an emigrant for statistical purposes ; moreover, there may from the nature of the case be greater facilities for observing and recording immigration than emigration. Be this as it may, in many, if not in all cases, the number of recorded immigrants greatly exceeds that of recorded emigrants, so that the reader will note some marked discrepancies between the statistics of emigration given in Chapter V., and those of immigration in the chapters following. Perhaps the truth lies between the two. The United States Census Report, 1880, is sadly deficient in method and arrangement, a fact that no doubt accounts for the conflicting statements as to the “total population ” in some of our best statistical hand-books. {a PREFACE My cordial thanks are due to the numerous persons who have given me kindly assistance or encouragement in my often tedious labours. More especially am I indebted to the late Dr. Farr, to Dr. W. Ogle, Mr. Noel A. Humphreys, and other officials of the Registrar-General’s Department ; to Dr. Buchanan; to my teachers, the late Dr. Murchison, Sir John Simon, K.C.B., and Dr. William Ord; to Messrs. Buchan and Mitchell for permission to make use of their figures and diagrams ; to the Councils of the Society of the Medical Officers of Health, the Epidemiological Society of London, and the Royal Statistical Society for kind permission to reprint papers originally published in their transactions; but, above all, to Dr. I. A. Dixey, of Wadham College, Oxford, who rendered me invaluable assistance in the earlier part of my work, and to Mr. A. C. Waters, of the General Register Office, who, besides assisting me in many ways with information and advice, has not only corrected the proof-sheets and prepared the Index, but has criticized the argument and checked and revised the figures and calculations in such a thorough manner as greatly to increase their value; yet, in spite of all that I owe to these gentlemen, the responsibility for facts, statements, and opinions lies wholly with myself. B. LONGSTATR. fTighlands, Putney Heat Aameonher 2515. 1R0N CONTENTS. CHAP, I. INTRODUCTORY - II. Ox Drearg-Rares, with a Diagram ... “ III. Tee BirtE-, DearH-, AND MarrIAcE-RATES OF ENG-LAND AND Wares ror Firry Years, with two Diagrams aes . es vv wre IV. Tune Growrx oF PoruraTioN IN Exeraxp, with two V. Tee MicraTioNns OF THE PgropLeE IN THE NINE TEENTH CENTURY, with a Map and two Diagrams VI. Tee Growra oF Now Narions—TrHE URITED States, with five Maps ee ee ow VII. Tare GrowrH oF Nw Namions (confinued)—CANADA VIII. Tee GrowrE oF New NasrioNs (continued)—SoUTH AMERICA ves ve. Ves ee aed IX. Tue Growra or New Nations (continued)—SoUTH ArRica AND AUSTRALASIA, with a Diagram ” X. Tae Growrm or MobperN CITIES ... ves a XI. Tur PoruratioN oF LoNDoN AND 118 MIGRATIONS, with a Map ... ee a _ XIL Foop SupPLY ves es cen vs cs XIII. Svecestions ror THE Cexsus or 1891 vr ww XIV. Tae Recent DecLixe IN THE ExeuisE Drath Rare CONSIDERED IN CONNECTION WITH THE CAUSES OF Dear rot a cn en XV. Tox Causation oF SuMumER DIARRHEA, with three Diagrams - CAGE 22 34 52 113 128 137 156 168 182 190 296 270 iy CONTENTS CHAP. XVI. Ox SOME STATISTICAL INDICATIONS OF A RELATIONSHIP BETWEEN SCARLATINA, ERvsiperLas, PUERPERAL FEVER, AND CERTAIN OTHER DISEASES, with a Map and two Diagrams _ ers wre -_ XVIL Tre GgeocrAPEICAL DISTRIBUTION OF DIPHTHERIA IN Excranp axp Wares, with a Diagram ... - XVIII. Prraisis, BrONCHITIS, AND PNEUMONIA: ARE THEY Ermpemic Diseases? with four Diagrams ve XIX. A CALCULATION OF THE PROBABILITY OF THE Acc DENTAL AND FATAL INCIDENCE oF PHTHISIS UPON BoTH HUSBAND AND WIFE .. I. ce XX. Tee SeasoNAL PrevarenceE oF CoNTINUED FrvEers iv Lonpox, with two Diagrams ... i XXI. HypropHOBIA STATISTICS, With two Diagrams 2 XXII. CoNCLUSION cee cee I. cer a. Nore oN AN EripEMic oF INFECTIOUS PNEUMONIA IN THE EIGHTEENTH CENTURY ... a INDEX vee PAGE 310 335 268 384 394 413 438 443 445 LIST OF MAPS AND DIAGRAMS. PLATE PAGE I Death-rates of Males and Females at various ages ,., I 7 II. Birth-rate, Death-rate, and Marriage-rate for Fifty years 13 III. Death-rates from certain groups of causes ... vee we 18 IV. Density of Population of the Counties of England and Wales « in 1831 .., vee fui —_ oe fo 99 V. Density of Population of the Counties of England and Wales ! in 1881 .., oe _— ve. po VI. Intra-migration in England and Wales, 1871-81 ... os VIL Emigration from the United Kingdom ,., _ vor VIIL Destinations of Emigrants from the United Kingdom wr IX. Distribution of “ Men of Colour” in the United States - X. Distribution of Natives of Scandinavia in the United States XI. Distribution of Natives of Germany in the United States _ XII. Distribution of Natives of Ireland in the United States “gt XIIL Distribution of Natives of Great Britain in the United States XIV. Birthplaces of the Inhabitants of the United Kingdom, some British Colonies, and the United States - aan XV. Map of “Greater London” sw ee “i _ XVI London— Average Weekly Deaths from Diarrhoea and from certain other Diseases ore ee wi «278 XVII, England and Wales—Death-rates from Diarrhea and certain other Diseases, with Temperature and Rainfall at Greenwich 281 XVIII. England and Wales—Death-rates from Diarrhea at various ages ee oe ee ee ve. en XIX. England and Wales—Death-rates from Scarlet Fever, Erysipelas, Puerperal Fever, and certain other Diseases, with Rainfall at Greenwich oo win are wi XX. London— Average Weekly Deaths from Scarlet Fever, Puerperal Fever, and certain other Diseases wv - XXI, Registration Divisions—Death-rates from Erysipelas, Puerperal Fever, and Rheumatism _ aes 330 LIST OF MAPS AND DIAGRAMS PLATE PAGE XXII. Distribution of Diphtheria in England and Wales ... ... B42 XXIII. England and Wales—Death-rates from Phthisis, Bronchitis, Pneumonia, and certain other Diseases _— we 371 XXIV. London—Weekly Deaths from Bronchitis and Pneumonia, in certain years, with Temperature ... es ... 376 ¥XV. London—Average Weekly Deaths from Phthisis, Bronchitis, Pneumonia, and certain other Diseases ces ... 377 £XVI. England and Wales— Death-rates of Males and Females from . Pneumonia, ab various ages oe yo we 379 KXVIL. Continued Fevers in London—Deaths and Admissions to Hospitals po wn ata I. ... 402 XXVIII, Scarlet Fever in London— Weekly Deaths and Admissions to Hospitals p> ~ sae woe we 405 XXIX. Cases of Rabies in Dogs, and Deaths from Hydrophobia in Loudon ... se ee oo _. ... 422 XXX. Distribution of Hydrophobia in England and Wales ... 433 STUDIES IN STATISTICS. CHAPTER IL INTRODUCTORY. MisTrUST of statistics is very general. On all hands one hears the remark, * You can prove anything by statistics.” The true and sufficient reply to this taunt is, “ Without statistics you can prove nothing.” Men as a rule form their judgments from a few instances that happen to have been brought prominently before their notice, but as these instances may have been exceptional in their character, the judgments formed are but too likely, as all experience proves, to be most erroneous. Social and political phenomena rarely allow themselves to be fitted into the rigid figures of the logician, but rather resemble indeterminate equations with many unknown quantities, and reasoning applied to such phenomena takes the modified form of probable reasoning. But although the dato, of sociology are not definite enough for exact mathematical treatment, they are often sufficient to admit of such an application of the numerical method as will yield a first approximation to a solution. It is unreasonable surely to say that, because a result mathematically accurate is unattainable, therefore we BR fo, STUDIES IN STATISTICS CHAP, should be content with mere guesses founded upon the personal experience of individuals. Should we not rather collect facts on the broadest basis available, and reason from these? Because the statistician cannot predict social crises as the astronomer predicts eclipses, is he to be precluded from demonstrating present tendencies and indicating their probable results? Surely again it is no more reasonable to condemn statistics, because many men cannot marshal figures correctly, than it would be to deny the utility of geometry, because many boys fail to master the first few propositions of Euclid. The truth "is that whereas almost every one thinks that he can understand figures, and read their meaning, the clear-headed statistician is a somewhat rare being, and by no means every one who writes papers full of tables of figures, even though he may write “ F.8.8.” after his name, is to be trusted as a safe guide in the matter. Because the statist as a rule makes little use of the higher mathematics, and performs most of his operations by means of the simple rules of arithmetic, it by no means follows that a sound knowledge of arithmetic, even if supplemented by readiness in using logarithms, the calculating machine, or the slide-rule, is a sufficient educational basis for the statistician. On the contrary, paradoxical though it appear, arithmetic has but a minor part to play. The primary requisite is a logical mind and a sound logical training; the second (and not less important) is a good general knowledge of the subject to which the figures under consideration relate. Ouly a chemist is likely to derive information from a new chemical experiment; in like manner the statistician must be now a banker, now a farmer, now a merchant, now a doctor, according as he is manipulating figures relating to currency, crops, tariffs, or causes of death. INTRODUCTORY After this short apology, I leave the reader to judge whether the following chapters prove the confessedly imperfect science of statistics to be capable of affording interest and profit, or rather confirm the popular opinion that such studies are ever barren of result. STUDIES IN STATISTICS CHAP. CHAPTER II. ON DEATH-RATES. THE vital statistician is constantly speaking of “rates,” it is therefore of prime importance to understand clearly what he means by the term. In the year 1881 there died in the town of Preston® 2139 persons; in the same year there died in Bristol no fewer than 4054 persons. It is conceivable that some one not versed in such matters might suppose from this that Bristol was a less healthy town than Preston, or that an individual living in Bristol would have run a greater chance of dying during that year, than an individual living in Preston. But such an inference would be wholly illegitimate, and would be contrary to fact. Bristol had in that particular year 207,522 inhabitants, but Preston only 93,982, therefore on the assumption that each individual Bristolian or Prestonian was squally likely (or unlikely) to die, it will be seen that as only one Bristolian out of every fifty-one died, but one Prestonian out of every forty-four, so far from the inference that Preston was the more healthy place of the two being correct, the precise contrary was the fact, and a Prestonian was more likely than a Bristolian to die in the proportion of 51 : 44. The simplest way to make the comparison in such a case, is to find out how many died in proportion to every 100, 1000, or 10,000 1 These figures are taken from the Registrar-General's Annual Summary of Births, Deaths, &e. for 1881, pp. xiv and xviii. ON DEATH-RATES living at the time in each town respectively. Arithmetically this is done by multiplying the number of deaths by 100, 1000, or 10,000, and dividing by the population. The result is called the death-rate per cent., per thousand, or per ten thousand, as the case may be, and these “ death-rates,” speaking generally, are comparable one with another. In practice the labour of working out these rates is usually shortened by making use of tables of logarithms, slide-rules, or the ingenious machine known as the arithmometer, invented by M. Thomas of Colmar, and since greatly improved by our countryman, Mr. Tate. In most cases it is found convenient to calculate birthrates, death-rates, and marriage-rates at so much per thousand per annum ; hence unless some other number is named, it may be assumed that the rate is per thousand.” In the instances given we get— DIX1000 _ 2189x1000 ; For Preston : Death-rate per 1000 — 5 = mar — 228 . Dex1000 _ 4064x1000 _ 1. For Bristol : Death-rate per 1000 — ES amg = 196 and these numbers are in the same ratio of 51 : 44. The formula shows that for every thousand persons living in the town of Preston in 1881, 22-8 persons died, whereas in Bristol only 19'6 persons in every thousand inhabitants died. In other words the death-rate of Preston in 1881 was 22°8 per thousand, in Bristol 19:6 per thousand, and other things being equal, we may fairly assume that in that year Preston was less healthy than Bristol. I have said other things being equal. It is assumed when the healthiness of towns is measured by their deathrates, that on the average the deaths bear a constant Proportion to the cases of sickness, or to the general health of the people, and when large populations are considered STUDIES IN STATISTICS CHAP. for long periods this is probably correct. Of course the proportion of deaths to cases of sickness varies enormously in different diseases; hydrophobia, for example, is always fatal in a few days; chicken-pox, a commoner ailment, very rarely kills. When, moreover, one population is compared with another, it is most important to know in what respects the populations differ. Now quite apart from occupation, surroundings, or social position, it is a well-established, indeed an obvious, fact that age has a paramount effect on liability to disease and death, while sex has also an important influence. In other words, in any given population the death-rates of males and females are different; still more different are the death-rates of infants, adults, and aged people. The death-rates at different ages vary slightly from time to time, as will be scen hereafter. The following ara the rates that prevailed in the ten years 1871-81— ENGLAND AND WALES. 10 vrars 1871-80. DEATH-RATES FROM ALL CAUSES PER THOUSAND LIVING AT VARIOUS AGES AND GROUPS OF AGES. Males. 1974 68-3 27-9 18:1 3 v7 °7 J 2 a 45 5—IC 10o—15 15—20 20—25 25—35 3 35—45 157 45—55 20-1 55-—068 348 65—75 ... % 69-6 7c and upwards ... 169-1 Females 157°3 636 27°5 79 rz'g 62 3-7 54 6'8 86 6 156 285 6o'T 1558 The same facts are represented graphically,” as it is ENGLAND & WALES. 10 YEARS, 1871-80. ANNUAL DEATH RATES FROM ALL CAUSES PER THOUSAND LIVING AT VARIOUS AGES & GROUPS OF AGES. MALES BLACK —_ _.. FEMALES RED DEATHS PER 100Q LIVING 200 — 50 00 - 4 N n AGE 8§ ON DEATH-RATES termed, in the diagram (Plate I.). It is the essential principle of this ‘graphic method” that quantities are represented by the lengths of lines. Thus in this instance ages are measured horizontally from the “origin,” at the left hand bottom corner, towards the right; distances measured along this line are termed in mathematical language ““abscissee.” The other quantities, here deathrates, are measured vertically, such vertical distances being termed “ordinates.” We see from the table that the mean death-rate for males between the ages of 45 and 55 is 20-1. Therefore from the point midway between 45 and 55 on the horizontal scale of ages, we measure vertically an ordinate equal in length to twenty divisions on the vertical scale of rates to the left hand, and make a mark at its extremity. Similarly for the ages 55 to 65 we measure an ordinate of length = 84'8 (say 85), and make another mark. In like manner marks corresponding to the other quantities are made, and then these marks are joined together by straight lines drawn from point to point. The resultant line is technically termed “a curve,” although it may be far from the idea commonly given by the word. If we measure an ordinate to a point in this curve midway between any two of our marks, we shall obtain the corresponding arithmetical mean value. This will not generally be the true value, but it will almost always be nearer to the true value than that corresponding to the mark on either side. A nearer approach to true intermediate values might in this case be obtained by connecting the marks with a curved line, instead of a succession of straight lines, but for simplicity’s sake I prefer in most cases to use the straight line, letting it be clearly understood that the marks for certain ascertained values are connected together, STUDIES IN STATISTICS CHAP. not for the purpose of ascertaining mean values, but to show that the phenomenon under investigation is a continuous one, and above all to render the facts more easy of apprehension through the eye. Some few persons easily grasp the signification of tables of figures, but very many more ave readily instructed by the aid of diagrams. From this diagram we see very clearly that in both sexes the mortality is by far the greatest at the two extremes of life, being very heavy in the first year, falling rapidly, but each year with decreasing rapidity, during the second, third, and fourth years, then more gradually until a minimum is reached in the period 10—15; from that time it rises slowly, but in an everincreasing ratio, until by the period 85—45 it is equal to the mortality of the fifth year of life, and by the period 65—75 it is equal to the mortality in the second year: The curve at both extremities, especially that corresponding to old age, is untrustworthy; on the one hand, the ages at death and the ages in the census reports of very young children, are stated by the informants in a confused manner; and on the other hand, in a majority of cases grave suspicion attaches to the extreme ages often recorded in death certificates, so that for many years the Registrar-General has ceased to divide into groups ages heyond 85, and in some cases even 75. As regards the difference between the sexes, it should be noted that from the third year of life to the thirtyfifth there is practically no difference between the mortality of the two sexes. After the thirty-fifth year females die less rapidly than males, and the difference in their favour increases with age. The superior longevity of women is well known, and to a large extent accounts for the excess of women in this country, four-fifths of that ON DEATH-RATES excess consisting of widows. Occupations, habits, and probably to some extent constitutional differences, account for the superior longevity of women; but it is not possible to give any good reason for the great excess of mortality of male infants in the first year of life; it would seem that it must depend upon some constitutional difference ; but whatever it may be, in the second year of life it has almost disappeared. It will now be clearly understood how important it is when comparing the death statistics of two populations, to take into consideration their “age constitution”; that is, the proportions per cent. of males and females at the several ages. As concrete instances, I may give the following—In the year 1883, the recorded death-rate of Bradford was 18-34, and consequently 1°20 below that of England and Wales; but when due allowance for the age ard sex constitution of its population had been made, it appeared that its corrected death-rate was 20°26, or *72 above that of England and Wales. The converse was the case with Norwich, its recorded rate being 19°64, or ‘10 above that of England and Wales, its corrected rate 18:79, or 75 below it. Manchester again came in that year at the bottom of the list of great towns with a recorded death-rate of 27°64, whereas the unhealthy condition of its population was cven worse than it appeared, since after correction its death-rate was 30°80. Males. Unmarried 7,828,21C Married -,376,80¢ Widowed 434," Total 12.639,7 JENS 1881. Females. 7,897,629 4,437,962 999,046 18.3834. E37 Excess of Females. 69,319 61,064 564,252 694.635 The excess of married women is partly to be explained by husbands absent at sea or abroad ; partly by many single women pretending to be married, while many men really married deny the fact. wr STUDIES IN STATISTICS CHAP. or nearly 60 per cent. above the average of the whole country. These are, however, extreme instances. In the majority of cases the corrections when applied to a town would only slightly raise its death-rate, and when applied to a country district slightly lower it; in short, correction for age and sex distribution usually accentuates the differences between healthy and unhealthy districts. The reason of this is not far to seek. It was pointed out in the annual summary of the Registrar-Greneral for 1883, that “The towns contain, as a rule, a much smaller proportion of aged persons and a much higher proportion of persons in the prime of life, as also a much higher proportion of females (at most ages the death-rate of females is below that of males) than does the country at large; and though these advantages are somewhat counterbalanced by an excess in the proportion of children, they are so to a limited extent only.” The social conditions which lead to more marriages taking place in towns than in country districts are sufficiently obvious: the young of both sexes are drawn into the towns by high wages and by the attractions of town life; the aged are left behind in the country villages. Young men leave their homes to emigrate to the colonies or the United States ; young women prefer to remain in England, but go into towns as servants or as shop assistants, hence the excess of women that is usually met with when town is compared with country. The peculiarity of the special incidence of mortality on young infants, so plainly shown in the diagram, has given rise to a curious fallacy which has exercised a charm over a certain class of mind. Sanitary reformers are constantly pointing to the high death-rates of many of our large towns as a proof of sanitary defects. Their opponents (sometimes also sanitary reformers) have called ON DEATH-RATES attention to the fact that most of these large towns have not only death-rates much above the average, but also birth-rates quite as markedly in excess. The high birthrate, the writers say, “it stands to reason” is the cause of the high death-rate, since as is well known the mortality of young infants is excessive. The answer to this argument is twofold : (1) The population of the town and its death-rate may be compared with that of the whole country age for age—if the town mortality proves excessive at each age, or at a majority of the groups of ages, cadit quastio ; but, (2) while it may be admitted that a high birth-rate is likely to lead to a proportionately large number of deaths of infants, the matter does not end there ; under the worst circumstances a majority (usually a large majority) of the infants survive the first few dangerous years, and as a result there is a very large propertion of the population living at the healthiest period of life, 5—45. Thus, under the conditions of modern life with a high birth-rate there will be associated a rapidly-increasing population, and consequently an unduly small proportion of elderly persons subject to a high death-rate, as a set-off against the excess of young infants. It will also be readily seen that a high birth-rate is only likely to be met with in a population containing a large proportion of persons at the reproducive ages, 4. e. at ages of comparatively low mortality. STUDIES IN STATISTICS CHAP. CHAPTER 111. THE BIRTH-, DEATH-, AND MARRIAGE-RATES OF ENGLAND AND WALES FOR FIFTY YEARS. HavING, as I hope, given the reader a clear idea of what is meant by a death-rate, I purpose, as an introduction to the more detailed and necessarily more technical discussions that follow, to make a few general remarks on births, deaths, and marriages in this country. The facts are set forth in a diagram (Plate I1.), which merits careful attention. Civil registration was established in England in 1837, the first year that we have a complete record of being 1838. As would be expected, it was at first far from complete, but the defects in the birth-register are far more numerous than those in the death- and marriageregisters. In addition to stringent legal provisions, it is the obvious interest of the parties concerned to see that a marriage is duly registered, and as regards deaths also the enforcement of registration is comparatively easy ; but with births it is far different. A new Act, designed to make registration more complete, came into force on the 1st of January, 1875, which rendered the registration of births compulsory, with penalties for non-performance, and aimed at securing greater accuracy in recording details as to deaths. As is plainly shown in the diagram, the recorded birth- 1838 40 RATE PER 1000 1 36 dERSONS LIVING 24 ENGLAND & WALES. BIRTH-RATE, DEATH-RATE & MARRIAGE RATE FOR 6&0 YEARS. > 44 "4s 48 “5p 52 54 56 B58 80 82 sa eq ‘ea 70 72 4 76 78 's0 ‘82 '84 ‘sc 88 Plate II oa RATE PER 1000 “* PERSONS LIVING 3 MEAN 23:.8¢ - 39 - LC ’8 x 264 £ 3 244%, 221 20 a [4 — 32 30 728 26 20 i8 + gm i eeee MEAN 21-GFE -— MEAN_ IG 4 _ {7 10 3 SHILLINGS PER OUARTER . 0 70 6C 50 30 30 20 10 0 ft I — I er pn -0 —+70 -60 50 10 7 ‘0 0 Nn SHILLINGS PER OUARTER « ! ender Stanfords Geog Estab? Ir BIRTH-, DEATH-, AND MARRIAGE-RATES 13 rate was in 1838 only 803" per thousand persons living ; it rose from that year by irregular steps, interrupted from time to time by sharp falls, to a maximum of 863 in 1876 ; since which time it has fallen, with one trifling exception, uninterruptedly to 306 in 1888. We may assume that from 1875 onwards the record is fairly complete, and on the other hand there can be no doubt that in 1838 it was very faulty. From the commencement of civil registration to 1875 the registration of births steadily improved, and probably for several years before the new Act came into force the number of births that escaped registration was, proportionately at all events, very small.! This uncertainty about the register deprives the birthrate curve of some of its value, but the following remarks hold good in spite of its imperfections. In no case has the birth-rate in any one year risen more than 1-4 per thousand. In 1846-7 it fell 2'3 per thousand, but this fall has never been approached before or since. It appears to be partly explained by the fall in the marriage-rate in the same year, which is also unprecedentedly large, both probably having relation to the failure of the potato crop, and the Irish Famine, possibly also to the commercial panic, while the widespread epidemics of influenza and typhus fever may well have had a share in the result, With this exception the birth-rate has never fallen more than 10 per thousand in a single year. Speaking generally, the birth rate corresponds to the ! My reason for saying this is, that the dlegitimate birth-rate wag ¢ractically constant (2:23) from 1845-65, from which time it has been slowly falling (to 1-4 in 1888). In addition to the carelessness which might attend the registration of births in general, there would obviously be strong temptations not to register illegitimate births, so that the Passing of a stricter Act might have been expected to cause a sudden increase in the number of illegitimate births registered ; but as a matter of fact, no such increase took place. dg STUDIES IN STATISTICS CHAP. marriage-rate, but with marked exceptions, e.g. 1852-3-4. As might be expected, the birth-rate curve tends to follow that of the marriage-rate. The most notable thing about the curve is the continuous fall from 1876 to 1886. The marriage-rate fell from 17:6 in 1873, to 14-4 in 1879; a total fall of 3:2 per thousand in six years. The birth-rate in the six years, 1876—1882 fell 26 per thousand. It is, however, very difficult to show in a simple manner the exact way and extent in which the birth-rate depends upon the marriage-rate. The question of illegitimate births has less effect than might be imagined ; the actual number of illegitimate births fluctuates very little from year to year (seldom more than from 1000—2000), while the relative number to population has been slowly but steadily diminishing for twenty years. The marriage-rate naturally calls for attention next. In general terms it forms a curve marked by a succession of waves (which are more marked and more regular in the last twenty years), periods of depression of two to five years alternating with periods of elevation of similar duration. In 1844-5 the rate rose 1'2 per thousand, and after two years fell 1-4 per thousand; in no other case has the annual fluctuation exceeded 1'0 per thousand, and has rarely reached that point. It should be noted that 1845 was a year of great commercial prosperity. At the bottom of the diagram is given the average price of wheat in England in each year from 1839, and also I Tt is well worth noting that the birth-rate fell from 1876-1886 also in Scotland, Ireland, Austria, Germany, Switzerland, the Netherlands, Belgium, and even France. No similar fall, however, occurred in Italy, Hungary, Norway, Sweden, or Denmark. In most of these countries the marriage-rate also fell, but in nearly all in a less striking manner than the birth-rate; while in the case of Austria, Germany, Belgium, Italy, Sweden, Hungary, and France the fall was inconsiderable. (1 BIRTH-, DEATH-, AND MARRIAGE-RATES 15 the most notable harvests in this country are indicated. It was at one time stated that the price of wheat had a great influence on the marriage-rate; that when it fell marriages became more numerous, and vice wversd. This very likely was the case at one time, but it certainly has not been the case for the last eighteen years. It will be noted also, as might be expected, that in recent times the quality of the harvest in this country has had little or no effect upon the price of corn. Still there is no doubt whatever, that the Russian war, and to a still more marked degree successively recurring periods of commercial depression have had a very distinct and very uniform effect in bringing down the marriage-rate; this is very clearly shown in the diagram.! The last wave in the marriage-rate, the crest corresponding to 1882, seems to have produced little or no effect upon the birth-rate, which has continued to fall. About the death-rate there is much more to be said. In the first place, as the diagram shows plainly, it is much more irregular in its fluctuations than either the birth-rate or marriage-rate. Its curve is of a more jerky character, rising or falling abruptly ; thus, fluctuations of 1'5 per thousand are common, a rise of 2-1 has occurred twice, a similar fall once, and on one occasion a fall of 4:3 per thousand. These sudden and extreme fluctuations are to a large extent caused by epidemics, of which several of the principal are indicated in the diagram. The recent epidemics which are most notable as measured by their effect upon the death-rate, or the rarity of their occurrence. are shown in the following tahle— : . Ogle read his : time before Dr. Og . 1 Th as written some So “4h special I ep x Marriage-Rates and Mariage em Sosioty reference o EW Growth of Population. to the Roya. in March 1890. STUDIES IN STATISTICS CHAP. { ear. Excess of Death-| Death-Rate Rate. | above average. Nature of Epidemic. | Death-Rate from Epidemic. 1838 847 1848 849 r854 1858 1863 1864 1866 i868 (870 871 il. 29:4 | 24-71 23:01 25-1 235 | 231 | 23-0 23-7 234 218 29-9 99-6 71.7 III. + o'1 + 24 + 07 + 28 + 12 + 08 + oy + 14 + 1'¥ LY . / Small-pox Typhus Influenza Cholera Cholera Scarlatina Scarlatina Scarlatina Cholera Diarrhoea Scarlatina Small-pox Small-pox 11 1-8 05 3-0 | L-2 ih [4 0-7 i4 14 1-0 0-8 1872 Average 1838—1872! &>-It will of course be borne in mind that the influence of an epidemic in raising the general death-rate largely lepends upon its being an exceptional occurrence, e.g. an :pidemic of cholera is wholly exceptional in this country, as the disease only visits our shores at distant intervals. Again, in these days of vaccination the average annual mortality from small-pox is but trifling, and considerable epidemics, such as that of 1871-2, ave fortunately of rare occurrence. As regards typhus fever, difficulties of nomenclature and diagnosis complicate the question too much for discussion at present. Measles and whooping-cough are most decidedly epidemic diseases, but they are never absent, and epidemics of these diseases, often severe, but usually circumscribed, occur so frequently in different parts of the country, that the difference of mortality for the whole of England and Wales in the most fatal and least fatal years is comparatively small. The cases of scarlatina and diarrhea occupy a somewhat intermediate position. Roughly speaking, each of these diseases caused {11 BIRTH, DEATH-, AND MARRIAGE-RATES 17 on the average of the thirty years 1850-79, an annual mortality of 09 per thousand, but in individual years it has varied from 0'4 to 1-4 per thousand, so that a large proportion of the mortality from these causes in especially fatal years may be termed exceptional. They kill large numbers every year, consequently the otal mortality caused by an epidemic of either of them cannot, at least in our present stage of sanitary progress, be regarded as an exceptional addition to the mortality of the vear in which it occurs. It has often been asserted by persons imperfectly acquainted with the facts that epidemics do not on the whole materially affect the general death-rate, since the individuals who succumb to their attacks would, had there been no epidemic, have died of some other cause ; it is moreover alleged that in the main they kill off unhealthy persons, and in consequence lead to a low mortality in the years immediately following. I will not deny that there is some truth in these assertions, but any benefit to the average health standard of the community that may result from an epidemic removing weakly individuals, is completely overwhelmed by the permanent injury that in too many cases results to the constitutions of those who are attacked by the epidemic, and indeed escape from death, but recover imperfectly, and have to go through life weakened or maimed. We have, in short, to deal not only with killed but with wounded. A glance at column III. of the table will show that in eleven out of the thirteen epidemic years the general death-rate exceeded the average, though in only six cases by an amount at all corresponding to the ravages of the epidemic. The cholera of 1849 was not only the most serious epidemic of any kind that has visited this island during the last fifty years, but the one that most STUDIES IN STATISTICS CHAP. directly increased the death-rate. This, however, is not without its compensations. To the fearful lessons of that summer and, autumn, and to ‘the still more terrible sufferings of our troops before Sebastopol in the winter of 1854-5, more than to any other circumstances do we owe that movement of public opinion which produced not only sanitary legislation, but, what is far more, that improved sanitary practice which has resulted in a remarkable freedom from serious epidemics and a striking fall in the death-rate in the last fourteen years.! The diagram shows, moreover; that it is by no means universally true that epidemic years are followed by years of low mortality ; in seven out of the thirteen instances they were followed by high mortalities, in four cases by spidemic years. The next diagram (Plate III.) throws more light. on this subject. The method of representation is here somewhat different, the death-rates being represented by zoloured areas. Confining our attention in the first place to the upper part of the diagram, it will be seen that the death-rate each year is represented by a column of colour, part blue, part red. The red portion corresponds to so much of the death-rate as is due to the ravages of Small-Pox, Measles, Scarlet Fever, Diphtheria, Whooping-Cough, Continued Fever (Typhus, Enteric, and Simple), and Cholera. The blue, much the larger portion, represents the deaths per thousand due to all other causes. The diminishing length of the red columns towards she right-hand of the diagram shows clearly enough that, what for shortness we may term the Epidemic Diseases, are far less fatal now than in the first twenty-five years bt Sunitary Fundumentuls : two Lectures delivered at the Royal Naval College, by J. Netten Radcliffe, The Practitioner, vol. xxi. pp. 140— 160, and pp. 221—240. Macmillan and Co., 1878. DEATHS ER 1000 LIVING. 76 1847-66 24 MEAN OF TOTAL 22-6 22 ENGLAND & WALES. FORTY YEARS. - DEATHS PER THOUSAND PERSONS LIVING FROM SMALL POX, MEASLES, SCARLET FEVER, DIPHTHERIA, NHOOPING COUGH, TYPHUS, ENTERIC & SIMPLE CONTINUED FEVERS, & FROM CHOLERA - COLOURED RED. FROM ALL OTHER CAUSES. COLOURED BLUE HARD WINTERS, MARKED BLUE. HOT SUMMERS, MARKED RED. 0 ‘sa =a se sp ‘eo 5a 66 8 70 M2 7a 7@ i mg 82 4 ma DEATHS PER 1000 oc LIVING. na 22 1867-86 MEAN OF TOTAL oo "209 20 8 3 i 2 0 3 Gc 5 MEAN 4 1847-66 & 3:5 3 a FrRITHDUSAND LIVING HROM|SMALL-POX MEASLES SCARLET *~&7 Fb HUUPIN UUGH, UUN| NURD VERS & UE That 3 MEAN 2 1867-86 -9:4 3) uv 4 MEAN 1847-66 2.1 3 “ated <= udu danD ving + — oy dry of "EATHS PER THOUSAND divine FROM BIgEASES TRAM DD" RRHGEA & DYSENTFR } . 4 Lt VE wf od i ] gy . | ™ RESPIRATORY ORGANS [EXCLUDING] PHT im —— +=0 ‘4 MEAN 1867-86 -36 8 g TEOUSAND 1 Aen - ShM REMAINING CAUS MEAN 1847-66 is 15h -15 _MEAN 1867-86 14-1 =+ OU 0 Torn 2 5 Ti i aid : = I wk Ho | F : SE i pais Lordy Cnn cd 4 48 BO 52 64 ‘Se B88 80 '&2 ‘e4 Be '&8 70 12 va a 78 BO ‘82 'A8a =e - 5 ) 11 BIRTH-, DEATH-, AND MARRIAGE-RATES 19 out of the forty illustrated. This is shown perhaps still more clearly in the middle part of the diagram, where the death-rates from these diseases are set out separately in red. Comparing the first twenty years with the second, the mean epidemic death-rate fell from 8:46 to 2:43 per thousand, that is to say, no less than 30 per cent.; similarly the mean death-rate from all other causes fell from 19:15 to 18°58 per thousand, or only three per cent., a striking contrast. As regards the death-rate from other causes, it should be observed that the decline has occurred mainly in the last six years. This subject is, however, treated in great detail in Chapter XIV. so I will say no more about it here. As early as 1848, and again in 1855, important Acts were passed by the legislature dealing with the public health, as well as a number of measures that indirectly tended to its improvement. In his report to the Privy Council for 1866, Sir John Simon said, “The date of this change—the passing of the Sanitary Act, 1866 —marks the beginning of a new era in the progress of sanitary reform.” This Act was amended in 1868, and considerably strengthened in 1872, and all the Sanitary Acts were consolidated (except for the metropolis) by the comprehensive Act of 1875. It would appear reasonable to infer that the decline in the mortality due to epidemic diseases, so notable in the last fifteen years, is closely connected with this activity in sanitary legislation. At the same time some part of it may well be attributed to more rational methods of medical treatment. and to improved nursing, Those who are familiar with these matters will note that T have not included among the epidemic diseases here ne that is so classed in the official returns—I mean liarrhoea—a disease which is mainly fatal to infants. and ey STUDIES IN STATISTICS CHAP, more especially in the summer months. It must not be assumed that I consider summer diarrhoea to have been wrongly placed by Dr. Farr in the class of ¢ zymotic” diseases; far from it. I differ strongly from those who hold that the summer diarrhea of great cities is a purely “climatic” disease. At the same time it is far more directly, and it might almost be said uniformly, affected by temperature than any other disease of the zymotic class {save perhaps cholera). It appeared, therefore, for my present purpose better to group diarrhcea with the other causes of death, many of which are especially influenced by meteorological conditions. Lower down in the diagram are shown the deaths from diarrhoea per thousand living, coloured yellow. Just above will be seen some red marks; these indicate unusually hot summers. Corresponding to cach of these will be noted a rise in the diarrhea mortality. Similarly the death-rates from diseases of the respiratory system” (but not including phthisis or consumption) are given in green. Above are blue marks, placed on the lines dividing certain years from the next; these indicate exceptionally hard winters, and correspond more or less closely with increases in the mortality from respiratory diseases. Finally in black arc given all the remaining deaths. It is plain that by successively subtracting (1) the chief epidemic diseases, (2) diarrhea and (3) respiratory diseases, we have produced a curve marked by much less conspicuous fluctuations than the one we started with, In the. original curve the fluctuation from year to year amounted to one per thousand and upwards in fifteen cases, even such fluctuations as 2'1 and 4'3 occurring, After abstraction of the epidemic diseases, fluctuations of one per thousand and upwards are scen only six times, and the maximum is 1'8; finally, after abstracting diarrheea au BIRTH-, DEATH-, AND MARRIAGE-RATES 21, and the diseases of the respiratory system, the maximum fluctuation is 1-3, and even that is exceptionally large. Speaking generally, the meaning of this is that two-thirds of the deaths in this country are due to causes which do not fluctuate greatly from year to year, although when long periods are taken into consideration, changes, which may be termed secular, are apparent. This question is, however, dealt with in some detail in the chapter on the decline of the death-rate in recent years. 2 STUDIES IN STATISTICS CHAP. CHAPTER 1V. THE GROWTH OF POPULATION IN ENGLAND. How the present century may compare with those that are to come we can have little or no idea, but that the contrast between the 19th century and those which preceded it will always be emphasized by the historians of the future, perhaps even more than by those of the present, is certain. What, then, are the cardinal facts that lie at the bottom of this contrast? What causes have led to the remarkable developments of the past fifty years ? The matter may, I think, be summed up in a sentence. The application of novel mechanical appliances has rendered possible an unparalleled increase of population, by bringing within reach unlimited supplies of food. I do not concern myself with Oriental nations nor with periods of remote antiquity, but with these limitations I maintain that such an increase of population as has been witnessed during the last fifty years in Europe and its daughter-states in the New Worlds of the West and South, has never been seen before. Those wonderful mechanical triumphs of the age, the locomotive and the steamship—in the development of which our country has played a leading part—have brought to our very doors, as it were, tracts of land so vast as to seem to the wearied imagination illimitable ; land for the most v THE GROWTH OF POPULATION IN ENGLAND 23 part with a fertile virgin soil, land moreover within the temperate zone, and fitted by its healthy climate for settlement by Europeans. This land has existed for thousands of years at the very least, much of it has been known to Europe (though very imperfectly) for several centuries, but it has only been readily accessible for a few decades. The “opening up of this land has afforded unlimited employment and unlimited supplies of food. If what I have here said be true—and it cannot be denied—it is evident that the great expansion of the 19th century is not only exceptional but absolutely unique, since nothing exactly like it can ever occur again. The growth of population may indeed go on—doubtless it will go on—applied science will provide our descendants with ingenious contrivances as yet not dreamed of, but no fresh mechanical inventions can discover for us vast regions of fertile soil lying under a temperate sun. Such lands will be discovered in detail, on a small scale ; new railways will be built, opening up corn-producing lands as yet unmapped ; but no such marvel is in store as the opening up of the great western prairies of North America, or the colonization of such an island as Australia. That the recent increase of population is absolutely enormous it will be easy to show, that it is exceptional I shall not attempt to prove. It would indeed require an antiquarian essay to which I am wholly unequal; the reader must accept my bare statement. England and Wales alone add 1000 a day to the population of the world!* That is to say, over and above reserve men who fill up the gaps caused by death, a fresh 1 Forty-Third Annual Report of the Registrar-General. Average of the five years 1876-80-—Daily births, 2425 ; daily deaths, 1426. The averages for the five vears 1882-6 were—Daily births. 2456 ; daily deaths, 1441. STUDIES IN STATISTICS CHAP, regiment at full war strength daily marches to the front. Where is this regiment ultimately quartered? What becomes of the people? The census of 1881, and the Registrar-General’s reports, supply the means of obtaining an answer to this question as follows— DAILY INCREASE OF THE PEOPLE OF ENGLAND AND WALES AND THEIR ULTIMATE DESTINATIONS. Daily Daily Daily increase. destination. migrants. London ia es cs 19 large towns oe wit 56 other towns oe on Rest of England and Wales Countries bevond the sea ... Total es a. 133 135 40 599 1000 '65 i 10 - 1004 32 - 21 59 155 43 0 Put shortly into words, and remembering that we are dealing on the one hand with the excess of births over deaths and on the other with excess of emigrants over immigrants (the latter comprising Irish from Ireland, Scotch from Scotland, as well as returned colonists and European foreigners settling in England), the table says that the natural growth of population in London is 133 a day, reinforced by 32 daily immigrants, mostly from the country districts; that of the 19 large towns of the Registrar-General, comprising together a population in 1881 L This table was constructed as follows—The 19 large towns are those given in the Quarterly Returns for 1881 ; the 56 other towns comprise the 50 towns given in the same Reports, with the addition of the rapidlygrowing towns of Croydon, West Ham, Stockton-on-Tees, Aston Manor, Burnley, and Ystradyfodwg. The populations of these places in 1871 and 1881 were taken from the census, and the births and deaths from the Reports. Certain adjustments were required in the cases of Barrowin-Furness and Middlesborough, which, as those registration districts were not formed till after 1871, involved taking Stockton as one of the towns. Some minor alterations of boundaries were neglected, hence the ficures Iv THE GROWTH OF POPULATION IN ENGLAND 235 of 3% millions, is 185 a day, reinforced by 21 countrymen ; that of the next 56 towns, comprising a population in 1881 of 3% millions (say another London), is 140 a day, reinforced by 59 daily immigrants; while the residue of England and Wales, comprising the smaller towns and more rural districts, increasing through excess of births over deaths by 592 a day, after sending 112 recruits to London and the 75 great towns, send yet another 43 emigrants to seek new homes across the seas. In recent years emigration has been proceeding. on an unusually large scale, but as a set-off against this, the daily excess of births over deaths tends to increase: indeed forty years ago it amounted to only 500 a day.! It does not therefore seem unreasonable to suppose that this rapid production of human beings may continue unchecked to the beginning of the next century. The in the following table have no pretension to exactness. but they are quite near enough for the purpose. TEN YEARS 1871.80 London ... 19 large towns 56 other towns Rest of England and Wales Total 0 Births 1871-80. Deaths 171.80 Natural necrease. ee RTT Population Population 1871 Te] Actual increase. ft or Tis hor migration +107,748 + 72,656 +198.074 245,441 nets 1,201,546 790,96€ 877,283 791.749] 454,47! 460,229 479,797 3,254,260 3,230,957 2 869.362 3,816,48¢ 3,763,842 3.547.935 562,225 532,885 677.871! +.304.283]2,788,312".015, 07012 a57.68* 7,480,104 |- 526,776 18.588.782/5.178.311113,410,471/129.712.266|25.974.439||3,262. 1 72 I— 148.298 (The balance, 148,298, represents the net emigration from England and Wales during the decennium.) Then taking the number of daily births and daily deaths (as in years 1876-80) such as to give a natural increase of 1000 a day, and assuming the same distribution of the people to be now going on as during the decennium ending with the last census, the figures in the table in the text were deduced. ' The Reports of the Registrar-General show for the 5 years 1848-52 the daily births to have been 1628: the daily deaths, 1101; daily tnerease. 5927 26 STUDIES IN STATISTICS CHAP. result of twenty years, that is from the census of 1881 to that of 1901, assuming the people to be distributed in the same manner, would be in round numbers— PROBABLE INCREASE IN 20 YEARS 1881—1901. London ... ves 1,200,000 75 large towns .., 2,600,000 Rest of England and Wales ... 3,200,000 Emigrants sin .- 300,000 Total ee 7,300,000 Presuming that accommodation will have to be found in England and Wales by A.p. 1901 for some 7 millions more persons than were enumerated at the last census, the first point to be noticed is that more than half (52 per cent.) of this large addition will be given to our already teeming cities and towns,! less than half (44 per cent.) to the hamlets, villages, and smaller towns that make up the population, that for convenience we may term rural, though perhaps strictly speaking it scarcely merits that designation. To any one who has compared the physique of militia and volunteer battalions recruited in the country with the like recruited in towns, this will be painfully suggestive for the future of the people of England. The concentration of the people in certain parts of England and Wales is illustrated by the adjoining table, which gives the number of persons to the square mile in each registration county as determined by the census of 1831, and again, fifty years later, in 1881. In two Welsh counties there was a trifling decrease; in twenty-one English and nine Welsh counties the increase was comparatively small, under 100 to each square mile. On the other hand, in London, Surrey, Middlesex, Stafford, Warwick, Lancaster, West York, Durham, Monmouth, and 1 The smallest of these towns, Colchester, had in 1881 over 28,000 inhabitants. Iv THE GROWTH OF POPULATION IN ENGLAND 27 Glamorgan the density more than doubled. Over 400 additional persons were in the half century placed upon every square mile of Stafford, Warwick, West York, Durham, Glamorgan, and the parts of Middlesex and Surrey outside the metropolitan ‘area; over 1000 upon every one of the 2,043 square miles of the large county of Lancashire, while in the three registration counties of which London is made up, the additional numbers crowded upon each square mile were no less than 6,000. 18,000, and 26,000 respectively ! DENSITY OF POPULATION IN THE REGISTRATION COUNTIES OF ENGLAND AND WALES, SHOWING THE CHANGES IN FIFTY YEARS. “1831. | 1881. Minerease. Metropolitan Middlesex'25,300 51,828 26,528 . Surrey 9,336{27,925 18,589 i. Kent 2,456| 8,550 6,094 Surrey (Extra Met.) 221| 651 430 Kent ,, 263 468) 205 Sussex ... _- 186] 334 148 Hampshire we 192] 3851: Lg Berkshire ... _— 193 276 82 Middlesex (Extra Met.) 439| 1,365 926 Hertford ... ... 214) 298 7¢ Buckingham ~~ ... 204| 242 ar Oxford ... . 201] 237 az Northampton .. 180] 280’ lor Huntingdon 158 167 Bedford ... 06| 321 Cambridge 165 213 Essex Ton 208 891 Suffolk... 205 242 Norfolk ... 88, 21. Wiltshire ... 821 201 Dorset .., 60 194 Devon ... 1921 237 Cornwall ... 220 288 Somerset ... 253 294 Floucester 24 479 | Hereford ... £33 46 Shropshire en 162 182 Stafford ,.. ve 354 840 Worcester we joo] 564 Warwick ... son 362 766 | Leicester ,.. ar 801 Rutland ... Lincoln ... Nottingham Derby vee Chester ... Lancaster ... West York Rast York North York Durham ... Northumberland Cumberland Westmorland Monmouth Glamorgan Carmarthen Pembroke ... Cardigan ... Brecknoek Radnor ... Montgomery Flint _— Denbigh ... Merioneth Carnarvon Anglesea ... England & Wales!270 1831.1 1881, ‘Increase 127 ' 137 10 115 | 171 56 258 | 465 207 241 | 444 203 316 | 61¢ 303 566 1,70¢ 1040 s60 | 70: 432 i74 332 189 93 171 78 207 732 525 118 215 er ‘12 165 53 70' 82 12 169! 851 182 147 5760 129 ar 147 28 31 151 on 98 102 70 78 50 54 gr 84 30¢ 2 34 192! 60 91 131 228% rod 189 1 14a 00% Decrease STUDIES IN STATISTICS - CHAP. The two maps (Plates IV. and V.) show some of these changes in a perhaps more striking manner; the gradual growth of the grey and black areas, representing the more densely populated counties with upwards of 300 persons to the square mile, at the expense of the green areas representing the counties of more purely agricultural character with sparser populations, is plainly seen, notably in the case of Durham, South Wales, and the home counties. Let us now consider for a moment what arrangements must be made for the reception of these 7 millions. They will all require educating ; a large but uncertain fraction will have spiritual wants; they will all have bodily wants, A small fraction will be afflicted in mind, a much larger fraction in body and estate, some also will be morally afflicted. Over 1} millions will be of school age, and for these schools will be needed. As regards spiritual wants, about 31 millions will be between the ages of 20 and 60.! Now the Official Year-Book of the Church of England gives detailed figures which prove that almost exactly one quarter of all the children in the country of a suitable age are confirmed. Making then every allowance for aged, sick, indifferent, and backsliders, it will probably be within the mark to assume that church accommodation will be required for one-fourth of all above ten years of age; this would give 11 millions; but to be on the safe side we will say one million more or less regular attendants on the services of the Church of England. As to Nonconformists we have fewer data to go upon, but the required accommodation for them can scarcely be put lower than half a million. Calculating then from the figures of the 1 Of course none of those born since 1881 will be of these ages, but the children of to-day will then be grown up, while the middle-aged of to-day will then be old. Plate IV - 6 Yo ND &WALES eer of English Miles a7 20 30 40 50 gp emp rgy ‘srsons to the Square mile Census 1831 4 "RFOLK | “Ol K "2g =} + “NT se counties UNDER 100 00 - 200 00 = 300 London: Edward Stanford . Vangse counties 300~500 500 -1000 OVER 1000 W. Gr. n° "EC London, Stanfords Geog? Kstad: Plate V. 56] LV ND & WALES € Scale of English Miles 057 WW 70 30 40 50 — ee Rr Persons to the Square mile Census (881 4 « ‘2 % | Mr = counties “DER 100 200 or London: Edwurd Stanford W Cop % counties 300-500 500-1000 OVER 1000 0 Cd 1 E Gn Tendon. Stonfords Ceooal Tete v THE GROWTH OF POPULATION IN ENGLAND 29 census and other sources,’ it may be inferred that by 1901 the following new public buildings will be required. Schools for 1,500,000 children Churches for 1,000,000 Anglicans Chapels for 500,000 Nonconformists Asylums for 15,000 lunatics and idiots 2 Hospitals for . 6,000 sick (at least) Institutions for 400 blind Institutions for 300 deaf and dumb Prisons for wo —_— 7,000 criminals Workhouses and schools for 48.000 paupers It is to be hoped that the last two items have been over-estimated, but on the other hand many public institutions have been entirely omitted from the reckoning, and we may in any case feel quite sure that during the next twenty years there will be ample scope for the exertions of the Religious, the Benevolent, and the Rate-collector. At this point I will pause for one moment to call the earnest attention of economists to one of the results of modern legislation, with its increasing tendencies to State socialism. Between 1875 and 1887 (that is, prior to Mr. Giéschen’s separation of the Loans for Local W orks) the National Debt was reduced from £768,945,757 to £736,278,688, that is to say, by some 382% millions sterling.* This is primd facie a very satisfactory result for twelve years. It is, however, not so well known as ‘+ should be. that during the like peried the indebtedness I For inmates of institutions, see The Chrrities Register and Digest. Longmans, 1884. 2 Tn the county of London alone the average increase in the number of lunatics and imbeciles to be maintained in the pauper asylums amounts to 258 annually (from 1st Jan. 1884 to 1st Jan. 1889). 8 Our hishops might well take into consideration the advisability of making more use of existing buildings by increasing the number of services and clergy—and by more adequate ventilation. t Qtatictiral Abstreet. 1880. STUDIES IN STATISTICS CHAP. of the local authorities of England and Wales alone increased from £92,820,100 to £186,821,642," or by no less than 94 millions sterling; in fact, the Local Debt was more than doubled during the same twelve years! Thus the total indebtedness, not to speak of Scotland and Ireland, had increased by 61% millions! During the eight financial years 1880-88 the amount levied in rates increased from £22,000,000 to £27,000,000, or by nearly 23 per cent.’ I do not for one moment deny that many and great benefits have resulted from this vast expenditure, but I do assert emphatically, that it behoves those among us who would do everything by Act of Parliament to bear constantly in mind that some one will have to pay, and usually to pay heavily, for a “ Progressive” policy. It may be noted in passing that, calculating from the average density of those portions of the metropolis that have ceased to grow in population, which we find to be about 200 persons to the acre, these 7 millions of persons will require for their habitations and streets, with but a very small allowance for open spaces, at least 85,000 acres of land, though, considering the fact that suburban populations cover far more ground, the space actually occupied will probably not fall short of 70,000 acres. which of course must be withdrawn from cultivation. To explain the extraordinary increase of population in modern times we must analyze the factors upon which it depends. We will take first the causes that tend directly to diminish population ; they are three—war, disease, and famine. Wars are now scarcely less frequent than in the past. and perhaps more severe, but they are far shorter 30 1 Nineteenth Report of the Local Government Board, pp. elxxx, clx xxi, 2 Ibid. n. 592, Vv THE GROWTH OF POPULATION IN ENGLAND 3 in duration.! With the enormous armies and deadly weapons of to-day the numbers killed in action may not be less, but modern campaigns are conducted with far more consideration for the men taking part in them, and above all they are more decisive, so that the pestilences which in former times followed armies in their long-drawnout campaigns, claiming far more victims than shot and shell,” are now reduced to insignificant dimensions. At the same time the longer intervals of peace and general prosperity, with ample supplies of food, have led to rapid multiplication of the species. As births are the one cause of increase of population, so deaths are the one cause of limitation of that increase. The death-rate depends upon a number of circumstances which are more or less mutually interdependent. Thus Season affects the death-rate, extremes of heat, cold, or drought increase it; mild weather and rain diminish it; Famane, partly dependent on season, is in some places and it some times a potent factor in checking population ; War, often leading to famine, still more often to pestilence, is in modern Europe a more frequently acting cause than famine, but hardly so potent ; Pestilence, spread by war, fostered by famine, and regulated to a great extent by season, has a very potent effect upon the death-rate. It is, however, very rarely nowadays that epidemics occur on a sufficiently large scale to produce an appreciable effect ! While these pages were going through the press, the greatest living strategist, breaking his habitual “silence in seven languages,” told the German Reichstag, in an eloquent speech, that the duration and end of the next war could not be foreseen; that no one of the greatest Powers of Europe, armed as never before, could be so completely shattered in one or two campaigns as to confess itself beaten ; that the next war might be a seven years’ war, or even a thirty years’ war.—Von Moltke's speech on the German Army Bill, Standard, May 15th, 1890, p. 5. ? It has been stated that the campaign of 1866 was the first in which more mei were killed in action than died of sickness. STUDIES IN STATISTICS CHAP, upon the population. Even in 1849, the great cholera year, the births in England and Wales exceeded the deaths by 187,000; but in the Middle Ages it was far otherwise; the Black Death, the Sweating Sickness, and later the Plague, undoubtedly exercised a most powerful and enduring influence in checking the growth of population. The death-rate in our great cities, though much reduced in recent years, still greatly exceeds that of country districts ; again, the heavy death-rates of such places as Marseilles and Naples, Munich and Buda Pesth, still more Madras and Cairo, in which the sanitary conditions are, or till lately were, sadly behind the age, give us some idea of the uncertainty of life in bygone times. The excellent means of communication that have been established within the last half-century make famines almost unknown in Europe, though they still occur in India and China, and to a less extent among the Indians of North America. Improved sanitary legislation and a greater appreciation by the mass of the people of the importance of attention to details in matters relating to drainage, watersupply, and ventilation, as well as better methods of isolating, nursing, and treating the sick, have not only proved effectual checks to the ravages of epidemics, but have greatly lessened the general death-rate. Population is increased by births alone, for migration is not an increase; it is but a transfer, a mere matter of account. Births depend primarily upon the number of women of the child-bearing age. As the population increases there are naturally more women of such age, and hence more parents. The increase itself increases; it is like compound interest, what mathematiclans term a geometrical progression. The young folk of to-day are not aszociated with a proportional number of aged persons, Vv THE GROWTH OF POPULATION IN ENGLAND 33 since the latter are the survivors of a much smaller population of a bygone time. But births also depend upon marriages, and marriages, as we have seen, are largely influenced by the general prosperity, so that the recent prolonged depression in trade is seen to have depressed the marriage-rate in nearly every country of Europe, and this in its turn is seen to have been followed by a fall in the birth-rate. Nevertheless the present century has been largely one of prosperity, and has enjoyed long intervals of Peace, hence it has on the whole been one of high birth-rate. STUDIES IN STATISTICS CHAP. CHAPTER V. THE MIGRATIONS OF THE PEOPLE IN THE 19TH CENTURY. Or course the mere movement of people from one place to another cannot of itself directly either increase or diminish the total number of people in the world; but in the case of individual countries, or smaller areas, it often has a great, it may be even a predominating, influence in regulating their populations. Moreover, when the settlement of a new land increases the supply of food, or when the occupation of a distant country by a military force prevents or delays a number of marriages, we can easily see that migration may even influence the sum-total of the population. Migrations may be divided into three classes. Fuorstly, migration occurring within the country itself, which might be called intra-migration. Secondly, migration from country to country, but within the limits of one continent ; this might be conveniently called wnter-migration. Thirdly, migration from continent to continent, ordinarily called, according to the direction of its flow, either emigration or immigration, but which might with more precision be termed wltra-migration. Migration of the first kind, that occurring within the limits of the country, or sntra-migration as I propose to term it, is a great and constant movement varying only in deoree. directed from the cottage to the village, from the MIGRATIONS IN THE NINETEENTH CENTURY 35 village to the town, from the town to the city, from agricultural to mining or manufacturing districts. Its main motive power is the greater demand for labour in the large centres of population, and consequently higher wages ; men leave their homes with a view of « bettering ” themselves, often making little allowance for the greater cost of living in towns. But, in addition to purely economic reasons, there is another scarcely less potent—the love of excitement and novelty. The noise and bustle of the street, squalid though it be, and the glare of gas-lights, even though mainly those of the gin-shop, are strangely attractive to the countryman. By a curious perversion, the advantage of towns is said to be “life.” There is, in truth, more life in a given space, more high pressure, more rush; but it is the rush of a clock running down. [f there is more life in towns there is assuredly more disease and death. Town-bred children are notoriously less healthy than those reared in the fields, and the healthiest town has a higher death-rate than the surrounding country. It must, moreover, be borne in mind that the death-rates of our towns would be still higher were it not that they are depressed hy the large number of immigrants from the country at the healthiest period of life, domestic servants and others. A table in the census! shows that the population of some counties is decreasing, though this decrease is very slight. Thus the county of Montgomery reached its maximum in 1841; Huntingdon, Rutland, Brecknock, and Anglesea in 1851; Cornwall, Pembroke, and Radnor in 1861 ; Cambridge, Dorset, Hereford, Shropshire, Westmorland, and Cardigan in 1871. I give in a table the amount of the loss of population since the maximum. t Censis of 1881, vol. iv. pp. 80, 81. 36 STUDIES IN STATISTICS CHAP. DIMINUTION OF POPULATION OF CERTAIN COUNTIES AS ASCERTAINED BY THE 1881 CENSUS. Huntingdon Cambridge Dorset Cornwall Hereford Shropshire Rutland ... Westmorland Pembroke ... Cardigan ... Brecknock ... Radnor ... Montgomery Anglesea ... Year of maximum. Amount of loss. 1851 7,096 1871 919 1871 4,028 1861 38.473 1871 3,838 1871 1,285 1851 1,265 1871 816 1861 4,011 1871 2,732 1851 5,038 1861 1,125 1841 i. 3,560 1851 ee 8,102 Total 89 988 The total loss of these fourteen counties amounted to only 82,000, of which nearly one-half consisted of Cornishmen thrown out of employment by the failure of the mines. There is here little evidence of the * depopulation of the rural districts,” of which so much is heard on political platforms. For further evidence on this point, the reader is referred to a careful paper by Dr. Ogle. The accompanying map (Plate VI.) shows the net results of intra-migration in England and Wales for the ten years preceding the 1881 census. The counties coloured red actually lost 57,206 inhabitants during. the period ; those coloured blue and purple both gained in numbers ; the blue counties, however, did not gain so much as the excess of births over deaths would have given them, whereas the purple counties gained more. A comparison with Plate V. will show that the densest counties 1 The Alleged Depopulation of the Rural Districts of England, by William Ogle, M.A., M.D., F.R.C.P., Journal of the Royal Statistical Society. 1889, vol, lii. p. 205. Plate VI 56 AND & WALES Secale of English Miles 0 6 0 © 20 30 40 50 a ee py Migration of the people in the ten years 1871-80 1F 159 ORFOLK FFOY ”, . 21’ Counties tr population which has diminished Counties the population o' which has increased less than the excess hirthe aver death 1'W Cr. London: Edward Stanford Counties the population of which has _— increased more than the excess of birthe aver dest hs 150 i m—— — London, Stanford Geogt EBsinb® pupal 1 F Cr 7 MIGRATIONS IN THE NINETEENTH CENTURY 37 attracted immigrants from the less dense; population attracts population. Excess of births over deaths. Gain or loss of opulation (census) Red counties ... 183,415 — 57,206 Blue counties vr 1,504,279 + 952,293 Purple counties ... 1,722,777 + 2,367,086 Total ... eee + 3410471 | 3,262,173 Immigration (+) ar Emigration{ —). — 240,621 — 551,986 + 644.309 — 148,298 Being emigrants who Jett the country. The red counties, therefore, sent out 240,621 men, women, and children, the blue counties 551,986 ; of these 644,309 settled in the purple counties, while 148,298 crossed the seas.! These figures are the net results; of course the actual amount of movement was much greater, as numerous persons may be said to have exchanged places with persons in other counties, but of them and their migrations there is no record. This form of migration has been dealt with in considerable detail by Mr. E. G. Ravenstein,® and is alluded to below in Chapters X. and XI. Migration of the second kind—inter-migration, or the movement from one country to an adjoining one—is generally by persons in search of work or advancement, and is therefore apt to be directed from the poorer to the richer country; as when Italians go into France, Austria, or Switzerland, or when Irish and Scotch betake themselves to England. This inter-migration is not usually directly the subject of official records, but indirectly its results are from time to time recorded in the tables of birthplaces in census reports. The countries that are ! The figures in the table are derived from the Census Report, 1881, vol. il. p. xxi. } 2 The Laws of Migration, by E. G. Ravenstein, FRGS.; Foe the Statistical Society, vol. xviii ( 1885) p. 167; also vol. lil. (1889) p. . CHAP. most notable as attracting inter-migration are France and England. Census 1851 1861 1866 1872 1876 1051 1886 FOREIGN-BORN IN FRANCE.! 380,831 497,091 635,495 740,668 801,754 1,001,090 1.126.531 The number of foreigners in Trance is not only exceptionally large, but it is rapidly increasing. Since 1872, the population of France has increased by only 2,115,982, while the number of the foreign-born has increased by 885,863. That is to say, no less than 18 per cent. of the total increase consists of foreigners. This influx of immigrants into a country where the native population is systematically restricted is a strange phenomenon. The chief nationalities of the foreigners in France are given in the following table ; the very rapid immigration of Italians and Belgians deserves especial notice— NATIONALITIES OF FOREIGN-BORN IN FRANCE? 1872. Belgians ... ... 1347,558 Italians ... ce 112,579 Germans ... ce 39,361 Spaniards & Portuguese ! 52,954 Swiss - we 42,834 Dutch - — 17,077 Enclish. Scotch and Trish: 96 003 1881. 432,265 240,733 81,986 74,633 66,281 21,232 37 006 1886. 482,261 264,568 100,114 80,842 78,584 37,149 36.134 {Increase in | 14 vears. 134,703 151,989 60,753 27,888 35,750 20,072 10,131 In the British Islands the movement between the several divisions of the kingdom is very interesting, and has had most important economical results. The following 1 Statistique Générale de la France, 1886. 2 Ibid v = MIGRATIONS IN THE NINETEENTH CENTURY 39 table gives the birthplaces of the inhabitants of the several divisions! of the United Kingdom as ascertained at the last three censuses. It will be observed that although the number of Scotch in England, of English and Irish in Scotland, and also of English and Scotch in Ireland has increased, the number of Irish in England has diminished in the twenty years 1861-81. It would be quite wrong to infer from this that the migration of Irish into England has stopped, since deaths alone would have caused a much larger diminution in the time. BIRTHPLACES OF THE PEOPLE OF THE UNITED KINGDOM. CENSUS 1861. } BorRN I> ENUMERATED TN Enoland. Seotland. Ireland, England Scotland 9,120,052] 169,202] 5603212 785 804! 601,634 204,083 [reland | 50.936 16,925!5,720,686 Unirep Kine.|19.227,0200 “71 931/6.526.403 CENSUS 1871. England |21,692,165) 213,254] 566,540 Scotland | 70,482 3001.51 207,770 [reland 67,881 20,31815,306,757 Unrrep Kina. |21.830 528 3,295,103! 6,081,067 CEN~US3 1571. England 24,855,822! 253,628" 562,374 Scotland 91,823 5,397,75C 218,745 reland 69.382 22.398 .063,33¢ Txiren Kine. [25.017 02" The Colonies r India. Other Places. ToTAL. 51,672 ee 20,066,224 oe) 8,816} 3,062,204 10,420 5,798,967 202.131 IR 007 485 70,812]169,495]122,712,266 9,740 10,95 3,360,018 8,367! 9,054' 5,412,377 88.919/189.044 (31,484,661 BLA als 716125,974/439 12,874! © "9! 3,735,573 8,325! 5,174,836 734.848 Acting on a suggestion of Mr. Noel £ ifumphreys, I have been able to calculate approximately the amount 1 Wales is here reckoned as one with England. STUDIES IN STATISTICS CHAP. of movement between the divisions of the United Kingdom.* The method adopted is as follows, taking the Scotch in England as an example— In 1861 there were 169,202 natives of Scotland enumerated in England and Wales, but at the next census 213,254 ; the difference, 44,052, shows the net result of (1) migration between the two countries in ten years, and (2) the diminution by deaths of Scotchmen in England during the period. Now we cannot be very much in error if we assume that the average Scotch population in England during the decade was the arithmetical mean between the number at its beginning and end, thus— 169,202 + 213,254 _ 191,298 But of this number many must have died annually. To take a convenient figure, and one that is not improbable, we will assume that the death-rate was 20 per 1000. This death-rate would in the decade carry off one-fifth of the average population— WLR But the population instead of diminishing by this number actually increased by 44,052. Therefore enough Scotchmen must have entered England not only to fill up the 88,246 gaps caused by death, but also to provide the 44,052, the increase enumerated at the census of 1871. We get then as the net number of Scotch immigrants into England and Wales during the decade— 38,246 + 44,052 = 82,298 By similar calculations this table has been constructed. As it cannot pretend to accuracy. the results are given in round numbers. 1 The following figures first appeared in an article on the census of 1881 iu The Charity Organization Reporter, 1884, p. 2. MIGRATIONS IN THE NINETEENTH CENTURY 41 MIGRATION BETWEEN THE SEVERAL DIVISIONS OF THE UNITED KINGDOM IN DECADES. cotch to England. Irish te England. English to Scotland. English to Ireland. Scotch to | Irish to Ireland. Scotland. (841-30 | 50,200 311,500 1851-60 | 69,000 | 193,800 —_ - - 1861-70 | 82,300 | 81,700 27,100 | 28,800 | 7,100 1 44,900 t871-80 | 87.000 | 108,700 1 37,600 15,200 ! 6,300 | 53,600 Our ignorance of the number of the population in each year 1s a source of error, our ignorance of its age and sex constitution, and consequently of its death-rate, is a much more serious one, but in spite of all uncertainties this approximation is better than nothing. - An examination of these tables suggests several remarks. The flow from Scotland into England increases from decade to decade, though at a decreasing rate. The migration of the Irish is not anything like as large as it was in the decade 1841-50, but was larger in the last decade than in the last but one. The balance of intermigration for the last two decades was as follows— ‘Gain of England. | Loss of Scofland, | Loss of Ireland. 186170 | + 108100 | —17400 [= 90,700 1871-80 + 142.900 — 2.100 — 140.800 The colonial-born and foreign-born increased in each decade in each division of the United Kingdom (except that the colonial-born in Ireland in the last decade remained practically the same as in the previous one). It is curious that although the population of Ireland is steadily diminishing, the number of foreign-born is slowly increasing. Mr. Leonard Courtney has called attention to this," but he has overlooked what is, in my opinion, the chief cause, 4. e. the generally increased facilities of com. ' The Nineteenth Century, March 1888, pp. 340, 341. STUDIES IN STATISTICS CHAP. munication, which undoubtedly tend to a greater mixing up of the various peoples of the earth. It will be noticed that the number of foreigners in the United Kingdom is very much smaller than in France. In 1861 about one-fifth, in 1871 and 1881 about one-third, of the natives of “foreign countries ” enumerated in England and Wales were British subjects, the remainder being strictly foreigners.” Of these foreigners one-half were (in 1881) enumerated in London, and one-fourth in other large towns. Of the European foreigners one-tenth were sailors in our ports. If the increase of population of England and Wales in the forty years 1841-81, ascertained at each census, be compared with the excess of births over deaths, it is surprising how nearly equal the two sums are; in other words, it would appear that the immense emigration that has taken place has had very little effect upon the population. ENGLAND AND WALES. 1851 1861 1871 1881 Excess of births over Increase enumerated deaths in previous decade at each census. “6,61h 2,013,461 PG "1728.6" Forty years NO) NQ80 ky, Net gain or loss by migration. + 276,845 122,111 78,968 164,307 88,541 From this table it would appear that the population would have numbered only 88,541 more persons had no emigration taken place! And yet we know that during this period Australia and New Zealand have been peopled by British settlers, not to mention the many thousands that have gone to Canada and the United States. The apparent paradox admits however of explanation. To begin with, the births were imperfectly registered in the earlier decades, especially in the first, so that the true axeess of births over deaths was somewhat greater MIGRATIONS IN THE NINETEENTH CENTURY 43 than appears in the table. But by far the more effective cause was the great immigration of Irish and Scotch, described above. While the native English have gone out by one door, Irish, Scotch (and to a minor extent foreigners and returned colonists), have come in at another. Migration, then, has not materially altered the quantity of the population of England, but has greatly altered its quality, and this in two ways, not only by the substitution of Irish labourers for English, but by continually drafting off the most enterprising of the latter to extend the imperial domain in far corners of the earth. The third form of migration, that from continent to continent, from the Old World to the New, is of the greatest importance as a determining factor in history, and is that which I shall treat of at greatest length. [t.is impossible to state accurately the amount of migration taking place between different countries. Although in almost all old countries statistics of emigration are published, and in new countries the numbers of immigrants are recorded, the returns are liable to many errors. In the first place, they are often in themselves incorrect, either overstating or understating the facts; then there is mo means of proving the bona fides of alleged emigrants (or immigrants), they may return whence they came, or they may pass on to another colony ; lastly, many who are merely passengers to and fro, travellers on business or pleasure. are recorded as emigrants or immigrants. More satisfactory are the indications derived from the statements as to birthplaces in the censuses of the countries receiving the immigrants. These numbers show approximately how many have actually settled down in a new country, and the increase or decrease of persons of any particular nationality from census to census CHAP. indicates the direction (but not the exact amount) of fluctuations in the movement. Obviously census figures have the disadvantage that they are only published at long intervals. In inquiring into the statistics of ultra-migration it is well to consider the countries concerned in two groups— those that give, and those that take. To begin with those that give. The great radiating centres of emigration, confining ourselves to recent times, and more especially to the decade 1879-88, may be taken as four—Great Britain and Ireland ; Germany; Sweden, Norway, and Denmark ; and Italy. The average numbers of emigrants stated to leave their shores annually are as follows— The United Kingdom 248,000! Germany _ cus 130,000 2 Scandinavia ... ee 62,000 8 Italy ... - i. 82,0002 522.000 (1) The United Kingdom. The Board of Trade returns show that the average gross emigration of persons of British and Irish origin only was for the ten years 1879-88, in round numbers, 248,000. By deducting the recorded immigration (70,000) from the recorded emigration we obtain the met emigration, viz. 178,000. The figures, it should be remembered, do not include any movement of population between the United Kingdom and Hurope or the countries bordering upon the Mediterranean ; but there is no doubt that such interchange with these countries, though locally it produces important results (as in the East End of London), is relatively quite unimportant. About 67 per cent. of these emigrants went to the United States, 16 per cent. to Australia and New Zealand, 1 Board of Trade Return, 1886. 2 Almanach de Gotha. 8 Statesman’s Year-Book. v MIGRATIONS IN THE NINETEENTH CENTURY 45 12 per cent. to Canada, and a very small number to South Africa. But when a long term of years is considered, it is found that the proportion of Irish is great among emigrants to the United States; of English among those to Australasia; of Scotch among those to Canada. There 18 no doubt that the great American Republic has an especial charm for Irishmen; but once there they seem to lose their passion for holding land, and congregate mostly in the vast cities of the Eastern States. During the 37 years 1853-89 the following numbers are said to have left our shores'— EMIGRANTS, OF BRITISH AND IRISH ORIGIN. Total. Yearly average. ’ of a dation. English ... 8,439,188 .. 02,050 ... .. 4.1 Scotch ve. 689,705 Sih 18,641 ... .. . bb Irish - ... 2,775,007 ” ME OOO © wer eee 139 When these are compared with the mean populations of England and Wales, Scotland and Ireland respectively (taking the population in 1871 as sufficiently near the middle of the period), we find that on the average England and Wales have annually sent out as emigrants 4:1 per thousand of their population, Scotland 55 per thousand, and Ireland nearly 14 per thousand. But, as has been already shown, this does not by any means represent the whole movement of the people, the ntermigration must not be forgotten; which has resulted in an Irish element in England exceeding half a million, a Scotch element in England exceeding a quarter of a million, and an Irish element in Scotland exceeding 200,000. Under the circumstances described above, the decrease of the population of Ireland should excite no surprise, and the cause of the migration is not far to seek. The 1 Board of Trade Return, 1889. STUDIES IN STATISTICS CHAP. standard of living has risen everywhere, and is still rising ; and whereas Ireland is too poor a country to allow its population to live up to that standard, its people must seek a happier fortune elsewhere, consequently the less adventurous go to Liverpool and Glasgow, the more adventurous to New York. UNITED KINGDOM. INCREASE OF POPULATION IN FORTY YEARS. England and Wales Scotland - Ireland es United Kingdom Census 1841. 15,914,148 2,620,184 8,196,597 96.730,929 Census 1831, 95,974,439 2,735,573 A174 R24 ww) Increase 10,060,291 Increase 1,115,389 Decrease 3,021,761 Inerease 8,153,919 - The population of Ireland reached its maximum in 1845 or 1846, and has since then declined, but the rate of decline since 1861 has been steadily diminishing. The upper black line in the diagram (Plate VII.) shows the gross emigration of persons of British and Irish origin from the: United Kingdom from 1853 to 1889; but, as Dr. Giffen points out in his Annual Returns, it must not be forgotten that these are not by any means all true emigrants; many are merely passengers to and fro on business or pleasure, and these form a constantly increasing number ; moreover, not a few emigrants fail, and return to the mother country. The number of immigrants, or passengers landing in the United Kingdom, has only been recorded since 1876. The true emigration is the net movement from this country after the deduction of all immigrants. The net emigration is represented in the diagram by the lower black line in the right-hand portion of the diagram, the area tinted gray representing the number of immigrants. The blue line gives the gross number of Irish immigrants. The first maximum of emigration from the United 3ROSS EMIGRATION FROM THE UNITED KINGDOM OF PERSONS OF BRITISH & IRISH ORIGIN, ALSO OF IRISH ONLY; ALSO NET EMIGRATION FROM THE UNITED KINGDOM SINCE 1876. 853 — '55 ‘re? 59 ‘Gt '63 65 67 — '69 71 73 '75 '77 79 81 ey ny Pg 27 '89 1 300.000 GROSS EMIGRATION UNITED KINGDOM 200.000 NET EMIGRATION 00.000 IRISH GROSS EMIGRATION 3 3 A —_— ee a Condom. Stanford Geogt Estaht DESTINATION OF EMIGRANTS FROM UNITED KINGDOM,WITH IMPORTS & EXPORTS OF UNITED STATES. 1853 55 ag > 6" se 81 ‘63 '65 '67 69 71 73 “ap _— -T — rr Tr 7 75 — 77 '79 sl ‘R2 '85 87 ‘an rr ? 1 ' & ify —t —— 200 000 TO | UNITED STATES ebm 100.000 TO BRITISH NORTH AMERICA Pie —— o 3 Nn -_—t re, London : Stanford’s (eoqt Estab® 3 ~ MIGRATIONS IN THE NINETEENTH CENTURY 47 Kingdom was in the years immediately succeeding the [rish famine (1847-54). The discovery of gold in California in the autumn of 1847, and in Australia in the spring of 1851, largely stimulated the flow of migration during those years. The nationalities of emigrants from the United Kingdom have only been distinguished since 1853 (apparently the maximum year). Emigration greatly diminished after 1854 till it reached the minimum in 1861 ; from that point it rose rapidly till 1863 ; then after a slight fall rose again till 1873; after this year it fell till in 1877 it nearly reached the mark of 1861, but at once began to risc again till it reached its maximum in 1883. [t has been unusually high every year since 1880 ; indeed this wave of emigration exceeds that of the earlier movement, though owing to the growth of population in the interval it is not proportionately so great. It should be noted that in 1853 the Irish emigrants amounted to 69 per cent. of the whole, but since 1864 the proportion has steadily declined till it is now only 26 per cent. of the whole. The Scotch have been pretty constantly about 10 per cent. of the whole ; latterly 12 or 13 per cent. The English proportion has risen from 23 per cent. to 63 per cent., but in 1876-7 was 67 per cent. of the whole. The other diagram (Plate VIII) gives the destinations of the emigrants; the blue line shows those going to the United States, a curve practically nearly the same as that of the total emigration; the red line shows the emigrants going to Canada—its chief features are those of the American curve, but its tendency of recent years has been more decidedly upward ; the yellow line applies to Australasia. The latter differs from the two others in many particulars. This, according to Dr. Giffen, is because emigration to Australia is largely governed by the assistance given by the colonies. and is not 48 STUDIES IN STATISTICS CHAP, regulated simply and directly by the laws of supply and demand as that to the United States appears to be. The dotted yellow line represents the imports and exports of the United States.! This, it will be seen, corresponds pretty closely with the curve of emigration to the United States, and proves that the activity of trade in that country is the main cause which regulates the number of immigrants; a slight improvement in trade appears to immediately stimulate emigration from this country to a disproportionate degree. (2) Germany, the second centre of distribution, is scarcely less important than the first. Germany alone has for ten years (1879-88) sent forth yearly 130,000 emigrants.” Of these no less than 96 per cent. go to the United States; the next most favoured country is Brazil ; then other parts of the American continent ; a few go to Australasia; a very few to Africa. There is indeed scarcely a corner of the world—no matter how remote, no matter what the climate—in which Germans will not be found. These are, however, remarkable more for their perseverance, industry, and thrift than for their absolute numbers. Together with German emigrants may be taken Austro-Hungarians ; their number is very uncertain ; more leave Hamburg and Bremen than the official statistics of the Empire Kingdom admit. According to American statistics there have landed in the United States in recent years (1881-6) some 30,000 * Austro-Hungarians annually, and these numbers tally fairly well with the figures of the German ports. Holland also sends small numbers of emigrants to the 1 Report of the Chief of the Bureau of Statistics of the United States, 1889, p. 787. 2 Almanach de Gotha. 3 Statesman’s Year-Book, 1888. 4 Ibid, v MIGRATIONS IN THE NINETEENTH CENTURY 49 United States.’ In Belgium the balance of migration is in favour of the home country, and results in an annual addition of some 3000 a year to the population.? This is the more remarkable, as Belgium is the most densely populated country of Europe. The Swiss, besides overowing into neighbouring countries, cross the ocean annually to the number of some 10,000. The great majority go to the United States, most of the remainder to South America. (8) The third centre consists of the Bultic countries— Norway, Sweden, and Denmark, called shortly for convenience Scandinavia. The great majority of Scandinavians go-to the United States, a fair number in recent years to Canada. EMIGRATION FROM EIGHT EUROPEAN COUNTRIES COM. PARED WITH NATURAL INCREASE OF POPULATION. ANNUAL AVERAGES, TEN YEARS 1877-86. United Kingdom | rermany Italy Sweden Norway Switzerland Denmark France Total Excess of births over deaths. 1 452,000 537,000 254,000 56,000 98,000 22,000 27,000 92.000 1 “AK DON Emigrants. 148,0002 108,000 3 56,0003 28,0004 15,5004 7,500 3 6,000 ¢ 4,700 3 272.700 Emigrants per cent. of natural inereace 32-7 20-1 22-0 50-0 054 34-1 22-2 Koy 95-5 L Fiftieth Annual Report of Registrar-General. * Board of Trade Return, 1886. Net emigration of British and Irish origin, 8 Almanach de Gotha. t Statesman’s Year-Book. ! Some 5000 a year. Report of the Chief of the Bureau of Statistics of the United States, 1889. ? Average of 1882-86. Almanach de Gotha. y: STUDIES IN STATISTICS CHAP. Concerning Russia there is no precise information, but the American census of 1880 showed 84,279" Russian born, more than half being Poles, an increase of 65,000 on the previous census, and the American immigration returns for 1881-85? claim an average of 15,000 immigrants from Russia annually. It will be observed from the table that the number of migrants from Sweden and Norway, though insignificant when compared with English, German, or Irish emigrants, is relatively very large, amounting in the case of Norway to 55 per cent., and in the case of Sweden to 50 per cent., of the natural increase, or excess of births over deaths. In 1881-2 when Scandinavian emigration was at its highest, the number of emigrants from Norway exceeded the natural increase, and in Sweden in the same years nearly equalled it. The proportion borne by the emigrants to the natural increase of the population is about the same in Switzerland as in the United Kingdom, viz. one-third; in Germany, Italy, and Denmark it is one-fifth. In France the excess of births over deaths is very small, but the emigration is still smaller, amounting to barely 5 per sent. of that very small increase. (4) The last centre of emigration to which I have to call attention is Italy and the Mediterranean, or Latin countries. The Italian emigration returns distinguish between those who leave Italy for countries in Europe, principally France, Austria, and Switzerland, and what most people more usually understand by the term emigrants, those who cross the seas to seck a fortune in distant parts of the world. The former number 85,000 annually, the latter 56,000 (on the average of ten years 1 Cunsus of United States, 1880, 2 Statesman's Year-Book. v MIGRATIONS IN THE NINETEENTH CENTURY 51 1877-86). Of this number nearly two-thirds go to South America, of which by far the larger portion go to the Argentine Republic, the others to Brazil and Montevideo ; the majority of the remaining third go to the United States. France does not part with more than about 4700 a year,! South America (especially the Argentine Confederation) and the United States being equally favoured destinations. Very few Frenchmen settle in Canada ; it is said that the ¢ habitans” of Quebec do not like them to do so, as they introduce freethought and revolutionary ideas into that most conservative and most Catholic province, The emigration from Spain does not appear to be recorded ; it is probably not great. It seems to be mainly directed to Cuba, Puerto Rico, the Argentine Confederation, and in a less degree to Brazil. From Portugal * in the years 1879-87 the number of emigrants averaged 16,000. They nearly all go to America, mostly to Brazil; there is, however, curiously enough, a considerable emigration to Hawaii. These numbers include the emigrants from Madeira and the Azores, who amounted to 4667 in 1885. and 3726 in 1887. Almanach de Gotha. 2 Statesman’s Year-Book. STUDIES IN STATISTICS CHAP. CHAPTER VL THE GROWTH OF NEW NATIONS. Tre UNITED STATES. From the fall of the Roman Empire until the discovery of America the history of Europe was practically, so far at least as regards the Western races of mankind, the history of the world. In the three centuries that followed, as Prof. Seeley has so eloquently urged upon us, the events of most real moment were those which decided who should possess the new worlds beyond the sea, and more particularly the parts of them suited by soil and chmate for settlement by Europeans. At one time Portugal, at another Spain, at another Holland, seemed destined to be the great empire-state, the Great Power of extra-Kuropean Europe. Lastly, the 18th century witnessed the desperate struggle between France and England, which gave one-fifth of the land surface of the globe to the latter power. At the present time there are but six countries which have colonies of any consequence, viz. Great Britain, Holland, France, Spain, Portugal, and Denmark. The areas and populations of these countries and their several dependencies are given in a table below, but in these modern days of rapid development and “ national movements,” it would be very fallacious to take those figures as measures of the power of the several countries. The Vi THE GROWTH OF NEW NATIONS. UNITED STATES 33 areas must be qualified by considerations of climate ; to a North-European country a small possession in a temperate climate may be worth more than thousands of square miles of tropical forest; on the other hand, in a population, ‘considerations of race greatly outweigh mere numbers. Then again in the table the United States, that country of marvels, and the many States of South America, formerly colonies of Portugal or Spain, find no place. In recent years German patriots, distressed at witnessing the loss of population which their country suffered from the great tide of emigration from the Fatherland to the United States, conceived the idea of founding a colonial empire of their own. The idea would have been a good one earlier, but they were too late in the race, and the Germans have hitherto only succeeded in laying hands on part of New Guinea, a few islands in the Pacific, and large tracts of country on the west and east coasts of Africa. All these new German colonies are tropical in character, and hence little suited for settlement by a North-European people, a fact which the Germans themselves are beginning to see, though they are loth to admit it. BRITISH AND FOREIGN COLONIES.! AREA (square miles). POPULATION. Great Britain Asiatic possns. Holland France | Spain Portugal Denmark HAR 20,7571 12,648 204,092 192,959 34,499 18,784! Slop i Colonies, &e. ! 5,241,482] 11,738,161 202,171,781 26,841,597 8,799,857] 8,175,467 3,723,967 127.192 8,059, 17¢ _ 595,440 1,172,971] 536,798 7,672,048 358,693 6,350,874 40.277) 4,160,315: 100.0081 1.969.039! Total. "249,159,424 31,014,568 46,394,905 24,526,341 7,584,282 2,006,161 ! From Inaugural Address to the Statistical Society, 18 Nov. 1884, by Sir Rawson 'W, Rawson, K.C.M.G., C.B. STUDIES IN STATISTICS CHAP. [ propose now to consider the results of migration, not, as is more often done, from the European point of view, but from the other side, that of the countries to which the emigrants go. - By an analysis of the populations of new countries as shown by the census, taking into consideration the flow of migration still going on, I desire to sce what sort of nations are growing up beyond the seas, and what results are likely to follow from the continuation of the process of settlement. Broadly speaking, those who seek a fortune in a new sountry choose one of four regions of the world, viz.— 1. North America. The United States, or Canada. 2. South America. The Argentine Confederation, Uruguay, or Brazil. 8. South Africa. The Cape Colony, Natal, or the Transvaal. 4. Australasia. The various Australian colonies, Tasmania, or New Zealand. Amongst all the countries of immigration the United States of North America stands facile princeps. Emigration from Europe to “the plantations” of the New World has been going on for more than three senturies ; but it is in the last fifty years only that it has reached such dimensions as to constitute an historical fact of the first importance. The future of the European races is indeed intimately bound up with that of the United States of North America. Possessed of the largest habit. able area of all the countries of the New World, and of unequalled natural resources which are being developed by a people characterized by restless energy and keen intelligence, it exhibits a rapid growth of population and accumulation of wealth that are without a parallel. Fifty years ago its population numbered but 17 vi THE GROWTH OF NEW NATIONS. UNITED STATES 5 millions, at the last census 50 millions, at the present time over 62 millions! The energy of the people is sufficiently proved by the fact that in five years of war a debt of 540 millions sterling was accumulated ; the still more extraordinary fact that in twenty years of peace it paid off 200 millions of the capital of this debt, and reduced the interest on the remainder from 5 or 6 per cent., at which it was contracted, to the 3 to 41 per cent. at which it now stands, is ample proof of its wealth.? That the marvellous increase of the population of the United States, whereby the number of its people has been doubled every twenty-five years during the present century, admits of explanation, by no means detracts from its significance ; quite the reverse. Probably no other country has room for so many people. Canada and Australia have larger tracts of land unsettled, but the vast sub-arctic forests of the one and the riverless deserts of the other, so far as can be foreseen, will never be able to support any considerable population; and the greater part of Africa and much of South America are too hot for the northern races of Europe to thrive in. Yet, while the absolute increase of the United States people will be larger and larger during each successive decade, the relative rate of increase may be expected to diminish. In the early days of a new country when its popula-Sion 1s small, immigration is the chief factor in its growth ; but as the native population increases, immigration becomes gradually of less and less importance by comparison to the natural growth of the people. For instance, during the ten years 1851-60, it is recorded that 1,495,243 persons left ‘the shores of the United Kingdom for the United States, of which number it may be assumed 1 62,480,640, Census Bulletin, Washington, 30 October, 1890. 2 Statesman’s Year-Book. 36 STUDIES IN STATISTICS CHAP. that 1,224,615 were natives of England, Wales, Scotland, or Ireland.! Now, in proportion to the mean populations? at that time of the United Kingdom and United States respectively, ¢.e. 28,533,619 and 27,278,121, the loss to the exporting, and gain to the importing, country were 4'3 and 4'5 per cent. During the eleven years 1871-1881, there left our shores for the United States 1,263,476 persons of British or Irish origin, or almost exactly the same number as luring the ten years 1851-60; the proportions were, however, very different. United Kingdom ) Mean population ~ 33,543,430 Census 1871 and 1881 ) United States Mean population 44,357,077 Census 1870 and 1880 Proportion iit by emigrants to > 38 this, per cent. § Proportion i by immigrants to 2-8 this, per cent. y The figures show clearly how a migration may be absolutely constant, yet at the same time relatively both to the absorbing and the dispersing country decreasing. They also show that this vast exodus of men, women, and children was not inconsistent with great growth in the mother country, although it contributed in no small legree to that yet more rapid increase in the United States, whereby their population came up to and passed urs somewhere about the year 1856. In 1850 the Great American Republic, so far as mere aumbers are concerned, stood seventh among the Powers, n 1860 fourth, in 1870 third. and in 1880 second. ' Board of Trade Return of Emigration. Tn 1851-2 the nationalities of emigrants were not distinguished, so I lave assumed that the same proportion held good between foreigners and those stated to be of British or Irish birth as in the eight years 1853-60. 2 Means of two successive censuses. V1 THE GROWTH OF NEW NATIONS. UNITED STATES 57 ORDER OF PRECEDENCE OF THE POPULATION. 1850. Russia. France. Austro-Hungary. Germany. United Kingdom. [taly. United States. GREAT POWERS BY 1880. Russia. United States. termany. Austro-Hungary. France. United Kingdom, [taly. The political significance of this table is evident; it is intensified by the fact that Germany and Italy, two congeries of petty states in 1850, are now solid nations, while the Austro-Hungarian monarchy is much more consolidated than it was. At the next census the population of the United Kingdom will be very nearly equal to that of France. Our American cousins, so go-ahead in most respects, cannot claim as yet to equal us in vital statistics. Almost alone among the great powers (Russia and Spain excepted) America does not give the world her yearly record of marriages, births, and deaths. As regards some of the chief cities of the New World, such records have been kept for several years, so that it is to be hoped we shall before long have returns for the whole country ; a heavy task truly to carry out over such a vast territory, but not one that a young, vigorous, and proud people need fear to grapple with. Emigration from the Old World to the New has for the last forty years been so striking in its variety and extent, that we are too apt to look upon the growth of the great Republic of the West as wholly, or at least mainly, due thereto. But whatever may have been the case in the past, this is not so now. In default of other data we must assume some rate of natural increase (by excess of births over deaths) for the United States. The population of STUDIES IN STATISTICS CHAP, the States, it should be remembered, is principally composed of men of British or Irish stock, of Germans, and of negroes. As regards the United Kingdom, the average rate of natural increase for the ten years 1871-80 was 13 per thousand; that of the German Empire for the nine years 1872-80, 12} per thousand. These rates produced in the two countries mean annual increments of 430,000 and 540,000 respectively, Assuming similar rates for the United States, with its population in 1880 of 50 millions, we should have an annual natural increase of about 650,000. But let us suppose that the rate is notably less than this, say only 10 per thousand, we should still have an annual increase of population due to its own inherent vitality of half o million. The natural increase of Victoria is said to be 17 per thousand per annum, that of New South Wales 23, that of New Zealand 27 ; it is, however, rather difficult to believe in these excessive rates. The circumstances are indeed peculiar, but although the marriage- and birth-rates seem probable enough, it is difficult to believe that the death-rates in the following table are not understated, at any rate in the case of New Zealand— MARRIAGE. BIRTH-, AND DEATH-RATES PER 1000 LIVING. CENSUS YFAR, 1881.1 re we Victoria ... ... 136. . 3? New South Wales ... 16:5 ... 3890 New Zealand ... ... 133 ... 380 Noatha Natural increase, we 170 ee 229 ee. 9269 Of course if the rate of natural increase in the United States were to be admitted to be 20 per thousand, or even 15 per thousand, this would amount to an annual natural ! From Colonial Reports. Substantially the same figures are obtained when an average of the three vears 1880-2 is taken. VI THE GROWTH OF NEW NATIONS. UNITED STATES 359 increase of 1,000,000, or 750,000.. But young though it be when compared with other countries, it is not so young as our Australasian colonies, and we should therefore not 2xpect it to have such an abnormal rate of growth. The mortality statistics of American cities—so far as they are accessible in the returns of our Registrar-General —are as follows— New York Brooklyn... ... Philadelphia .. wey Baltimore “ a. Population, 1880. t,206,299 566,663 847,170 332.313 Average death-rate. 273 1874-86 234 » 20-7 ” 214 1879-86 The very imperfect returns given for Boston, Cincinnati, and Chicago would seem to indicate that the two former are healthy cities, the latter unhealthy. New York is an unhealthy city, and may be classed with Manchester and Liverpool. Philadelphia and Baltimore, on the other hand, are very healthy cities; they have a lower mortality than London, and may be compared with Brighton, Bristol, and Bradford. Brooklyn occupies an intermediate position ; its death-rate is higher than that of London, and it takes rank with Sunderland and Oldham. = So far then as these towns may be regarded as a fair sample of the American Republic, it would appear that the death-rate of the whole country is much about the same as that of England and Wales. Concerning the bLirth-rate I know of no information directly available. In the Annual Summary for 1887 published by the Registrar-General, the birth-rate of New York in that year is given as 23°0, and that of Baltimore as 20'7, but these low numbers seem incredible. There is, however, a means of approximating to the rate of natural increase of the United States by another quite independent method. %0 STUDIES IN STATISTICS CHAP, in the ten years 1870-80! the census showed an increase of 11,556,736 persons (excluding Indians). The United States immigration statistics assert that 2,812,191 immigrants landed in the country during the decade ;? the difference between these numbers would give a natural increase of 8,744,545. This on a mean population of 44,311,008 gives a natural increase of 19-7 per thousand per annum, a number that tallies fairly well with the figures for our leading Australasian colonies, But it will be asked, can the immigration statistics of the United States be trusted ? As regards the alleged enormous immigration of Canadians, it is well known that they are almost worthless. Moreover, the figures only profess to represent the gross immigration, no account being given of emigration. When a comparison is made with the statistics of the Ruropean countries from which the greater part of the migrants go we get these results— EMIGRATION TO UNITED STATES3 FIVE YEARS 1882-86, Taped States | British and German Excess of U. 8. rt gures. gures for same years, | & ic] United Kingdom ... 688,865 | 572,003 ¢ 204 Jermany -p ... 833,938 666.4215 251 There is then good reason to believe that the United States figures are excessive. Now it is obvious that if we assert that the United States immigration figures are too large, to that same extent the natural increase figures must be too small, and we are thus driven to the econslusion that after all the true rate of natural increase of the United States may be something like 20 per thousand ' Census of United States, 1880, vol. i. pp. 378, 379. ¢ Report of the Chief of the Bureau of Statistics of the United States, L889, p. exx. 3 Statesman’s Year-Book, 1890. ! Net emigration. Board of Trade Returns. 5 Almanuch de Gotha. 1890. v1 THE GROWTH OF NEW NATIONS. UNITED STATES 61 per annum, or an absolute addition to the present population of 1,000,000 a year by excess of births alone! Again, in the same 10 years the coloured population increased from 4,880,009 to 6,580,798, corresponding to a natural increase of 29-7 per thousand per annum, since the migration was probably inconsiderable. Yet such a rate scems scarcely credible, since there is reason to believe that intermarriages between whites and persons of colour are less frequent now than in the old days of slavery (race hatred having increased rather than diminished since the war), so that the half-caste population is not added to from the outside by fresh infusion of white blood to anything like the extent that was formerly the case. As regards the additions made annually to the population by immigration, the chief of the Bureau of Statistics states, that in the ten years 1880-89 no less than 5,248,568 immigrants landed on the shores of the United States!' Even if we deduct 20 per cent. from this number, we should still have a formidable army of 419,885 men, women, and children annually arriving in the ports of America with a view to settlement! That something like this number does settle there is certain, and is in truth the strongest possible proof of the vast resources of that great territory. In the intercensal interval 1870-80 the average annual crowth of population amounted to 1,150,000. From what has been said above it is pretty plain that it cannot now be much under 1,250,000, and may be more. The United States, which in 1790, at the first census, numbered under 4 millions of inhabitants, has now (1890) something like 521 millions, and is still growing at the rate of something like 8,400 per diem. It already exceeds in population every European country except Russia, and if there be; as 1 Report of the Chief of the Bureau of Statistics, 1889, p. exx. 52 STUDIES IN STATISTICS CHAP. I think there is, little doubt that it is mow the most powerful state in the world, before another quarter of a century is past its predominance will be beyond question, [ts native population is keen and enterprising, while the reinforcements that it receives from Europe are mainly sent by the most vigorous peoples of the Old World. [IMMIGRANTS INTO UNITED STATES! TEN YEARS 1880-89, Germans... ... .. 1,445,181 Swiss ... British ... ... ... 810,898 Dutch .. Irish... .. .. .. 674,061 French .. Swedes, N orwegians, Danes 671,783 Belgians Austrians, Hungarians 314,787 Chinese (talians ... ... ... 267,663 All others Russians, Poles, Finns 225.609 Total .. 81,151 52,715 18,189 18,738 65,806 571,987 ... 5,248568 Even if we admit that the census of the United States is not so trustworthy (or in American, reliable) as it might be, still the figures derived from it give a more solid foundation to go upon than the Immigration Returns, The population is therein classified both according to race and according to birth-places. The figures are so remarkable that I have given them in the following table (which is compiled from the census directly) in some detail, and in such a way as to show at a glance the two methods of classification. As regards the division into races, it will be noted that the proportions of Chinese and of civilized (or tax-paying) Indians is quite insignificant. To the latter should, however, be added the uncivilized Indians (who have not yet learned to appreciate the delights of paying taxes), amounting in 1880 to a little over a quarter of a million, It has not yet been ascertained with certainty whether the Indians are increasing or decreasing in numbers; with Y Report of the Chief of the Bureau of Statistics of the United States, 1889, p. exxi. Vi THE GROWTH OF NEW NATIONS. UNITED STATES 63 regard to the negroes, including Mulattoes, Quadroons, &c., there is unfortunately no doubt on this point, Since TABLE A.—CENSUS OF THE UNITED STATES, BIRTH-PLACES OF THE PEOPLE, AND POPULATION BY RACE.! Population by Race. Whites ... wie doe Men of colour ... wh Chinese ve —_ Civilized Indians Total ves vee Birth-places of White Population. United States ... ee Foreign-born 2 .., swe Details of Foreign-born. Born in Germany i ws Ireland oe i England and Wales ,» British N. America » Sweden & Norway s Scotland ver » Austro-Hungary france RE. Switzerland | Russia and Poland Mexico ed Denmark vie Holland os Italy ... a Belgium oe Spain & Portugal Other countries | and not specified y 1860. Tor cent. 26,922,638) 85-7 4,441,830] 141 34,933 ‘1 44.020 1 31,443,321] 1000 Per cent, of whole ropulation. L200) 1,101.242 72'6 12°1 1,276,075] 4°06 1,611,304] 5°13 477,455| 1°52 249,970 "80 62,620] ‘20 108,518 25,061 109,870 53,327 10,458 4 27,466 9,962 28,281 11,677 9,072 8,360 21.766! ‘35 08 '35 17 09 03 09 oa 1880. (3,402,970 6,580,793 105,613 66.407 ‘~155,783 36,843,291 6.559.679 ,066,742 854,571 745,978 717,157 376,066 170,136 135,5508 106,971 88,621 84,279 4 68,399 54,196 70,926 14,230 15,535 13,259 37.063 Per cent. 866 131 ‘2 100°0 Per cent, of whole »opulation. 73'5 12°1 392 370 1°49 1.43 75 32 27 ‘21 1° an 1; i: 3 os oy the war they have increased by some 2 millions, and a coloured population of over 63 millions remains as a fearful United States Census Reports, 1860 and 1880. 2 This number excludes Chinese and foreign-born slaves. 5 Two-thirds of these were Bohemians, t More than half of these were Poles. STUDIES IN STATISTICS CHAP, legacy from the greedy and unscrupulous traders of Liverpool and Bristol. Whatever the capacity of the African race for civilization may be, and it is to be hoped that it is considerable, there can be no doubt that the presence of these millions of “ men of colour” will for long enough be a constant source of political and social difficulties, the end of which it is hard to foresee. The “men of colour” at present .form .13 per cent. of the whole population; it is remarkable that the proportion of foreign-born is exactly the same.” But the 61 millions of foreign-born by no means represent the whole of the foreign element of the American nation. Another table In the census gives the parentage of the people, as follows— TABLE B.—CENSUS OF THE UNITED STATES, 1880. PARENTAGE OF THE PEOPLE.! 4 Percentages, Native-born, both parents native-born ... 35,190,787 70:33 Foreign-born, both parents native born 22,259 07 Having native fathers, foreign mothers TR 4 1-14 Having native mothers, foreign fathers . 27,664 2-67 Having both parents foreign-born wo 12,907,178 25-79 Total ... ... 50,051,315 100-00 From Table B it appears, that out of the 6,559,679 given in Table A as “ foreign-born,” 33,252 were so only in the sense that their birth-places were beyond the boundaries of the United States, both their parents having been natives. There remain then 6,526,427 foreign-born foreigners so to say. But these are of course included in the last item in Table B, so that we get 12,907,178 — 6,526,427 =- 6,380,751 who were born in the States, but both of whose parents were foreign-born. To these it will be fair to add onehalf of those who had one parent foreign-born, BREA TATA — 955.549 t Excluding 104,468 born in China. Census of United States, 1880. vi THE GROWTH OF NEW NATIONS. UNITED STATES 65 so that we get for the total foreign element of the population— Foreign-born of foreign parents ... wie - 6,526,437 Native-born of foreign parents ... vie ven 6,380,751 Half-of the native-born having one foreign parent 955,549 Total ves . es 13,862,727 Being 27-7 per cent. of the whole population, or 31'9 per cent. of the white population. We see then that nearly one-third of the whole white population is more or less foreign in sentiment and associations, Then it must be remembered that the average annual number of immigrants is stated to be 524,857! and even this number has often been exceeded, e. g. in 1882 it amounted to 788,992! It has been well said, “The United States of America are a great alembie, nto which the emigrant vessels of Europe are constantly pouring a vast quantity of unknown, doubtful, and even explosive matters ; the raw material of the American neople that is yet to be.”? It is the great aim of patriotic Americans—and by Americans I mean men who have been for at least two generations in the country, and whose patriotism s mot tempered by any lingering sympathy for nother fatherland beyond the sea—to nationalize completely, not only in name, but in fact, all this mongrel horde, to imbue alike German and Irishman, Scandinavian and Italian, with a thorough belief in and love for the “ Constitution of the United States.” The possibility of this task depends upon whether these new-comers will continue to hang together, nation by nation, preserving t Average of ten years 1880-89. Report of the Chief of the Bureau of Statistics of the United States, 1889, p. exxi. % Old World Questions and New World Answers, by Daniel Pidgeon. London : Kegan Paul & Co., 1884, p. v. STUDIES IN STATISTICS CHAP. each their language and customs, forming individualized communities within the State; or whether they will be absorbed into the mass of citizens of orthodox belief. Geologists describe the little-known process by which metals or other minerals separate themselves from the originally fluid mass of igneous rock into crystals, grains, threads, or nuggets as one of segregation. This suffiriently graphic term may be used to designate such hypothetical separation of one racial element from the mass of the nation we are discussing. It is pretty obvious that segregation would be both more probable and more signifiant in proportion as the number of those of the race in question increased. ‘Thus, if in the past 60 years no more than 40,000 Spanish and Portuguese have migrated ‘0 the United States,’ it is plain that even if the two races were to unite (an event not very probable in itself), they could achieve nothing whatever against the 50 millions by whom they are surrounded; in fact, such a small qumber must almost of necessity be soon absorbed and lost. But surely it is a very different matter with the frish or Germans, who are counted by millions; for amongst these large numbers, language and customs, sraditions and sentiments will be much more strongly retained, while at the same time, owing to the great facilities for marriages between persons of the same race, she purity of the stock will be likely to be kept up. 1. The Indians.—It is but fair to consider first the aboriginal race, the Red men so familiar in romance. There can be no question that they were once far more numerous than now, but it is far from equally certain that their numbers are still on the decline. For the moment taking Canada and the United States together, we get the following figures— 1 1821-—1881. Almanach de Gotha, 1884. 56 11 THE GROWTH OF NEW NATIONS. UNITED STATES 67 INDIAN POPULATION—UNITED STATES 1880, AND CANADA 1881. Tax-paying cee 66,407 In the Indian Sonlbery | 943 597 and other “agencies ’ Alaska cen ce vee 8,655 Canada ee -. A. 108,547 Total . . 427.136 . In the north-west territory of Canada, at the census of 1885, there were 4,848 half-breeds enumerated, nearly three-fourths of them French. There is no doubt that in the province of Quebec tribes of Indians have been completely absorbed into the French population, Indians of pure blood being almost unknown there; but the traces of Indian blood in the “ habitans” are often evident. A very large proportion of the population of Mexico consists of Spanish half-breeds, ‘though it is not possible to give numbers. The Red men, the original freeholders of the American soil, are then still numerous, but relatively insignificant. At the present time they are mostly to be found in the central and western portions of the country, scattered over the prairies, among the foot-hills of the Rocky Mountains, or in the sub-arctic forests of the far north. "Though they number 427,000, it should not be forgotten that this sum is made up of many nations, between which chronic warfare has been a tradition for generations past—nations that have no bond of union save suspicion, hatred, or fear of white settlers. The elements needful to. unite the scattered tribes do not seem to exist, and without union they must remain utterly powerless to resist the encroachments of the advancing waves of civilization—a civilization which, it must be confessed, has too often exhibited a barbaric exterior to the Indian. Nor is there any reason to suppose that the Indians of the north are likely to 58 STUDIES IN STATISTICS CHAP. ally themselves with the numerous Indians and still more numerous Spanish-Indian half-breeds of Mexico. The Red man, it is admitted, has many fine qualities, not the least of which is the difficulty of taming his proud nature. From what I have myself seen in Canada, I have hopes - that the wise policy of the Dominion Government will be rewarded. - Canadian pioneers are naturally impatient in waiting for results, but that the Blackfoot [ndians, until a very few years ago the fiercest and most dreaded tribe, have taken the first steps in civilization, [ am convinced by what I saw in 1884. But while speaking thus hopefully, I do not mean to imply that in the future the Red man will play more than a very subordinate part. He is completely out-numbered, and has to begin a new mode of life from the very beginning. The Indians of the east that early explorers found culiivating the soil, growing “Indian” corn, have disappeared. A war of extermination, the diseases of the white man, and above all his “fire-water,” have proved too much for them. A wiser and more generous policy may keep alive the hardier and more vigorous Indians of the prairies, but they have no political future. Numerous though they are, and troublesome as they are likely to se for many years to come, they can scarcely be regarded 2s a powerful body. Their ancient divisions into many nations, that lived in a state of chronic warfare onc with the other, taken in connection with the immense area ver which they are spread, make their prospects of political power in the future infinitesimal. Ab the same time, it must not be thought that the description of the Red man given by the average western settler is much nearer to the truth than that of Fenimore Cooper. 9. The Chinese.—From the aborigines we will turn to the newest comer. In 1880 there were 105,613 Chinese 1 THE GROWTH OF NEW NATIONS. UNITED STATES 69 in the United States, as against 85,000 in 1860. The great majority (92 per cent.) are to be found west of the Rocky Mountains, in the cities of California (especially San Francisco), and working in the mines. In Canada in 1881 there were but 4,383, most of them in British Columbia. It would appear, then, that in spite of all the sensational statements that have been made on the subject, the Chinese settlements on the Pacific coast of the American continent are not as yet, numerically at all events, very formidable. When the Chinaman cmigrates he leaves his women-kind at home; the natural consequence is that he does not stay long. He is, in fact, not a settler, but a mere bird of passage. His manner of life, and above all his thrift, may lead to very serious social and economic results, but until he breaks through custom, and, like the Irishman, takes his family with him on his wanderings, he can have no political future in America. Should the Chinese once really take to emigration, it is difficult to say what the results might be, not only to California and British Columbia. but to Australia and throughout the Pacific. The labour markets both in California and British Columbia have already been much disturbed by Chinese competition, and enactments have been passed both by Congress! and the Dominion Government to check their immigration. It may be noted here, that any possible future arrangements giving greater autonomy to the regions west of the Rocky Mountains would not result in improving the position and prospects of the Celestials, 8. The Negroes.—These, including all of mixed blood, aumber 61 millions. and. as has been said. thev have 1 Since this was written a treaty has been signed between China and the United States, by which the immigration of Chinese labourers info the United States is ahsolutelv forbidden for twenty vears 70 STUDIES IN STATISTICS CHAP. increased by upwards of 2 millions in 20 years. During the decade 1870-80 the coloured population increased by nearly 35 per cent., while the white population, in spite of the enormous immigration, increased by only 29 per cent. To some extent this increase is doubtless owing to intermarriages between whites and mulattoes, quadroons, octoroons, &ec., but it is asserted that such unions are far less frequent than they were before the war. As is well known, the distribution of the coloured population is fairly clearly defined. In New Jersey, Pennsylvania, West Virginia, Ohio, Indiana, Illinois, the Indian territory, and New Mexico, and all states and territories to the north or west of these, the coloured population is under 5 per cent. of the whole, and in Missouri ander 7 per cent. In all the states to the south of these the proportion of men of colour exceeds 16 per cent. In Virginia, Georgia, Florida, and Alabama the proportion exceeds 40 per cent. In Louisiana it amounts to 51, in Mississippi to 58, and in South Carolina to 61 per cent. The accompanying map (Plate 1X.) shows by gradation of colour these proportions.’ ‘The very rapid increase of the coloured population, and its concentration in certain states, constitute the gravest social and political danger to the future of the Republic; in more than one sense a black cloud may be said to hang over it. Experiments in self-government by negroes have been carried on in Liberia since 1822, and in Hayti since 1795. A republic in the former country, an empire, a monarchy, and finally a republic in the latter, have alike proved, I will not say failures, but rather disasters. A constant succession of revolutions has resulted not merely in anarchy L i : se on om ship this, and similar maps in this work, it should be bom nind that absolute numbers me Sly. are not represented, but proportions 3" £™y ws uo 105 1000 Mp SA = - = ~ue 60" ASga. Mont DAK. parr Wyo 2, NEBR ri CoLg. : i KANS CENSUS 1880-8! $ COLOURED POPULATION ! Proportion per cent. of the population of each PROVINCE, STATE X TERRITORY oF CANADA AND THC UNITED STATES 1 2 7 40, Scale of English Miles WOS0 0 100 200 300 ADO I pgm rt ——— WEEE “on™ el percent. IL i 5 per cent 10 per ce. London: Xdws marr cer a] London. Stanford G+ i 5p per cent. oTALT. ores 7 vi THE GROWTH OF NEW NATIONS. UNITED STATES 71 but in barbarism, a barbarism that is not figurative but real, in which fetishism has spread, and even cannibalism has been resorted to. That the blacks have now a large share of political power in some of the states is evident; that they will have yet more in the future is to be expected ; it only remains to hope that the strong hand of the Federal Government, and the influence of example, will keep those states from the fate of Liberia and Havti. Before proceeding to discuss the part now taken and in the future to be taken by some of the European races in the United States, it will be as well to give in somewhat more detail the teachings of the census as to the parentage of the people. TABLE C.—UNITED STATES CENSUS, 1880. FOREIGN PARENTAGE! (INCLUDING CHINESE). Parent Native of 1. Father. II. | Mother III. | Percentage of whole Moan. Population. Ireland —_ vhs Germany — ves (ireat Britain... - Scandinavia ... vo British America - Other Foreign Countries 4,529,528! 4,448,421 4,488,972 4,883,842 14,557,629 72) ,735 2,039,808| 1,790,200] 1,915,004 635,405 63.309 633,357 939,247 931,408 935,328 1.891 485! * 99g. 118" 1.9%2 vor ~N 4 8 (3 L Census of United States, 1880, vol. i. p. 674. The figures in this table are partly estimated in the following manner— The details of parentage were “ tallied” for 28 states, 7 territories, and the district of Columbia, which together comprise a population of 26,354,124, including 2,673,217 foreign-born, made up as follows— Per cent. of all those ‘n United States. 370 336 EN 6 75 1 Total 14.349.310113.585.0801 12.967, 19 Natives of Ireland ve Germany wo Great Britain | Scandinavia wet British N, America Other Countries ... Total ... D6 0.17 00 72 STUDIES IN STATISTICS CHAP. This table gives for the predominant nationalities the numbers of those who had foreign fathers and mothers. I have added a column giving the mean of these two, as the best available measure of the strength of each national element ; also the proportion per cent. of these mean numbers to the whole population. The above figures may advantageously be combined with those relating to birth-places (see Table D), so as to show for each nationality the foreign-born, the native-born of foreign parents, and the total, this last being the number in column III. of Table C. Firstly, a few general remarks as to Tables C and D. It will be seen from Table B that there are more natives of the United States having a foreign father and a native mother, than with a foreign mother and a native father, nearly two and a half times as many. This is easily The following results were obtained—-Persons with foreign fathers ves a 5,529,331 ” mothers vee 5,223,497 both parents foreign ai : 4,994,017 ’s one or both parents foreign ... - - 5,758,811 Foreign-born, both parents native-born ... ve 13.307 A further analysis showed that— To 1000 born in Ireland there were with Irish fathers, 2442; with Irish mothers, 2387. To 1000 born in Germany there were with German fathers, 2483; with German mothers, 2306. To 1000 born in Great Dritain there were with British fathers, 2223; with British mothers, 1941. To 1000 born in Scandinavia there were with Scandinavian fathers, 1690; with Scandinavian mothers, 1671. To 1000 born in British North America there were with British North American fathers, 1310; with British North American mothers, 1292. Te 1000 born in other foreign countries there were with other country fathers, 1559 ; other country mothers, 1439. It was then assumed by the census authorities that the same ratios would hold in the states for which the figures were not “ tallied.” As the states “tallied ” seem to be fairly representative states, the assumption was perhaps a fair one, and the figures in tables C and D may be accepted as approximations. vi THE GROWTH OF NEW NATIONS. UNITED STATES 73 explained. In nearly all returns of emigration it is found that many more men than women leave their native country. In the case of the Chinese, as stated above, scarcely any women emigrate. French and Spanish women are very unwilling to leave their homes; in the case of the Irish, on the contrary, there seems to be no such repugnance. The excess of foreign men will naturally seek wives in a large number of cases among the women of the country. TABLE I.—FOREIGNERS IN THE UNITED STATES! (INCLUDING CHINESE). i of | N: a of foreign parents. foreign parents 2,634,401 2,753,993 998,890 193,095 218,171 626.4922 Ireland ... 1,854,571 Germany ... . 1,966,742 Great Britain wis 916,114 Scandinavia 4 440,262 British America ... 717,157 Other Foreign Countries | 647.377 Total ... 6.542.923 7.4%4 972 Total. Ratio of IL.:1. 1-42 1-40 1-09 44 30 97 13,967,195! 1°13 From Table D we learn, that in the case of the Irish, German, English, and Scotch races the children of settlers born in America outnumber the surviving immigrants, the excess amounting to 40 per cent. in the case of the Irish and Germans. With respect to the Scandinavians, Canadians, and other nationalities, quite the reverse is the case—the immigrants outnumber the second generation of settlers. The children of Canadian parents amount to only 30 per cent. of the number of natives of Canada, and the children of Scandinavian parents to 44 per cent. of the natives of Norway, Sweden, and Denmark. In these latter cases two explanations are possible, —(1) the immigrants may not be true settlers, but rather, t The figures in this table aro partly estimated. See note to Table C. 7 STUDIES IN STATISTICS CHAP. like the Chinese, birds of passage attracted to the country by high wages, but returning home when they have “ made their pile,” as American slang so pithily expresses it ; or, (2) the migration may have been so recent that enough time has not elapsed for many children to be born, while for the same reason comparatively few of the original migrants have as yet had time to die. As regards the Scandinavians, the second explanation is probably the true one, as regards the Canadians the first. A large proportion of the natives of the Dominion living in the United States are French Canadians from the province of Quebec. They are attracted by the high wages offered by the New England factories, but their catholic and conservative feelings are repelled by the school regulations and democratic institutions of the Republic. They find their religious scruples are more respected under monarchical than under republican institutions, and so they gladly return with their earnings to their native country. The French Canadian, it should be remembered, is a staunch Catholic, and views with suspicion and distrust alike the agnostic and socialistic opinions so common in France! This explains the fact that the emigration from France to Quebec is so very small. 4. The French.—This is a natural place to allude to the French section of the American people. Although the whole of Canada and a large portion of the United States once belonged to France, and the achievements of some of the early French explorers, notably La Salle and Verandrye, fill some of the most glorious pages in the history of geographical discovery, it is remarkable what a small share Frenchmen are now taking in the 1 See Old World Questions and New World Answers, by TDanlel Pidgeon, 1884, pp. 173—182. vi THE GROWTH OF NEW NATIONS. UNITED STATES. 75 development of the country. In Canada it is true the 11 millions of French Canadians hold their own as regards numbers, and include some eminent politicians, but it must be acknowledged that they are sadly lacking in that spirit of enterprise which marked the lives of their ancestors in the 17th and 18th centuries. In the United States there were in 1880 only 107,000 natives of France, or 02 per cent. of the whole population ; in 1860 there were 110,000. The annual number of French immigrants recorded is under 5,000. In Canada in 1881 there were but 4,389 natives of France enumerated. These small numbers are partly accounted for by the slow rate of increase of the population of France (which in its turn is due to the low birth-rate), partly by the fact that half the small French emigration is directed to South America. As remarked above, an unknown but certainly very large proportion of the natives of Canada in the United States are of French origin, and of these the majority are comparatively new-comers, who have moreover little intention of permanently settling. Unfortunately the census makes no distinction between Canadians of French and British origin. The distribution of “natives of British North America,” as they are termed in the census, gives us, however, some clues. The cities most favoured by the Canadians are Boston, Chicago, Detroit, Lowell, Fall River, and New York. It is known that within recent years very large numhers of French Canadians have been attracted to the mills of New England, and on geographical grounds alone we might safely conjecture that emigrants from the province of Quebec would largely predominate in New England, while in Michigan and the North-Western States we should expect to find more natives of Ontario. This opinion is confirmed by the very interesting analysis given 76 STUDIES IN STATISTICS CHAP. in the Census Report of the parentage of natives of the United States enumerated in certain states of the Union. NATIVES OF BRITISH NORTH AMERICA ENUMERATED IN CERTAIN STATES AND TERRITORIES OF THE UNION? Per cent. of population of State. Maine us Vermont ... New Hampshire Massachusetts Connecticut Rhode Island : Total in New England Michigan ... New York Illinois ... Minnesota Wisconsin Towa ne California ... Kansas ... Dakota ... ius Total in 9 other States Remaining States 37,114 24,620 27,142 119,302 16,444 18.306 148,866 84,182 34,043 29,631 28,965 21,097 18,889 12,536 10.678 242.9928 388,887 85.349 37 7'4 78 67 26 66 [0 Jr ry I'n 38 22 I'3 2'2 13 79 ce 2'5 .. 0'3 In the four New England States, Connecticut, Massachusetts, New Hampshire, and Rhode Island, it is found that for every 10,000 natives of the United States having fathers natives of British North America, as many as 6,966 have mothers also natives of British North America; 2,180 have mothers natives of the United States; 661 mothers natives of Ireland, and only 193 mothers natives of any other country. When, however, we investigate the parentage of the natives of the United States having British North American fathers in Minnesota, Wisconsin, and California. we get a very different result. L Census of United States, 1880. vi THE GROWTH OF NEW NATIONS. UNITED STATES 77 NUMBER OF NATIVES OF THE UNITED STATES HAVING MOTHERS BORN IN CERTAIN COUNTRIES! TO EACH 10,000 HAVING FATHERS BORN IN BRITISH NORTH AMERICA. Mother a native of British North Ame-’ United States Ireland ... Great Britain Scandinavia Germany ... Other Countries Total Eo ‘The exclusiveness of the Canadians in the New England States as compared with the others strongly confirms the other evidence that they are mainly of French extraction, since ¢& prior one would expect men not only of a different race but professing the Roman Catholic religion to be averse to mixed marriages. Unfortunately the figures for Michigan are not available, as that state is not among the “ tallied.” As there must be many French Canadians outside New England, it will, I think, be safe to say that the French Canadians in the States number not less than a quarter of a million. If this be the case, taking the French in Canada and the 107,000 natives of France in the States, together with the undoubtedly much more numerous Americans of French extraction (especially in Louisiana), we may safely put the French element in North America at something like 2,000,000. The French in Louisiana are separated by too vast a distance from their cousins in Quebec and New England for any union between them to be possible. The natives of France are most numerous in New York, Obio, Louisiana. California. Illinois, and Pennsylvania. The . Census of United States, 1880, vol. i. pp. 680—692. CHAP. citics containing most French are New York, New Orleans, and San Francisco. 5. The Dutch.—Thesc have no doubt impressed some of their national characteristics upon the American people, and the way in which Dutch names constantly rise to prominence in America shows the important influence still exercised by the early settlers of New Amsterdam. The migration from Holland to the States is, however, small, though increasing! In 1860 there were but 28,000 natives of Holland enumerated ; twenty years later 58,000. In Canada in 1881 there were stated to be 30,000 people of Dutch “origin.” In any political catastrophe it might be fairly assumed that the Dutch would side rather with the Germans or English than with the Irish or French, but in any case their action could have but a trifling effect on the issue. The Dutch-born are most numerous in the states of Michigan, New York, Wisconsin, Illinois, Iowa, and New Jersey. The city of Patterson (New Jersey) contains a notable number of Dutch inhabitants; there are also many Dutch in Chicago and New York. 6. The Spaniards.—History reminds us of the once great nation that played so conspicuous a part in the discovery and early “development of the New World. The early Spanish settlers in Florida are probably completely lost in the mass of the population, and in 1880 the natives of Spain in the United States were very few; in Canada, as might be expected, they are still fewer. The natives of Mexico are far more numerous than the natives of Spain; they are found in the States and territories along the frontier, and are probably for the STUDIES IN STATISTICS 1 The number of Dutch emigrants in the five years 1882-86 averaged 4,011 per annum, “most of whom went to North America.” —Statesman’s Year-Book, 1888. The American reports claim 53,000 Dutch immigrants in the ten vears ending June 30th, 1889, vi THE GROWTH OF NEW NATIONS. UNITED STATES 79 most part of Spanish extraction, but with much admixture of Indian blood. DISTRIBUTION OF SPANIARDS AND THE UNITED STATES. MEXICANS IN Svante” Mexicans. New York Louisiana - California Illinois ... Texas ... . Massachusetts ... 1 Rest of the Union ver 1,852 Total ... 5,121 1.216 937 Ca 3 oN Texas ... Arizona . California .. New Mexico ... Louisiana va 43,161 9,330 8,048 5,173 494 Rest of the Union... 1,663 68.399 In Mexico, where the Spanish race is still predominant, the blue blood of Castile is sadly diluted by Indian admixture, and it is much more likely that the Anglo-Saxon American will absorb Mexico, than that the Spaniards will extend their sphere of influence. 7, The Portuguese.—Of 8,138 natives of Portugal in 1880, as many as 4,705 were found in California. There were also 1,161 Portuguese in Massachusetts. The Portuguese seem to be still as fond of making long voyages as in the days of Vasco di Gama. In 1884, among the inhabitants of the Sandwich Islands were 9,000 Portuguese, many of them natives of the Azores and of Madeira.’ 8. The Scandinavians.— The migration from Sweden, Norway, and Denmark has greatly increased of recent years, and is very large in proportion to the population of those countries; it is mainly directed to the United States. The tables of birth-places show the increasing importance of the Scandinavian element in the American population.® L Census of United States, 1880, vol, i. p. 495. 2 Almanach de Gotha. i 3 The movement of Scandinavians into Canada has only just com- 30 STUDIES IN STATISTICS CHAP. UNITED STATES CENSUS. BIRTH-PLACES. Natives of 1860. 1880. Sweden ai ce 18,625 97.332 194,337 Norway ... 43,995 1. ,240 181,729 Denmark ... es 9,962 . 30,107 64,196 Total Scandinavia 72,682 ... 241,685 a 440,262 Since 1880 the number of emigrants reported. to have left Sweden, Norway, and Denmark has greatly increased, as has also the number of immigrants reported to have landed in American ports. From Table D we learn, that in addition to the 440,262 natives of Scandinavia in the United States, there were computed to be also 193,095 persons born in America of Scandinavian parents, making a total of 633,357. The importance of these Norsemen is greatly increased by their somewhat singular geographical distribution. This is shown in the next table; also by the accompanying map (Plate X.), in which the gradations of blue colour indicate the proportion of natives of Scandinavia to the total population of each province, state, and territory in 1880-81. In the areas left white the proportion was under one per cent. menced, and is abt present of small proportions, as shown by the census. Census of Canada. Birth-places. 1861. 1871. 1881. Sweden. \ Norway. oe 515 I. 588 - 2,076 Denmark. ( The Chief of the Bureau of Statistics states that in the ten years ending June 30th, 1889, no fewer than 671,783 Swedes, Norwegians, and Danes landed in the United States, so that we may expect to find that the eleventh census will show more than double as many Scandinavians as the tenth. qr 907 108 J CC ry® 51 od 51. Na LE Ra, CENSUS 1880-8i NATIVES OF SCANDINAVIA Proportion per cent. of the population of each ~A3€ PROVINCE, STATE & TERRITORY oF CANADA UNITED STATES 25 Scale of English Miles 100 50 0 no 200 300 400 RT Pr I RET C ng A&W. Gr nar cent. 2.5 ner cent 5-0 per cent. 7-5 per cent. London :- Edward Stanford I London, Stanford Geogl Estab: 10 per cent. onaon, oras Geog” Lis Vi THE GROWTH OF NEW NATIONS. UNITED STATES 81 DISTRIBUTION OF NATIVES OF SCANDINAVIA IN CERTAIN STATES AND TERRITORIES. CENSUS. 1880. Sweden. Norway, Denmark. Total. Minnesota Wisconsin lllinois ... Towa ce Dakota ... Nebraska New York I. Michigan “en Kansas ,.. wi Utah cee ee California ... Total of these eleven States and Territories: 160,371 175,723 | 53,791 380,885 Rest of the Union 33,966 6,006 10,405 50,377 Total ... | 194,337 | 181,729 | 64,196 | 440,262 107,768 66,284 65,414 46,046 17,869 16,685 16,494 16,445 14,403 £2,755 0,722 , Per cent, of whole vopulation. 13-8 50 21 2:8 13-2 37 03 1-0 14 89 1:1 24 01 0-9 This remarkable concentration of Scandinavians within such a comparatively small area is one of the most striking facts in the distribution of nationalities in America. Nearly one-fourth of the whole number were found in the state of Minnesota where they amounted to nearly 14 per cent. of the whole population. It will be observed, that the Norwegians are even more concentrated than the Swedes and Danes, the two states of Minnesota and Wisconsin containing considerably more than half of them. The Swedes are absolutely most numerous in Illinois, but that is a populous state, and the proportion is larger in Minnesota. For the Danes, Utah seems to have great attractions. The cities containing most Scandinavians are Chicago and Minneapolis. The Scandinavians in many cases form compact settlements, their language being apparently a great bar to fusion with the Americans ; thus the statistics of parentage show that 32 STUDIES IN STATISTICS CHAP. up to 1880 there had been remarkably little intermarriage between natives of Scandinavia and natives of the United States. NUMBER OF PERSONS BORN IN THE UNITED STATES HAVING MOTHERS BORN IN CERTAIN COUNTRIES, TO EACH 10,000 PERSONS HAVING FATHERS BORN IN SCANDINAVIA, ENUMERATED IN THE STATES OF MINNESOTA AND WISCONSIN. Mothers born in Scandinavia . » 5 United States . 5» British North Amer’ .. Great Britain ,, Germany ,, Ireland ... ,, Other Countries Total 33 [0.CO0 Where the proportion of Scandinavians to the total population of the state is less than in Minnesota and Wisconsin, e.g. in California, Maryland, and the city of New York, intermarriages are relatively somewhat more frequent, wives being chosen by preference from natives of the United States, Ireland, and Germany. Under these circumstances, and bearing in mind the great increase of Scandinavian immigration in recent years, it seems likely that the Norse element will play an important part in the future history of the Republic. The Norsemen have many good qualities, and race affinity should make their assimilation by the Anglo-Saxon majority an easy matter. Language is a very potent agent in preventing the fusion of races, both directly by limiting intercourse and indirectly by preserving national sentiments and traditions. Much, therefore, depends on the ability of the masters of the elementary schools to make English the language of the rising generation. 9. The Germans.—Far more important than that L Census of United States, 1880, vol. i. pp. 680—692, So 105 100 ay ® un MONT Wyo. 35 NERR CoLo. . © : KANS. N.ME.. CENSUS i880-8i NATIVES OF GERMANY Proportion per cent. of the population of each PROVINCE, STATE & TERRITORY OF CANADA AND THC UNITED STATES 3r aq 25 Scale of English Miles 0050 © 100 200 300 400 _r—t—t———taowoay Long HH ; . | per cent. 2.5 ner cent. 5-0 per cent. _ 7-5 per cent. London: Edward Stanford 1} 0° London, Stanfords Geogl Esta 1 10 per cent. VI THE GROWTH OF NEW NATIONS. UNITED STATES 83 played by any of the nationalities that we have hitherto considered is the part of the great Teutonic people in the New World. In 1860 the natives of Germany in the United States numbered 1} millions, while in 1880 they were but very little short of 2 millions, or a number just equal to the population of the kingdom of Wiirtemberg. Their importance, moreover, is increased by the fact that a large proportion of them are concentrated in a few of the states, although their segregation is not so marked as that of the Scandinavians. (See Plate XL) DISTRIBUTION OF NATIVES OF GERMANY IN THE UNITED STATES, 1880: Per cent. of Population of the State New York 355,913 Illinois ... 235,786 Ohio... 192,597 Wisconsin 184,328 Pennsylvania 168,426 Missouri 106,800 Michigan 89,085 Iowa ... 88,268 [ndiana ... 80,756 Minnesota 66,592 New Jersey 64,935 Maryland 45,481 California ven 42,532 Texas ... i. 35,347 Nebraska ‘oe 31,125 Kansas ... 3 0 nee 28,034 Total of these 16 States ... 1,816,005 Rest of the Union vo 150,737 Total ... ver 1.966.742 70 77 60 1'0 ”'9 J 4 I 8g 57 a9 4'9 2'2 6-9 28 3 08 w It will be seen, that while the Germans are relatively most numerous in Wisconsin and Minnesota, and so far resemble in their distribution the Scandinavians, they are absolutely most numerous in the state of New York. The peculiarities in their distribution are better shown thus : 1 Census of United States, 1880, vol. i. 34 STUDIES IN STATISTICS CHAP. DISTRIBUTION OF GERMANS IN TWO GROUPS OF STATES, 1880. Four Eastern States. Number. New York, Pennsylvania, New orsog, and Maryland ! 634,755 Eight Central States. Ohio, Michigan, Wisconsin, \ Minnesota, Towa; Missouri, \ 1,044,212 Olinois, and Indiana ~~ | Total sxe ... 1,678,967 or 85-4 per cent. of all the natives of Germany in the United States. Per cent. of Population, 56 5G The political significance of the fact, that some six per cent. of the population of these twelve important states consists of natives of Germany is increased by the fact, that the number of persons having one or both parents natives of Germany is still greater. To the 2 millions of natives of Germany enumerated in the United States should be added 2% millions of natives of the United States having their parents natives of Germany.! The precise facts as to foreign parentage are available for certain of the states and territories only. In the next table I have given the numbers relating to German parentage in 12 states and one territory comprising a total population of 13,856,493 persons. Out of these 1,348,733 had both parents German, and 226,051 either father or mother German, making a grand total of 1,574,784 persons having one or both parents German. The census also shows us, that'in these same states and territories where one parent only was German, in two cases out of three the other parent was a native of the United States; and since the tide of German immigration has been flowing in a considerable volume for nearly 40 years, it 1s reasonable 1 Census of United States, 1880, vol. i. p. 674. (It is remarkable that the Germans born in Canada amount to only 25,000. Those of German origin are said to amount to 250.000.) PERSONS OF GERMAN PARENTAGE IN CERTAIN OF THE UNITED STATES. Both Parents German. Father only. ! Mother only, Born in Foreign United Wy States . Born in United Rtates Foreign Born. Born in Jaited States. Foreign Born. Total one or oth parents German. Wisconsin Missouri Minnesota Maryland California Texas ... oe. Kentucky ... Nebraska ce. Louisiana a. ve Massachusetts Som Connecticut ... a West Virginia you Dakota -— 226,325 144,629 78,255 59,167 33,393 40,650 45,126 27,849 23,066 14,432 14,043 10,176 5.956 191,524 109,311 71,009 45,947 44,083 36,706 30,669 32,812 17,512 16,967 15,801 7,089 6,936 28,297 38,580 12,625 15,334 12,702 10,190 11,837 6,979 7,697 5,553 3,868 4,389 1101 492 424 413 127 581 248 86 179 100 337 68 36 aalr 13,425 12,117 5,817. 5,033 5,540 4,069 3,749 2,456 1,355 1,681 1,298 1,175 549 438 258 300 83 387 118 75 163 127 167 61 22 118 460,501 305,319 168,419 125,691 96,686 91,981 91,542 70,438 52,857 39,137 35,139 22,887 14,18% ry | ns Total _ 722,367 | 626,366 — —— seers 1.348.733 I 61,264 | 2,317 226,051 Census of United States. 1880, vol. i. pp. 680—699. [59,152 | 3,318 1,574,784 Per cent. of Population of State. 350 141 216 13-4 11-2 58 56 156 56 2:2 56 37 10-5 [1'4 = wn 2f | 3 nS x 2 Rg 1 3 3 \ > @ 2 o-= © ® or - Zw SB =~ a © 5 n tA oO 5 2 & Co 2 i — 3 26 > 2 PE to r=27 5 3 = — ~ “4 3 nN N bi wy pi = 2 = = rt 0 ry 36 STUDIES IN STATISTICS CHAP. As might have been expected, it will be noted that a very large majority of the persons of German origin sorn in the United States had both parents natives of Germany. That there are more persons having a German father than having a German mother is of course explained by the fact, that among the immigrants there is always a more or less considerable excess of males. In the two states of Wisconsin and Minnesota, where Germans are relatively as well as absolutely very numerous, there were in 1880 no less than 845,502 persons born in the United States having a German father, and 323,822 persons having a German mother. It is interesting to observe the extent of intermarriage indicated, and the degree of attraction between different nationalities exhibited. With Wisconsin and Minnesota may be compared California, Maryland, and Missouri, in which states the Germans, though numerous, are not so notably predominant, GERMAN PARENTAGE IN CERTAIN STATES, 1880.1 Natives of the United States having German fathers Mothers being German ., not German Natives of the United States having German mothers Fathers being German ., not German Wisconsin. Minnesots. Per cent. Palifornia. Maryland. Missouri. Per cent. 304,580 40.999 88 2 (1'8 781 21°Q 303.805 | Toon 345.509 Too'o | 304,580 | ogg | 937,189 | 91.2 19242 go | 22.600 | 88 393,829 | 1000 | 259,879 | 100% As might have been expected, it is seen that the amount of intermarriage with persons of other nationalities differs ‘n different states,” also such intermarriages are naturally U Census of United States, 1880, vol. i. pp. 680—692, ) 2 Those given in the table have been selected on account of these lifferences being great. VI THE GROWTH OF NEW NATIONS. UNITED STATES 387 more numerous in those states in which there are fewer Germans, and in which consequently the Germans will be more brought into contact with persons of other races. The difference is much more noticeable in the case of German men chosing foreign wives. It is evident that in Wisconsin and Minnesota the total amount of intermarriage is very small. The intermarriages are analyzed further in the following table. PERSONS OF MIXED GERMAN PARENTAGE IN CERTAIN STATES, 1880.1 Wisconsin, | Minnesota Ter cent. ~alifornia Taryland. Wissouri Per cent. Natives of the United States having German fathers. Mothers being natives of United States .,. Ireland ... » Great Britain ... British N. America Scandinavia ... Other Countries Total 27,186 2,486 1,294 859 748 8.349 40,9292 664 61 32 2°1 i'8 20°4 1000 | 53,485 4,153 1,957 404 152 6.465 606.616 | &og 62 29 6 2 97 IO00'0 Natives of the United States having German mothers. Fathers being natives of United States ... Ireland ... wet Great Britain ...-British N. America Scandinavia ... Other Countries Total 3,297 1,072 ,578 869 569 5.869 19.242 43°1 56 82 4'5 0 3586 tr 000 12,257 1,389 1,782 241 325 5.746 541 61 76 I'1 i'4 207 22.690 | 1000 It is evident that of the intermarriages that do take place, the majority are with natives of the United States ; that is to say, no doubt in a considerable number of cases with persons also of German extraction. It is remarkable that intermarriage with Scandinavians even in the two states in which these are most numerous is very rare. LU Census of United States, 1€80, vol. 1, pp. 680—692, 38 STUDIES IN STATISTICS CHAP. From general considerations of race affinity the reverse might have been expected. It will be noted, that natives of Great Britain find more favour relatively to men of other countries as husbands of German women, than as wives of German men ; the same is also true of natives of British North America, Scandinavia, and * other countries”; in fact, the husbands of German women when not German are a somewhat more miscellaneous lot than the corresponding wives of German men. Bearing in mind what has been said above, this is probably explained by saying that German women arc less indisposed to mixed marriages than German men. As there are fewer natives of Scandinavia and of British North America in California, Maryland, and Missouri than in Wisconsin and Minnesota, it is not surprising that intermarriages with persons of those nationalities are less common in the same states; intermarriages with natives of ‘other countries” are also notably less, these deficiencies being made up by a great excess of marriages with natives of the United States. Where the Germans are in a smaller minority, naturally intermarriages will be relatively more numerous. The American cities in which Germans are most numerous are New York, Chicago, Philadelphia, Brooklyn, Saint Louis, and Cincinnati! In New York City the Germans are almost as numerous as the Irish, and make ap considerably more than one-fourth of the population. NEW YORK CITY, 1880.2 Persons with both parents German ... oe 331,589 sy fathers only yy eee ces 25,146 ,»» mothers sy eee oe ee _10,557 Total ... a ce 367,292 or 304 per cent. of the population. - Census of United States, 1880, vol. i. pp. 538—541. ? Ibid. pp. 675, 676. vii THE GROWTH OF NEW NATIONS. UNITED STATES 89 Difference of language is probably the greatest bar to the amalgamation of races, and is doubtless one of the chief reasons why intermarriages between Germans and natives of other countries are, as we have seen, relatively so few. The aggregation of large numbers of German-speaking people together, whether in agricultural settlements or in large towns, tends naturally to preserve their national language ; the numerous German clubs and German newspapers add to and strengthen the other influences, while certainly the yearly addition of some 140,000 immigrants fresh from the Fatherland must still further increase the difficulties of the American schoolmaster. Whether in the future any poltitical circumstances will ever bind together these millions of Germans so as to make them into a separate nation within a nation it is impossible to say; but it is well to keep in view, that as to-day there is an Irish vote to be reckoned with, so some time there may be also a German vote. Even now one hears that the temperance party in the United States finds the beer-loving Germans the staunchest opponents of “prohibition.” 10. The Irish.—In many ways a marked contrast to the Germans are the Irish. Though in but few cases separated from the English Americans by language, they are nevertheless clearly marked out by peculiarities of race, by ‘sentiments, habits, and for the most part by religion. Unlike the Germans, they already exercise a powerful political influence, even greater perhaps than is due to their large numbers. The Irishman, like the Spaniard, is a born politician, and this is to both alike a misfortune. The figures of the census afford a good deal of interesting information illustrating the “Irish Question” as it affects America. STUDIES IN STATISTICS CHAP. DISTRIBUTION OF NATIVES OF IRELAND IN THE UNITED STATES, 1880.1 Number i [Per cent. of total SUMBETn 1 Population of State. State New York Pennsylvania Massachusetts [llinois ... New Jersey Ohio ... Connecticut California Missouri ... [owa ne Michigan Wisconsin Rhode Island Minnesota .. [ndiana ... ve Marvland - . Total of these 16 States Other Statés and Territories otal 199,445 236,505 226,700 117,343 93,079 78,927 70,638 62,962 18,898 £4,061 £3,413 £1,907 35,281 25,942 25,741 21.865 672,707 181.864 I 854 5% 5'5 (27 38 82 2°'8 12 7" ey 32 28 33 "3 5 34 2'Q 37 The first point that strikes one in glancing over these figures is the remarkable difference between the distribution of the Irish and that of the Germans. The Irish are far more numerous in Massachusetts, Rhode Island, and Connecticut ; far less numerous in Wisconsin, Minnesota, Illinois, and Michigan. There are very many both of Irish and Germans in New York, Pennsylvania, Illinois, and Ohio, but the Irish largely preponderate in the two first, the Germans. in the two latter states. In Plate XII. the percentage of natives of Ireland in cach province, state, and territory is indicated by the depth of green colour. It will perhaps contrast better the distribution of the LU Census of United States, 1880, vol. 1. pp. 492—495, Towra Ws 100 105 1000 ev 90° o% tv I rT T7771 - "r £5 0’ pn ~ I, o - A rg DAK NER®™ CoLo, N.MEX £ KANS. 3 CENSUS 1880-8! NATIVES OF IRELAND Proportion per cent. of the population of each PROVINCE, STATE & TERRITORY oF CANADA IINITED STATES 1 Tel 25] Scale of English Miles 00 50 O 00 200 300 400 050 0 100 200 300 40 oer ct 2.5 per --20 per cent. {.omdon: Edward Staptornd 7 pes cent 0 : Lotto, Startomilly Bore" 0 ner cent Vi THE GROWTH OF NEW NATIONS. UNITED STATES 91 Germans and Irish to compare two groups of states, each comprising a little over 13 millions of inhabitants— the one being mainly agricultural, the other largely manufacturing, NUMBERS OF NATIVES OF GERMANY AND OF IRELAND IN TWO GROUPS OF STATES. CENSUS 1880. Ohio, Michigan, Wisconsin, Minnesota, Iowa, Minis, . and Indiana! Per cent. of whole population ... New York, Pennsylvania, ma, Se ea [sland.? Per cent. of whole population ... 47 en 8-8 Total vee es wv 1,561,151 1,538,982 79 9 of total Germans. 83 9 of total Irish, From these figures it is evident that the Irishman, in spite of all his land-hunger, more often seeks a new home in the manufacturing centres of America than on its prairies. Can it be that the “ horrors of the mid-passage ” destroy his appetite ? or is it rather that this much-talkedof land-hunger loses its political value in his new home ? The cities which find most favour in his eyes are New York, Philadelphia, Brooklyn, Boston, and Chicago. In New York City there were in 1880 no less than 388,920 persons both of whose parents were Irish, in addition to 23,048 persons who had an Irish father, and 23,469 persons with an Irish mother, so that considerably more than one-third of the whole population of the city had either one or both parents Irish; and nearly 200,000 of these, or more than one-sixth of the whole population, L Aggregate population, 13,612,056. 2 Aggregate population, 13,179,194. 3 Census of United States, 1880, vol. 1. pp. 538—5b41. Germans. Irish. 937,417 . > 52 STUDIES IN STATISTICS CHAP. were actually natives of Ireland. It is not difficult to understand how it is that the Irish under these circumstances manage to manipulate municipal politics to their own advantage. It is curious that in New York, while [rish men rarely marry German or Scandinavian women, it is comparatively common for Germans or Scandinavians 0 take Irish wives. In either case, however, only a small portion of the frish intermarriages are with the Teutonic races, the great majority being with natives either of the United States or Great Britain, and of these again no doubt a large proportion are with persons of Irish extraction. CITY OF NEW YORK, 1880. PARENTAGE OF THE PEOPLE. Irish feathers. Mothers German pp 508 | » Scandinavian .., 9 Total mothers not Irish 22.008 Irish mothers. Fathers German vee 2,747 » Scandinavian... 324 Total fathers not Irish ~~ 22.494 The distribution of persons of Irish parentage in several of the states is shown in the next table, which emphasizes the remarkable figures given above. In Massachusetts and Rhode Island the Irish form more than 80 per cent. of the population, and in Connecticut 27 per cent. It is very unfortunate that the figures of the United States census relating to parentage were not “ tallied” (analyzed) for the important states of New York, New Jersey, Pennsylvania, Ohio, and Illinois. It is also a great pity that French Canadians have not been distinguished from Canadians of British origin, as it would have been very interesting to study the intermarriages of the French Canadians, more especially with their co-religionists the righ. L Census of United States, 1880, vol. i. pp. 675-6. <¥) = = —_— So an oF = = = © + : = DD = PO we = Po - 2, eh » © + = 1 = 5 = 3 w sz Br 3 o > 2 D = c+ £5 > @ Tic 3 hon 2 BR Pd PERSONS OF IRISH PARENTAGE IN .CERTAIN OF THE UNITED STATES. Both Parents Irish. Father only. | Mother only. Born in United States. Foreign Born. Born in United $tates. Foreign | Born. Born in United States. Foreign Born. Total one or both parents Irish. Massachusetts Connecticut California Wisconsin Missouri vi Minnesota ... Rhode Island... Maryland ... Kentucky ... Louisiana Vermont se Nebraska ce New Hampshire 247,822 82,952 58,315 68,681 56,546 £0,459 36,916 25,412 23,529 15,767 15,874 12,680 13.600 240,015 72,218 66,440 16,613 51,289 29,834 37,550 | 22,240 18,704 13,926 | 12,652 11,407 14.094 23,535 6,631 11,527 15,420 17,335 8,872 4,110 7,920 6,910 4,281 3,027 1,247 1.264 2,955 315 | 1,193 1,184 | 578 1,052 487 187 153 106 375 146 269 25,918 8,078 17,733 9,961 8,598 5,964 4,822 4,532 3,268 5,103 2,338 2,611 1.659 2.030 301 1,239 996 543 808 404 176 112 92 276 320 158 542,275 170,495 156,447 142,855 134,889 86,989 84,289 60,467 52,676 39,275 34,542 31,711 31,044 Total of these 13 States 698,553 | 636,982 1.335.535 9,300 | 100,585 | 7,455 232.419 Census of United States, 1880, vol. i. pp. 680 —693. 115,079 1,567,954 Per cent. of population of State. 30-4 27-4 18-1 10-9 6-2 111 30-5 6:5 32 4-2 10-4 7-0 8:9 126 2 — y p 2 o 3 7 n 2 23 % Ba STUDIES IN STATISTICS CHAP. with persons of other nationalities) in the three characteristically Irish states, Massachusetts, Connecticut, and Rhode Island, and also in three states in which Germans prevail, viz. Missouri, Wisconsin, and Minnesota. IRISH PARENTAGE IN CERTAIN STATES, 1880.1 Natives of the United States having Irish fathers Mothers being Irish not Irish 8 Natives of the United States having Irish mothers Fathers being Irish .. not Irish sassachusetts. Connecticut. Rhode Island. 367,690 34.976 401,966 367,690 38,818 106.508 "Per cent. Missouri. P Wisconsin. | ot Minnesota. cent. I's gs 165,686 | 41.627 799 20°I 1000 207.31 3 | 100'C 90°5 95 165,686 | 87-1 24,623 1279 190.209 | 100-0 1000 As before in the case of the Germans, we see that in those states where the Irish are relatively most numerous intermarriages are least frequent. It will, however, be noted that while Irish men appear to be more exclusive in their attentions than German men, on the other hand, Irish women are more attractive than German. On the next page is a detailed analysis of the results of intermarriage. This table shows the relatively considerable number of ‘ntermarriages between Irish and natives of Great Britain or British North America, but of these doubtless a considerable number are of Irish extraction, so that, strictly speaking, they are not intermarriages in the sense in which the term is here used. - Census of United States, 1880, vol, 1. pp. 680—692, VI THE GROWTH OF NEW NATIONS. UNITED STATES 93 PERSONS OF MIXED IRISH PARENTAGE IN CERTAIN STATES, 1880.1 Natives of the United States having Irish fathers. Mothers being natives of United States Germany Great Britain British N. America Scandinavia Other Countries Total assachusests Connecticut. Rhode Island 22,666 344 5,384 4.446 81 355 34,276 Per cent. 56°2 Io 186 130 2 10 O00 Missouri. Wisconsin. |Per cent. Minnesota. 31,631 1,749 8,635 3,871 153 588 76-0 42 87 93 ‘4 TA 41.627 100°0 Natives of the United States having Irish mothers. “athers being natives of United States Germany Great Britain British N. Americe Scandinavia Other Countries {otal 19,884 2,646 9,358 1,783 396 1,751 RE 81% 51°3 63 24°1 12°3 To "5 1000 12,252 3,770 4,606 2,656 139 100 24.523 49'9 15°4 188 10'8 6 4'5 1000 It is an important fact to keep in mind, that whatever may be the ultimate success of the Americans in absorbing she Irish and the Germans, these two important foreign slements show as yet very little tendency to mix, a fact which may have sinister results for the Republic. Meanwhile the Irish continue to pour in to the extent of some 67.000 a year. 11. Natives of Great Britain.—If I am not much mis-‘aken, even the Yankee most given to “Spread-eagleism ” (especially when catering for the Irish vote) must in his heart of hearts feel more interest in the immigrants from Great Britain than in any others. Whatever ill-feelings may have resulted from political differences, the British are still his kith and kin, united to him by history, language, and literature ; Germans and Scandinavians may make Census of United States, 1880, vol. i. pp. 680 —692, STUDIES IN STATISTICS CHAP. quiet, law-abiding citizens, but they differ from him in language and in traditions, and are not so easily assimilated by the body politic. The natives of Great Britain amount to 916,114, made 1p as follows— NATIVES OF GREAT BRITAIN IN THE UNITED STATES, 1880. Proportion per cent. to population of mother countries in 1881. Natives of England » Wales 5 Scotland Great Britain... ,, Ireland ry 369 *7¢ ) 7 *.9 6 1 ax The first thing to be noticed is, that whereas the natives of Ireland in the United States equal fully one-third of the population of Ireland, the British born equal 8 per cent. only of the population of Great Britain. © The Welsh are relatively most numerous, the English least 30, the Scotch occupying an intermediate position. The following table shows the distribution of natives of Great Britain in those states in which they are absolutely or relatively most numerous. It is not much to the credit of our country that the proportion is highest in the territory of Utah; the English and Welsh in the United States are represented about in proportion to their numbers, but the Presbyterian Scotch and the Roman Catholic Irish would appear to be proof against the seductive preaching of Mormon missionaries. In Plate XIII. the percentage of persons born in Great Britain in each state, &e. is indicated by the depth of red colour. vows 1200 Ws WoT 105 re C ca “iN DAK. MINN NEBR. 25 O00 oq Long TIE" W. | per cent oo COLO. N.MEXx 2.85 per cant KANS »-0 per cer. London - Edward Stanford 2 per cent CENSUS 1880-8I NATIVES OF GREAT BRITAIN Proportion per cent. of the population of each PROVINCE, STATE & TERRITORY oF CANADA AND THC IINITED STATES ¥ 25 Stale of Englisch Miles 00 50 © wo 200 300 400 gman 0 rer cent Tomidor. Stamford Gonnl Fopit a 2 0 3 vo ow = = =e Tn 2 3 3 ap © je] ry gs = 8 5 3 = ZB > ct E c+ a Ho D ot QE @ 37 ¢ BH 3de 2 3S, 5 5 = 2 E 5 2 3 B=) he] 3 v ed — o B © ob + oF 7 => om DY e+ 2 8 x = B® oO a * — =, > =o - NATIVES OF GREAT BRITAIN IN CERTAIN STATES, 1880. WALES. Per cent. Persons. lof State. Persons. Per cent, of State. P [Ber conti ersons. lof State. New York ... Pennsylvania llinois .-Dhio ve Massachusetts Michigan ... New Jersey Wisconsin ... California ... Lowa i. Utah ve. Missouri ... Kansas i" Connecticut Rhode Island Colorado... Other States 116,362 80,102 56,318 | 41,555 47,263 43,202 31,285 24,916 24,657 22,519 19,654 15,798 14,172 15,453 12,500 8,797 88.123 2°29 1-87 | 1-83 1°30 2°63 2°64 277 0°89 285 I°39 3°65 73 1°42 2:48 I aa 28,066 20,735 16,645 | 8,946 12,507 10,731 7,633 5,770 6,465 5,885 3,201 3,641 3,788 1,157 3,039 1,673 27 954 ‘55 ‘48 51 28 70 66 | 67 ‘44 75 42 2°22 ‘17 38 67 [‘I0 86 — 7.993 29,447 3,694 13.763 873 830 863 5,359 1,920 3,031 | 2,390 1 1,766 2,088 | 407 167 1,212 8976 ‘14 "69 12 43 03 05 ‘08 ‘22 ‘19 1°66 ‘08 21 ‘07 "06 62 ‘04 ras Total A. a ae wn 662,676 1:39 1 170,136 “34 83,302 17 Census of United States, vol. 1. n. 494. TOTAL. "Per cent. of State, Persons. 151,651 130,284 75,657 64,264 60,643 54,763 39,781 36,038 33,042 32,435 25,245 21,205 20,048 | 20,017 | 15,706 11,682 193.653 30 3° 2°5 20 4 3'3 35 2'7 3'8 2'0 75 I'0o 2'0 32 5'7 50 6H 216.174 1: 3 CN )8 STUDIES IN STATISTICS CHAP. With the exception of a predilection for Utah (which, however, does not attract a number of settlers that is absolutely very great), the British scttlers are very evenly spread over the northern and western states. The English are about four times as numerous as the Scotch, and it will be observed that this proportion is maintained in nearly all the states except Utah and Colorado, in both of which the English are relatively, as well as absolutely, the more numerous. The distribution of the Welsh is, however, very peculiar. In the aggregate they are just half as numerous as the Scotch, but this proportion is preserved in very few of the states—in Rhode Island, Massachusetts, Michigan, Connecticut, and New Jersey there are both absolutely and relatively very few Welsh, but in Colorado, Ohio, Wisconsin, and especially Pennsylvania, they are remarkably numerous. In the city of Scranton, Pa., there are 3,616 Welsh, and in Pittsburgh, in the same state, 2,012. It is possible, though I do not know it to be the fact, that they are largely employed as coal-miners in the last-named state. The cities in which most English and Scotch are to be found are New York, Philadelphia, Brooklyn, and Chicago. From the three following tables we learn that, as might have been expected, the British settlers intermarry very freely, and they are no doubt soon absorbed into the body politic ; this is as it should be, and cannot but be a cause of rejoicing to Englishmen and Americans alike, since sverything that tends to draw together the United States and the United Kingdom must be for the advantage of both. The United States is one of the grandest products of British enterprise; and if we cannot but regret that we do not owe allegiance to a common government, we must be glad to see anything tending to strengthen the British element and the British way of looking at things PERSONS WITH PARENTS NATIVES OF GREAT BRITAIN IN CERTAIN OF THE UNITED STATES.1 Both parents G. B. Father only. —— Mother only. 3orn in United States. Foreign Born. Born in United States. Foreign Born. | Born in United States. Foreign Born. Total one or 10th parents British. Per cent. of population of State. Massachusetts Wisconsin... California... Missouri we Rhode Island... Connecticut ... Minnesota Nebraska Maryland Colorado Texas... 26,246 36,144 18,381 14,724 7,336 10,578 9,644 7,968 5,325 5,412 3,593 56,023 37,219 35,467 21,134 13,598 18,877 14,615 11,862 8,205 11,935 8.474 21,532 20,988 16,178 16,431 4,490 7,689 9,624 6,810 6,386 | 2,811 6.313 3,971 1,746 2.166 837 618 389 1,441 671 240 407 375 13,701 11,912 9,893 7,517 3,139 5,084 1,955 3,780 3,259 1,650 9.290 2,372 916 1.580 565 566 332 382 $41 211 302 2483 123,845 108,925 83,665 61,308 29,747 42,949 41,161 35,532 24,626 22,517 21,218 70 83 9-7 2-8 10-8 69 53 7-0 26 11-6 1-3 146,351 | 237,4c0 383.760 119,252 | 12,861 67,110 8,510 207,733 591,493 3 Total of above States Census of United States, 1880, vol. i. pp. 680—692, Mt ee oo BB D c+ TE 3. @ = n U2 oo & + aD #2 & Ie] po <Q 2 ieee ox SE oO 3 = 2, B = re - =~ 2 So Ry 2, D 5 ™~ ™ t = oo @€n = = > = oe + yl rH 5 Ie rp ~~ — » 0 OO 00 STUDIES IN STATISTICS CHAP. BRITISH PARENTAGE IN CERTAIN STATES, 1880.1 Natives of the United States having British fathers Mothers being British .. not British Natives of the United States having British mothers fathers being British .. not British Massachusetts. Rhode Island. Connecticut. 414,160 | 33,711 77.871 14,160 21,924 26.084 Per cent. 567 13'3 “O00 668 332 ron n Wisconsin. Missouri. Minnesota. 50,512 17.043 107.555 60,512 24.384 R4 ROB Per cent. 56°3 437 [O00 71°3 287 TOO PERSONS OF MIXED BRITISH PARENTAGE IN CERTAIN STATES, 1880.2 Natives of the United States having British fathers. Mothers being natives of United States Ireland Germany British N. America Scandinavia Dther Countries )» y 3 9 9 Natives of the United States having! British mothers. Fathers being natives of United States : Ireland { Germany British N. America Scandinavia Other Countries | Total Total .assachusetts. | | Wiscors® Rhode Island. Per cent, Missouri. Connecticut, Minnesota. 20,542 610 9,358 278 | 387 Er 2,008 | 89 76 2 350 10 33,711 | 100°0 35,230 4,60 6 2,305 3,042 869 991 47.048 12,672 | 8,384 984 1,114 129 641 578 29't 45 yr ‘6 20 15,908 3,635 (,953 607 597 ARG 21.9924 1000 24 5R4 Per cent. 74'9 98 4'9 6°35 1'8 21 [000 653 14'9 80 6'6 24 28 100°0 [f these figures be compared with those relating to he Irish in the same states, it will be observed that the 1 Census of United States, 1880, vol. i, pp. 680—692, 2 Thad. vi THE GROWTH OF NEW NATIONS. UNITED STATES 101 British show no exclusiveness in their matrimonial relations. As regards the nationalities selected, there is much less difference than might have been expected. The British are even rather less disposed than the Irish to unite with the Germans, though somewhat less averse to Scandinavian partners. British men are more disposed to take Irish wives than Irish men are to take British wives. The Irish, however, intermarry more freely with the Canadians. In proportion to the natives of Ireland, the natives of Great Britain in the States are slowly gaining ground. The apparent decline of the Irish in the United States (and the same thing is true in England) is partly due to the gradual diminution by deaths of the large numbers of Irish who left their native country for America (and for England) in the years 1847-53, partly to the fact that emigration from Great Britain has been more active in late years. An average of 81,000 British immigrants have, according to the Bureau of Statistics, landed on the shores of the United States annually since 1880. NATIVES OF THE SEVERAL PARTS OF THE UNITED KINGDOM IN THE UNITED STATES AT THREE SUCCESSIVE CENSUSES. English ... Welsh... | Scotch... frish- ... Total ... 1860. 131,692 45,763 108,518 1 611,304 2,197,277. Per cent.] 1870. 20 | 550,924 2 | 74,533 5 | 140,835 va 1,855,827 100 ,2,622.11=Per cent. «I ” L w ICO 1880. 662,676 83,3021 170,136] {,854.571 “T0.685 Per cent. 24 3 6 67 100 12. The Cuanadiaiis.—As has been said above, the United States Census does not make any distinction between Canadians of British and French origin. For his reason it would appear to be uscless to analyze the 02 STUDIES IN STATISTICS CHAP. figures relating to natives of British North America. What I have to say on the subject has been given in connection with the French in America! It should, however, be borne in mind, that there are in the United States probably 400,000 natives of Canada of British origin. HE FOREIGN-BORN GENERALLY. We learn from the detailed figures in the accompanying table several facts of importance— (1) The foreign-born formed a smaller proportion of the population in 1880 than in 1870. This, as was explained above (p. 55), is due to the fact that, as a new country increases in population, a time comes when the natural increase exceeds the immigration ; equal annual increments bear a smaller and smaller proportion to the whole population; the older immigrants die off, and their places are taken by their children born in the country. [n some individual states, e.g. in Maine and New Hampshire, the proportion of foreign-born has increased, owing to the immigration of Canadians during a period of great commercial depression in Canada ; also in Florida, owing so the impulse recently given to fruit-culture in that state. (2) The distribution of the foreign-born is very irregular. They are relatively most numerous in the far west. That portion of the country which includes Dakota, Nebraska, Colorado, Arizona, and all the states and territories to the west of these, is, excepting California (which contains more than one-third of the whole population of this vast area), very thinly populated, containing less than 2% millions of people; but of those 640,000, or 29 per cent., are foreign-born, the proportion in Nevada and Arizona reaching 40 per cent. In Minnesota and Wis-1 See above, page 74. vi THE GROWTH OF NEW NATIONS. UNITED STATES 103 consin the foreign-born are about one-third of the population; in New York, Massachusetts, and Rhode Island about one-fourth ; in Ohio, Pennsylvania, and Illinois they range from 12 to 19 per cent. Un the other hand, in Virginia, Tennessee, the Carolinas, Georgia, Alabama, Mississippi, and Arkansas they constitute but 1 per cent. of the population or less. Within this latter area, however, the blacks make up from 26 to 61 per cent. of the population. THE PERCENTAGE OF FOREIGN-BORN IN EACH OF THE STATES AND TERRITORIES IN 1870 AND IN 1880; ALSO FOR A NUMBER OF THE STATES THE PER-CENTAGE OF PERSONS HAVING BOTH PARENTS FOREIGN-BORN.* Nevada Arizona Dakota Minnesota Malifornia Wisconsin [daho Utah Montana Wyoming Rhode Island Massachusetts New York Michigan Nebraska Washington Connecticut Colorado New Jersey (linois Oregon (owa, Pennsylvania New Hampshire Ohio PERCENTAGE OF Foreign- | born. 870.180 Persons having both parents Zoreign-horn. IRR 14:2141°¢ 601/397 33:9|38-3 36°5(34-3 37°5]33°9 '4+6/30°8 52-6306 35°4(30°6 38°7120+4 385/281 255 28 24-2(24-9 26°0(23°8 227(28°7 25°0(21°5 21°0{21-0 1°120°9 66/205 20'9119+6 0-3{19°0 128/175 17°1]16-1 54[18°7 9-3]13°3 4-0/12-38! 509 50°7 59'8 550 324 h5°1 45°3 11°8 128 47°0 44°8 37°3 3077 40°5 3175 25°0 20°11 PERCENTAGE OF Persons Foreign- | having both born, parents foreign-born, .870.11880 1580. 0) 13-3111°1 12:91 9-8 {2°31 96) 0'7' 89 78 91 84 7-8 76" 72 8:1! 67 7.31 65 3:5! 58 2:6' 37 4:81 36 3-9! 3-0 Lor 1-3 1-6 1-1 t-1 1-0 14 ‘8 0 °F v2. 8 ol 7 -al gl Termont es {ansas ... ce Wissount ... I District of Columbia Maryland . laine ... . udiana ... ro Texas ... oe New Mexico ... Delaware ow Louisiana rhe Florida ... - Kentucky . W. Virginia ~ Arkansas - I'ennessee - Virginia ... wr Wississippi ey Alabama ... _ South Carolina ... Georgia ... es North Carolina ... United States... 144/133] States “ tallied” [10.6!101 12°5 RR 13°5 12°2 6 85 69 2’ 22 Rd | 8 £6 1°5 13 5 259% 189 1 Calculated from tables in Census of United States, 1880, vol. i. pp. £26, 678. 2 Jistimated. (C4 STUDIES IN STATISTICS CHAP. (3) The territory of Utah derives an especial interest from containing the country of the Mormons with their peculiar tenets. It is therefore worth noting the nationalities of the population of the territory, though of course it does not follow that these are the nationalities also of the Mormons themselves, since there are many Gentiles” in the serritory. : TERRITORY OF UTAH, 1880.! BIRTH-PLACES OF THE PEOPLE. 58% “ 5584 et EE SE. 1085 :.5 5 ach HE Eom i= 8 2825 Vr 2 ceSs SBE gEHE 3 Lo [EN PERS fren Natives of the United States » England os Denmark hd Sweden Scotland Wales ... {reland vee Norway ‘ee Switzerland ... ee British North America ... Fermany ... hp Ither Countries es 19,969 19,654 7,791 3,750 | 3,201 2,390 1,321 1,214 1,040 1,036 885 719 ‘3.963 69°5 2 137 3:0 5°4 12-1 26 1-9 2:2 1-9 oy 29 '9 07 ‘8 7 7 12 ” 14 6 ‘04 2 3 TOON 2 The most striking fact is, that 12 per cent. of all the Danes in the United States were found in Utah territory ; the English and Welsh were also present in notable excess ; but the Irish, Canadians, and especially the Germans, were proportionately to their numbers in. the Republic very poorly represented. (4) Where the facts for foreign parentage are available they intensify the results given for foreign birth; on an average the number of persons having both parents foreign 1 Census of United States, 1880, vol. 1. pp. 492—495, v1 THE GROWTH OF NEW NATIONS. UNITED STATES 105 is nearly double the number of the foreign-born, but of course there are wide deviations from this average; thus in Wisconsin the proportion is exceeded, and we see that actually two-thirds of the population is either foreign-born or born in the United States of foreign-born parents; in Minnesota likewise two-thirds of the people have foreign parents, We also learn from the census something as to the relative attractions of town life for the different nationalities. NUMBER OF FOREIGN-BORN IN FORTY-FOUR LARGE CITIES OF THE UNITED STATES IN 1880, AND ALSO PROPOR-TION OF THE WHOLE OF EACH NATIONALITY FOUND IN THESE CITIES.? Natives of [taly yw Hungary ... Poland ... Ireland” ... Germany Bohemia ... France ... Austria proper Scotland ... England ... . Switzerland ver Holland ... won Russia ... . China ... ib British North America Wales ... - Sweden... - aii Denmark an wn Norway ... en 20h Number, 26,893 6,644 25,459 839,345 761,742 33,041 40,954 13,081 51,060 197,974 22,152 14,293 8,777 23,150 127,644 13,417 27,367 8,088 14,583 Per cent. of Nationality. 60-8 57-6 52-4 45-3 387 38-7 38-3 33-8 30-0 29-9 25:0 24-6 24-6 22-2 17-8 : 16:1 wwe 14-1 po 12-6 oo R:0 Of the more important nationalities (from an American point of view), the Irish are the most fond of towns, the Germans following not far behind them. The very small proportion of Scandinavians found in the cities is noteworthy. The Italians, Swiss, and Poles are mostly found in New York, but the Italians are also numerous in San - Census of United States, 1880, vol. 1. p. 470. (06 STUDIES IN STATISTICS CHAP. Francisco, the Swiss in St. Louis, and the Poles in Chicago. The Bohemians especially affect Chicago, New York, and Cleveland. Mr. Crouch, the U.S. consul at Milan, informs me that the immigration of Bohemiuns and Hungarians has assumed large proportions in recent years, and that they form a very undesirable addition to the population, since they are far behind most European nations in civilization. The difficulty of their language leads them to congregate together in cities and prevents their ready assimilation. Many of the Bohemians are employed in coal-mines. TABLE E.—IMMIGRATION TO THE UNITED STATES FROM THE PRINCIPAL EUROPEAN COUNTRIES, TEN YEARS ENDING JUNE 30rm, 1880, COMPARED WITH NINE YEARS ENDING JUNE 30rH, 1889.2 From Teun years | Nine years 1871-80, 1881-89. England and Wales ... Scotland i 4 .. Ireland ves vee Germany ... a Norway and Sweden Denmark ... po Austria-Hungary ... Russia _— . Finland ... Poland ... Holland ve Switzerland ... ae Italy ... a -p France —_ A. Belgium ws wire Spain and Portueal ... 460,458 87,564 136,871 718,182 211,245 31,771 79,969 39,135 444 12,675 16,541 28,293 55,762 72,201 7,221 9.893 599,797 137,828 602,458 1,360,543 527,360 78,766 297,520 168,227 9,458 40,733 49,375 74,995 255,309 43,876 17,506 5.563 Total, principal | European countries ! 2,207,298 4,260,314 Increase Absolute. |Per cent. 30 57 38 89 150 148 308 330 221 199 165 358 -393 142 — 448% | 2,008.089 89 Furthermore, the statistics of immigration give us some 1 United States Census, 1880, vol. i. pp. 538-541, Reports of the Chief of the Bureau of Statistics of the United States. 3 Decrease. 11 THE GROWTH OF NEW NATIONS. UNITED STATES 107 insight into the changes in the constitution of the nation that are now going on, and which will very shortly be registered in the eleventh census of the United States. Table E compares the immigration during the nine years since the census of 1880 with that of the fen years that preceded it." Among the French and Spanish immigrants there has been a notable decrease, but in the case of all the ther countries given in the table there is an equally notable increase. The absolute increase is greatest in the case of Germany ; but it is very great in the case of Norway and Sweden, Austria-Hungary, Italy, Ireland, England and Wales, and Russia. Proportionately the increase is greatest in the case of Italy, Russia? and Austria-Hungary, least in the case of the British Islands. These figures, therefore, confirm the impression that the people of the United States is becoming less and less British in race. Indeed, when Table E is compared with Table A (p. 63), it will be seen that more immigrants from Norway and Sweden, Denmark, Austria-Hungary, Russia and Poland, Italy and Belgium entered the country during the nine years since the census of 1880 than were respectively enumerated as living in the States at that time. Table F gives for thirteen years the proportions of females among the immigrants of various nationalities. That the Irish head the list, the number of males and females being practically equal, is a striking proof of the strength of ! The figures in the Almanac de Gotha differ materially from those given in the Statesman’s Year-Book. A reference to the original reports shows that in the former the figures are given for calendar years, in the atter for fiscal years ending June 30th. I prefer the latter, as so nearly coinciding with the census which is taken in June. 2 A new movement from Finland has originated within the last eight years, but whether the migrants are by race Swedes or Finns I have no means of ascertaining. 3 STUDIES IN STATISTICS CHAP, the family tie with them. Then follow the great Teutonic nations, with about two females to three males. France, Switzerland, and Belgium occupy a medium position, with one female to two males. In the case of Spain and [taly, the proportion falls to one female to four males. The proportion of females is a fairly good test of the TABLE F.—IMMIGRATION TO THE UNITED STATES FROM THE PRINCIPAL EUROPEAN COUNTRIES AND CHINA. PERCENTAGE OF FEMALES IN THE THIRTEEN YEARS ENDING JUNE 301m, 1889.1 [reland Germany Netherlands Austria England Norway Sweden Scotland Wales Denmark Russia France Switzerland Belgium Portugal Poland Hungary Spain Finland Ltaly China permanency of the stay of immigrants of each nationality. The Irish go over bodily, or if the men go alone they subsequently send for the other members of the family. The Italians, on the other hand, to a great extent go not so much to seek a new home as to seek work, so they do not so often take their women with them, but after a hard-working and frugal life return to the fair country of their birth, intent in many cases on buying land. The L Reports of the Chief of the Bureau of Statistics of the United Ntertes. TABLE G.—IMMIGRANTS TO THE UNITED STATES FROM CERTAIN EUROPEAN COUNTRIES. NINETEEN YEARS ENDING JUNE 30Ta, 1889.1 Great Britain. Ireland. Fermany. Norway, s Sweden, &| Austria: Nenmark. gary. 21. 28,575 35,481 19,178 "4.392 19,323 1,274 12,254 21,820 85 657 Russia, Finland, XPoland. i . Italy. "France. Switzer! Holland. land. (871 (872 1873 874 875 :876 1877 878 879 88a 85,455 34,894 39,499 39,021 17,904 29,291 23,581 22,150 29,954 v3 073 7,439 58,732 77,344 33,707 37,957 9,575 4,569 | 15,932 20,013 - 71.603 82,554 141,109 149,671 37,291 47,769 31,937 29,298 29,313 34,602 24 638 .887 1,410 7.112 3.850 7,658 5,276 5,396 5,150 5,963 17 267 £,208 2,816 2,665 4,190 £,972 3,757 5,868 7,667 3981) 3,631 5,700| 3,017 7,132| 3,195 2595 4,344 1.942] 5,791] 7191] 12.354 3,137 9,317 14,798 9,643 8,321 8,002 5,856 “159 4.655 1313 2,269 3,650 3,107 2,093 814 ,549 ,686 1/808 161 6.156 993 1871 1,909 1872 3,811 :873 2,444 1874 1.237 18753 855 1876 591 (877 608 (878 753 1879 3.340 ' 1880 1881 1882 :883 884 :885 1886 1887 (888 1889 Total ten years 1871-1880. | Total nine years | 18811889. | 81,376 102,991 ! 76,606 65,950 57,713 32,929. 93,378 108,690 87.999 TENTH CENSUS O01 72,3421 210,485 T6402) 250,630 31,486 194,786 33,344 179,676 31,795 124,443 19,619 84,403 38,370 106,865 73,518 109,717 85.557 99.538 TIE U 81,582 105,326 71,994 32,728 .0,704 £6,735 37,629 31,924 52 504 “hs 27,985 29,150 27,625 36,571 27,309 28,680 10,265 15,811 34 174 1,655 21,590 11,920 17,226 20,243 21,739 36,894 39,313 38.838 15,4011 32,160" 3,792 16,510 13,644 11,315 17,622 51,558 25 307 TNE, 1{70.) 5,227 [10,295 6,003 [10,844 4,8211°2751 3,608. 9,386 3,493 5,895 3,318 £805 5,024 5 214 6,4 7,057 591% 7.07 8.597 11881 9,617 1882 5,249 1883 1,198 1884 2,689 188s 2,314 1886 L506 1887 5,845 1888 6.460 188q 548,022 436,871 737,625 602,458 718,182 242,016 | 72.069" £2,284 55,762 172,201 128,293 16,541 260,842 306.126 209%.€20 218.418 255,300 143,870 [74,005 140,375 Reports of the Chief of the Bureau of Statistics of the United States. 1.8 Q | = 52,5 7 rE. Fi ee NN +8 |W Loo % bv = t ye wom y=< @ gE IK o - 2B ® 8 8% ® = =+ = 2 EB > 2 oa * 2 Q © Fh h .B 0 fy ~~ = Yo mn Ye gg TB ® ~ 4d ty Sie Th to =o - ow BB 8 STUDIES IN STATISTICS CHAP. In Table G it is possible to follow out the tide of immigration year by year. It will be noted that the maximum occurs in different years with different countries. Thus in the case of France it was in 1878, and doubtless was related to the war and the cession of Alsace and Lorraine. In Germany, Scandinavia, and Holland it was in 1882; in Ireland and Switzerland in 1883. These and the two years immediately preceding were years of great prosperity in the United States. In the case of Great Britain, Austria-Hungary, Russia, Italy (and I may add Belgium, although it is not in the table) the maximum was in 1888, and was perhaps associated with depression in those countries, but it coincides with a revival of trade in the United States. Emigration is caused by either a centrifugal or a centripetal force, by discontent at home or prospects of success abroad—at one time one, at another another will prevail SUMMARY. [10 In conclusion, I would urge that it is important to bear in mind that if we think of the United States as an Anglo-Saxon nation, a mere off-shoot of ourselves, we make a great mistake, and may come to very false conclusions. In 1880 that great nation consisted of 50 millions of people, but of these only 431 millions were of European stock, and only 862 millions natives of the country they live in. Moreover, we have also learned that only some 85 millions of the people were native born, and with both parents native born; from this number again may be deducted 65 millions of men of colour, leaving as real Americans® only 281 millions. 1 Tt has been often stated, and there is statistical as well as general evidence to support the statement, that, from whatever cause, the old American stock does not multiply at anything like the rate maintained by vi THE GROWTH OF NEW NATIONS. UNITED STATES 111 or but 57 per cent, and of course a very large number of these must have been of foreign extraction, Irish, German, French, Dutch, Spanish, &c. But if we assume these to be thoroughly nationalized, we get the following approximate constitution of the people :— Prue Americans, whites, native born of native parents ... British ~~... sui sis ie British Canadians sin vie Total more or less of British Stock 928,500,000 2,000,000 500,000 31,000,000 4,800,000 600,000 5,000,000 600,000 1,300,000 12,300,000 6,600,000 200,000 50.100.000 [rish ve. ee 5% I iin French and French Canadians ... on -— Germans ... an ce. eens er Scandinavians —_ _ a _ . Other European races ... . Total Europeans other than British Men of Colour ... ee . Chinese and Indians vee . These numbers have no pretensions to be accurate, but they are quite sufficiently near the truth for my purpose, and may probably be accepted as the nearest round numbers. In still more general terms I would say, that about 52 per cent. of the inhabitants of the United States may be looked upon as of Anglo-American stock, about LO per cent. as Irish, 10 per cent. German, 18 per cent. African, and 5 per cent. a mixture of nearly every European race, Scandinavians and French preponderating. If there be any important inaccuracy in this statement, it consists in overrating the Anglo-American element, but I regard it as an undoubted fact, that many persons of German, Dutch, or Irish descent have been as completely the newly-arrived immigrants, If the New Englanders, the descendants of the old Pilgrim Fathers, are from this cause losing ground, it is a matter greatly to be regretted. STUDIES IN STATISTICS CHAP. Americanized as the Huguenots in England have been Anglicized. Indeed, these elements in his ancestry, together with the results of climatic influences, constitute the differentia, of an American, whereby we distinguish him from an Englishman of the old country. There seem to be three weak points in this young but vigorous nation— the German Socialists, probably not very numerous, the Irish anarchists, and the negroes; but of these the last is by far the most serious source of trouble for the future. California and New England, Wisconsin and South Carolina are now in many respects very different ; will time diminish or intensify these differences? Canada appears at first sight to lie at the mercy of the United States, but is it certain that the States would be unanimous in annexing it? Is it not conceivable that a state of war, arising say in the next century out of tariff disputes, might lead to a rearrangement, whereby the union of Canada with the great Republic might cause the disruption of the latter? Might not the Germans and Scandinavians, for example, object to an Irish proposal to ¢ liberate” the French Canadians from their English “oppressors,” and prefer to side with the Anglo-Canadians, while in the moment of difficulty the negroes in the South took matters into their own hands 2! 1 In the Century Illustrated Monthly Magazine for September 1884 (vol. xxviii. p. 761) will be found a paper by Mr. J. E. Chamberlin, entitled “The Foreign Elements in our Population.” This is written from the American point of view, and is a thoughtful paper pregnant with suggestions. 11 THE GROWTH OF NEW NATIONS. CANADA 113 CHAPTER VII, THE GROWTH OF NEW NATIONS (continued). CANADA. FEW countries have had their population counted so many times and during such a long series of years as Canada. The first official census of ““ La Nouvelle F rance,” of which the results are now known, was taken in 1665, that is, fifty-seven years after the foundation of Quebec, when the population was found to be 8,215. Eight other censuses taken in the 17th century still exist, the population in 1698 having increased to 15,355. Nine censuses of New France, or Lower Canada, as it was called later, now the Province of (Quebec, were taken in the last century ; five years before Wolfe's triumph and death on the heights of Abraham, the population of New France was 55,009; in 1765, two years after the close of the Seven Years’ War, and definite establishment of British ascendancy in North America, the census gave its populalion as 69,810. The first census of the United States was taken in 1790, the result showing a population of very nearly four ! See the Census of Canada, 1870-1, vol. iv., which contains a most interesting account of the statistics of Canada from 1605, summaries of all the censuses being reprinted therein. STUDIES IN STATISTICS CHAP. millions, including slaves. The census of Lower Canada in the same year showed 161,311 inhabitants, taking no account of Nova Scotia and the other maritime provinces, which may have contained between 50,000 and 60,000 people. In 1881 it had more than eight times as many inhabitants, having doubled in the previous thirty-seven years. Previous to 1780 Upper Canada was practically uninhabited ; it is believed that in 1784 about 10,000 of the United Empire Loyalists were settled there. In 1824 the first census of Upper Canada was taken, showing 150,066 inhabitants. It is remarkable that from that year a census was taken annually for nineteen years. In ten years the population had doubled, in a little over ten years it had again doubled itself, in another sixteen years it had again doubled itself, being in 1861 as much as 1,396,091 ; since then the rate of increase has diminished, the census of 1881 giving 1,923,228, the population having doubled itself in thirty years. Manitoba is now more populous than New France was in the hour of Wolfe's victory, while New Brunswick and Nova Scotia each contains double the population that Lower Canada did at the termination of the war of American Independence. In spite of these facts, it must be confessed that Canada does not prove so attractive to emigrants as one might wish, and it is but natural that Englishmen should view with a jealous eye the large numbers of their countrymen that year by year settle in the United States. Prior to 1840 the emigration returns (which commence with the year of the battle of Waterloo) show that the state of affairs was the reverse of the present, in each period of ten years a greater number of emigrants left these shores for Canada than for the Great Republic. In the decade 1841-50 this was entirely changed, the number of emigrants more than doubled, the great migration from Ireland vit THE GROWTH OF NEW NATIONS. CANADA 113 consequent on the potato famine commenced, and since then the number whose destination was stated to be the United * States has constantly excecded those going to settle in Canada ; moreover, about the same time emigration to our Australasian colonies began to assume large proportions, and has in many years exceeded that to British. North America. In the last few years, however, Canadian immigration has shown a tendency to increase. The Dominion Government has at last succeeded in convineing people that Manitoba and the North-West Territories of the Dominion are not the frost-bound wilderness that interested fur-traders had represented them to be. Although the rush of immigrants to these prairie lands has been exaggerated by emigration agents and others, the following figures taken from recent censuses show that, although the province is remote, and but in its eighteenth year, a great settlement. has really been made. CENSUS OF MANITOBA, AS LIMITED BY AWARD OF PRIVY COUNCIL IN 1884. Increase in five years. 1881, 31 July, 1886. Actual. Per cent. Population 62,260 108,640 46,380 74 [nhabited Houses 12,102 22,180 10,078 83 But of the above increase of population 31,211, or no less than 67 per cent., were Canadian born, and only 12,078, or 26 per cent., natives of the United Kingdom. Of the population in 1886, 5 per cent. were born in Russia or Poland, 2 per cent. in Iceland, while the natives of (Germany, Scandinavia, and France together scarcely amounted to 1 per cent. When the natives of Canada were traced back to their “origin,” it was found that of all the Manitobans, onefourth were English, one-fourth Scotch, and one-fifth Irish. The Canadian Pacific Railway entered the North-West [16 STUDIES IN STATISTICS CHAP. Territory (District of Assiniboia) in the summer of 1882, placing it for the first time in communication with the outer world ; by the summer of 1884 it had traversed it and the higher prairie of Alberta to the Rocky Mountains. The effect of this remarkable enterprise on the population of the “ Provisional Districts” of Assiniboia, Saskatchewan, and Alberta may be scen by the following census figures, in which some 20,000 Indians are not included. Co CENSUS OF ASSINIBOIA, SASKATCHEWAN, AND ALBERTA. Increase in 4} years. 1881. 1885, Actual. Per cent. Population } . Whites and Half-breeds 5,624 28,171 22,547 401 [nhabited Houses ow. 2,064 - 9,460 7,396 358 Of the total white population in 1885 (amounting to 23,323), no less than 11,074, or one-half, were natives of Canada (Ontarians predominating largely), and only 7,158 natives of the United Kingdom, mostly of Great Britain. In Canada even more than in the United States, it would appear that new regions are mainly settled by natives of the American continent, work for which they are especially adapted. Not only is there a flow of population from the Old World to the New, but in America itself the cry is still “ Westward ho!” The descendants of the early settlers move on to begin again the work of subduing the wilderness as their forefathers did, while the blank spaces left in the more settled districts of the sast are filled by fresh immigrants from Europe. The growth of cities in the far west is not absolutely so great as in the suburbs of London, but in some respects it is far more striking. Tor instance, in the Report of the Minister of Agriculture for 1882, the Brandon (Manitoba) immigration agent reports that 80,270 acres vi THE GROWTH OF NEW NATIONS. CANADA 117 of ‘prairie were “ broken ” during the summer of that year. It is hardly surprising to be told by the same gentleman, that where eighteen months before was rolling prairie, there now stands Brandon “city,” with its 4000 inhabitants, ten miles of streets, eighteen miles of planked footways, three fine churches, a ‘ veneered ” school-house two stories high (1), sixteen hotels, two banks, and a good fire brigade with a steam fire-engine ! Some of these figures, especially the estimate of population, are probably much exaggerated (there seems indeed to be something in the atmosphere of new countries that excites men to exaggerate), but the writer was sufficiently astonished at what he saw in Brandon in 1884; it should be remembered, too, that in a very thinly populated country undergoing rapid development, a place that, so far as size goes, would be counted in Europe but a large village, has the importance and many of the characteristics of a county town in the old country, or, as the Canadians love to say, “at home.” There were in the Dominion at the census of 1881, after deducting 6,714 persons born at sea, or whose Dbirthplaces were not recorded, 3,715,492 natives of Canada, and 602,604 persons of foreign birth ; details of their several nationalities are given on the following page. From these figures it appears that the foreign-born are 14 per cent. of the Canadian population as against 13 per cent. in the United States. Similarly the Irish form nearly the same proportion of the foreign-born in the two countries, while the proportion of natives of the United States in Canada is Lut little larger than that of Canadians in the States. There are, however, some remarkable differences in the other nationalities. In the United States the French (curiously emough) are proportionally swofold, and the Germans sevenfold more ‘numerous, while < STUDIES IN STATISTICS CHAP. the Scandinavians are more than twenty times as many. Per contra the natives of England and Wales are (proportionally) two and a half times as‘numerous in Canada, and the Scotch seven times as numerous. Thus we see that though the proportion of foreign-born is practically the same as in the States, yet these foreigners do not DOMINION OF CANADA. BIRTH-PLACES. CENSUS 1881. Per cent, 860 14'0 100°C Natives of the Dominion ... ve Foreign-born oe Se . 3,715,492 602,604 1,318,096 6,714 4,324,810 Born at Sea, or Birth-places not given Total population ... ... NATIONALITIES OF FOREIGN-BORN 1881. Natives of Ireland we wi England and Wales ... Seotland I. oe Other British Possessions United States ... ¢ Germany io Russia ... —_— France ce Scandinavia ... All other countries tes Per cent, 308 281 19°1 14 i2°9 1°2 1 302.6 1000 cause such a race-mixture as there; since in place of the Germans and Scandinavians, making up nearly 37 per cent. of the foreign element, they form but 41 per cent. in Canada, the difference being fully made up by an excess of English and yet more of Scotch. The great strength resulting from this purity of race is self-evident. It is to be hoped that the Dominion Government will see this, and t More than half Chinese. ‘iu THE GROWTH OF NEW NATIONS. CANADA 119 not take steps to artificially stimulate immigration from the continent of Furope. CANADA. CENSUS 1881. ORIGINS OF THE PEOPLE. Per cent. 300 22°1 20°60 16°2 59 7 3 ng 2'5 ‘5 I 3'r 100°C French er ver frish ... - ... i. English and Welsh ... Scotch vee ver $5 erman I .e .-Dutch A. vas . Scandinavian es Other Europeans and * not given ’ Indian wet Eo African ce - Chinese I. er “ Total non-European ; Grand Total ee us . The Canadian census also distinguishes the people by “origins,” although I am not able to state what the principles are on which they are determined. From these figures we learn several points of the greatest importance. In the first place, in Canada there is no ““ Negro question,” the population is in fact almost wholly European, the men of other races, including Indians, accounting for only 3 per cent. of the population. Secondly, the predominant racial element, no less than 30 per cent. of the population, is French— French, moreover, not only by origin but by language, religion, and customs ; not, however, French in political sympathies, or at least only very partially so. Thirdly, the Scotch do not fall far short of the Irish in numbers, while in energy, wealth, position, and influence they far exceed them. To a great extent Canada is a Scotch country, there is no country outside their native land apon which the Scotch have so sionally set their mark as [20 STUDIES IN STATISTICS CHAP. Canada.’ In Montreal Scotch names meet the eye everywhere ; most of the ships sailing between Liverpool and Quebec are built in Scotland, owned by Scotchmen, and manned and officered by the same enterprising race ; positions of importance throughout the country are held by Scotchmen ; and finally, the respected Premier is Scotch, as was also his late rival. Fourthly, the Irish, although more numerous than either English or Scotch, are not, as in the United States, twice as numerous as the English and Scotch taken together. Lastly, the countries of the continent of Europe taken collectively (but excluding France) account for only 8 per cent. of the population. Canada, then, has a population characterized by a very large French element, a relatively small Irish element, and a relatively large Scotch element. The French Canadians, though not very numerous, are a very distinct and homogeneous people, and must on no account be forgotten in forecasting the future of North America. Thrifty and industrious, they patiently cultivate their small farms, using the methods of their grandfathers ; ill-educated and unenterprising, they are the devoted children of the Church of Rome; lastly, unlike their kinsfolk in Europe, the ‘habitans” of Quebec marry early and produce large families. It is by this last trait that they hold their own? in spite of the fact that while the 1 Proportion of persons born in Scotland to 100 persons born in Eng. land and Wales, In Great Britain, 146; in the United States, 22-8; in Australia, 259 ; in New Zealand, 435 ; in Canada, 67-9. The persons of Scotch “ origin ” in Canada are to.those of Enelish and Welsh “ origin ” as 78-4 : 100. 2 PERCENTAGE OF PERSONS OF FRENCH ORIGIN. Leer Tanada. Upper Canad-~ Upper and Lower Canada. 76-3 . 2-4 35:1 780 oh 47 _— fom 357 79-1 53 es 359 Census. 1861 1871 .. 1881 VII THE GROWTH OF NEW NATIONS. CANADA 121 English, Scotch, and Irish receive constant reinforcements from the old country, but few recruits come from France;' on the contrary, the French Canadians pour into New England to work in the mills until they have earned enough money to return to Canada and buy land. In Lower Canada their political power is absolute, since the suffrage is practically universal, and they form fourfifths of the inhabitants. It is a great error to suppose that the French have been absorbed by the general population, they live quite apart from the English and Scotch, and show but little more sympathy with the Irish; indeed, until quite recently French-Irish riots were of annual occurrence in Quebec. In Montreal ‘and Quebec French society and English society stand aloof from one another more than: an outsider would believe to be possible; very few French families enter, or desire to enter, English or Scotch houses. It is a matter of great regret that of late years two circumstances have tended to diminish the intercourse between French and British Canadians. While the vast sums spent upon the improvement of the navigation of the St. Lawrence, by giving greatly increased facilities for shipping at Montreal, have led to the transference of much ! NATIONALITIES OF IMMIGRANTS REPORTED AT QUEBEC, NINE YEARS, 1881—1889. (From the Report of the Minister of Agriculture, 1889.) English +5 139,098 ' Teelanders Th wis 5,292 Scotch ... . 28 77 Taws (all in 1882) _ 1,375 Irish ... ve. ... 39,6 french and Belgians ... 1,382 Scandinavians... ... 51,221 All others yas us 2,600 Germans _ ... 6,745 Total... ‘ee «+ 975.765 It is probable that the great majority of the Scandinavians went on ho the United States, since they do not appear in the recent censuses of Manitoba and the North-West. 122 STUDIES IN STATISTICS CHAP. British capital from Quebec higher up the river; the withdrawal of the British garrison from its famous fortress has undoubtedly diminished the social attractions of that city. Hence Quebec is not what it was ; commercially and socially it is going downhill; it is now no longer the terminus of the great ocean route, and no longer a meeting-place for the two sections of Canadian life, but is almost exclusively a French city. Montreal, on the other hand, is growing at the expense of Quebec; but here too an unfortunate state of affairs exists, and the “ French difficulty ” is very obvious. While by far the greater part of the wealth, intelligence, and enterprise of the city is Scotch, English, or Irish, the large majority of the inhabitants is French, and an extended suffrage gives them complete command at the polls ; hence the municipality is French, and its officials, from the mayor down to the policemen and postmen, are nearly all French nominees. Meanwhile the British population lives almost secluded in the fashionable (northwest) quarter of the city. National differences are largely maintained, and national animosities stimulated by the societies, such as those of St. George and St. Jean Baptiste, which celebrate with undesirable warmth and enthusiasm the national festivals. The Société de St. Jean Baptiste in particular appears to exist especially for the purpose of keeping alive the French idea, and fostering in all possible ways the French language and French customs. Language of course has a good deal to do with this coldness, but I think religion has more. The Scotch, who are in the ascendant, give a puritanical tone to religion, which is eminently Protestant, and there is little doubt that the prejudice against the Roman Church is far stronger in Canada than in England, and the feeling is reciprocal. The catholicism of Quebec is a narrower yi THE GROWTH OF NEW NATIONS. CANADA 123 faith than the catholicism of Europe’; it offers uncompromising resistance to the modern spirit, which it would exclude by any possible means. Politically this has one advantage, since it tends powerfully to counteract any political influence which the French Republic might exert. The sympathy of Quebec is not with modern, revolutionary, agnostic France, but with the Catholic France of the last century; and the priesthood know well enough that they have more to fear from republican France than from conservative England. There is perhaps no part of the world in which the Church of Rome is so little interfered with as on the banks of the St. Lawrence, and in return for this liberty her French-Canadian subjects give to their Queen and her representatives a heartily loyal allegiance. But while I gladly recognize this, and willingly admit the many good qualities of the French-Canadian, it is impossible to help regretting that there are so few signs of real union into one people ; and the existence of a million and a third of men of French race on the lower St. Lawrence, increasing more rapidly, it is believed, than any known people, cannot but be a source of anxiety and possible trouble in the future. It is worth while examining somewhat carefully the distribution of the French in Canada. The province of Quebec contains a little over a million, 83 per cent. of the whole, and in that province they constitute four-fifths of the population. In New Brunswick they number 57,000, and make up rather more than one-sixth of the population. In Manitoba and the Territories the recent censuses show that the other nationalities (as the result of recent * Sir Charles W. Dilke, in Problems of Greater Britain, says that in Canada Roman Catholicism has a puritanical side, thus e.g. it discourages dancing. I would remark that this chapter was written before the appearance of Sir Charles's new work. r24 STUDIES IN STATISTICS CHAP; immigration and settlement following the opening up of the country by railways) are rapidly gaining on the French, and it is plain that those provinces are not destined to fall permanently under French influence. As regards the. rest of the Dominion, in Prince Edward Island and Nova Scotia the French form barely one-tenth of the people, in Ontario but 5, and in British Columbia but 2 per cent. The concentration of the French in Quebec cannot, however, be made by the most radical politician an excuse for separating CENSUS OF CANADA 1871 AND 1881. PROPORTION PER CENT. OF PERSONS OF FRENCH ORIGIN IN THE SEVERAL PROVINCES TO THE TOTAL OF EUROPEAN ORIGIN, ¢.e. EXCLUDING INDIANS, AFRICANS, AND CHINESE. Prince Edward Island Nova Scotia New Brunswick Quebec - Ontario ves Manitoba ... British Columbir The Territories Total 1886. 10-9 2 1885. 1742 that province from the Dominion by giving it independence or yielding it to France. The Province of Quebec, owing bo its geographical situation, commanding the lower St. Lawrence, that unequalled water-way, must belong to the power that holds Ontario. It must always be governed either from London, from Ottawa, or from Washington. This is so obvious that it seems perhaps needless to affirm it, but I do so for this reason. If it be granted that Quebec can never rule itself, or again ) 1 No figures available. Including all French half-breeds as French, vil THE GROWTH OF NEW NATIONS. CANADA 125 be ruled by France, two alternatives remain—that it shall continue a part of Canada, or be annexed to the United States. Owing to questions of religious privileges, especially in reference to education, it is likely to be the last part of the Dominion to assent to annexation. Lastly, it is now generally admitted that a Canada independent both. of Great Britain and of the United States is an impossibility. It remains then that the existence of the French in Canada once again, as in the last century, tends to increase the loyalty of a British colony, and whatever difficulties may possibly occur in the future, the French-Canadians will not facilitate the annexation of Canada by the United States. The other foreign elements in Canada are, as has been shown, very small compared to what they are on the other side of the international boundary, and the more recent censuses of Manitoba and the North-West show that as - yet the: Germans, Dutch, Russians, Poles, Scandinavians, and Icelanders, all taken together, make up therein respectively but 15'5 and 26 of. the white population. We come then to the very satisfactory result, that, as far as numbers can prove such a point, the population of our largest colony is more homogeneous, more solid, and (from an English point of view) better constituted than that of its great rival to the south. This should be a matter of great satisfaction when we consider how long and severe the competition between Canada and the United States must be. The Canadians are undoubtedly an energetic people ; they have overcome great difficulties in the past, both political and physical, they have made progress in all directions,' and never did the future of Canada appear . . ir whole ! The Canadian Universities are admirable, and indeed their sdueationg] system, 126 STUDIES IN STATISTICS CHAP. so brilliant as at this moment. Their present political difficulties are but the result of the impetuous energy of the people, and the tact of their prime minister, one of the greatest living statesmen, Sir John A. Macdonald, may be trusted to overcome the Manitoban difficulties as he has overcome many others. The adverse circumstances in Canada are mainly 1hree —(1) the climate, (2) the geographical structure of the country, and (3) the existence of a large population speaking a different language. Concerning the last enough has been said above. The climate of Canada has been much abused, but there is no doubt that it has also been much exaggerated. As regards health and the enjoyment of life, it has been amply proved to leave little to be desired ; but the compulsory suspension of many kinds of work for several months in each year leads to serious social and economic difficulties, while in Manitoba and the North-West the agricultural season is so short as in many years to make it difficult, in spite of the magnificent weather of the northwestern summer, to get the harvests saved in good condition. I do not say these difficulties are insuperable, but that they are great difficulties is undeniable. At the same time, over a large area of the northern states on the other side of the international boundary the same climate holds sway, and has not proved incompatible with a degree of progress that is the wonder of the world. What I have termed the geographical difficulty is twofold. Firstly, putting to one side all consideration for the land west of Winnipeg and north of the fiftieth parallel, a very large area of the province of Ontario and a still larger proportion of the province of Quebec has a very sterile and rocky soil, only the lowlands along the rivers being of much value. This, again, is a drawback that part CII THE GROWTH OF NEW NATIONS. CANADA 127 of the state of New York and much of New England suffer from equally, and yet prosper. Seeondly, and to some extent associated with the last, the vast extent of Canada is in some ways a source of weakness, its territory lacks compactness, and its provinces are isolated by natural barriers; the barren lands of the north of New Brunswick and the Gaspé peninsula cut off the maritime provinces from Quebec; the isolation of the latter having been aggravated by the supinencss of the Home Government in the matter of the Maine boundary. The * muskegs ” or swamps about the Lake of the Woods separate Ontario from Manitoba; and lastly, the Rocky Mountains cut off British Columbia from Assiniboia. These undoubted inconveniences Nature has herself to a large extent counteracted by the magnificent stream of the St. Lawrence and the great lake system, while the 2nergy, skill, and perseverance of the Canadian people have so improved - the navigation of their noble river, that the largest ocean-going steamers now discharge their cargoes at Montreal, five hundred miles above Anticosti, a thousand miles from the Straits of Belleisle—a feat in inland navigation without a parallel in the world. The Intercolonial and the Canadian Pacific railways have united Ottawa with Halifax on the one hand, and Vancouver on the other ; while lastly, when the new canal and lock at the Sault St. Marie are completed, vessels of a considerable size will be able to sail from Port Arthur on Lake Superior to Quebec without leaving British waters. These great works done by a people as yet not more numerous than the Dutch, are but a foreshadowing of the mighty part to be played by Canada in the time to come. 128 STUDIES IN STATISTICS CHAP. CHAPTER VIII. THE GROWTH OF NEW NATIONS (continued). SouTH AMERICA. Brazir, Uruguay, and the Argentine Confederation are the three portions of South America to which there is a flow of European emigration ; this may be insignificant indeed in comparison with that to the United States, but yet is far too large to be lost sight of. 1. Brazl.—This fine country, the greater part of which lies within the tropics, once a dependency of Portugal, till lately an empire, now the newest of republics, is still essentially Portuguese. Under the rule of Pedro II Brazil made great advances and did much to develop its resources. By a system of gradual emancipation it cast off the reproach of being the last slave-owning country of America. In.1850 the slaves were estimated at 21 millions ; in 1872 they numbered 1,511,000; in 1886 they were estimated to be 1,188,000, and their emancipation was finally completed in 1888. The emancipation of the slaves has naturally proved a stimulus to immigration, but the recent revolution may not unlikely cause a temporary check. The last census of Brazil was taken in 1872, and was, moreover, only a partial one, giving results of doubtful value ; according to its figures (and there are no others available) we get the following results :— Il GROWTH OF NEW NATIONS. SOUTH AMERICA 129 CENSUS OF BRAZIL, 1872.1 Free Population Slaves wx Total ... Slaves. Natives of Brazil ... 5, other countries Total ... San Free Population. Natives of Brazil ... . ,s other countries ... Total ... die ve Population by Race. Europeans ... . Africans cee he [ndians a oe Mulattoes and half-breeds Total ... ves Nationality of Foreign-born. Portuguese ... Germans : Africans French Others : Total 8,419,672.2 1,510,806 7,930,478 Per cent. 848 15°2 OCS 1,372,246 138,560 1.510.806 go'8 92 100'0 8,176,191 ror 97'1 243,481 en 2'9 8,419,672 ves 100°0 3,787,289 1,954,452 386,955 3,801,782 0.930.478 381 197 39 383 1000 121,246 49'8 45,829 188 44 580 183 6,108 25 25,718 es 10'6 94348] ... 1000 It" is said that in the early days when Brazil was a dependency of Portugal, but few Portuguese women settled there ; as a result of this, the very large number of half-breeds and mulattoes should be noted. In 1872 the foreign element was very small, under 3 per cent.; of the foreign-born one-half were natives of Portugal, onefifth Germans ; the presence of 45,000 natives of Africa shows how recently the slave-trade was actively carried on. L Almanach de Gotha. } ? Excluding wandering Indians, variously estimated at from 250,000 Ho 600.000. 130 STUDIES IN STATISTICS CHAP, The returns of immigration for the five years 1883-7 state, that nearly 160,000 persons entered the country. The principal nationalities were as follows :— IMMIGRANTS TO BRAZIL. 5 YEARS 1883-87. Italians 53,458 Portuguese 17,651 Germans 9,437 Spaniards : 7,569 Others ver 41,508 Total ... 159.623 Per cent. 335 29'9 59 4'7 26'0 1000 If these returns are at all to be trusted we may safely say, without going into figures, that the foreign element in Brazil is now much more considerable than at the time of the census, eighteen years ago; also that besides a notable addition to the Portuguese and a smaller addition to the Germans, a new element has risen to importance—the Italian. The census of Italy taken in 1881 gave the number of Italians then resident in Brazil as 82,196. Tt is stated in the Statesman’s Year-Book, but I do not know on what authority, that there are in South Brazil 189,600 German colonists. It should be noted that more than half the foreign-born in Brazil in 1872, and two-thirds of the recent immigrants, belong to the Latin races of Southern Europe. 2. Uruguay—At the partial census in 1880 it appeared that one-third of the population was foreign-born, the remaining two-thirds being Uruguayans of mixed race. Among the foreigners Spaniards and Italians predominated, next in number being natives of the adjoining states, and then French. . Almanach de Gotha and Statesman’s Year-Book. 2 Censimento degli Italiani all’estero, 1881. VIII GROWTH OF NEW NATIONS. SOUTH AMERICA 131 CENSUS OF URUGUAY, 1880.1 Natives Foreign-born Total Foreign-born. Spaniards [talians Brazilians Argentines French English Fermans Dthers Total 39,780 36,303 20,178 15,546 14,375 2,712 9 1 9 140,222 298,023 140,222 438.245 Per cent. 68 32 100 283 25'9 14'4 IT'X 0'3 2'0 Ts 3 100'0 There has been a considerable immigration, mainly from [taly and Spain, during the last twenty-five years. The following table gives the immigrants, emigrants, and net migration of the port of Montevideo for the six years 1882-87. MONTEVIDEO. [talians Spaniards French English Germans Others Total [MMIGRATION AND EMIGRATION, 1882-87.2 Immigrants. Emigrants. 33,719 7,682 ER --4 Excess of Immigrants. 18,413 10,830 1,765 1,350 1,140 2,664 36.162 i 73.909 ~ po] 37,831 Thus it will be seen, that nearly two-thirds of the foreign-born, and five-sixths of the excess of immigrants ver emigrants, are of the Latin races. Uruguay is not important by the amount of its population, but it is one of the parts of South America that Y Almanach de Gotha and Statesman’s Year-Book. ¥ Report of Bureau of Statistics of United States, 1888-9, 132 STUDIES IN STATISTICS CHAP. seems to be making real progress, and is interesting from our present point of view as a place of destination for Italian, Spanish, and French emigrants. It is stated that the Spaniards and French in Urnguay are chiefly Basques. According to the Italian census of 1881, there were at that time 40,003 Italians in Uruguay. 8. The Argentine Confederation. — This, formerly known as La Plato, is a republic with a very large territory, nearly the whole of which, like that of Uruguay, lies in the temperate zone; probably it is for that reason that this is of all South America the part that shows the most unmistakable signs of progress. The last census of the whole Confederation was taken in 1869, with the following result— CENSUS OF THE ARGENTINE CONFEDERATION, 1869.2 Per cent. 887 11°3 1000 Natives ves ”-Foreign-born ... ” Total a 1,665,497 211,993 1.877.490 Nationality of foreign-born :— Ttalians Spaniards French English Swiss ... Germans Others 2 Total 71,442 34,080 29,383 700 R60 =~ 336 16°1 I5°3 51 2:8 2°4 247 1000 By far the most populous and most important province of the Confederation is Buenos Aires, capital La Plata. This province should be carefully distinguished from the capital city of Buenos Aires, which, like Washington in the district of Columbia. forms a separate territory of the 1 Almanach de Gotha. 2 Chiefly from various American states. vill GROWTH OF NEW NATIONS. SOUTH AMERICA 133 Republic. Of this province then a census was taken in 1881, which enables us to measure its progress since 1869. CENSUS OF THE PROVINCE OF BUENOS AIRES. 1869. Argentines ... 254,205 All others ... 63,115 317,320 Per cent 8o ro Details of Foreign-born -— [talians R729 Spaniards +536 French 2739 English RYE Germans TIA Swiss . All others .. 7,99 63,115 1000 1881. Per cent. 393,482 ay 747 133,099 - 253 h26,581 .. 1000 57,128 33,692 1,738 9.089 Ldn an 42°9 25°3 156 6'8 I 13 7'0 1000 9,277 135.099 We see that in this one province the foreign-born bave in 12 years doubled in numbers; they have also increased considerably in proportion to the whole population. . The increase has occurred in all nationalities, but more especially among the Italians (who have been trebled) and the Spanish. In 1881 the three chief Latin races, Italians, Spaniards, and French, made up 84 per cent. of the foreign-born, and one-fifth of the whole population. According to the Italian census of 1881, there were then 254,388 Italians resident in the Argentine Confederation. The population of the federal capital, Buenos Aires, has been ascertained at various periods. 1855 350 vee 91,548 } Quoted in Censo General de la Provincia L869 ce wv 177,787 de Buenos Aires, 1881. [887 ain ... 404,173 2 QCenso Municipal de Buenos Aires, 1887. Censo General de la Provincia de Buenos Aires, 1881. 2 ¢« Antiouo Municipio.” 5%; STUDIES IN STATISTICS CHAP. From these figures it will be seen that Buenos Aires differs from Washington in that it is one of the quickestgrowing cities of the world; a table on page 6 of vol. ii. of the Censo Municipal shows that its growth in the last 18 years, 7-3 per cent. per annum, exceeds that of Chicago and the other great cities of America. The great foreign immigration began in 1854; its offect is very clearly seen in the birth-places of the inhabitants. CITY OF BUENOS AIRES! BIRTH-PLACES. 1869. Per cent. 1887. Per cent. 94,968 .. 507 204,734 ... 47°2 92,158 49°: 228,641 ... 528 187.196 .. 1000 . 433,375 ... 1000 Argentines .. Foreign-born . While the Argentine population increased in 18 years by 116 per cent., the foreign-born population increased by 148 per cent. Another table of the census shows us that the proportion of females to males among the foreign-born has improved considerably during the same period. POPULATION BY SEX. CITY OF BUENOS AIRES. To each 100 males, there were females 1869. 1887. Among Argentines - i. 139 —- “. 116 Foreion-born ... . 45 \- hd . BIRTH-PLACES. CITY OF BUENOS AIRES, 1887. Natives of Argentine Italy ... Spain... France - Other Countries Per cent. 47-3 319 9-1 6 nM 100-0 Chiefly South Americans. L ¢¢ Actual Municipio” (including ¢ partidos 7). m1 GROWTH OF NEW NATIONS. SOUTH AMERICA 135 FOREIGN-BORN. CITY OF BUENOS AIRES. [6alians .. Spaniards French. English Germans Swiss - Austrians ... 544 Portuguese ... 798 1869. 14,233 14,609 [4 180 15 1887. 38,16¢€ 29,5@° Increase in 18 years. Absolute. Per cent. 93,933 . 212 24,953 171 5,851 41 = 31 88 CRE 84 2,127 I,53; .. 291 1,057 ven ar) ee 32 We learn also from the same census some interesting details as to these foreign-born. Thus, of the children in bhe capital of school-age (6—14 years) there were Argentines I. ... ... . 78 per cent. Foreign-born ... a . 29 “ I'hen, again, of the persons of eight years of age and upwards, the proportion who could read and write was 2s follows— CITY OF BUENOS AIRES. PERCENTAGE OF PERSONS ABOVE EIGHT YEARS OF AGE ABLE TO READ AND WRITE. 188%. Germans English French Argentines Spaniards [talians oe Dther nationalities ales J} 64 80 nl 4 amaler Go 54 ”E Total 100 00 88 83 79 61 78 The unexpectedly favourable position occupied by the English should be highly flattering to our educational reformers. It promises well for the Republic that the Argentines appear to be better educated than the Italians and Spaniards. In the eight years ending with 1878 the net immigrasion into the Argentine Republic amounted to 223,200 : [23 STUDIES IN -STATISTICS © CHAP. in the eight years ending with 1886 it had increased to 411,066. The following table shows the nationalities of the immigrants who landed at the port of Buenos Aires in the ten years ending with 1888. NATIONALITIES OF IMMIGRANTS WHO LANDED AT BUENOS AIRES DURING THE TEN YEARS 1879—1888. Italians Spaniards French Austrians Germans Swiss ... English Belgians Others . Total a 109,306 80,695 53,890 14,020 0,255 10,805 10,508 6,067 1,170 607.316 Per cent. 674 £3'3 8g 2'3 8 8 i Ic 8 Co ... 1000 It will be seen that the Italians form two-thirds of the whole number, the Spaniards and French being the next most numerous, these three nationalities together ~onstituting nearly nine-tenths of the immigrants. The migration from Italy has increased greatly in the last four years, but proportionately the increase in the case of France, Spain, and Belgium has been still more marked. From the facts quoted, it is plain that an extensive movement from Southern Europe to South America is taking place, and it seems likely that the future of the latter continent will lie in the hands of the Latin races. \ Reports of the Bureau of Statistics of the United States; also Almanach de Gotha. YX GROWTH OF NEW NATIONS. SOUTH AFRICA 137 CHAPTER IX. THE GROWTH OF NEW NATIONS (continued). SouTH AFRICA AND AUSTRALASIA. TaERE is so little statistical information in reference to the Cape Colony and Natal, that it is not worth while liscussing the matter at much length. The last census of the Cape Colony was taken in 1875, with the following results® :— CENSUS OF CAPE COLONY PROPER, 1875. European population . Native population ” Total ... er European Population. Natives of the Colony ... ,, Great Britain . Germany ... Holland ... ,» Other Countries Total ... wn Native Population. Kaffirs and Bechuanas ... Hottentots cer “es Fingoes sos os Malays ... -. - ce Mixed and others es ee . Total ... Co Lo 236,783 484,201 790.984 203,463 22,000 4,685 883 5,752 236.783 214,133 98,561 73,506 10,817 87,184 ... 484.201 Per cent. 32'8 672 LOO'G 8579 93 2‘0 ‘4 24 Fears 44'2 204 15°2 a 2'2 we» 180 Statesman's Year-Book. 8 13 STUDIES IN STATISTICS CHAP. The native whites are chiefly the descendants of Dutch, German, and French settlers,’ so that not only are the blacks twice as numerous as the whites, but among the latter the Dutch largely outnumber the English. The estimated populations of the dependencies of the Cape Colony were as follows in 1885 * :— Europeans. Griqualand West (1881) 16,927 Transkei ... ve 82: Griqualand East ... 3,06 Tembuland “on &. Total we oo 89,187 Per cent. ... ar Other. 32,174 T7332 0 Total. 49,101 119,552 96,180 122,638 - 387,471 100°0 In these territories it will be seen that the whites are in a very small minority. Bechuanaland, Pondoland, and Basutoland are estimated to contain a population of 846,000, of whom probably but a few hundreds are whites. It may be incidentally remarked, that in the Orange Free State the .census of 1880 gave a total population of 183,518, of whom 61,022, or 45'7 per cent., were whites, and 72,496, or 54'8 per cent., natives. Natol.—The population in 1886 was returned as follows 3 :—-Europeans Kalffirs ... Coolies ... 37,437 374,915 30,345 449.697 Per cent. 85 c.f TOO'OQ Emigration to South Africa is not so active as might he wished, Several caunses—the numerous native wars, ! The descendants of the Huguenots who settled at the Cape in the 17th century have long lost their language, and speak “Cape Dutch.’ See Problems of Greater Britain, by Sir Chas. W. Dilke, 1890. 2 Statesman's Year-Book. 3 Thad. X GROWTH OF NEW NATIONS. SOUTH AFRICA 139 the preponderance of Dutch in the white population, the ample supply of black labour, the great cost of the passage (until recently even greater than to our Australian colonies |)—have contributed to this result. But more than all these the vacillating policy of the Home Government, culminating in the surrender of the. Transvaal after the disaster of Majuba Hill, has destroyed our prestige, and caused a general feeling of uncertainty and insecurity. The Board of Trade Returns give the following figures :— EMIGRATION * FROM GREAT BRITAIN AND IRELAND TO ae SATE OF GOOD HOPE AND NATAL. SEVEN YEARS, Emigrants Natives of Great Britain and Ireland ee J 20857 Foreigners ... Sun 6,700 Total ... .. 47.008 . Immigrants. Net Emigration. 7,444 1,529 38,035 ... 8973 {zu In 1883, '84, and ’'85 the immigrants exceeded the emigrants, and the net result for seven years 1882-88, viz. 3,973,% was very trifling. Unless the amount of emigration and, what is more important, permanent settlement be very largely increased, and that soon, the “Dutch diffi culty ” and the ‘“ Kaffir difficulty ” are likely to endure, and not improbably lead to serious troubles in the near future. The emigrants of British origin were divided as follows + — 1 This table is constructed from the point of view of Great Britain; of course what we term ¢ emigrants” are at Cape Town “immigrants.” ? The net emigration from Great Britain and Ireland to the Cape in 1889 amounted to 9,966, as against 3,599 in 1888, and 2,034 in 1887. The great majority of the emigrants were Englishmen.— Board of Trade Return on Emigration and Immigration, 1889. 40 STUDIES IN STATISTICS English ... Scotch... aie Trish a os Total ... Lo. 35,688 3,702 909 40.299 CHAP. Per cent. 88's 9'2 2'3 1000 These proportions at any rate are satisfactory, and give no indication of an * Irish difficulty ” being added to the others. Insularity and ignorance in Downing Street seem to have made South Africa a plaything, and have much to answer for. Itis very doubtful whether, under the peculiar circumstances of the case, it was wise to give self-government to the Cape Colony in 1872. The annexation of the Transvaal in 1877 may have been unwise; but folly assuredly culminated in the surrender of the Transvaal in 1881, an act that has shaken British prestige throughout the land to an extent that it will take years to recover from ; but a certain class of our legislators do not seem to understand what prestige is, and its vital importance as a factor in dealing with inferior races. The recent discovery of valuable gold-fields in the Transvaal, and the consequent rush of immigrants, emphasizes in an unexpected manner the singular ineptitude of the surrender.! Should, however, these gold-fields hold their own, the predominance of English among the diggers, and the approach of railways on all sides, foreshadow possible changes, as a result of which once again the Union Jack may float over Pretoria. The charter wisely given to the South African Company opens another vista of great possibilities. 1 Sir Frederick Young, K.C.M.G. (4 Winter Tour in South Africa : Pethrick and Co., 1890), calls attention to the serious results of the absence of municipal government in Johannesburg and other Transvaal towns. and contrasts their condition with that of the towns in Natal. 1X GROWTH OF NEW NATIONS. AUSTRALASIA 141 AUSTRALIA AND NEW ZEALAND. Fifty years ago the population of Australia and Van Diemen’s Land was under 80,000,! and New Zealand had aot been colonized ; that is to say, a continent exceeding that of Europe in area comprised a white population about squal to that of Wolverhampton at. the present day. South Australia was colonized in 1836, New Zealand in 1840; Victoria was separated from New South Wales in 1851, and Queensland in 1859. At the census of 1881 the continent of Australia was found to contain 2,134,489 inhabitants, Tasmania 115,705, and New Zealand 489,933, making a grand total of 2,740,127 (excluding aborigines), or, if we also exclude Chinese, Polynesians, Malays, and other coolies, about 2,690,000 persons of Huropean races; ¢. e. a population larger than that of Paris, considerably larger than that of Norway, Denmark, or Greece, and equal to that of Switzerland ; sufficiently large, at all events, to completely disturb the “balance of power” in the South Pacific. It cannot be other than a source of the greatest satisfaction to Englishmen, that 95 per cent. of these people are essentially British, having been born either in the colonies or in the old country. A census was taken in 1886 in Queensland, and also in New Zealand ; these showed an increase in the white population in five years of 109,328 and 88,549 respeectively, or a total increase of very nearly 200.000 for these two colonies alone. The growth of Australia from the establishment of the first penal settlements at Botany Bay, New South Wales, towards the end of the last century, and in Van Diemen’s I Census, 1831. 142 Land in 1804, to the present time is unexampled. When America was settled, difficulties of locomotion by sea and land made any such rapid progress towards settlement quite impracticable. Where less than a century ago savages of a very low type reigned not only supreme, but absolutely without competitors, there are now a number of provinces enjoying all the advantages of the military power and the prestige of an old and powerful country, while at the same time, so far as domestic affairs are concerned, they have all the privileges of free states. New South Wales and Victoria are the most completely organized, have the largest populations (now exceeding a million in each colony), the largest revenue and expenditure, and the largest debts; Queensland is in process of rapid development, and has a great future before it; in South Australia progress has been less rapid, but doubtless its time will come ; Western Australia is but just emerging from the embryonic stage, yet it offers a magnificent field for emigration ; Tasmania has less capacity than the other colonies, but its progress, if less startling than that of some of them, has been steady. New Zealand has shown a most extraordinary rate of progress. In 1861 its white population was under 100,000 ; the recent census (1886) showed it to be 578,482; its revenue (in 1885) amounted to 3% millions, but its expenditure exceeded that amount by + a million, so that it is not surprising that, with a smaller population, and much smaller territory, it has a debt practically equal to that of New South Wales. Nevertheless, if only its finances be kept in prudent hands, there is no reason why its credit should not soon stand as high as it once did. ‘Whatever may happen in the future, whether the process of disintegration or federation prove the stronger, STUDIES IN STATISTICS CHAP, X GROWTH OF NEW NATIONS. AUSTRALASIA 143 there is scarcely room for doubt that the dominant power of the South Pacific will be British, even if not in name and allegiance, at any rate in race and language. POPULATION OF AUSTRALASIA (EXCLUDING ABORIGINES) AT VARIOUS PERIODS. The figures, unless otherwise stated, are taken from the various census reports, © 1821. Je 1841. | 1851. | 1861. | 1871. | 1881. L tm nn Iv. 7. vi. | vin Queensland ... New South Wales| { 20,783160,861 Victoria ... ... South Australia —- — Western a — | 500 I'asmania ves A 4°7716.954 New Zealand Total ax | 36,263]78,316] 213,176] 420,747 1,750.21211.910,432/ 2,740, 127 ; 8,575] 30,059 178,668] 350,860 77,345) 338,628 63,700] 26,830 4,622 14,837 70,130) 89,977 26,707) 99.021 218,525 749,825 861,566 279,865 29,708 115,705 489.033 1 Notes on this table, mainly from the Statesman’s Year-Book, 1888. New South Wales.—First penal settlement in 1788. In the next year the population was 1030, in 1810 it was 8,293. The first constitution was established in 1843. The number given in column IT is that of the census of 1833. Victoria.—Formerly Port Philip, settled in 1835; in the next year t contained 224 persons; separated from New South Wales in 1851. Queensland.— Formerly Moreton Bay, settled by convicts in 1825; he census of 1846 gave it a population of 2,257. It was separated from New South Wales in 1859. At the census of 1886 the population uwnounted to 322,853. Tasmania.—Formerly Van Diemen’s Land, the first penal settlement was established in 1804; a dependency of New South Wales till 1825. The number in col. I is from Stanford's Compendium of Geography, dustralasia, 3rd edition, p. 248; that in col. IT from the Stafesman’s Yewr-Book, 1866 ; that in col. ITI from the census of 1842 : that in column VI from the census of 1870. South Australic.—Colonized in 1836. The number in col. I11 is from ihe census of 1844. Western Australic.—Settled in 1829. The figures in columns 1V, V, and VI are from the censuses of 1848, 1859, and 1870 respectively. New Zealand.— Colonized in 1840. The number in ¢ol. IV is from the Statesman’s Year-Book. At the census of 1886 the population amounted to 578.482, 114 STUDIES IN STATISTICS CHAP. According to the statistics published by the Government statist for Victoria, it would appear, on the average of the eight years 1881-88, that the Australasian colonies collectively increase each year—(1) through excess of births over deaths by 65,000, and (2) through excess of immigrants over emigrants by 66,000, giving a total annual growth of 131,000. Particulars of this growth are given in Table H, from which we learn that New South Wales is now growing much more rapidly than Victoria, also that immigration into Queensland is very active. The Polynesian coolie immigration fluctuates extremely, from several thousands in the year to nothing; Chinese immigration has ceased. The loss of South Australia by emigration is due to a prolonged period of depression ; probably most of the emigrants have gone into other Australasian colonies. The experience of the last census showed that the official estimates of population founded upon the excess of births over deaths, and of immigrants over emigrants, gave results that were not realized by actual enumeration. Probably the emigrants are not all recorded, or the deaths, or both. We must therefore consider that the figures in the table somewhat overstate the real facts. In the case of Queensland, however, the estimate was very close, especially as the periods do not quite correspond, thus :— INCREASE OF QUEENSLAND IN ™7 7 7TVE VEARS 1881-851 By excess of births ... By excess of immigrants Total estimated increas: Census Population, May 1, 1886 April 3, 188. Actual increase in 5 years ... 4 Difference ... son 2,634 or 2'3 per cent. 1 Report on Australasian Statistics by the Government Statist of Victoria. and Census Reports. x GROWTH OF NEW NATIONS. AUSTRALASIA 145 In New Zealand the estimate was less satisfactory, thus :— INCREASE OF NEW ZEALAND IN THE FIVE YEARS 1881-85.1 By excess of births ... oe o By excess of immigrants - Total estimated increase ... Census population (whites), 1886 -... ans ” 1881 ... ws Actual increase in 5 years Difference ... _ 7.818. or 8-8 per cent. TABLE H.—AVERAGE ANNUAL INCREASE OF THE AUSTRALASIAN COLONIES. FOR THE EIGHT YEARS 1881-88 2 New South Wales ... Victoria or don Queensland ... Pp New Zealand ae South Australia ... Tasmania ... ce Western Australia .., Total vee o ee By excess of Sirths over deaths, 20,308 15,572 5,915 13,346 6,940 2,452 637 By excess of ; mmigrants over! Total estimated emigrants. increase. 27,562 18,436 16,199 2,728 ~~ 1,0473 1,402 961 47,870 34,008 22,114 £6,074 5,893 3,854 1,508 65.170 1 66.941 137.417 The quality of the population of Australasia is no less satisfactory than its rapid growth. (See Table J.) Of all our colonies Australia is most nearly a reproduction of the mother-country. In both Australia and New Zealand the proportion of foreigners is insignificant. In Australia the aborigines are few by comparison, and rapidly dying v Report on Australasion Stutistics by the Government Statist of Victoria, and Census Reports. } j * Report on Australasian Statistics by the Government Statist of Victoria. * Excess of emisrants, 46 STUDIES IN STATISTICS CHAP: out (the last Tasmanian died twelve years ago), while in New Zealand the “Maori” difficulty seems to be at an end, and even Lord Macaulay, were he now alive, would scarcely be likely to prophesy great things of the New Zealander of to-day. In Table K are given the proportions of natives of Scotland and Ireland to the natives of England and Wales in several countries; it will at once be seen that the Irish, though relatively twice as numerous in Australasia as in the United Kingdom, have a far lrss predominance than in Canada, @ forteori than in the United States, where they are in proportion to Englishborn people twice as numerous as in Canada. The Scotch, like the Irish, are more enterprising than the English in pushing their fortunes into remote parts of the world,’ and New Zealand seems to have especial attractions for them, though it is not - quite such a favoured colony as Canada. But enterprise is not the sole reason that sends the Scot abroad; like the Irishman, he has to admit that his country is poor in natural resources, and if he does not find an opening in England (there were a quarter of a million natives of Scotland in England and Wales at the last census), is apt to go far afield in search of ane. A point to be noted in Table J is the disparity of the sexes. This, as might have been expected, is but trifling among the natives of Australasia, but is considerable among the natives of England and Scotland, and indeed among the natives of every country except Ireland. 1 For each 1000 English and Welsh in the United Kingdom -in 1881 there were 7 in Canada, 20 in Australasia, and 30 in the United States. For each 1000 Scotch in the United Kingdom there were 31 in Canada, 41 in Australasia, and 46 in the United States. Similarly for each 1000 Irish there were 32 in Canada. 45 in Australasia, and 317 in the United States TABLE J.—BIRTH-PLACES OF THE INHABITANTS OF AUSTRALIA, TASMANIA, AND NEW ZEALAND AT THE CENSUS OF 1881. } Australia and Tasmania. - New Zealand.2 Males. Females. Persons. | Per Males Females. Persons. | Jer Total Australasia Persons. Natives of — Australasia ... 714,403! 710,240 1,424,643 633 121.192 119,489 240,681 | 49°2 1.665.324 499,922 151,027 261,996 14,644 6,274 42,203 6,941 7,583 4,401 2,740 2,363 43,430 16,741 14,538 Total ve vee} Lz 2,250,194 220,328 489,933 . 100°c 2,740,127 Aborigines 8 ~- | 17,560] 14,140 31,700] 1°44 24,368 | 19,729 | 44,097 831 75,797 ! The facts are chiefly taken from a table in the Census of Victoria, 1881, General Report, p. 47, but some details are taken from the censuses of the several colonies. 2 From the Census of New Zealand, 1881, p. 191. 8 The numbers for Australia are chiefly estimates, and very defective. There are no Aborigines in Tasmania, The percentages of the Aborigines are calculated on the total, including them. England and Wales Scotland a wih [reland wae vis Other British Possessions United States as Germany ... Ben Denmark ... ve. Sweden and Norway France yp Switzerland ... .. Italy... oe ee China I. ves Other countries .. Not specified and at sea 229,017 56,656 106,542 7,196 4,017 24,161 3,617 4,594 2,770 2,044 1,720 38,241 12,280 7.399 149,718 41,618 106,091 3,434 1,416 13,223 1,125 454 783 364 160 156 766 5.058 378,735 93,274 212,633 10,630 5.433 37,384 4,742 5,048 3:553 2,408 1,880 38,397 14,054 12,380 16°8 70,704 31,005 27,666 2,429 637 3,188 1,409 1,872 614 269 385 017 950 1.268 50,483 21,748 21,697 1,585 204 1,631 | 790 663 234 63 98 16 737 890 121,187 | 52,753 49,363 4,014 841 4,819 | 2,199 2,535 848 332 483 5:032 247 10°8 10°1 8 +H v x aN A ~ N 148 STUDIES IN STATISTICS CHAP TABLE K.—PROPORTION OF NATIVES OF SCOTLAND AND NATIVES OF IRELAND TO 100 NATIVES OF ENGLAND AND WALES IN 1880 AND 1881. Natives of Scotland. Natives of Ireland. [n United Kingdom... ... 14:7 | In United Kingdom ... ... 234 , United States... ... 22-8 New Zealand ... ... 40-7 Australial ,.. ... 259 Australia? I J | New Zealand... ... 435 Canada ... ... ... 1094 Canada ... ... ... 679 United States... ... 2486 As we saw in treating of the United States, it is remarkable that the Irish do not leave their women behind them when they emigrate ; either the women are less averse to emigration than those of other countries, or, what seems suggested by the facts, the ties of family are stronger. The figures given in the Board of Trade Returns of Emigration are very striking. It appears that, taking the ten years 1878—1887, for every 100 adult Englishmen that emigrated 53 adult English women left these shores ; for every 100 Scotchmen 55 Scotch women ; but for every 100 adult [rishmen as many as 95 adult Irish women. The proportion of children to total emigrants was as follows— English, 19-2. Scotch, 22:2 Irish, 13-5 per cent, In Australia and New Zealand there is a decided deficiency of women ; for every 100 males there are in the former continent but 85 females, and in New Zealand but 82. In Canada the disproportion is not so great, yet there is a total deficiency of 53,000 women. The deficiency is greatest in Manitoba and British Columbia, amounting together to 18,000. It seems a great pity that more cannot be done to remove our surplus women to countries where there is a demand for them. 1 Among the Australian colonies the English-born are most numerous in Western Australia and South Australia, the Scotch in Victoria and Queensland, the Irish in Queensland and Victoria, NATIVES OF ENGLAND, SCOTLAND & IRELAND ENUMERATED IN VARIOUS COUNTRIES AT THE CENSUS OF 1880-81. ALSO PERSONS OF FRENCH ORIGIN IN CANADA, ALSO NATIVES OF GERMANY AND MEN OF, COLOUR” IN THE UNITED STATES. Absolute numbers.20000 to each small square. Percentages measured vertically. each division gives 2 per cent. 2 3 & £3 £2 UNITED KINGDOM. ud 5 7 *ER CENTAGE 100 rE J 20 SE Ee MT OS ot rate = a NSH ERS BE HVE ES Tee salam SEDER RG ENE NN rhe SITTSRRENE ET TE TERT ORE y 5 Li il lL REI ED ) = % A 57 A A Sy SA A 7 i i a . re 2 lina RATT A 7% Ci a pon ord gear 50 oo 7 “ i i 7 aan me ; A a wl ie am Ci a A yet 7 in ow i 4 a ay oo - 0 30 20 nr i hh 0) 5 Oa 7% 204 V0 0 477 7 BO a es a Yd wi” 1% a 100 a0 80) te | — i -70 G0 50 Ci 40 i 20 120 iD 0 MILLIONS 0 NATIVES OF ENQLANNZZ TZ, gd A) Ld IRELAND Stun ov GERMANY OF FRENCH ORIGIN 224 i= . 3 3 COT AND ROO MEN OF COLOUR FF X GROWTH OF NEW NATIONS. AUSTRALASIA 149 Mr. Ravenstein! has called attention to the interesting fact that, although men are more prone to ultra-migration than women, on the other hand, women are more addicted than men to that form of intra-migration which he terms “short distance migration.” Moreover, he finds that this holds good in most countries. TABLE L.—NATIVES OF THE UNITED KINGDOM, 1880-81. United K. + Per 1881. gent. Canade. Australasia. o. Per _ Ter 1881. cont.| 1881. ient. Tnited States. a. ‘er aant Total Per cent, oe |25,017,027] 72°5[169,504] soc dad 4’ THE 26+026,432,491) 68-3 Scotch | 8,673,615] 10°6/115,062] 24°5 191,077) 16's] 170 1 4,109,220! 106 Irish - - 5,843,406] 16-9/185,526 a 28°71,854,571) 670; 8,145,499] 21°1 United K.124,534,048'r00°01470,092 10001912, 945/000 2.770.685 1000 38,687,770 1000 It is not possible to number accurately the people of British and Irish stock now living in the world, but in Table L are given the numbers of natives of the British Islands in the principal English-speaking countries, as enumerated at the last census. After the United Kingdom there are most of these in the United States, then in Australasia, Canada coming last. Some of these facts are exhibited in the diagram (Plate XIV.), which shows by various colours the number and percentage of the natives of England, Scotland, and Ireland in the United Kingdom, New Zealand, Australia, Canada, and the United States respectively; also the persons of French origin in Canada, and the “men of colour” and natives of Germany in the United States. The uncoloured spaces represent in the case of the United Kingdom British subjects born abroad, colonial-born living in the United Kingdom, and foreigners of various 1 The Laws of Migration, by E. G. Ravenstein, F.R.G.S., Journal of Royal Statistical Society, vol. lii, (1889) pp. 249. 253. &e. 150 STUDIES IN STATISTICS nationalities. It will be noted, that the total number of persons living in the United Kingdom in 1881, but not born in it, is much smaller than might be expected; it has no doubt increased since the last census, but the fact that an enormous majority of the total foreign population has settled in London makes it unduly obvious to Londoners, and hence leads them to exaggerate its importance. The uncoloured spaces in the case of the colonies represent mainly persons of British and Irish stock; in the case of the United States also persons of British and Irish stock (the latter largely preponderating), but mixed with them a large white population recruited from every country of Europe—Germans and Scandinavians being by far the most numerous. The diagram incidentally shows the difference in character of a new country like New Zealand principally peopled by recent emigration and a more settled country like Canada. It also shows very clearly the immense total population of the United States as well as the fearful incubus of the negroes of the South. The number of persons of African race in Canada would be represented by one small square. Looking at this migration question from the point of view of the British Empire, it is important to bear in mind that the centre of gravity of that empire is slowly shifting. The populations of our great self-governing colonies in North America and Australasia increase much more rapidly than that of the mother-country; so that if the population of the United Kingdom at each census be taken as 100, it will be seen that the united populations of British North America and Australasia, which in 1841 corresponded to 7, in 1881 corresponded to 21. © CHAP: X - GROWTH OF NEW NATIONS. AUSTRALASIA ISL PROPORTION BORNE BY POPULATION OF NORTH AMERICAN AND AUSTRALASIAN COLONIES TO THAT OF MOTHER-COUNTRY. Census, 1841 1851 1861 1871 1881 United Kingdom. 100 “ng _— 177 100 Colonies. LL 16 18 21 It may fairly be assumed that this disparity will still further decrease, and the question is now exercising the minds of many of our ablest public men of both political parties, whether it will be possible much longer for our so-called Imperial Parliament, which represents only the inhabitants of the United Kingdom, to direct alone the destinies of our wide-reaching empire. The alternative vaguely shadowed forth is, that imperial defence, peace and war, postage, and other matters which interest all the empire alike, should be dealt with by some body on which the various colonies should be represented, the local administration of Great Britain and the various colonies remaining untouched in the hands of the local parliaments. This grand conception, which appears to be steadily gaining ground, is termed Imperial Federation. Without discussing either the advantages or difficulties of any suggested scheme, I will merely add, as it appears to be germane to my present subject, that the composition of such a Federal Parliament would be—on the basis of one member for every 100,000 of white population as enumer ated at the census of 1881—somewhat as follows 162 STUDIES IN STATISTICS IMPERIAL FEDERAL PARLIAMENT. Members. 260 37 52 England and Wales Scotland ... - Ireland ... .. Canada... ve Newfoundland . ... West Indies -_ Cape of Good Hope Natal vee “e New South Wales Victoria ... vin South Australia ... Queensland _— Western Australia Tasmania, _ - New Zealand ... cer vi Total Co Co Co 2 LOT CHAP. Per cent, 61 9 rag TOO The distribution of power in the present House of Commons 1s as follows :— Per cent. England and Wales = 495 . 74 Scotland i. 72 ee 11 Ireland ina 103 wo 15 870 100 It will be seen that in the suggested body the English members would still command a majority. India would have to be dealt with specially; probably a few members might be nominated by the Council to represent Indian interests ; in any case it could not be represented on a basis of population. In concluding this long review of the growth of new nations, I shall say but few words. We have seen that a strong tide of migrants flows from Germany, Ireland, Great Britain, and the Scandinavian countries to the great North American Republic, and that the main stream IX GROWTH OF NEW NATIONS. AUSTRALASIA 153 is joined by a tributary rivulet of greater or less volume from nearly every country of Europe. We have also traced a weaker but a rising tide flowing from Italy and Spain to Brazil, Montevideo, and Buenos Aires, the stream to Brazil being joined by a goodly river from Portugal, those to Uruguay and the Argentine Republic by small streams from France. There are yet two other tides both flowing from the British Islands, one to the Dominion of Canada, and the other to the great islands of the South Pacific, a relatively large number of Irish and Scotch going by the former, of English by the latter. We have also looked into the composition of the new nations built up as the result of these tides. In the United States we found a people largely British in origin, but comprising in its northern portions at least 5 millions of Germans and nearly 5 millions of Irish, with over half a million of Scandinavians, and a like number of French; but sadly hampered in the south by 61 millions of “men of colour.” We found, moreover, that these various elements do not mix readily, and saw reason to believe that trouble might consequently arise in the more or less distant future. In Canada we found a people one-third French ; quiet, plodding, exclusive, wishing to be let alone, and though not mixing with the British, seemingly not anxious to pass under either French or American rule. The other two-thirds of the people of the Dominion we found to be vigorous and homogeneous, with a large proportion of Scotch, and as yet fortunately with but small foreign admixture, In South Africa we found a state of affairs as yet ansettled—the British colonies containing a majority of blacks, and of the few whites a majority of Dutch extraction and Dutch in sympathy. E54 STUDIES IN STATISTICS CHAP; In Australia and New Zealand we found rapidly growing populations of a thoroughly British type. The Irish fewer in proportion than in North America, and but few foreigners. The native troubles seemed to be at an end, and the Chinese difficulty ” not as yet serious. Germans we found everywhere, but only in large proportion in the United States, though even there not strong enough to dominate the British element. The emigration from Sweden, Norway, and Denmark we noted as not only large in itself, but bearing an exceptionally large ratio to the natural increase of those countries. Moreover, the settlers were found congregated together in one restricted portion of the United States. The movement of the Latin races, especially the Italians, appeared to be of increasing importance, and pointed to the possible predominance of Italy in South America in the next century. Finally we attributed this mighty movement of modern peoples to two causes— (1) a redundancy of population in Europe pressing closely upon the means of subsistence; (2) vast areas of virgin soil of great fertility, much of it lying under a temperate climate, brought almost suddenly within reach by the railroad and the steamship. The lessons to be learned from the long array of facts seem. to me to be as follows :— 1. The stream of migration should not be checked (save perhaps where it flows into London), but rather encouraged to flow on. 2. Where practicable the flow of British emigrants should be directed to British soil, especially to Canada and to South Africa. 3. Our colonies should not artificially stimulate the migration of foreigners to their shores. xX GROWTH OF NEW NATIONS. AUSTRALASIA 155 4. All legitimate means should be adopted to encourage the use of the English language by subjects of the English crown, and by citizens of the United States. The’ valour and determination of our ancestors have placed under the guidance of Englishmen this great movement in at least three continents ; what they accomplished by the arts of war, let not Britain fail to maintain by the arts of peace; nay more, let her insist, peacefully if possible, but by force of arms if need be, that she means to carry out her trust to the end. 156 STUDIES IN STATISTICS CHAP, CHAPTER X. THE GROWTH OF MODERN CITIES. AMoxGsT the results of modern civilization the tendency of people to flock into cities is one of the most striking. It is observed alike in old countries with increasing populations such as England and Germany; in a country with a stationary population like France ; in a new but populous country like the United States ; but perhaps most strikingly of all in very newly settled countries with small populations, such as Victoria, New South Wales, Canada, and even Manitoba, one of the very newest. The causes of this are manifold—on the one hand the application of machinery to agriculture has diminished the demand for agricultural labourers, and at the same time the increased use of meat as an article of diet has greatly augmented stock-farming, which requires few hands; on the other hand, the applications of machinery and steam have led to an unexampled development of manufacturing industries attracting labour from all parts to the chief centres of industry, the migration of the labourers being stimulated by education and the newspaper press, and immensely facilitated by improved means of communication. ! For facts proving this to be the case in Germany, Holland, Austria, Hungary, Sweden, France, and Italy, see Mr. E. G. Ravenstein’s valuable paper on “The Laws of Migration,” Journal of the Royal Stutistical Society, vol, lil, (1889) p. 241. THE GROWTH OF MODERN CITIES 157 Commencing with some of the youngest countries. In 1870 the woods and prairies of the far north-west of Canada passed from the rule of the “ Governor and Company of Adventurers of England trading into Hudson’s Bay,” under the sway of the Government at Ottawa—exactly 200 years after the grant of the charter to the company by Charles II. At the census of 1870 the population of the new province of Manitoba was but 12,228, half of them Indians, and Winnipeg, its present capital, but a village of 241 inhabitants. In 1881 the population of the province was 65,954, of the capital 7,985. At the census of 1886 the population of Manitoba was 108,640, and that of Winnipeg 20,288. Where twenty years before were but the establishment of a fur-trading company, Fort Garry, and a cluster of wooden houses with a few Indian tepes, there now stands a flourishing city, with all the paraphernalia not only of municipal government, but with a provincial House of Commons. Nevertheless, the good people of Winnipeg must have been disappointed to find their fellowcitizens in 1886 many thousands fewer in number than sanguine Manitobans had been wont to estimate them. The point I wish to call attention to is, that the capital contains one-fifth of all the inhabitants of the province. Queensland was thrown open for settlement in 1842; it was separated from New South Wales in 1859 ; in 1861 it had a population of 30,059 ; at the recent census (1886), with a total population of 822,853, Brisbane, its capital, had no less than 32.567 inhabitants. or including suburbs 79 649 New Zealand, with a white population in 1886 of 578,482, has no city larger than Auckland, with 57,048 inhabitants, but its four chief towns contain together nearly one-third of its total inhabitants. New South Wales had in 1881 just over three-quarters (58 STUDIES IN STATISTICS CHAP. of a million of people, of whom more than a quarter (220,984) lived in the capital and its suburbs! At the same time the total population described as urban (which, however, included all towns and villages containing 100 inhabitants and over) greatly exceeded that described as rural.’ But perhaps Victoria is the most striking case. Its population in 1881 was 861,566. Its four largest towns with their suburbs contained more than two-fifths of the whole number, Melbourne alone accounting for 282,947.° Melbourne is the largest city in the British colonies, and probably the most remarkable in its growth, but its especial distinction amongst capitals is that it comprises within its boundaries a larger proportion of the inhabitants of the country it dominates than any other. Montreal (140.747) and Toronto (86.415) are fine I CENSUS OF NEW SOUTH WALES, 1881. 12Q1 (ar Per cent. of Inhabitants.| whole population. Inhabitants. | Per cent, of whole population. Sydney city (excluding shipping) ... .. “rs Sydney with suburbs ... Total Urban population... ,, Rural 2 - 74,423 134.736 151 13:8 9%67 29:5 4677 433,391 57-9 53-3 | 314850 | 42-1 234,162 967 417 2 Census of New South Wales, 1881, p. xvL 5 ORENSTIS OF VICTORIA. 1881. 1871. 1881. Increase, Melbourne city en 54,993 ... 65,859 ... 10,886 with suburbs ... 206.780 ... 282.947 ... 76.167 In 1861 ¢“ Greater Melbourne ” comprised 26 per cent. of the whole population ; in 1871 a yet larger proportion. 29 per cent.: and at the last census 33 per cent. < THE GROWTH OF MODERN CITIES 159 cities of rapid growth, but the first contains only about one-tenth of the inhabitants of the province of Quebec, and the second not one-twentieth of those of Ontario. But the country par ewcellence of large and rapidly growing cities is the United States of America. In the following table are given particulars concerning the twelve American cities in which the absolute increase of population in the thirty years between 1850 and 1880 has been greatest, and also of four much smaller cities in which the rate of growth has been extraordinarily rapid ; in Chicago it has been hoth absolutely and relatively marvellous, INCREASE OF CERTAIN AMERICAN CITIES IN THIRTY YEARS. Census 1850.| onus 1880. | Absolute, Relative Increase. New York, with Brook- ) lyn, Jersey City, and | Hoboken f Philadelphia Chicago ... St. Louis ... Boston .. San Francisco Baltimore ... Cleveland ... Cincinnati ... Buffalo .,.. Pittsburg ... Washinoeton 621.909 1.694. 6R3 1,302,774 31 fold 340,045 29,963 77,860 136,881 34,776 169,054 17,034 115,436 42,261 16,601 40.001 847,170 | 503,185 350,518 369,839 233,959 | 339,313 160,146 255,139 155,134 156,389 147 9293 507,125 473,222 272,658 225,958 199,183 163,259 143,112 139,703 112,873 | 109,788 107,202 25 168 45 1 27 67 2:0 9-4 2-2 37 3-4 a.m Kansas City Toledo i. Minneapolis St. Panl 1,418" | 3,829 2,564 1 1 aagq 55,785 50,137 46,887 A]. 472 51,367 46,308 “gyn 126 131, 183 310 2 A0. 1% The city of Buenos Aires, federal capital of the Argentine Confederation, has increased in thirty-two years from 91,548 to 404.178. or 44 fold. { Census of 1860. i6o STUDIES IN STATISTICS CHAP. It must not, however, be supposed that the rapid growth of cities, or even the birth of new cities, is confined to the New World and the Antipodes. ENGLAND AND WALES. GROWTH OF POPULATION OF LARGE TOWNS IN FIFTY YEARS. Increase. Absolute. | Relative. 2,161,489 23 fold 430423 | 31 31 London ... ee Liverpool with Birkenhead } Birmingham with Aston Manor | Manchester with Salford - } Sheffield - Leeds ... ys Newcastle with Gateshead } Bradford ver West Ham es Nottingham ... Bristol ... - Hull ... a 233.148 91,692 193 393 517,649 984,508 309,119 911.162 284,501 192,816 185,726 142,130 22 31, 25 31, £1, ui, | 2-4 ’ 20, 3-0 I” 110 63 145 985 959, 69.032 44,198 11,580 76,190 104,408 51.911 183,032 128,953 186,575 006,874 154.240 138,834 117,373 110,385 102,466 102,220 Cardiff ... ee Croydon I. Barrow-in-Furness (Dalton sub-district Middlesborough Ystrad-y-fodwg ... Norwich . Stockport oe | Bath ... ah 7,545 19.447 82,761 78 053 78,216 66.506 4.333 63.037 568 2145 55,934 55 639 58,704 55,366 £2,487 61,116 40,603 50 800 87,842 59,553 51.814 26,726 18,950 1.014 144 147 1-02 In the above table are given particulars concerning the increase in 50 years of twelve cities and large towns of England in which the absolute increase has exceeded one hundred thousand, also of three English and two Welsh towns in which the increase, though moderate absolutely, has been relatively very rapid; thirdly, of three towns THE GROWTH OF MODERN CITIES 161 which have altered remarkably little in the half century. It will be noted that the large towns of England have grown about as much in fifty years as those of the United States in thirty ; but even America cannot produce anything to outdo the growth of London, although it must be admitted that in the last thirty years the decennial increase of New York and its associated cities has approached that of London. The growth of our so-called metropolis during the present century has been treated ably and in great detail by Mr. R. Price-Williams,! and from his paper I have taken the figures to compile the following table, which shows the remarkable growth of certain districts of London during the last fifty years— Poplar . Hackney ... Camberwell ... Lambeth ve Wandsworth... Kensington ... Islington Co £831. 25,066 34,597 28,231 87,856 33,090 35,442 37.316 L8sL. Absolute. Tnerease. lative. 156,519 6-2 fold. 186,462 -I,G vd ,, 186,593 258,362 6% , 253,699 165,843 29 210,434 177,344 64 270,369 234,927 - 76 ,, 282.865 245,540 76 .. The parish of Battersea in the Wandsworth district, numbering 5,811 inhabitants in 1831, increased by no less than 101,951 in the fifty years; and of this increase no less than six-sevenths was in the twenty years 1861-81. West Ham and Croydon mav be considered as suburbs of London. Aston Manor and Birkenhead were small villages at the beginning of the century, and were little more in 1831, but they must be regarded as suburbs of Birmingham and Liverpool rather than as independent growths. Large towns seem to grow with abnormal rapidity, like - Journal of the Statistical Society, vol. xviii. p. 349. 162 those plants that send out in all directions ‘‘ runners” which root at every joint, so that at length it is difficult to distinguish the mother-plant from its offspring. They have, however, one point in which they differ from plants— all soils seem equally fertile and adapted to their growth. The most remarkable development de novo of a large bown in this country is certainly that of Middlesborough. A tiny village with but 412 inhabitants in 1801, it had by 1831 only increased to 568; but in the next ten years its people were multiplied by ten, the census of 1841 giving it 5,917 inhabitants. In the next ten years only 2,000 citizens were added; but in the ten years between 1851 and 1861 the population more than doubled, and the same extraordinary phenomenon repeated itself in the next decade. Between 1871 and 1881 its growth was more reasonable, from 39,882 to 55,934, an increase of 16,052, or 40 per cent. The growth of Barrow-in-Furness is scarcely less remarkable. It was separated in 1876 from the sub-district of Dalton. At the census of 1831 the Dalton subdistrict, comprising several villages and a number of hamlets, numbered only 4,333 inhabitants, and in 1851 but 6542. Then it began to grow in earnest; at the next scnsus its people numbered 11,243; in 1871 no less than 30,099, having increased between four and five-fold in twenty years. Not content with this, it doubled again in the next decade! The result was, that after throwing off the new sub-district of Barrow-in-Furness with 47,259 inhabitants, what was left of the Dalton sub-district in 1881 still contained 15,778, or nearly four times the population of the whole sub-district in 1831. In still more recent times, in the once quiet vales of Glamorganshire there has sprung up a place with the alarming name of Ystrad-y-fodwg. The population of STUDIES IN STATISTICS CHAP. THE GROWTH OF MODERN CITIES 163 that cacophonous village numbering at the beginning of the century 1,285, had reached 3,824 in 1851; then the development of the mines led to a rapid growth; in the next ten years the population was more than doubled ; between 1861 and 1871 it nearly trebled, and during the next decade it again considerably more than doubled, so that the Ystrad-y-fodwgites at the last census numbered 55,632, and they are doubtless still multiplying rapidly, and before long the congeries of mining villages which already constitute an “urban sanitary district,” will be completely organized into a municipal borough, with mace and chain and all other essentials. Of our large towns Bath is the only one which appears to have passed its prime (so far as numbers are concerned), for whereas in 1851 it had 54,240 people, it had in 1881 only 51,814. It is true that at the last census Manchester had nearly 10,000 fewer citizens than at the last but oue, but then Manchester, like the city of London, is filled up and has outgrown its limits, so that its increase of people must be enumerated in the suburbs—not within the ancient boundaries.’ A few words as to the growth of great cities on the continent of Europe. In France, where the population, as is well known, increases but very slowly, we nevertheless note the growth of large cities. This growth is, however, not as in England dependent mainly upon the two factors—(«) natural increase of the town population, and (b) migration of the superfluous population from the country. In France, according to M. Toussaint Loua,’ the natural increase of the rural population oreatly exceeds that of the urban. indeed in a Y Journal de la Société de Statistique de Paris. March, 1885. ? The boundaries of the municipality have been recently enlaraed so as to follow the retreating ponulation. 164 majority of the towns the deaths exceed the births [Paris and Lille are the most notable exceptions to this rule]. The population of many of the departments is increasing ; in seventeen departments a decrease has shown itself with rifling exceptions at every successive census from 1851— 1886.) Where the population of the rural districts is not decreasing it is increasing but very slowly. I would call particular attention to the fact, that over a large extent of France, republican and democratic as it is, with its sub-divided land, and its much-talked-of petite culture, a real depopulation of the rural communes is taking place, although there it can surely not be maintained that unjust land-laws and a grasping aristocracy oppress the peasantry. According to M. le Vicomte Rovric de Beaucaire,” there js in Germany a similar though not so marked a migration from the rural districts into towns; General Walker, the Director of the United States Census, told me that the same movement was very obvious in America. Any impartial mind will infer that the tendency thus generally apparent to quit the country homestead or-village, and seek employment by preference in a large town, is due to other causes than the size of holdings or the laws regulating the tenure, transfer, and inheritance of land, since where all these circumstances vary we find the same migration going on. I will now give a few instances of the growth of large towns in France. STUDIES IN STATISTICS CHAP. In the North-east : Yonne, Jura, Haute-Saone. In the North-west : Eure, Orne, Calvados, Manche, Sarthe. In the South-west: Tarn-et-Garonne. Lot. Lot-et-Garonne, Gers. Arigge. In the South-east : Drome, Vaucluse, Basses Alpes, Var. Bulletin du Ministore de I Agriculture de France. February, 18886. THE GROWTH OF MODERN CITIES 163 FRANCE. POPULATION OF CERTAIN LARGE TOWNS. Paris cee Lyons ... Marseilles Bordeaux Lille ee RK Toulouse 115,229 Nantes ... 113,625 St. Etienne 8 992.9250 1861.1 1,696,141 318,803 260,910 2730 1886.7 23° “A0 nN 3% *8% 73.272 147,617 127,482 117.87H Increase in 25 years. Alito, Per ord 248 409 38-2 er,127 26-1 19,533 43 5.7232 47-8 56,445 42-8 34,388 30-4 13,857 12:2 25 625 27:8 Another point should be kept in mind when considering the increase of French towns. At the census of 1886 France contained 1,126,631 foreigners, or 2'9 per cent. of the whole population, a proportion that is not approached in any other European country. 43 per cent. of the foreign-born were Belgians, 23 per cent. Italians; the nationalities next most numerous were Germans, Spaniards, Swiss, Dutch. and English.* Belgians [talians Germans .. Spaniards ... Swiss wi Dutch _ ... English, Scotch, and Irish Nthera 182,261 264,568 "00,114 79,550 78,584 37,149 36,134 18.97] The Belgians, Spaniards, and Italians are chiefly found in the departments adjoining their native countries ; the nther foreigners are mainly congregated in the capital. In Belgium, as might be expected, there is a rapid orowth of large cities. of which the following are examples— Y Census of France, 1881. Stutesiman’s Year-Book, 1888. 8 The population of St. Etienne diminished between 1881 and 1886 by close upon 6,000 persons, or 4-8 ver cent. 4 Densus of France. 1886. C6 STUDIES IN STATISTICS CHAP. GROWTH OF BELGIAN CITIES: Increase in 34 years. Census 1846. Census 1880. Absolute. Per cent Brussels and suburbs 188,458 394,940 206,482 109-6 Antwerp ... es 88,487 169,112 80,625 91-1 Ghent vu ... 102,977 131,431 28,454 27-6 Lidge ee ves 75,961 123,131 47,170 62-1 One might go on indefinitely reciting the statistics of the growth of European cities, but it will be suflicient to give the facts for half a dozen German cities, the Austrian capital, and the Hungarian capital —well known to some of us as the abode of the famous statistician, M. J. Korosi, who has written largely on the growth of modern cities. GROWTH OF GERMAN AND AUSIRIAN CITIES.2 Berlin ... Hamburg ... Breslau op Miiuchen ... Dresden... Diisseldorf ... 1858. 163,645 132,440 129,747 137,095 117,750 46.846 1885. 1,315,287 305,690 299,640 261,981 246,036 115.190 Increase. Absolute. Pur cent. 851,642 184 173,250 131 169,893 131 124,386 91 128,336 109 63,341 146 132 1851, 1881. Buda Pest ... 178.062 370,767 192,705 108 1857. 476.222 1880. 1.103.857 [ believe the true explanation of the remarkable growth of cities in all parts of the world in recent times is that, under modern conditions, with improved culture, and above all improved means of communication, a much smaller fraction of the people is able to provide—and U Annuaire Statistique de la Belgique, 1888. 2 These figures are all taken from the Almunach de Gotha, except the last, which come from Die Hauptstadt Budn-Pest 2m Jahre 1881. von Joseph Korosi, § 7. THE GROWTH OF MODERN CITIES 167 provide more amply than in the past—the food required by the whole body. Then again, those set free from the pursuit of agriculture apply themselves to. supply the innumerable new requirements of a people that is living ap to a standard of comfort far higher than that which contented their fathers. 163 STUDIES IN STATISTICS CHAP. CHAPTER XL THE POPULATION OF LONDON AND ITS MIGRATIONS. As far back as history extends, we find men living in large cities, some very ancient ones, such as Nineveh and Babylon, large even compared with most of the capitals of modern Europe ; but it may be asserted, with little fear of contradiction, that such an agglomeration of human beings as is now living, toiling, suffering, pleasuring, entering the world and quitting it, to north and south, west, and more especially east of us, making up modern London, is absolutely without precedent in ancient and modern civilization. Numbers fail utterly to give us an idea of its vast size, which can only be grasped by comparison with other populations. The table shows the populations of the four largest cities of the Western World, together with the twelve countries of Hurope that have fewest inhabitants, British North America, and our colonies in Avstralacia. POPULATION IN MILLIONS, 1880—1881 Belgium ee soa ... B58 Australasian Colonics Roumania, ... i ... B'4 Switzerland nh Ireland wos .. ... B92 Paris... Sweden op er . 46 Denmark Canada and Newfoundland £ % Greece Portugal . er ... <3 Norway Holland . A. ... 47 Servia London I. ee. 32 Few You Jeotland » L _. 53% Bera 2:7 2:9 2:3 2:0 2:0 . T.0 i ¢ 1 A paper read at a conference at the International Health Exhibition, T.ondon. June 4, 1884. THE POPULATION OF LONDON 169 From this it is at once apparent that London far 2xceeds in population not only Paris, the city that comes nearest to it, but Scotland, Switzerland, and even our colonies at the Antipodes. It has double the population of Norway or Servia, nearly double that of Denmark or Greece. Its inhabitants are three times as numerous 1s ‘those of New York or Berlin. London contains more people than Paris, Berlin, and Brussels taken together, while it nearly doubles the united inhabitants of New York and the neighbouring cities of Brooklyn, Hoboken, and Jersey City.. So far, by London I have meant “the metropolitan area,” as it is called, the London of the Metropolitan Board of Works—now the county of London —and of the Registrar-General,' and unless the contrary is specified, this will be the meaning attached to the word London throughout this paper. But outside this area there is fast closing in a circle of towns such as Bromley, Croydon, Kingston, Brentford, Tottenham, Stratford, and West Ham, all of which are included within the Metropolitan Police District, and go to make ap what the Registrar-General calls Greater London. This London has a population but little short of five millions, and therefore exceeding that of Sweden and approaching that of Ireland. London has certainly been a large and important city since the time of the Roman occupation, but we have no means of estimating, even approximately, the number of its inhabitants before the beginning of the present century. [t is stated that, A.D. 859, some 800 vessels were employed in the port of London in the export of corn? Truly, ! These areas are mot, strictly speaking, identical; the county of London includes the hamlet of Penge, in addition to the “London” of the Registrar-General, ? Haydn's Dictionary of Dates. 170 STUDIES IN STATISTICS CHAP, matters have changed since then! Whatever was the size of London in her time, it is quite certain that good Queen Bess thought it was quite large enough, since she issued a proclamation from Nonesuch Palace, on the 7th of July, 1580, forbidding the erection of buildings where none had before existed in the memory of man, as she deemed the extension of the metropolis likely to increase the plague; bo create a trouble in governing such multitudes ; a dearth of victuals; the multiplying of beggars, and inability to celieve them ; an increase of artisans, more than could live together ; and, moreover, as likely to impoverish other cities for lack of inhabitants! The decree stated that lack of air, lack of room to walk and shoot, &c., arose out of a too crowded city. In 1631 Sir R. Ducic, the Lord Mayor, reported to the Privy Council, by special command, in view of an impending scarcity, that “the number of mouths esteemed to be in the city of London and in the liberty,” comprising London within the walls, London without the walls, and the Borough of Southwark, to be 130,168." But as the great plague, only thirty-four years later, is said to have killed very nearly 70,000 persons, there must have been a considerable population outside the range of Sir Robert Ducie’s estimate. “In 1682 Sir William Petty, a very careful and acute observer, reckoned the population at 672,000, inhabiting 84,000 houses.” * With the beginning of the present century we get on firmer ground. According to the census of 1801, the people of London mustered 958,863; by 1841 this figure was more than doubled, viz. 1,948,417, a million of inhabitants having heen added in forty vears; in the next twenty years \ Natural and Political Observations on the Bills of Mortality, by Captain John Graunt, T.R.S., fifth edition. 1676. Quoted in Report on the City Duy Census, 1881. 2 Toftie's History of London, 2nd edition, 1884, vol. i. p. 352. INCREASE OR DECREASE IN TEN YEARS 1871-81. Central Area. ... -— 74,824 [nner Ring w+ 636,547 County of London + 562,223 Outer Ring ... + 318,797 Greater London + 87° Central area } (population decreasing, nner ring \ population Outer ring | inereasin: LONDON. POPULATION 1881. Central Area... Inner Ring sar ... 878,556 ...9 087.997 County of London Quter Ring o ...8,816,483 .. 050,178 Greater London ... av ...4,766.661 = 3 >, TT.» a. THE POPULATION OF LONDON 171 another million were added, and again in the next twenty yet another million; in other words, in forty years the additions to London equalled the collective present populations of Liverpool, Birmingham, Manchester, Leeds, and Sheffield, or the entire population of Norway, bringing up the numbers in 1881 to 8,816,483. But in spite of Queen Elizabeth’s proclamation, London is not yet grown up; it is still growing at the rate of 50,000 a year, 1,000 a week, 150 a day! At this moment,! by the trifling addition of a Hull or a Salford, London probably numbers at least 170,000 more inhabitants than at the last census. This growth is far from uniform. The map (Plate XV.) shows by the light shading outside the metropolitan area, what the Registrar-General calls the Outer Ring, extendng from the outskirts of Local Government London, to the boundaries of Metropolitan Police London. In the centre is a portion (white in the map) called in the census the Central Area, and comprising the City, Strand, Westminster, &ec.; the part between the Central Area and the Outer Ring (indicated by the dark shading) may be called the Inner Ring. The census shows that during the last ten years the population of the Inner Ring has increased by 636,547 persons, or a Liverpool and 2» Birkenhead. The Outer Ring, during the same time, increased by 818,797 persons—say a Leeds; but it was otherwise with the Central Area, where there was a lecrease of 74,324, equal to the population of Halifax. This needs some comment. As it is of great importance that a census should be taken simultaneously all over the country, it is convenient that a time should be chosen when but few persons are moving about: such a ‘ime is a Sunday night.: But obviously on a Sunday June 4, 1884. 2 Census, 1881, vol. iv. p. 13. 172 STUDIES IN STATISTICS CHAP. night the numbers in the business parts of a town will be at the lowest, whereas the residential suburbs will then have most inmates. The city authorities, for reasons which may be divined, were greatly incensed at the results of this method of numbering the people, and took a day census on their own account.’ Through the very valuable and interesting report on the results of this there runs a spirit of indignation which is quite amusing. I am by no means convinced that the proposed new munieipality would be the great success that its promoters love to fancy, since new machinery does not necessarily provide new men ; but the Report of the Day Census of the City, when arguing against the proposed enlargement of the corporation, seems to me to prove too much. If it be true that whereas, according to the Imperial census, only 74,897 persons slept within the city, whereas the corporation enumerators found 261,061 persons to be “ residing, occupied, or employed ” therein * during the working hours of the day,” and in addition counted 797,563 passengers into the city by the various inlets during a single day ; and if, moreover, the first number is decreasing, whereas the other two numbers are increasing, surely the conclusion is, that a city which is yearly less and less capable of providing sleeping accommodation for its workers should enlarge its bounds. The causes of the diminished number of sleepers in Central London are chiefly clearances for railways, public buildings, and new streets, and the conversion of dwelling-houses into places of business. The growth of a population is made up of two factors : (1) the natural increase of the people, and (2) migration. The natural increase of a population is measured by the axcess of births over deaths, which would be the only cause lL Report on the Day Cons of the City of London, 1881. Longmans. <I THE POPULATION OF LONDON 173 of change if there were neither emigration nor immigration. As a matter of fact, all populations are affected by both of these movements; that of London more especially. The admirable Weekly Return of Births and Deaths in London, now issued by the Registrar-General’s department, is the lineal descendant of the old bills of mortality issued by the parish clerks with more or less regularity since 1603, a pedigree which should place the weekly return very high indeed in the aristocracy of periodical literature. From this high-born but unassuming paper, we learn that in each year some 185,000 persons are born in London, and some 80,000 die, giving 370 daily births and 220 daily deaths, a natural increase of 150 per liem. The population of London as a whole increases considerably faster than the excess of births over deaths would account for, that is to say, there is extensive ¢mmigration. On the other hand, the number of Londoners who are also London-born does not grow nearly so fast as would result from natural increase left to itself; therefore there must be considerable emigration from London (using emigration in a general sense, and not necessarily implying thereby a long sea voyage). This is also shown by the fact, that the Londoners living in the provinces in 1881 exceeded those >numerated in 1871 by 148,336. There is no record of the numbers who annually come from Ireland and Scotland, from various parts of England and its colonies, from Europe or America, to settle in London, nor of those who leave its smoky streets for clearer skies. In the absence of the requisite data, we must be content with the balance struck by the census every ten years, which expresses the net result of all the changes :— 74 STUDIES IN STATISTICS CHAP. Ten vears, 1871—1880. Tixcess of Births over Deaths. 78.486 252 GLO Enumerated Increase. Balance of Migration. 152,810 + 260,558 +107,748 -198,887 + 300,635 Central Area a Inner Ring ... ve Metropolitan Area Outer Ring... g we Greater London ... y 51 C881 030 Vv The table shows that as the final result of ten years of destruction, reproduction, and locomotion, the inhabitants of the central area decreased by 74,324; but since the excess of births over deaths during the same period was 78,486, it is plain that at least 152,810 persons must have changed their abodes to some place without the central area. But it is certain that an additional unknown number went into this area: these may be said to have ““ paired ” with an equal number who “ went into the other lobby "—emigrated. In like manner we see that at least 260,558 immigrants settled in the “Inner Ring,” and at least 198,887 in the « Outer Ring.” We learn something as to whence the immigrants come by studying the tables of birth-places in the census. TONDON—BIRTH-PLACES OF THE PEOPLE. Per 1000 Inhabitants Numbers. Natives of London ai 0 other parts of England and Wales ok [reland -— cee Foreign Parts gro Scotland ee. ves the Colonies ... ee Islands in the British Seas 2 401.955 610 1,172,570 80,778 79,709 49,554 26,520 5.397 307 nT 27 13 7 Tatal 3.0" 1000 As compared with 1871, the Irish in London formed at the last census a somewhat less fraction of the whole, THE POPULATION OF LONDON 175 while provincials and foreigners were somewhat better represented. Doubtless many will be surprised to see that Scotch, Irish, foreigners, and colonists, all put together, number only some 6 per cent. of the London people; but it should be kept in mind that children born in London of Irish or German parents will be classed among the London born. The wood-cut on p. 177 is reduced from a large cartoon designed and drawn by the late Mr. Arthur Burgess, which was intended to show to the Conference, by the proportionate height of the two figures, the two chief component parts of the annual increase of London according to the average of the ten years 1871-80. Of those born abroad or at sea, 19,457 were British subjects (either naturalized or the children of British parents born abroad), leaving 60,252 foreign-born, who were also foreign subjects. There were fourd to be residing in London in 1881 one-half (51 per cent.) of the whole number of true foreigners in England and Wales, although the population of London amounts to only a little more than one-seventh (14'7 per cent.) of that of the whole country. This fact alone is a strong proof of the paramount importance of our capital. The most numerous among the foreigners were Germans, 21,966 ; French, 8251; Poles, 6931 (mostly Jews, and very poor) ; Dutch, 4193 ; and Italians, 3504. A few facts as to the occupations of Londoners will be of interest. To begin with, there is no great staple manufacture in London, save those of furniture and books ; consequently there is not found any preponderance of operatives in any one trade. London is the scat of government of a vast empire: hence we find the Civil Service employs of all ranks 16,670 individuals, including 762 women. The paid officers of the Local Government 176 STUDIES IN STATISTICS CHAP. and of the Poor Law numbered at the last census 3,570 ; 624 of them women—a surprisingly small body, soon, perhaps, to be greatly increased, for the benefit of the ratepayers. The police numbered 9,354, the soldiers about the same : so that less than 20,000 men sufficed to keep in order 3% millions of people—roughly, one policeman and one soldier to every 400 inhabitants. The clergy of the Established Church numbered 1,961, the ministers of other denominations 1,134; as assistants to these may be added 1,302 missionaries, bible-women, &ec., and 1,131 nuns and sisters of mercy. Per contra, the lawyers, besides an army of clerks, numbered 5,905—as yet all men. The doctors were 3,715 strong, including 10 women. Schoolmasters and mistresses were as many as 17,963— a number 2} times as great as that recorded in 1871, the increase being, of course, due to the operation of the Education Act. The most numerous class in London seems to be domestic servants, of whom 801,188 were enumerated, five-sixths of them women. Next come the building trades, employing 121,182—caring little, I fear, for Queen Ilizabeth’s proclamation ; the furnishing and decorating trades employ 42,366. The London printers are no fewer than 26,226; the railway servants being hardly less numerous—22,088. The publicans amount to 7,731; their natural associates, the pawnbrokers, to 2,378. On the other hand, coffee and eating-house keepers number only 8,686. Among the poorer classes may be mentioned 78,115 general labourers, 8,564 costermongers or hawkers, 48,559 washerwomen, and 19,334 charwomen. The magnitude of the metropolitan charities may be judged of by the fact that 4,186 persons, four-fifths of them women, are returned as in “ hospital and institution service.” GREATER LONDON GAINS PorurarioN BY @) Births. : (6) Immigrants from the pro-7inces, colonies, or foreign countries; by far the larger number coming from various parts of the United Kingdom. GREATER LONDON LOSES PopuraTiON BY (¢) Deaths. (d) Emigrants to the provinces, colonies, and foreign countries. The births always exceed the deaths—this excess is called natural increase. The immigrants exceed the emigrants—this excess may be termed balance of migration. In the diagram the height of the baby bears the same oroportiontothe height ofthe ountryman as the ¢“ natural ncrease ” to the ‘balance of nigration” during the 10 years 1871-80. During that period the internal growth of reater London was to its growth by accessions from outside, as 74 to 4. (The actual figures were 574,385 and 306,635.) 3 3 - ~ nf ) ~ STUDIES IN STATISTICS CHAP, The trades most commonly followed by foreigners in England, and therefore presumably in London, are those of sailor, servant, tailor, teacher. clerk. merchant, musician, and baker. There is no direct means of ascertaining how many out of the four millions of Londoners can be rightly called poor, meaning by poor those who are constantly or at frequently recurring intervals in absolute want of food and clothing. Nevertheless, some indications of the extent to which distress exists may be gleaned from the census. Thus, the blind numbered 8,214, the deaf and dumb 1,972, lunatics and imbeciles 6,082 (more than half being inmates of asylums).! Of these 10,218 persons, belonging to what are known as the “afflicted,” many would of course be in good circumstances ; but all who have worked among the poor must have observed in how many cases an “afflicted” child, much more an afflicted parent, drags down a whole family to the verge of poverty. The struggle for existence is severe for the healthy—a little oxtra trouble makes it insupportable. Then, again, there were no less than 173,143 widows. Dut there is some more direct evidence. The workhouses and workhouse schools contained 85,511 paupers, while at the same time nearly 50,000 persons were in receipt of out-door relief. In the hospitals 7,375 sick were under treatment. Juvenile offenders gave 1,790 pupils to reformatory and industrial schools, whereas the prisons contained 6,924 full-blown criminals. These numbers will justify us in saying that the very poor cannot number less than 100,000 ; there must, however, be a much larger number but a very little removed from this class. As regards housing, the fioures 1 The number of pauper lunatics chargeable to the county of London on January lst, 1889, was 10,369, according to a return prepared by the Asvlums Committee of the London County Council. THE POPULATION OF LONDON 179 show that of these 100,000 very poor, some 50,000 are the inmates of public institutions. The Dwellings Committee of the Charity Organization Society ascertained in L881 that there were at that time in London nearly 40,000 persons accommodated in various “improved dwellings” of one kind or another. Also nearly 2,000 persons received the benefit of supervision on Miss Octavia Hill's admirable system.) Of the 40,000 in Industrial Dwellings, many, indeed the large majority, would be above the very poorest class, and, as I think, quite rightly so. The fact that by the City census it was found that no less than 176,009 passengers entered the city in the single day by the twelve railway stations, gives some idea of the extent to which the outskirts of London supply sleeping accommodation for those who have to earn their bread near its centre. One word as to the health of this mass of humanity. The metropolitan area extends over some 12 miles from north to south, some 17 miles from east to west; within this space 89 governments, supposed by common consent to be utterly incompetent, so rule the houses, roads, and drains of their 82 millions of subjects, that for the last ten years they have suffered a death-rate of only 22% per thousand,” or but 1 per thousand in excess of that of the whole country. Only seven of the great towns of England an’ show a better account, whereas, on the continent of Europe, Christiania and Geneva, mere villages by comparison, alone excel London; Philadelphia and a few other American cities are also our superiors. Have all the governments that are supposed to be capable accomplished { The number under each of these heads is much larger now (1890). * Ranging from 20-8 per 1000 in the western districts to 21-9 in the northern and southern districts, 24'9 in the central, and 25:0 in the astern districts 180 STUDIES IN STATISTICS CHAP. their task as well? Do not suppose that I bring this forward as an argument for the maintenance of the status in quo,—far from it: I wish to impress on you that by quiet, steady perseverance on similar lines we have a reasonable hope of accomplishing yet more. That we raise a “ bitter cry ” does not mean that we are worse than we have been ; it rather means that we are realizing more fully that we might be better. It is believed that in the 17th century the deaths in London exceeded the births, and that the death-rate then was something like 49 per 1000 per annum, or more than double what it is now, and worse by far than the unhealthiest cities of Europe of bo-day.! You will now, I hope, be able, to some extent at least, to realize the enormous scale upon which social problems present themselves to the London philanthropist, the vast mass of raw material that he has to work upon. I must leave to others the more cheerful task of giving practical suggestions, of describing schemes to improve the dwellings of the toiling myriads of our metropolis. On my own part, I have but one more word to say, and must confess myself a disciple of Queen Elizabeth, and wish that something could be done, if not to diminish the number of inhabitants, at least to prevent their increase, in which case it might be possible to work off the vast mass of arrears that neglect has allowed to accumulate. As it is the task seems very hopeless, since there are to-day 200 people more than yesterday, to-morrow yet another 200 will be added, and so on, and so on. There is but one remedy that seems at all practicable, that is, emigration ; only beware of aiding emigration from London itself to t See The Sanitary Condition and Laws of Medieval and Modern London, by J. W. Tripe, M.D., Trans. Soc. Med. Off. of Health, 1881-2, pp. 9-—12. This paper contains much valuable and curious information. KI THE POPULATION OF LONDON 181 any great extent; if you do so, for every family sent to Canada from Whitechapel or Poplar, two families will rush in from Norfolk or Devon, or even Ireland. Some eighty persons come in from the country to settle in London every day. Stop these people en route. Kidnap them and ship them off to the Antipodes. These countrymen will make far better emigrants than townsmen. In dgther words, do all that can be done to discourage immigration into large towns, and tell Hodge that if he is dissatisfied with his own parish, he will do best to cross the seas. 82 STUDIES IN STATISTICS Sr CHAP. CHAPTER XII. FOOD SUPPLY. NEARLY a century has elapsed since Malthus’ famous essay first appeared. The rapid development of America and Australia, with their seemingly boundless supplies of food, made men forget his solemn warnings. It will be interesting to endeavour to see whether modern statistics throw any more precise light on the matter. For this purpose we must know something of the number of mouths to be fed, and the quantity of land required to produce food for them. The following table is intended to show in the light of the experience of the last thirty years what the growth of the population of Western Europe amounts to annually. In constructing the following table I have, so far as seemed practicable, framed the earlier estimates in accordance with the present territorial boundaries, that is to say, I have anticipated the transfer of Savoy and Nice from Italy to France, of Lombardy and Venetia from Austria to Italy, of Schleswig-Holstein from Denmark to Germany, and of Alsace-Lorraine from France to the same power. The difficulty or impossibility of making correct allowances for these transfers, and the irregular times at which the first set of censuses were taken, make the results only roughly XII FOOD SUPPLY 183 approximate for many of the individual countries, but the figures for the whole may be taken as fairly accurate. INCREASE OF POPULATION OF WESTERN EUROPE IN 30 YEARS! (Three 000s omitted.) Country. United Kingdom ! France .., es Belgium oe Holland - Germany Denmark Sweden wo Norway -— Austro-Hungary Switzerland ... Italy ... vo Spain... vo Portugal i» Date. 1851 1851 1846 1849 | 850 850 ‘850 | 845 1850 | 1850 1 ‘851 1857 1848 Donulation.’ 27,391 34,592 4,337 3.247 35,231 1,408 3,483 1,328 30,727 2,393 24,138 15,220 3,452 186.947 Date. 1881 1881 1876 1879 880 1880 "880 | ‘875 1880 '880 881 RRYT {R78 Increase in 30 years. Population. - "A psolute.: Per cent. 27°4 89 23°'0 30°I 284 39°8 31° 361 230 18g 17°9 [3°4 20°5 54 21'0 01 | 39,254 96.9 1226.2 34,885 37,672 5,336 4,223 15,234 1,969 4,566 1,807 37,786 2,846 28,460 17,257 4.160 7,494 3,080 999 976 | 10,003 561 1,083 479 7,059 453 1,322 2,037 708 Although we are without positive data, it is practically cerfain that for a long period the population of Europe increased very slowly, receiving great checks from famines and pestilences as well as devastating wars. We may, however, take it as proved that at the present time the countries in the table, apart altogether from the emigrants they send abroad, are within their >wn boundaries increasing by 1} millions per annum. A simple calculation shows clearly that the rate of increase (6'6 per cent. in each ten years) cannot possibly have been as great in the past, and that if it should be as great for another century, the population will be ! Figures in clarendon type are from the Census, those in ifulics are astimates derived in part from the Almanuch de Gotha, but mainly from Stutistique Internationale. Etat de la Population, Stockholm, 1875-76. The Canary Islands are not included in Spain, nor Madeira and the Azores in Portugal. 84 STUDIES IN STATISTICS CHAP increased by 200 millions, or more than doubled, a result not indeed impossible, but one that taxes our powers of belief considerably. CALCULATED POPULATION OF WESTERN EUROPE, ASSUMING THE SAME RATE OF INCREASE AS IN THE THIRTY YEARS 1850-80 A.D. 500 A.D. 1000 A.D. 1500 A.D. 1750 A.D. 1800 A.D. 1850 A.D. 1880 AD. 1990 35,000 844,000 20,230,000 99,037,000 36,069,000 186,947,000 226,201,000 454.995.000 In connection with the direction of flow of currents of migration, it is interesting to divide the countries of the world into those which import and those which export wheat. The following table is derived from the statistics of trade given under each country in the Statesman’s Year-Book for 1887. I have thought it best not to give figures, since they might lead the reader to suppose that the facts are more accurately known than is in truth the case. Toa notable degree (though with marked exceptions which admit of obvious explanation) the tendency, as might have been expected, is for food and men to travel in opposite directions. It should not be forgotten that the United States exports in addition enormous quantities of animal food ; the same is true, but to a much less degree, of Canada, the Argentine Republic, and New Zealand. Queensland and New South Wales are small exporters of meat. As regards Europe, many countries that import wheat export large amounts of animal food and vegetables, e.g. France, Belgium, Holland, and Germany. Denmark, which imports small quanties of wheat, exports on a much larger scale cattle, swine, hams, and butter. {11 FOOD SUPPLY 18% Countries exporting wheat. Amount. very large ‘arge zonsiderable considerable considerable small Countries importing wheat, Amount. very large large 1arge arge arge ‘arge arge considerable onsiderable small sonsiderable small small Russia... son Roumania ... Austro-Hungary Turkey ee Bulgaria — Servia... . Jnited Kingdom ... France on Jelgium ... dolland ... Jermany ... Switzerland ... [taly... vn Spain ee Portugal ... Sweden wie Norway 3 9 denmark ... Natal [ndia ... . Egypt ‘es Tunis... N very large small small South Australia Victoria oo New Zealand... considerable small small New South Wales ... small Queensland... ... small United States very large Canada wis considerable Chile ... a small Argentine Republic considerable But it will be observed, that of all the countries named in the table of increase of population, Austro-Hungary is alone in exporting wheat, all the others require to supplement to a greater or less extent their home supplies by imports from other countries. According to writers on Hygiene, an adult man taking active exercise requires, in addition to animal food, about 14 Ibs. of bread per diem. It will probably not be much above the mark, therefore, if we say that a population of European race will consume on an average about 1 Ib. of bread per head per diem. This, taking Dr. Parkes’ figures, amounts to an annual consumption of nearly 51 bushels of wheat! To be within the mark then, we will ' A Manual of Practical Hygiene, by E. A. Parkes, M.D., F.R.S,, 5th edition, 1878, pp, 222, 241. I have taken the bushel of wheat as weighing 60 lhs., yielding + of its weight in flour, of which 100 lbs. yield 140 lbs. of bread. 186 STUDIES IN STATISTICS CHAP. say that each individual requires 5 bushels of wheat per annum,! and taking the crop to average 30 bushels to the acre, we see that each acre of corn land will support 6 people. Taking the yearly growth of population in the United States to be, in the decade 1880-90, nearly 1,250,000, we find that 208,000 acres of land require to be brought under the plough annually to supply their increasing requirements en wheat alone. Now 208,000 acres is not far short of the area of Huntingdonshire. But in Fssex, Norfolk, and Suffolk in 1884 2% a little more than one-third of the total area was under corn crops, and in Lincolnshire a little less than one-third ; these four being the chief corn-growing ~ counties of England. Let us, however, suppose one-half the land devoted to the growth of wheat; it will certainly then be within the mark if we say that to feed the new mouths added to the United States in each year of the decade 1870-80 required that at least 400,000 acres of land (= 625 square miles) should be brought into complete cultivation. In other words, to make the figures simpler, each million of population requires af least 500 square miles of fertile land for its food supply. But in the United States the yield is stated to be only 13° bushels to the acre, so that, keeping well within the mark, we may again double the last figure, and allow 1000 square miles for the support of each million, If this be so, we may say that the present population of the United States, reckoned in 1890 at 624 millions, could be almost supported by the possible agricultural produce of the single state of Iowa. 1 A calculation from the figures given in the United States’ Census, 1880, shows that, after exporting nearly one-third of the wheat produced. shere remains for home consumption about 6 bushels of wheat per head. 2 Stanford's Parliamentary County Atlas. 1885, 8 ensus of United States. 1880, vol, iii. FOOD SUPPLY 187 But again, from the analogy of Europe, it would appear that we have much understated the amount of land required by such a large population. La Belle France is rightly considered to be a very fertile country; its population, a little over 38 millions, cannot, however, find sustenance on its 204,092 square miles, and has to import considerably more food in the way of grain and flour than it exports in other forms, such as wine, fruit, vegetables, dairy produce, and the like. It seems fair then to say, that under present conditions it contains at least as many inhabitants as the soil will support. If this be so, in France each million of population requires 5,340—say 5,000, square miles for its support. By simple proportion we find that the 62% millions of the United States would require at this rate not the produce of Iowa only, but in addition that of Minnesota, Wisconsin, Michigan, and Illinois, which, as a glance at the map will show, occupy a considerable portion (more than 10 per cent.) of the whole territory. But while the population of the United States increased as we have said by 1,200,000 in each year of the decade 1870-80, the population of Western Europe, as we have seen, also increased by 1,400,000. This gives then 22 millions of new mouths to be fed annually, and this at the very lowest estimate. And if we say that each million requires 1000 square miles of land for the purpose of growing wheat, we get to the result that 2,600 square miles of virgin land must be brought under the plough every year ; now 2,600 square miles contain 1,664,000 acres, or more than the area of Devonshire. It is, therefore, no exaggeration to say that the time will come, and that before the ond of the next century, when men will once more think seriously of Malthus and his warnings. We are too apt to forget that most of the best land in the United States 188 STUDIES IN STATISTICS CHAP, has already been taken up; of this we have had a striking proof twice recently in the land-rushes or “booms” that have taken place when an Indian reserve has been thrown open to settlement. The truth of the matter is, that while the railway and the “self-binder” have been increasing so largely the wheat supply, a vast city population has been growing up in the eastern and middle states, a population as dependent on imports of food as any in Europe. Further, it is important to bear in mind that the growth “of population proceeds not in an arithmetical progression, that is by equal increments in equal times, but in a geometrical progression, like compound interest, so that larger and larger amounts are added in each successive period. Now the population of the United States during the past century has doubled itself in every twentyfive years. Supposing the same rate of growth to be maintained, what would this imply ? No less than this: twenty-five years ago the population of the United States was thirty millions less than it is now; but twentyfive years hence it will be increased by twice thirty millions, and in the next succeeding quarter of a century by an additional 120 millions. It is true that the area and resources of the country are very great, but they are not without limit.! Dr. Giffen, in his inaugural address as president of the Statistical Society in 1882,% gave good reasons for saying that « perhaps in the course of ten, and almost certainly in the course of twenty years,” we should begin to feel in various ways the consequences of the ! Professor Hayden (Stanford's Compendium of Geography, North America, London, 1883) estimated that there were in the United States 660 millions of acres (say one million square miles) of arable land still uncultivated. Prof. Macoun (Manitoba and the Great North-West, London, 1883) in like manner gave Canada still 200 millions of acres. 2 Journal of the Statistical Society, vol. xlv. p. 545. X11 FOOD SUPPLY 189 complete settlement of the Western states. One consequence we may venture to hope will be that rents in this country will begin to rise again (and what is more, to be paid) with a rise in prices, so that the English landlords and farmers, or at least their children, may look forward 50 a good time coming. These speculations are involved in so many uncertainties, and are subject to so many possible modifications, that they are of perhaps little value, nay rather may lead to evil results. This, however, is the point on which I wish to lay emphasis : the demand for food increases every year not merely in Europe, but even more rapidly in those countries which we are too apt to look upon merely as the granaries of Europe. "It is, however, likely that timber will fail us long before our wheat supplies fall short. See W. Little in Forestry for August 1883. 190 STUDIES IN STATISTICS CHAP. CHAPTER XIII. SUGGESTIONS FOR THE CENSUS OF 1891. ONCE more we are on the eve of a census. Next year, 1890, that greatest of all statistical operations will be carried out in Austria, Belgium, Denmark, Germany, Hungary, Norway, Sweden, Switzerland, and the United States ; and in the year following in France, Italy, and throughout the British Empire. In the eloquent words of M. Korosi, « Millions will again be spent ; the great army of snumerators, an army which includes in all half a million, perhaps a million of agents, will be mobilized once more. Battalions of this army of enumerators will invade again every inhabited place, from the snow-clad peaks of Alaska and of Iceland to Cape Town and even to the south of Australia. In order to draw up under uniform rules the vast inventory of the human commonwealth, patrols of this great army will penetrate anew into the palaces of the great capitals of the world, and into the lonely shanties of the squatters in the wilderness; they will omit neither the tents of the Bedouins, the huts of the Eskimo. nor the wigwams of the Red Men.” ? { Read before the Royal Statistical Society, 25th June, 1889. 2 Bulletin de Institut Indernational de Statistivue, 1887, 1% Livralson, p. 215. XIII SUGGESTIONS FOR THE CENSUS 191 To the Royal Statistical Society every census, even of the smallest or least powerful state, is full of interest, and each has its special value ; but at the same time that of the British Empire, comprising as it does the largest number of individuals of any, and concerned as it is with every continent and every climate, has naturally an overpowering interest. If the marvellous growth of population in »ur own islands, as recorded from decennium to decennium, touches us most closely, yet it scarcely interests the thinking Englishman of to-day more than the progress, yet more marvellous when all the conditions are considered, of our colonial empire. The census of 1881 revealed to as the existence of populations in Canada as numerous as, and even in Australasia not far short of, that of the United States at the time of the war of Independence, and in civilization, wealth, and organization far in advance of it. The enthusiasm of colonists leads them to put trust in. extravagant estimates of their rate of progress, and another census will serve to test these estimates and enable us once more to touch firm ground. Indeed there is reason to believe that our own estimate of population, even in England and Wales, is beyond the mark, and needs orrection, or verification. The earliest census of which we have any account is that of Moses, recorded in Numbers.! This was followed nearly five hundred years later by that carried out by King David through the unwilling agency of Joab. Both of these seem to have been undertaken for military purposes; the latter took nine months and twenty days to accomplish, and whether’ we follow the figures given in the account in Samuel,? or in the Chronicles? we see that the number of fighting men in Israel and Judah had ' Numbers i. 1—46. 2 9 Samuel xxiv, 1—17. 3 1 Chronicles xxi, 1—17. (92 STUDIES IN STATISTICS CHAP. somewhat more than doubled in the interval ; the usually accepted chronology makes this interval 473 years, giving a rate of increase far slower than any of which we have recent experience, but it is obvious that there are so many points of uncertainty in the accounts that it would be rash in the extreme to draw any inferences from them. What is, however, of more practical interest is, that the record of the Divine wrath at David's census produced such a powerful impression upon the Christian world, that we hear of no census during the Middle Ages, nor indeed until the 17th century. This is the more remarkable in that the census was a regular Roman institution, one of the chief duties of the Censors being to record the number of Roman citizens every five years for purposes of assessment. In the New Testament® we read of a census of “ the whole world,” undertaken at the command of Cesar Augustus. A census of Gaul is alluded to by Tacitus.? Unfortunately no reports of these Roman censuses have eached our day, the records having all perished. They are, however, in a few instances quoted by historical writers, e.g. Livy says that the number of “citizens ” at the census taken B.C. 457 was 117,319°%; at the census of B.c. 193 it was 143,704*; five years later, B.c. 188, it was 258,318; and the historian notes that the Campanians were on that occasion registered at Rome, and various others were included for the first time®; at the census B.c. 173 the number given was 269,015, with the explanation that the figures were small because the Latin allies had been entered in their own states ®; lastly, Tacitus says that at the census taken during the reion of the Emperor Claudius (died a.p. 54), - St. Luke ii, 1. 2 Ann., Lib, I. cap. 31, and Lib. IL cap. 6. 8 Liv, iil. 24. ¢ Id., xxxv. 9. 5 Id., xxxviii. 36. 6 71d., xlii. 10. XIII SUGGESTIONS FOR THE CENSUS 193 the number of citizens was found to be 5,984,072 ;* probably this number included all the citizens throughout the empire. Gibbon, writing of this census, calculates that this number of citizens probably corresponded to a population of 20 millions, counting women and children. ? I have recently been informed by Mr. H. H. Risley, of H.M. Bengal Civil Service, that a census of Thibet was taken in the 13th century by Kubla Khan, but know no details. The earliest census in modern days which I have come across is that of La Nouvelle France, as the part of Canada now known as the province of Quebec was then called, in 1665. This, it will be noted, was a census taken in a Roman Catholic country in which the Church has always been especially powerful, so that one may presume that no religious objection was made. Canada is indeed the most remarkable of all countries for the number of its censuses. No less than fifteen regular censuses?® of New France, some or all of them nominal, were taken between 1665 and 1754; nine of them in the 17th century. During the same period seven censuses were taken of Acadia (now Nova Scotia), and six of Newfoundland. In France official estimates of the population were made in 1700 and again in 1784, the latter being founded on the mean annual number of births, but the first regular census of France was taken in 1801. In Sweden the «Table Commission,” which was charged with the duty of collecting through the clergy periodical returns of population and annual returns of + Tacitus Ann. xi. 25. 2 Decline and Fall of the Roman Empire, vol. i. ch. 2. Gibbon gives the figures as 6,945,000. 8 Exclusive of numerous official ¢* Statements of Population.” See Census of Canada, 1870, vol. iv. Introduction. A “nominal” census is one in which every individual is recorded by name. 194 STUDIES IN STATISTICS CHAP. births, marriages, deaths, and causes of death, was founded in 1749, the celebrated Linnseus being one of its founders. Under its auspices a census was taken in 1751, then every third year (with three omissions) to 1775, then every fifth year to 1860, since which the census has been decennial. In the United States of America a decennial census is provided for by the Constitution ; the first was taken in 1790, the tenth in 1880. The first proposal to take a census in this country was contained in a bill introduced into the House of Commons in 1753, “for taking and registering an annual account of the total number of the people, and of the total number of marriages, births, and deaths; and also of the total number of the poor receiving alms from every parish and extra-parochial place in Great Britain.” The bill, though supported by the Ministry, was violently opposed. Its chief opponent in the Commons, Mr. Thornton, member for the city of York, said, “I did not believe that there was any set of men, or, indeed, any individual of the human species, so presumptuous and so abandoned as to make the proposal we have just heard. . ... I hold this project to be totally subversive of the last remains of English liberty. . . . The new bill will direct the imposition of new taxes, and indeed the addition of a very few words will make it the most effectual engine of rapacity and oppression that was ever used against an injured people. . . . Moreover, an annual register of our people will acquaint our enemies abroad with our weakness.” Mr. Matthew Ridley, member for Newcastleapon-Tyne, stated that he knew by letters that ““ the people looked on the proposal as ominous, and feared lest some public misfortune or an epidemical distemper should follow the numbering.” It was further urged that the scheme was costly and impracticable ; that it was an Imitation of X11 SUGGESTIONS FOR THE CENSUS 195 French policy, borrowed from our natural enemies; and that it would not only be a basis for new taxation, but for a conscription. On the other side, Mr. George Greville, a lord of the Treasury, said, “It will be extremely useful at all times for many useful purposes; and in the case of a long war it will be absolutely necessary; for the usual methods of raising recruits for our army would not then be sufficient ; we should be obliged to have recourse to that of obliging each parish to furnish a certain number.” ! The Bill passed through all its stages in the Commons by large majorities, but was thrown out on the second ceading in the House of Lords. The House of Lords, be it observed, as on many occasions in our history, taking the side that was believed to be that of popular rights and liberties. “ Nearly half a century passed away before the proposal was renewed ; but when the new Bill was introduced, in November 1800, into the House of Commons, it had the advantage of a great change which had apparently occurred "in public opinion on the subject of population. The old fear that the number of the people was falling off, and that an enumeration would betray the inability of the country to furnish a due supply of soldiers for the army, had given place to a new and opposite form of alarm, namely, that the people were increasing so rapidly as to outstrip the means of subsistence. Among the causes which may be supposed to have brought about this change of opinion, probably the most powerful was the great dearth which prevailed in the country at the time when the Bill was brought forward, much of the time of both Houses of Parliament being occupied in the year 1800 in discussions on “ the present high price of provisions’; while-a second L This paragraph is taken, with slight abridgment, from the General Report, Census, 1881, vol. iv. p. 1. 6 (9 STUDIES IN STATISTICS CHAP. cause that may fairly be assumed to have had some influence on the matter, was the attention excited by Malthus’s great work, of which the first edition was published anonymously in 1798, and taught its readers that there were other aspects of the question of population than the military one. “The Population Bill was brought in by Mr. Abbot, member for Helston, on November 20th 1800, and passed through all its stages without opposition. The enumeration was made on March 10th in the following year, and has been repeated ever since, without omission, in the first year of each successive decennium.”? The first census of Ireland was taken in 1812, but was a failure. Of the Irish census of 1821 a committee of the Statistical Society said, in 1840, it “ appears to have been by far the most perfect in its machinery and method of any that has yet been executed in these islands.”? It was the first nomanal census taken in the United Kingdom. It is well to bear in mind, though it is often forgotten, that in the last century there was no knowledge whatever of the number of the English people (such as was to some extent supplied on the continent of Europe by the plan of police supervision), and that our first census, imperfect as it was, was taken ten years after the death of Adam Smith, twenty-five years after the publication of the Wealth of Nations. The early censuses were taken by the agency of the overseers of the poor in England, by the schoolmasters in Scotland ; the results were abstracted under the superintendence of Mr. Rickman, assistant clerk to the House of Commons. From 1841 onwards the censuses of Great 1 This paragraph is taken, with slight abridgment, from the General Report, Census, 1881, vol. iv. p. 2. 2 Journal of the Statistical Society, vol. lil. p. 90. X 111 SUGGESTIONS FOR THE CENSUS 197 Britain have been taken through the agency of the Registrars of births, deaths, and marriages, those of Ireland by the Constabulary, with very great advantages in each case. Opposition to the census on Biblical grounds is now a thing of the past: opposition on the ground that the census is inquisitorial has also, there is good reason to selieve, gradually lost its force. But while it is now agreed among all civilized nations that a census is a useful and desirable thing, many points still remain open for discussion ; these practically resolve themselves into the three following questions, viz. — I. How frequently should the Census be taken ? IL. What Questions should be Asked in the Schedule ? III. In what way should the Information Obtained be Abstracted and Tabulated ? There is an immense deal to be said under each of these heads, and the Statistical Society has since the year 1840 appointed a committee before every census to inquire into the various questions and report to the council. In consequence of the reports of these committees, the Society has from time to time made important representations to the Government of the day, which, while they have disappointed many by not effecting all that the Society bad hoped to achieve, yet have resulted in very important improvements. The most notable was in 1840, when the action of the Society led the Government to withdraw its Bill and introduce another modelled on the lines suggested by the council L.— How frequently should the Census be taken ? General experience appears to point to the unanimous conclusion that the interval from census to census should In no case be greater than ten years. Considerations of 19% STUDIES IN STATISTICS CHAP. expense have in the majority of cases prevented the census from being taken at shorter intervals; thus, in Austria, Belgium, the greater part of the British Empire, Denmark, Holland, Hungary, Italy, Norway, Sweden, Switzerland, and the United States, the census is decennial. However, the exceptions to this long interval are both numerous and important, In the Sandwich Islands a census has been taken every six years since 1853. In France a census was taken in 1801, 1806, and every five years since 1821. In Sweden every five years from 1775—1860. [n Germany in 1867, 1871, 1875, and every five years since. In Finland every five years since 1875. In Sweden every three years from 1751-75. In Baden every three years from 1846-67, and from 1821-48 more frequently. In Hanover every three years from 1830-64. In Prussia every three years from 1834-67. In Saxony every three years from 1834-67. In Bavaria every three years from 1834-67. In Wirtemburg every three years from 1834-67. In Upper Canada (now the province of Ontario) a census was taken annually for nineteen years from 1824-42, In New Zealand nine censuses have been taken since that of 1851, at intervals of from three to seven years, the last interval being five years. [n Queensland there is also a quinquennial census. In the North-West Territory of Canada it was found desirable to take a census in 1885. and in Manitoba in 1886. U Sir Rawson W. Rawson in MS. See also Etat de lu Population (Congres International de Statistique, 1876). XIII SUGGESTIONS FOR THE CENSUS 199 Fourteen of the States and Territories of the Great American Republic took an intermediate census in 1883, and it is rumoured that in future the census of the United States is to be quinquennial. If a census is useful to the Executive Government for a variety of purposes, such as taxation, military service, or parliamentary representation, it is to the working statistician something more: it is not merely the great measure of the rate of progress of nations, but it is a prime necessity for all measurements. The population as recorded in the census is, as it were, a sealed standard, or rather it is like the base line or the datum level of the surveyor, to which all distances and all elevations have to be referred. But the statistician is not concerned only with heights and distances, his world is one of many dimensions. Land and houses ; railways and ships; cattle, grain, and timber ; gold and silver, iron and coal ; imports and exports; taxation and revenue; births, deaths, and marriages ;—all measurable things are measured by him in relation to population. The merest tyro in statistics knows that crude gross numbers are of little value for purposes of instruction, nay, that the most erroneous conclusions are but too easily drawn from them; a unit of comparison is required in all figures, and as the statistician ‘or the most part deals with ever-changing quantities— “variables” as the mathematician terms them—he needs scale for comparative measurement. As, moreover, the igures with which the statistician deals are always related, either directly or indirectly, to men, his scale must be me showing numbers of men, so that it is of paramount necessity that he should know at any time the number of men corresponding to the number of things ander discussion, that is, the population of the country or listrict concerned. Further, as all human things are, like 200 STUDIES IN STATISTICS CHAP. the atoms of Democritus, in “a constant state of flux,” in default of continuous information these populations must be ascertained at short intervals. Those who hold that ten years is not too great an interval between census and census, will doubtless maintain that no useful inferences as regards Great Britain ean be drawn from the experience of the countries named above. They will say that on the one hand in great military countries, such as France and Germany, the requirements of the conscription necessitate a frequent enumeration of the people, while, on the other hand, rapidly growing but sparsely populated new countries, such as New Zealand, Queensland, or Manitoba, afford no precedents for old countries like this. Without denying that there is some force in these objections, I would submit that the population of this country grows very fast, and that its pressure is strongly felt ; that, moreover, such districts as Barrow-in-Furness and parts of Glamorganshire (not to speak of the suburbs of London) grow with a rapidity which rivals that of the mushroom cities of the southern hemisphere and the far west. On the other hand, if we have no compulsory military service, we have school boards, sanitary boards, poor law boards, county councils, and other authorities which are all in various ways interested in knowing from time to time the exact populations of their administrative areas. Population is now recognized by both parliamentary and local government Reform Acts to be the basis of representation, whereas under the existing arrangements a very few years after the taking of a census no one knows what the population is. Thus, for example, the Local Government Act, 1888, gives the status of county boroughs” to towns (other than those which are * counties of themselves ”’) which on the 1st day of June, 1888, had XIII " SUGGESTIONS FOR THE CENSUS 201 populations of not less than 50,000. Among such county boroughs the following will be found in the schedule : Barrow-in-Furness, Coventry, Devonport, Dudley, Great Yarmouth, Hastings, Reading, and Wigan. None of these had in 1881 a population amounting to 50,000; it is indeed probable enough that on the day named some of these towns had populations of that size, but it is almost equally probable that others still fall short of that number. By what means these towns were included in the schedule is, I suppose, an official secret.” Doubtless it is a matter of small moment whether the population of Reading, for example, was on Ist June of last year 48,000 or 50,000, but I cite this as a fair instance of the state of uncertainty resulting from the long interval that now separates one census from another. Again, objectors will call attention to the fact that triennial censuses and even annual censuses have been tried and abandoned again after trials extending over many years. True; but in Germany triennial censuses have given place to quinquennial, not to decennial, censuses. In other countries the decennial period has equally been tried and found wanting ; thus Canada, after giving up annual censuses for decennial, is now, in part at any rate, reverting to quinquennial censuses. In fact, as in most things, a reasonable mean should be fixed upon, a mean which balances the evils of long delay and excessive expense. Not only is more interest taken in statistics than was the case when ten years was fixed upon as the census period—not only are statistics used for many practical purposes then undreamed of—but the changes which it is the object of the census to measure are far more rapid than then. Perhaps the rate of natural increase 1 Sir Hugh Owen informs me that the increase of population was deduced from the increase in the number of inhabited houses. STUDIES IN STATISTICS CHAP. of populations is not very different, but the railway and the steamship, and indirectly the penny post and the telegraph, greatly increase migration—every degree of migration, that from continent to continent, from country to country, from county to county, even from village to town ; all these forms of migration, the last not the least, are of great interest to the politician and the political sconomist ; yet when an inquiry was recently made by Parliament into the subject of the immigration of foreign paupers and its influence on the labour market of this country, the only official information available as to the number of foreigners resident in England was no less than seven years old, and therefore quite inconclusive, if not absolutely valueless, for the purposes of the inquiry. Then again the changes in industrial conditions brought about partly by new inventions, but largely by improved facilities of communication, are far more rapid than formerly : so rapid that there is an unquestionable want of more frequent information as to these points than is oow given us by the Government. But perhaps the chief argument in favour of a quin-Juennial census is its desirability as a basis for the more sorrect calculation of birth-rates and death-rates. In this country the registration of births, deaths, and marriages is very complete, and the Registrar-General publishes weekly, quarterly, annually, and decennially, elaborate reports of the highest value. In addition to these returns, the medical officers of health throughout the country report in great detail every year (and at shorter intervals) to the sanitary authorities of the various districts. These various reports are by universal consent admitted to have con-:ributed materially to the marked improvement in the public health from which the nation has benefited in recent years. Since, however, the medical officers are in too 202 X11 SUGGESTIONS FOR THE CENSUS 203 many instances in complete ignorance of the true populations of their districts for at least five years out of every ten, the published birth-rates and death-rates lose much of their value, and the large amount of money expended upon their preparation and publication is to that extent wasted. In all statistical computations as to any matter under investigation in the interval between two successive censuses, it is obviously necessary, whatever that interval may be, to interpolate corresponding values for the population. It is pretty evident that these interpolated values will be more likely to be close approximations to the facts when the intervals are short than when they are long. The official method of estimating the population at any given time is based on the assumptions (1) that the rate of increase during the decennial period is uniform, and (2) that the rate of increase in the years succeeding the last census is the same as in those immediately preceding it. In some cases the results obtained are very satisfactory thus when the census of 1881 was published it was found that the actual populations of Birmingham, Brighton, Liverpool, Maidstone, Middlesborough, Southampton, Tynemouth, and Wolverhampton were within 1 per cent. of the estimates. But in the case of the following towns, among others, the estimates exceeded the actual populations by from 6 to 11 per cent., viz. : Portsmouth, Bolton, Leicester, Salford, Sheffield, Rochdale, Chatham, and Stoke-on-Trent; by 15 per cent. in the case of Halifax, 16 per cent. in the case of Merthyr Tydfil, and 22 per cent. in the case of South Shields. On the other hand the estimates fell below the true populations in the case of London by 8 per cent., Nottingham 5 per cent., Hull 8 per cent., Lincoln 10 per cent., Barrow-in-Furness 15 per cent., and Stockport 18 per cent. Again, the same method of caleulation applied to the districts and sub-districts into which STUDIES IN STATISTICS CHAP. London is divided for purposes of sanitary administration, would have over-estimated the populations of Paddington, Kensington, and Battersea to the extent of 15, 27, and 39 per cent. respectively. It is obvious that such considerable errors must render death-rates or other figures founded upon them entirely deceptive and valueless. The Registrar-General himself, in the Weekly Return, has for some time past called attention to the fact that as regards the twenty-eight great towns, the estimated populations of Leicester, Salford, Bradford, Newcastle-on-Tyne, and Cardiff are not to be trusted ; that those of the first three towns are probably over-estimated, of the other two underestimated ; more recently he has spoken of the estimates for all the great towns as untrustworthy. Further, he is so mistrustful of estimates in the case of the smaller towns, that in recent quarterly reports he has ceased to publish estimated populations (and consequently deathrates) of any of the “fifty other towns.” These limitations naturally detract considerably from the interest and value of the respective returns. I have shown elsewhere® that partly owing to the great fall in the birth-rate exceeding the fall in the deathrate, but more especially to the sustained large flow of emigration, there is good reason to believe that the actual population of the United Kingdom in April 1886, or the middle point between the census last passed and the census next to come, was between 400,000 and 500,000 less than the number given by estimation according to the official method. The error it is true amounts to only about 1} per cent., still if the present conditions are maintained it will amount to little short of a million before we have another census to set us right. If so great an error can be made in the estimation of a population of ! Journal of the Statistical Society, vol, xlix. pp. 760-63. XIII SUGGESTIONS FOR THE CENSUS 205 36 millions, it will be readily understood that when small areas are dealt with, such as counties or towns (in which intra-migration plays an important part), errors absolutely indeed less, but proportionately far greater, may be expected to occur. A great deal has been recently written to prove that gross death-rates, that is, death-rates calculated on the total population without any allowances for the age and sex constitution of that population, are very imperfect tests of sanitary conditions; but of what avail is it to undertake laborious calculations with a view to apply the requisite corrections, when the gross death-rates themselves are perhaps over or under-estimated by 2:0 or 3-0 per thousand ? The last objection, that which will probably have most weight with the Government in deciding the question, is that of expense. Let us see what this amounts to. The cost of the last four censuses of the United Kingdom, according to a Parliamentary Return (No. 877 of 1875), which I have brought up to date, is given on p. 206. It will be noted that the cost per head in Scotland, and in 1871 in Ireland, has been much greater than in England and Wales ; this is doubtless partly due to the populations in the former countries being more scattered, but largely to the fact that a greater amount of information is sought to be obtained. A mere counting of heads by a simpler intermediate census would be less costly, and yet extremely useful ; besides its direct use, it would be a valuable check on estimates as to number of children of school ages, number of married persons, &c.; but I understand that in the opinion of the authorities of Somerset House the saving to be effected by thus curtailing the inquiries would not be so considerable as an outsider might suppose. Whether 206 STUDIES IN STATISTICS : CHAP. COST OF THE CENSUS. Cn” "Local Expenses. Total. Cost per 1,000 Enumerated. ExcgrLaxp AND WALES. [841 ... 1851 ... 1861 ... 1871 ... L8&1 ... E 28,400 33,062 28,803 41,674 | 38,900 £, 58,328 60,070 66,915 78,304 86.972 £ 86,728 93.132 95,720 119,978 122,878 £ sd. 3 9 - 5 4 4 15 5 5 14 SOOTLAND. 1861 ... 1871 ... 1881 ... 5,218 18,464 23,682 714 8 6,340 | 20,764 | 27,104 | 8 1 4 6,165 21,582 27,747 7 8 6} IRELAND. 1851 ... 1861 ... 1871 ... 1881 ... 24,281 4,289 28,570 26,702 2,259 28,961 35,527 | 3,067 38,594 17,451 4,164 21,6015 47 2 419 11 7 2 T* 4 3 6} Umitep KINGDOM. 1861 ... 1871 ... 1881 ... 60,725 87,638 148,363 5 2 7 83,541 | 102,135 | 185.676 | 519 2 59,522 112,718 172,240 418 9 any appreciable saving could be effected, in view of a quinquennial census, by the establishment of a census office with a small permanent staff, as recommended by this Society in 1880, and again in the present year, is rather a matter for the Local Government Board to decide, but there would certainly be some advantages in such a plan. The regular work of registration would not be periodically disturbed, and the staff of highly-trained * The Irish tables in 1871 were much more extensive than in other censuses ; also the commissioners were called upon, when their work was partly done, to alter their tables so as to make them harmonize with those of the English census. X11 SUGGESTIONS FOR THE CENSUS 207 specialists would find employment in the comparatively short intervals between the publication of one census and the preparations for another, in preparing as required estimates of populations and returns of various kinds for parliament, royal commissions, select committees or government departments, and in conducting such inquiries as those desired by Messrs. Booth and Marshall, to which I shall shortly recur. But any way, suppose the cost of an additional census of the United Kingdom in each of the years 1896, 1906, &c. to amount to £200,000, what is such an expense to the richest country in the world ? Great Britain has always taken the lead in matters relating to the public health, and has been an example to other nations ; let her not lag behind in this matter, for the census is in reality a process subservient to sanitary operations. After all, the expenditure proposed is small in comparison with that granted (and wisely granted) by Parliament for agricultural returns. The present census may be taken to cost £20,000 a year: is it such a very serious thing to raise this amount to £40,000? In this connection it will be interesting to refer to the Act passed by Congress last year authorizing the eleventh census of the United States. This enacts that the enumeration shall be commenced on 1st J une, 1890, and completed before 1st July; but within fourteen days in the case of all towns which had in 1880 upwards of 10,000 inhabitants. In no census previous to that of 1880 was the enumeration completed under five months, a period approaching the nine months and twenty days of King David's census! The cost of the earlier censuses in America had risen from $44,000 in 1790, to $833,000 in 1840, and $3,361,000 in 1880; while the amount provided in this Act is, exclusive of printing, engraving, binding, &ec., $6,400,000, or say £1,280,000! The 208 STUDIES IN STATISTICS CHAP. snumerators are to be paid by results on the following scale— 2 cents for each living inhabitant, ? 2 death reported. ” surviving soldier, sailor, marine, or widow of such. 3 What effect such a method of payment may possibly have upon the population, the death-rate, or the pension list, an American might be better able to judge than an Englishman. The great elaboration of the United States census is largely owing to the fact, that under the constitution the federal authorities have no power to obtain statistical information, and attempts to obtain it would be opposed through the jealousy of States’ rights; therefore the powers conferred on the executive, by that provision of the constitution which necessitates a decennial census, are stretched so as to comprise inquiries into all manner of subjects not necessarily comprised in a census. The inference to be drawn from all this is, that, as recommended by this Society in 1880, and again in the present year— The Census Act of 1890 should be so framed as to establish the census upon a permanent quinquennial basis. II.— What Questions should be Asked in the Schedule ? Statisticians may be divided into two classes, (a) those who clamour for much information cn many subjects, even shough such information be confessedly very imperfect ; and (b) those who, being of a more sceptical turn of mind, prefer to ask for very little, and to concentrate their efforts on getting that little with the greatest attainable accuracy. The latter school has, perhaps fortunately, hitherto prevailed on English census commissions; the Scotch and KIT SUGGESTIONS FOR THE CENSUS 209 Irish officials have been more sanguine and more enterprising, perhaps for that reason more expensive. The English schedule, 1881, is divided into eight columns, viz., (1) name and surname, (2) relation to head of family, (3) condition as to marriage, (4) sex, (5) age last birthday, (6) rank, profession, or occupation, (7) where born, and (8) if deaf and dumb, blind, imbecile or idiot, or lunatic. As to the first seven columns it 1s, I think, plain that none of them is superfluous ; the eighth, however, I object to. People do not like to admit defects in them. selves or their children, and I should think that this demand that any afflicted members of the family should be specified cannot fail to irritate. Experience accordingly proves that the figures as regards idiots are very imperfect, so imperfect as to be valueless.! For this reason I am of opinion that inquiries as to those mentally afflicted might well be confined to the inmates of public institutions. The inquiries as to the blind and deaf-mute are perhaps less offensive, and, especially in the case of the blind, may be useful. Many additional points for Inquiry have been from time to time suggested ; such are— A. Particulars as to House Accommodation —The Statistical Society has asked for more information on this head in 1850, 1860, 1870, and 1880, asking especially for a return of the number of rooms, urging that the proportion of persons to rooms would be a valuable index of sanitary condition. The number of rooms with windows Was ascertained in the census of Scotland in 1861, 1871, and 1881, though the question was not included in the householder’s schedule. The housing of the working classes is a subject that now attracts a oreat deal of L Census, 1881, vol. iv. pp. 68, 69. 210 STUDIES IN STATISTICS CHAP. attention, and if the number of rooms can be readily ascertained without adding another question to the schedule, it might be well to do it. The proposal of the International Statistical Institute to classify families according as they dwell on the ground floor, the first floor, &ec., is evidently framed on lines that are suited to the lofty houses of continental towns, and perhaps to blocks of artisans’ dwellings in this country, but scarcely to the ordinary conditions of English life. It has also been proposed that dwelling-houses “to let” should be distinguished from other unoccupied houses.’ B. Divorced Persons.—The International Statistical Institute insists that divorced persons should be so returned.’ I do not myself in the least believe that such persons would enter themselves correctly, and I do not see that any material point would be gained if they did. ¢. Occupations.—It is in reference to these that most suggestions have been made. In the instructions on the back of the 1881 schedule, persons engaged “in trades, manufactures, or other business,” who employ others, are told to return themselves as “ masters,” and to state how many they employ. No use seems to have been made of this information, which was perhaps very imperfectly obtained. Mr. Booth, Professor Marshall, and a number of other political economists who are greatly interested in improving the statistics of labour, look upon the distinction between employer and employed as one of prime importance ; they therefore propose that a mew column should be inserted after that relating to “rank, profession, or occupation,” that in this should be stated whether the individual was “ employer,” ¢ employed,” or “ neither.” Their L Journal of the Statistical Society, vol. xliil. pp. 137, 138, 2 Bulletin de Institut International de Statistique, 1887, Tire Livraison, p. 366. XIII SUGGESTIONS FOR THE CENSUS 211 idea is that shortly after the census a special, more detailed, inquiry should be made as to all who returned themselves as employers. This suggestion strikes me as practical ; for while I think that the kind of information which these gentlemen seek could never be satisfactorily obtained through the census directly, it would be otherwise with a special inquiry made of the comparatively small number of persons who would be returned as employers. p. The Unemployed.—In the instructions on the back of the 1881 schedule, persons who may be at the time of the census out of employment, are instructed to return themselves as “unemployed.” Either the question was not answered, or the answers were unsatisfactory (as in my opinion they must always be); any way, no use was made of them. The Statistical Society does not appear ever to have asked for this information. E. Wages.—It has been proposed that men earning wages should be asked to state their amount. This also I do not find among the Society’s recommendations; and personally I am strongly of opinion that no answers of any real value would be obtained. F. Religious Profession.—A religious census is taken in most foreign countries : it is taken in Ireland, and in most of our colonies and possessions. The principal exceptions are Great Britain, France and Italy, and the United States. Before the unification of Italy a religious census was taken in Sardinia. The proposal to ask people in England their religious profession, a proposal that has been recommended by this Society in 1840, 1860, 1870, and 1880, has always excited an amount of feeling which I confess I find it difficult to understand. The Dissenters, it is stated, object that persons of no religion would be returned as members of the Established Church, while 212 STUDIES IN STATISTICS CHAP. many persons of all sorts of opinions say that a man’s religious opinions are his own affair, and that no one has a right to inquire into them. This last objection is, T think, largely founded on the misconception that the object of the census is inquiry into individuals, whereas inquiry is made of individuals solely with a view to obtaining results as to aggregates. However, while I think that a column should be added to the schedule for religious profession, T do not agree with the recommendation of the Society in 1880, that the filling in of this column should be compulsory. To make it optional meets, to a great extent, both the objections given above; the cnumerator is not compelled to return something, and is not therefore tempted to put down “ Church”; while the householder who objects to the question need not answer it. It would probably be necessary, where no entry was made opposite the names of children, to put them down invariably to the father’s religion (if any); otherwise small and enthusiastic denominations might get a fictitious increase of numbers by looking after their members’ children. It should be borne in mind that, according to Dr. Grimshaw,' no difficulty is experienced in Ireland, where very few persons avail themselves of Mr. Forster's conscience clause.’ I fully admit that the returns obtained would not be absolutely correct, nevertheless I cannot but think that the results of successive censuses would be most valuable for comparisons, and that most interesting tendencies to increase or decrease, in one direction or another, would be brought to light. Y Jubilee Volume of the Statistical Society, 1885, p. 154. 2 In the discussion which followed the reading of this paper, Mr, Martin J. Griffin, speaking of the Canadian census, said that they found no difficulty in that country in rerard to the relicious nuestion. X11 SUGGESTIONS FOR THE CENSUS 213 6. Education—Particulars as to schools have been asked for in days gone by, but I take it that all such information is now to be found in the reports of the Education Department. In 1870 the Society recommended that the schedule should ask, in reference to all persons aged seven years and upwards, whether they could read and write. ~The information might be interesting if it could be obtained, but as there seem to me to be points of greater importance, personally I should not press for it. H. Nationality.—TIt has been proposed to ascertain the nationality (or race) of foreigners, apart from their birth-places. More important is the suggestion to record the children of foreigners born in this country. Everything that throws a light upon the composition and changes in blood of a nation is to me of extreme interest. To carry out this proposal would, however, involve asking of every one the birth-place of his parents, as is done in the United States. Incidentally the replies would throw much light on the origin and constitution of our town populations. I. Sickness—It has been proposed to ask, in the present infirmity column, for a return of all those sick at the time of the census. This would no doubt give formation of interest to friendly societies and to the medical profession, but I see a serious fallacy in the fact that the amount of sickness, usually excessive at the time of the year when the census is taken, would be greatly influenced by the state of the weather immediately preceding the census day. The permanently infirm might be correctly returned. J. Consanguineous Marriages.—Lastly, it has been proposed by Sir John Lubbock and others, to settle once for all the disputed point of the alleged causation of STUDIES IN STATISTICS CHAP. idiocy, deaf-mutism, &c., by the intermarriage of near relatives. The point is one of undoubted interest, nay of real practical importance ; to an outsider, moreover, it might seem susceptible of easy solution. It would require a special column in the schedule in which should be entered against every married or widowed person whether or no there was any relationship between husband and wife, since the point to be ascertained is, what number of consanguineous marriages are contracted. It is not at all likely that this information would be correctly given ; moreover, we have seeri that the return of idiots is extremely defective, so that on more than one ground it would be impossible to draw valid conclusions from the returns obtained. [IT.— How are the Results to be Tabulated 2 It was the especial work of the International Statistical Congresses to secure as far as possible uniformity in the questions asked in the census. In this they obtained a very large measure of success, but M. Korosi, of Buda Pest, has shown what a little way uniformity in schedules goes towards securing uniformity in results ; the methods of abstraction and tabulation are at least as important, while much depends upon questions of interpretation and classification of replies. M. Korosi has made it a great object of his life to endeavour to secure such uniformity of method. M. Korosi’s admirable work may perhaps be said to err on the side of thoroughness, it is too logical and too complete for most English statisticians; I shall not attempt to discuss M. Korosi’s suggestions in detail, but since they have recently received the imprimatur of 4 Jubilee Volume of the Statistical Society, 1885, pp. 1569 and 171; also Bulletin de Institut International de Statistique, 1887, 1tre Livraison, p. 196. XIII SUGGESTIONS FOR THE CENSUS 215 the International Statistical Institute, I think it best to give them here. It should, however, be borne in mind that these recommendations have in view international comparisons, not what one may term domestic statistics, and that therefore they refer to the whole country, not to counties, towns, or other sub-divisions. As many of the questions alluded to do not enter into the English schedule, many of the recommendations are inapplicable. “ Conclusions voted by the General Meeting on the Proposal of the Census Commattee “To render possible international comparisons of the results of the enumerations of the populations of different countries, the committee proposes to adopt the following rules for tabulation— “1st. As regards the statistics of /ouses: the numbers of families and of persons of which they are composed to be tabulated according to the following categories: living in basements—on the ground floor—on the first floor—on the second floor—on the third floor—on floors above the third—families occupying several floors. “2nd. As regards statistics of famalies: families composed of a single person—of several persons—special establishments ; indicating in every case the numbers of families and of their members. “ Special establishments to be distinguished according to the following sub-divisions: Hotels—hospitals—schools and educational establishments—religious houses (other than educational)—jails and prisons—almshouses and charitable institutions — military establishments — other establishments. The numbers of establishments and of inmates to be given under each heading. v Bulletin de IInstitut International de Statistique. 1887, 1% Livraison, pp. 365-67. 216 STUDIES IN STATISTICS CHAP. “ Information as to Indwiduals. “3rd. Number and state of aggregation of the population : classification of parishes (communes) or other administrative units, according to the following classes : number of parishes with not more than 500 inhabitants— of 501 to 1,000 inhabitants—of 1,001 to 2,000—of 2,001 to 5,000—of 5,001 to 10,000—0f 10,001 to 20,000—of 20,001 to 50,000—of 50,001 to 100,000—of 100,001 to 200,000—0f 200,001 to 300,000—o0f 300,001 t0 400,000 —of 400,001 to 500,000—above 500,000. (It is desirable to give the enumeration of centres of aggregated populations composed of several communes, such as Manchester with Salford, Brussels or Vienna with its suburbs; this will give the aggregated populations of neighbouring communes which—although separated as far as their local government is concerned—represent nevertheless from the economic and social point of view only a single centre of population.) “4th. The distinction of the sexes should enter into all combinations in the tables of population. “5th. Ages should be tabulated by annual periods. (It is desirable to give in the table of annual periods of age a recapitulation in quinquennial groups.) “6th. Civil Condition should be tabulated under the following categories at least : unmarried ; married (including the separated) ; widowed ; divorced. “ Coil Condition should be combined with age in the following divisions: 0—15, 15—20, and so throughout in quinquennial groups of ages. “7th—8th. Religions and Nationalities— “(a.) All denominations returned in the schedules should be reproduced in the tables. “(b.) For the principal religions and nationalities. XII1 SUGGESTIONS FOR THE CENSUS 217 and more particularly for those which exceed 10,000 individuals, the following combinations should be shown :— “(a.) Age in decennial periods. “(B.) Total number of persons above fifteen years of age unable to read or write. “(c.) Principal religions combined with principal nationalities. “9th. Birth-place: The tabulation will be made separately for every one of the states indicated in the schedules. “10th. Number of persous over fifteen years of age in the total population wnable to read or write. “11th. Statistics of the population according to the Jloors of houses inhabited by them (see Resolution No. 1), and according to age (by quinquennial groups—see Resolution No. 5), to be given in the following groups :— “(a.) Country (places not exceeding 2,000 inhabitants). “(0.) Towns (of from 2,001 to 100,000 inhabitants). “(c.) Large towns (above 100,000 inhabitants). “12th. Infirmities, that is to say statistics of the blind, deaf-mute, lunatic, and idiotic. There should be shown : “(a.). The numbers of persons living in families or in institutions. “(b.) Infirmitics in the following combinations at least: number of blind who are also deaf—or dumb—or lunatic—or idiotic : number of deafmutes who are also either idiotic—or lunatic ; “(e.) Civil condition according to the four categories ; “(d.) Principal religions : ‘“(e.) Principal nationalities : “(£) Age in annual periods. 218 STUDIES IN STATISTICS CHAP. “Jt should be remarked that the questions proposed do not concern the statistics of counties (départments) or parishes (communes). It will be sufficient to reply to each question by a single number relating to the whole country.” As regards these recommendations, I would remark that the first recommendation, as I have said above (relating to floors of houses), is not applicable to England.’ The same remark, that it is unsuited to England, applies to the third resolution (as to the aggregation of population), at any rate so far as its form is concerned. I am in some doubt as to what English term is most akin to commune, but certainly English parishes differ so very much in area, that I can see no object in the table proposed, unless rural parishes are separated from those forming integral parts of towns. It seems to me that what is wanted is rather a classification of districts according to the density, not the mere amount of the population. A small urban parish of 2,000 inhabitants might be most densely, a large rural parish with a like number of inhabitants most sparsely, populated. As regards the fifth resolution, that referring to ages, the question has been exhaustively discussed in various reports of the Registrar-General, also in the report of the last census,’ where, with a view to giving as nearly as may be the information required, and at the same time to avoid fallacies, two tables showing the estimated number living at each year of age are given.’ Resolution No. 6—civil condition : I do not myself in the least believe that in this country the divorced would 1 Dr. Balfour reminds me that it 7s applicable to some of the Scotch towns, notably Edinburgh. 2 (Census, 1881, vol. iv. pp. 17—20. 3 Ibid. pp. 85, 86. X11 SUGGESTIONS FOR THE CENSUS 219 return themselves as such, neither do I see any practical object in that part of the resolution. As regards Nos. 7 and 8, the first part of the recommendation is evidently directed against unwise grouping. Mr. Risley informs me that the last census of India loses much of its value by the grouping together of religions and castes. The conditions in India are of course special, and the religious sects and castes have a practical importance from an administrative point of view. Of course so long as we in England do not ask people’s religious opinions, their ability to read and write, &c., we cannot classify these things. Quite apart from the recommendations of the Institute, there are several points concerning the abstracting of results and making of tables which deserve consideration. I am informed that no part of the work of tabulation 1s so laborious and so difficult as that relating to the occupations of the people,’ and certainly no part gives so little satisfaction to the large number of persons interested therein. Thus, for instance, a well-known labour representative in the House of Commons has complained that the important trade of the iron-moulders is not even mentioned. There is no distinction between employer and employed, a point of much interest. There is also a confusion between dealers and manufacturers. The reply of the experts of the Registrar-General’s department, is, that the schedules ask that employers should be distinguished from employed, but that the distinction appears in the replies in only a very small minority of cases; that so long as it is left to householders to fill in their own schedules it will be impossible to get them filled in 1 For this reason M. Korosi and the International Statistical Institute have as vet not dealt with it. 220 STUDIES IN STATISTICS CHAP, correctly ; nay more, that even on the Continent where the schedules are filled in by the police, a method very repugnant to the instincts of our people, the information obtained is very far from correct. Moreover, even with the best information, classification is very ditlicult— thus, for example, a man returned as a bootmaker may keep a small shop and deal in machine-made boots: at the same time he works at the last himself, and in busy times may employ one or two hands. How should he be described ? Then again, every one is familiar with the modern practice of dealers describing themselves as manus facturers, when in reality they make nothing. Men following two trades also cause difficulties. Thus A.B. returns himself as “farmer, brewer”; perhaps he has a large brewing business and a very small farm; whereas C.D. returns himself as “miller, farmer,” whereas in truth his mill is but a poor affair, while his farm is a large and valuable one. It would appear as if the difficulties of the last two classes of cases might well be met by a system of two-fold entry. Thus under farmers it might be stated in italics, “of whom . . . are also returned under other classes 7; and similarly after bootmakers, ¢“ of whom . . . are also returned as dealers in boots.” Then at the end there would be given a gross total from which second entries would be deducted, leaving a net total that would correspond with the total population. A point which admits of far easicr remedy is some sort of subdivision of the large mass of persons returned as of no occupation. If only the children and the married women were returned separately, it would minimize the evil of rash inferences drawn by ill-informed people. This alteration is such a simple one, that I trust it will be adopted in the following census. Mr. Booth and Mr. Marshall, as mentioned above, are X11 SUGGESTIONS FOR THE CENSUS 221 anxious that the schedule should indicate in the case of every person whether or no he was an employer—then that subsequently a second schedule should be delivered to all employers, by which much valuable additional information might be obtained. The only objections to this scheme that are apparent are (1) its expense, (2) its inquisitorial character. As regards the first it should be remembered that the second inquiry would only involve a minority of the schedules, and as regards both objections it should be borne in mind that similar detailed inquiries as to commercial matters are effected without much difficulty in the United States and in our colonies. Many have complained of the unsatisfactory lumping together of men of all degrees of skill, or want of skill, under the general heading of “labourers”; also that the labourers in the census form a class of themselves, and are not apportioned to the various trades to which they really belong. The farm labourer is so fur as possible kept apart, but it is otherwise with the bricklayers’ labourer. These objections are fully admitted by the officials, but it has hitherto proved impossible to get these men to describe themselves otherwise than as “labourers.” In Denmark! the women and children dependent on those engaged in each occupation are given, a point much desired by economists, but again one that presents great practical difficulties. In occupation statistics one uniform method of classification cannot be suitable for all countries; the attempt, for example, in 1881 to apply the English system to India led to the most absurd results. In any case there can be no doubt of the advisability of taking into consideration the way in which the workmen classify v Journel of the Statistical Society, vol. iil. p. 81. 222 STUDIES IN STATISTICS CHAP. themselves. If at the last census the great trade organizations had been consulted the iron-moulders would not have been omitted. As regards birth-places it has always struck me as very unsatisfactory that 56,378 persons are described as “natives of foreign parts, but British subjects,” while nothing is said of the countries of their birth, and no distinction is made between children of English parents born abroad, and true foreigners who have become naturalized. The census committee of the Society in 18801 suggested, to meet the case of the children of British parents born abroad, the addition of these words to the schedule : * If of ¢ British parents,” add those words.” The Statistical Society also recommended in 18807 that the Government should order returns to be made by various public departments giving detailed information as to the subjects under their purview at the date as near as might be of the census. Such for example as— Local Government Board, as to pauperism. Education Committee of Privy Council, as to education and schools. Charity Commissioners, as to charities. Prison Commissioners, as to criminals, Lunacy Commissioners, as to lunatics. Board of Trade, as to agriculture, &c. Inspectors of factories and mines. Also to ask for reports from railway and dock companies. Failing a proper religious census, it might be well to repeat with proper precautions the inquiry made in 1851 as to the numbers attending places of worship. Lastly, the want is ever more and more generally felt of a short abstract giving in a convenient form the 1 Journal of the Statistical Society, vol. xliii, p. 138. 2 Ibid. vol. xliii. pp. 135, 136. X11 SUGGESTIONS FOR THE CENSUS 223 results of the census of all the divisions of the United Kingdom together, but not in any way to supersede the several separate reports The increasing interest of late years in the development of our colonies has caused a very natural demand for something even more than this, for an abstract of the principal results of the census of the whole Empire, something more detailed than that given at the end of the fourth volume of the last English census.’ Such an abstract, for example, should give distinctions of races. If everything were prepared beforehand it might be issued a very short time after the returns were made up, and while the mass of details was passing through the press. i The conclusions then which I submit to you for discussion are as follows :— I. That the Census Act of 1890 should be so framed as to establish the census permanently on a quinquennial basis. II. That no inquiries should be made as to persons of unsound mind other than inmates of institutions. ITI. That the number of rooms in dwelling-houses should be recorded, but this should be ascertained by the enumerator, and not be upon the schedule. IV. That a column should be provided on the schedule in which each person should he entered either as “ employer,” “employed,” or “ neither.”? This question to be in view L Census, 1881, vol. iv. pp. 74, 75, and 112. ® Mr. Waters has called my attention to a point of detail which his experience teaches him to be of very practical importance. The words “employer ” and “employed” differ only in the last letter; it will therefore often be difficult or impossible to distinguish one from the other, and many errors and much loss of time will result. It would Le better to use such words as “ master” and “journeyman” (which may Le abbreviated into “M.” and “J.”), or some purely conventional signs, e.q. “A” for “emplover.” ©“ B” for “employed.” 224 STUDIES IN STATISTICS CHAP. of a subsequent special inquiry as to the employers. Should this special inquiry be refused by Parliament, of course the question on the schedule would fall with it. V. That there should be a column on the schedule for religious profession ; the filling up of this should, however, be optional. VI. That persons following more than one trade, or being at the same time manufacturers and dealers, should be entered under both headings, and the requisite corrections made in totclling the occupations. VIL That in the classification of occupations the great trade organizations should be consulted. VIII. That the large mass of persons now placed under the heading “ persons returned by property, rank, &e., and not by special occupation,” should be broken up by separating all children under a certain age, and all married women not otherwise specified. IX. That the country of birth of ‘ natives of foreign parts, but British subjects” should be given; also that those born of British parents should be given separately from those naturalized. X. That an abstract should be published giving the principal results of the census of the entire British Empire. Doubtless many more committees will present reports, and on many future occasions the council of the Royal Statistical Society will make representations to the Government, before a parsimonious Treasury will undertake to meet all the wishes of ever-grasping statisticians, Still we must persevere ; according to the laws of our being we must perforce continue to ask for that which we think we ought to have. If on the one hand we look back and see how many disappointments we have suffered in the past, possibly we may feel sorely discouraged ; but on the other X11 SUGGESTIONS FOR THE CENSUS 225 hand if, with the Government statistics of to-day in our hands, we try to picture the materials with which the statisticians of 1835 had to work, we ought to feel truly grateful for our improved resources. If in this paper I have dealt with the question of a quinquennial census at what may to many appear to be undue length, I have done this with intention, feeling as I do that this is not only the reform most wanted at the present moment, but the one that we are most likely to obtain, and one that if obtained would tend more than anything else to the gradual perfection of the census of the British Empire. 126 STUDIES IN STATISTICS CHAP. CHAPTER XIV. THE RECENT DECLINE IN THE ENGLISH DEATH-RATE CON-SIDERED IN CONNECTION WITH THE CAUSES OF DEATH, [Read before the Statistical Society, 18th March, 1884.] OnE of the most striking facts of the day, from the statistician’s point of view, is the remarkably low deathrate that has prevailed in this country during the last eight years. From 1838 to 1875 inclusive the death-rate averaged 22-3 per thousand, but during the eight years 1876-83 inclusive, it has averaged only 20°83. The only approach to such a low mortality was in the five years 1841-45, when the death-rate averaged 214; so that the recent low rate is both considerably lower and extends over a greater number of years.’ Mr. Noel A. Humphreys, in his paper last session,’ proved conclusively that if the English death-rate should continue at the low average of the five years 1876-80, the mean duration of male life in this country would be increased by 2:0 years, and that of female life by no less than 3'4 years, as compared with the English Life Table. Moreover, among males 70 per cent., and among females 65 per cent., of this increased life would be lived between L Forty-fourth Annual Report of the Registrar-General, p. xliii; Quarterly Return of Marriages, &e., No. 140, p. ix. ? The Recent Decline in the English Death-Rate and its effect on the Duration of Life.—Journal of the Statistical Society, vol. xlvi. n. 189, XIV THE DECLINE IN THE DEATH-RATE 227 the ages of 20 and 60 years, or during the most useful period. At Mr. Humphreys’ request I have taken up the question where he left it, and have examined the changes that have taken place in the death-rates at different ages in the two sexes from various causes of death. The subject is intricate, surrounded with difficulties, and though entailing considerable labour, is perhaps not productive of adequate results. I have to thank Mr. Humphreys and his colleagues at the General Register Office for help in many ways, but must take upon myself the entire responsibility for the accuracy of the figures put forward, and the conclusions drawn therefrom. The somewhat voluminous tables no doubt contain errors, but I feel confident that these cannot be of much importance, since the whole of the calculations have been checked, and all unexpected results carefully re-examined. Moreover the calculations have been made by the aid of the Arithmométre of M. Thomas, which, in my hands at any rate, gives more accurate results than logarithmic computation, besides being infinitely less laborious. As T shall have to deal entirely with the Registrar-General's Returns, it is well that I should say in limine that after carefully reading all that has been advanced in recent years against “The National System of Vital Statistics,” I am thoroughly convinced of the soundness of that system and the fallacies inherent in all attacks upon it. Moreover, having studied for several years the figures relating to “Alleged Causes of Death,” I have been more and more convinced of the value of those figures, and I fully believe that they may be taken as, on the whole, a fair approximation to the truth. At the same time it is hardly necessary to say that like all other statistics they require care and knowledge in handling. STUDIES IN STATISTICS CHAP. Without doubt the figures relating to alcoholism, venereal diseases, and perhaps insanity, are almost valueless ; but that does not prove that those relating to scarlet fever, pneumonia, or cancer are equally valueless. Neither does the fact that a large number of certificates are carelessly filled up, invalidate the far larger number that are more trustworthy : indeed these very sources of error are subject to laws, and are more or less constant factors of the whole. When it is possible, as I have elsewhere’ proved it to be, to find general laws regulating many of the causes of death, and especially mutual relations between these causes, and relations between some of them and various external phenomena, the only possible inference that I can deduce is that the figures dealt with are the expression, more or less accurate, of facts in nature. The national importance of the subject must be my excuse for this digression. But in spite of what I have just stated, it must be admitted that there are numerous fallacies to which the classification of deaths according to their alleged causes is liable. They are— Firstly, in a number of cases no cause of death is specified at all. In 1847 about 4 per cent. of the deaths came under ““ cause unspecified,” but in 1880 this proportion was reduced to less than 1 per cent. Secondly (and this is a much more important source of error), in a larger number of cases the cause is stated in such a vague or indefinite manner as to be almost valueless. The terms “atrophy,” “debility,” “old age,” “ heemorrhae,” ““ dropsy,” &c. give but little information ; but it is satisfactory to note that, from the commencement of the reports in 1838 to that last issued, the improvement in this respect is steady and continuous. For example, 1 See chapters xv., xvi,, xviii, and xx. of this work. X1V THE DECLINE IN THE DEATH-RATE 229 very few cases are now returned simply as “ dropsy,” but the dropsy is referred fo disease of the heart, kidneys. liver, or ovaries, Thirdly, where the cause is definitely assigned, there is good reason to believe that in a large number of cases, owing to carelessness, ignorance, a desire to spare the feelings of relatives, or even less worthy motives, the causes of death assigned are barely approximately correct. The returns themselves indicate that this source of error is also diminishing. Fourthly, medical science like most other things is more or less under the dominion of fashion, hence the same disease is at different times given different names. By the aid of a little medical knowledge this can to a large extent be allowed for. Before entering into fuller detail it will be convenient, without taking into account sex and age, to ascertain what alleged causes of death appear to be growing more and more potent, and what on the other hand are less and less frequently assigned. For this purpose we may utilize Table 84 in the Registrar-General's Forty-Third Annual Report. In this are given the mean annual rates of mortality for each period of five years from 1850 to 1879. From it the table on the following page is constructed. This table suggests several remarks. In the first place it will be noticed that six so-called ““ zymotic discases,” viz, fever, cholera, small-pox, scarlet fever, diarrhoea, and measles, between them contribute a fall of 1,375 deaths per million, against which there is a set-off of 114, being the rise under the heads of diphtheria, croup, and whooping-cough, leaving a net fall of 1,261 deaths per million from the principal zymotie discases. But as the total fall 230 STUDIES IN STATISTICS CHAP. TABLE A.—RISE OR FALL IN THE DEATH-RATES PER 1,000,000 PERSONS LIVING IN ENGLAND AND WALES, FROM VARIOUS CAUSES OR GROUPS OF CAUSES. AVERAGES OF QUINQUENNIUM 1875-79, COMPARED WITH AVERAGES OF QUINQUENNIUM 1850-54. Risen per 1,000,000 Living. Fallen per 1,000,000 Living, Lung diseases ... ... 1,213 Phthisis aha 694 Heart , ve es 696 Developmental diseases 617 Brain diseases, excluding } Fever ... ee 569 . 378 convulsions ... Dropsy ... we 405 Kidney diseases ... 229 Convulsions... . 342 Cancer ... vn 191 Cholera ... ves ve 287 Diphtheria and croup 69 Small-pox en i. 197 Tabes mesenterica 65 Scarlet fever ... ver 173 Whooping-cough ... 45 Hydrocephalusor tuber-Rheumatism ee 45 cular meningitis It Liver diseases! ... 32 Sudden death — cause) unascertained Diarrhoea cee wi Diseases of stomach and intestines 258 | Measles ... ee . All other causes . 2,963 Total ... a a fall of 1,049. from all causes is 1,049, it is evident that the deaths from non-zymotic diseases must have risen 212 per million. The increased mortality from lung diseases is remarkable. It will be shown in the sequel that this is not to be explained by the fall in phthisis, even were that fall large enough to explain it. Doubtless, however, some of the cases (but not very many) now classed under bronchitis would have been attributed to phthisis thirty years ago; also some cases formerly attributed to old age alone are now more properly classed under diseases of lungs, brain, 1 The rise of 32 per million under “liver diseases” is made up of a fall of 25 under hepatitis, and of 20 under ascites, and a rise of 77 under « other liver diseases.” X1V THE DECLINE IN THE DEATH-RATE 231 or kidneys. Among the diseases of the nervous system there is a fall under the vague term “convulsions” in the first few years of life, nearly as great as the rise under the other members of the group at later ages. Now it is- obvious that errors due to all the four causes mentioned above, and more especially errors due to changes in nomenclature and classification, will be minimized by comparing two periods as near together as possible, and for my present purpose it is convenient to take the periods 1861-70 and 1876-80. The facts for the first period are at once available in Tables 2 and 3 of the Supplement to the Thirty-Fifth Annual Report of the Registrar-General, which give the death-rates for males and females separately, from all causes, and twentyfive separate causes, at twelve groups of ages. My tables differ from these in the following respects :— (1.) The rates are per 100,000 living at each age, instead of per million, as I believed that the essential features would be rendered somewhat more clearly by the ase of fewer figures. (2.) T group all ages above 75 together, since the returns of ages of very old people are untrustworthy. and the numbers are small. (3.) In place of twenty-five causes or groups of causes I have used only sixteen. This was done mainly to save labour, but partly because some of the causes given by the Registrar-General, such as *‘ diseases of the skin,” were unimportant for my purpose. If it had been practicable without very great labour, I should have preferred to separate ‘‘ convulsions” from other brain discases, and liver diseases from other diseases of the digestive organs, the group, “ Discases of the Digestive System,” being too heterogeneous to have much value. With the figures so obtained I have compared like 232 STUDIES IN STATISTICS CHAP. rates for the five years 1876-80. These rates were laboriously calculated from the numbers of deaths given in each annual report, which had to be cast and the averages taken. At first I had wished to take advantage of the 1881 report, issued since Mr. Humphreys paper, but decided not to do so, for it seemed best to consider the same years as he did; moreover, the classification of causes of death was altered in 1881 in such a way that I found the difficulties greatly increased when I began by including that year. Tables I, II, V, VIII, form a double set, Ia, 11a, &ec. relating to males, 1B, IIB, &c. to females. All four tables relate to the same facts, namely, the average annual numbers of deaths in England and Wales from certain causes at certain ages. The french type refers to the ten years 1861-70, and the roman type to the five years 1876-80. Tables Ia and IB give the actual average deaths. Tables IIa and IIB give the death-rates per 100,000 living at each age. Tables Va and VB give the percentage rise or fall of the death-rates in the second period as compared with the first. Tables VIIIA and VIII give the gain or loss in lives per annum consequent on such fall or rise in the death-rates.! By the use of the four tables together it is hoped that the real importance or otherwise of the fluctuations will 1 Tables Ta and Is are derived from the Supplement to the Thivty-fifth Annual Report of the Registrar-General, pp. 2, 3, and from the Annual Reports for 1876-80, pp. 146 ef seq. in each report. Tables T1a and IIB are derived in part from Tables 2 and 3 in the Supplement to the Thirty-fifth Annual Report, in part are calculated from Tables Ia and Is by the aid of the populations at each age given in the census, from which mean populations for the period in question were calculated. Tables Va and Vs, VIIa and VIII are derived from the preceding tables by calculation. Riv THE DECLINE IN THE DEATH-RATE 233 be shown. An example will make my meaning clear. Among females over 75 years of age the death-rate from diphtheria fell 40 per cent. : the consequent gain was only one life per annum. Among females between 45 and 55 years of age the mortality from lung diseases rose 5 per cent.: the consequent loss of lives amounted to 126 per annum. In the examination of these tables it will be convenient to consider those causes of death first which Table A shows to have been long decreasing in fatality, and subsequently those which appear to have increased. Phthisis, or pulmonary consumption, is very fatal at all ages from 15 to 75, but most so between 20 and 55 ; at ages 5—20 1t is more fatal to females than to males; after 45 it is more fatal to men. The death-rate has declined at every age with one insignificant exception ; the decline being for males 14, for females 22 per cent. For both sexes the decline was greatest (28 and 27 per cent.) between the ages of 15 and 20, when deaths may be considered, in an economic sense, the greatest loss, since the individual of that age has been supported by society for a great number of years, but has as yet done but little in return. The decline has been not much less considerable among both males and females at all other ages below 25, varying from 18 to 26 per cent. The mortality of males above 85 has diminished very little, but that of females has fallen between 14 and 24 per cent. until the age of 75, after which there is a trifling rise. The lives saved amounted to 3,966 males, 6,806 females ; not less than 2,885 of the former and 4,233 of the latter were of ages 15— 85, that is to say, were among the most useful of the community. The question has often been considered as to how far this well-known fall in the STUDIES IN STATISTICS CHAP, mortality from phthisis is due to “medical fashion,” that is, how far it may be a mere question of classification, since it is equally well known that the mortality assigned to the diseases of the respiratory organs (principally bronchitis) is increasing. Now there are two answers to this question. Firstly, it would appear that the discovery of the so-called bacillus of tubercle tends to confirm the opinions of the older writers that phthisis is a distinct disease sus generis, with many forms of manifestation, and not an ill-assorted mixture of unrecognized diseases as has been more lately taught. Secondly, there is an argument that is more in place here: the mortality from phthisis is shown by the table to have diminished at the ages 5—35 to an extent varying from 15 to 28 per cent., whereas at all these ages the mortality from diseases of the lungs remained comparatively stationary. This is especially well seen in Tables VIIIa and VIIIB. I have elsewhere® shown other reasons for believing that substantially the diagnosis of phthisis from bronchitis corresponds with nature. We must therefore credit medicine, whether preventive, curative, or both, with another triumph here ; and let it be remembered that this triumph means the saving of many lives when they are most valuable. There is still ample material to work upon, plenty of room for further improvement. Developmental diseases come next in the list of those that have decreased in fatality during the last three ycars. In the tables these fall under the head “all other causes.” The most important of them are “atrophy and debility,” relating almost entirely to young infants, and “old age,” which explains itself. «AU other causes” account for about one-third of the deaths of children under 5, about half the deaths over 75 years. The death-rate has 1 See chapter xviii. of this work. X1V THE DECLINE IN THE DEATH-RATE 235 fallen 11 per cent. for males and 18 per cent. for females, and the fall has takén place at every age of both sexes; this fall is tolerably even throughout, at least it does not exhibit the marked preference for the two extremes of life : which might have been expected. The “lives saved” number 8,425 males, 9,837 females; of these numbers 4,346 males and 4,372 females are in the first five years of life, and may be fairly attributed to the better management of children and improved sanitary surroundings, which have diminished the deaths under the vague headings “ teething,” “atrophy,” and the like. Of the rest, 2,431 male and 3,086 female lives saved are at ages over 65, at which period there is a large increase of deaths from local diseases,” as they are called. Diagnosis is now more accurate, and certificates are more carefully filled up, hence fewer deaths are attributed to “ old age.” : “ Fever” comprises typhoid or enteric fever, now the most important by far; typhus fever, which is still prevalent to some extent in Ireland, and lingers in the Irish quarters of Glasgow, Edinburgh, and Liverpool ; and some less defined diseases numerically not very important. Fever kills males and females about equally, and at all ages, the two extremes of life suffering somewhat more severely. The percentage fall in the death-rate is very remarkable, being no less than 57 per cent. for each sex, the fall being spread over all ages, though old people gain most ; the number of “lives saved” is no less than 6,188 males and 6,510 females yearly ; the salvage being greatest under 5 years of age, but very considerable at each age below 75. Tt seems impossible to dissociate this great saving of life from the operation of the Public Health Acts, and more especially the indirect results of this as shown in the proceedings of such bodies as the Sanitary Institute, the exhibitions of sanitary appliances under various auspices, 236 STUDIES IN STATISTICS CHAP. and other proofs that public attention has been drawn to these matters. The fall in the death-rate from fever is without doubt the great triumph of the sanitary reformers. Typhus has been driven out from place after place by measures taken to check overcrowding and want of ventilation, until it now lingers only in the lowest quarters of a few large towns. Things are not perfect yet, far from it; but they were once much worse, and that not so long ago. On the other hand, measures directed to improve drainage and water supply have been most successful in restraining enteric fever within comparatively narrow limits. Dropsy.—All that need be said about this is that in 1850 no less than 9,981, or 2:7 per cent., of the total deaths had this cause assigned, whereas in 1880 only 3,192, or 0°6 per cent., of the total deaths were so classed. This does not mean that fewer persons die of dropsy now than then, but that the dropsy itself is now more commonly attributed to its proximate cause—disease of the heart, kidneys, liver, or ovaries. In the tables “dropsy ” is included under “Heart diseases,” but it has now ceased to play such an important part among the assigned causes of death as seriously to affect the correctness of the returns. Convulsions are included in the tables under brain diseases. Scarlet fever has diminished notably, but from its usually occurring in epidemics, which are possibly iufluenced by meteorological conditions, it is not safe to draw conclusions as to the future. It is of course a disease of childhood, most of its victims falling before the age of puberty. In both small-pox and diphtheria’ there has been a L Since 1880, diwphtlizria has shown a decided tendency to increase. X1v THE DECLINE IN THE DEATH-RATE 237 saving of life, but these diseases now cause such small mortality that their’ total extinction would only reduce the general death-rate by 0:2 per 1,000. The diarrhea figures are confused by the occurrence of the epidemic of Asiatic cholera in 1866. Diarrhea kills mostly at the extremes of life, cholera has not this marked preference. This explains the fact that under diarrheal diseases the fall of the death-rate is greatest from 5 to 75. There is room for great improvement under this head. The mortality from measles has diminished notably, but whooping-cough - has been stationary. Whoopingcough is much more fatal to girls than to boys. Measles causes nearly five times, -and whooping-cough more than six times as many deaths as small-pox, while the amount of mischief to be attributed to these little-thought-of ailments in the way of impaired general health, permanent lung disease, and even blindness and deafness, will probably never be known. It would appear that as yet preventive medicine has failed to control these diseases; can this be because the people will not co-operate to put down diseases which affect only young children ? Turning now to the other side of the account; first in importance come lung diseases. Like diseases of the brain, and in contrast with phthisis, they are most fatal in the first five years of life and after 45. Again, lung diseases are 21 per cent. more fatal to males than to females, but this difference does not apply to the ages 5—20, as boys do not appear to suffer more than girls. In the case of phthisis, it may be remembered, the mortality of females from 5 to 20 exceeds that of males at the same ages. Lung diseases have increased in the period under consideration, among males 13, among females 14 per cent.; the increase being mainly in the first five years of life and after STUDIES IN STATISTICS CHAP. 35. There is a slight fall among boys of 10—15, a more marked fall among girls and young women from 10 to 25. The annual “loss of lives” from the increased prevalence of “lung diseases” amounts to 6,490 males and 5,404 females; of these 8,486 males and 2,557 females were under 5 years old, while 2,919 males and 2,952 females were aged 85 years and upwards. The extra deaths among women of 65 years of age and upwards amounted to 2,267 against 1,544 among males of like ages. Doubtless at both extremes of life there has been a transference from such assigned causes as “atrophy ” and “old age,” or “debility of age,” to bronchitis and other diseases of the lungs. Heart Diseases and Dropsy.—It is necessary to take these together, since Dr. Farr did so in the table which I have adopted as my standard. In the first period dropsy made up 25 per cent. of the whole, in the second period only 9 per cent. But, ‘as in a large proportion of the cases (Dr. Haviland put it at three-fifths of the whole number), the dropsy was certainly caused by heart lisease, the error due to grouping the two together is not so important as might at first sight appear. Females suffer from heart disease somewhat more than males, but in both cases the death-rate is much higher after 35 than before. The mortality of males has increased 14 per cent., of females only 11. In both cases there has been a notable fall at the earlier ages, but a considerable increase, greatest in males, after the age of 35. Whereas 344 male and 242 female lives have been saved before the age of 20, at all ages over 20 no less than 2.909 male lives and 2,363 female lives have been lost annually in the latter as compared with the former period in the tables. Corresponding with this increase is an increase (not shown in the tables) in deaths from X1v THE DECLINE IN THE DEATH-RATE 239 rheumatic fever, which is a most fruitful cause of heart disease. Brawn Diseases.—These affect males considerably (21 per cent.) more than females; they are most fatal at the extremes of life. The mortality at all ages has not materially altered, but it is very different when the separate ages are considered. In the first five years the male death-rate has fallen 13 per cent., the female 15 per cent. This we know from other sources! to be very largely due to the fall in “convulsions.” In contrast with this, at all ages over 35 there is a rise among both males and females of from 5 to 22 per cent. On one hand 2,448 male children and 2,227 female children that would have died annually from diseases of the nervous system under the death-rate prevailing in 1861-70, now survive under the rates prevailing in 1876-80; Lut on the other hand at ages over 35 no less than 2,098 males and 2,078 females now die that would have survived, or died of some other cause under the former régime. Kidney diseases are in both sexes most fatal in middle and advanced life; they are about twice as fatal to men as to women. Female mortality has, however, increased nearly twice as much under this head as male, although the increase is very great in all cases, varying from 20 to 81 per cent. The total loss of life from increased renal disease amounts to 1,702 males and 1,417 females, mainly in middle and advanced life. Cancer, in contrast to renal disease, is twice as fatal to women as to men; it is rare in early life, but steadily increases in frequency from the age of 25 upwards. Cancer has increased 38 per cent. in males, 24 per cent. in females, the greater increase in males being probably due to the fact that cancer of the stomach and liver, V Forty-fourth Annual Revort of Reaistrar-General. Table 33. STUDIES IN STATISTICS CHAP. which is commoner in men than women, is much more difficult of diagnosis than cancer of the female breast or of the uterus. Hence improved medical skill affects the returns for it more. The loss of life due to the increased mortality from cancer amounts to 1,187 males and 1,661 females, of whom seven-eighths are above the age of 45. A recent writer! said he was convinced that the longcontinued and steady increase of cancer was not apparent only, and accounted for by increased accuracy of diagnosis and registration, but was an undoubted fact; the cause is quite unknown, but must probably be sought in some abnormal circumstances of our artificial existence. It should be remembered that very few, if indeed any, recover from this much-dreaded disease, and also that it chiefly attacks after the reproductive age is past. Irom these facts we may draw the consolation that if there are many killed, there are no wounded, and that although the tendency to cancer may be handed down bo offspring, they are not born enfeebled in consequence of their parents’ ailment. As regards the other diseases in the tables, I will only remark that the increased number of deaths from violence at the higher ages, accompanying diminished death-rates at the lower ages, is a curious fact that is difficult to explain. 240 To make the various relations and changes above indicated more striking, I have constructed the summary Tables IIIa, IIIB, IVa, IVB, VI and IX,? showing the same 1 H. P. Dunn, F.R.C.8., in British Medical Journal, 1883, pp. 708, &e. Z See also an article, “ The Recent Increase of Cancer,” by Dr. Herbert Snow, Nineteenth Century, July 1890. 8 These tables are all obtained from tables Ia and Is by caleulation. XIV THE DECLINE IN THE DEATH-RATE 241 facts as Tables I, II, V, VIII, but for three groups of ages only. First, all under 20 years of age, or the dependent period of childhood and youth; second, active working life, the productive period, from 20 to 65; and third, the dependent period of old age, after 65 years. Mr. Humphreys, in his corresponding period, drew the line at 60 years; I could not have followed him exactly without great trouble. These short tables show firstly, that in both sexes the death-rates from “all causes ” below 20 had decreased considerably, between 20 and 65 they had also decreased, but much less, whereas over 65 there was a very small increase. Secondly, the rates of decrease were largest among females, whereas the increase was largest among males. Thirdly, the most marked increase in the death-rates from the several causes was from diseases of the kidneys, more especially among females, and next from cancer in men. So that it would appear there is a tendency for the mortality of the two sexes from these causes to approximate. It is possible that increased drinking among women may account for some of this increased fatality from kidney diseases. Fourthly, the great decrease of mortality from fever at all ages is shown in a very striking manner. Fifthly, Table IX. shows conclusively that increased precision in diagnosis, and care in registration, will not account for all the increased deaths from diseases of the lungs, brain, heart, &c., since the increase under these heads greatly exceeds the decrease under “all other causes.” The same tables show in an equally striking manner that ‘ phthisis” and “lung diseases” in the reports indicate, in the main at any rate, essentially different causes of death, since the fluctuations are in no way reciprocal. 242 STUDIES IN STATISTICS CHAP. After this analysis it is possible to consider the bearing of the facts ascertained on the following problems. Supposing that the last twenty years may be taken as a fair index of what we are to expect in the future, ¢. e. that those causes of death which have increased in fatality will continue to do so, or remain stationary ; and conversely, those which have decreased in fatality will either become stationary or continue to decrease— what will be the effect on— (1.) The duration of life ? (2.) The sex and age distribution of the population ? And (3.) The rate of increase of the population ? In the last case the question of emigration will have co be considered. (1.) Mr. Humphreys’ paper, although appearing to do so, did not give a complete solution of the first problem, at least as far as a forecast of the future is considered. If an aggregate variable be made up of several component variables, which latter vary at different rates and in different directions, then, given two terms of a series, it is possible to get an approximation to subsequent terms by two different methods, viz. either a calculation from the known variations of the aggregate, or a similar calculation from the known variations of the components. To take as an instance the estimation of populations from the known increase in the interval between the last two censuses. The first method is that officially in use in this country, the largest available areas are taken for calculation. Probably this gives on the whole the best results, provided always that it is applied to a eentre of population complete in itself with vacant space on all sides for expansion. Where these conditions are not Xv THE DECLINE IN THE DEATH-RATE 243 fulfilled," a better result may in some cases be obtained by estimating separately for the constituent parts and adding these partial estimates together. It is of course possible in this second method to give undue importance to local changes of a special and possibly exceptional character, so that perhaps the best plan of all is to combine the two methods by taking into consideration the past actual and future possible growth of each ward or other district of which the area is composed. The same holds good with death-rates; the gross death-rate of either sex at a particular age is made up of a number of death-rates from different causes, some of which are falling, others rising, while others are stationary. It is therefore a point of special importance to remember that whereas a diminishing death-rate can only fall to zero, an increasing death-rate may conceivably rise indefinitely. This is well illustrated by the cases of fever and kidney diseases. In the aggregate no less than 12,698 fewer persons (in proportion to population) died from fever annually during 1876-80 than during the period 1861-70. But as the total number of annual deaths from this cause is now but 9,576, an equal further saving of life is no longer possible. On the other hand, there is no such obstacle to the mortality from kidney diseases rising again by 3,000, or indeed by 80,000 per annum. It is important therefore, in ¢ looking ahead,” to remember that great saving of life can only be accomplished where there is now heavy loss of life. Small-pox bas happily been reduced by efficient vaccination within such narrow limits that if it were extinouished to-morrow, 1 A good instance is that of the county of Essex, whose peculiarity consists in the fact that it is contiguous to London along a considerable portion of its western boundary. Dr. John C. Thresh in Public Health, February 1890, p. 323. STUDIES IN STATISTICS CHAP. the result would be trifling compared to a fall of 50 per cent. in the mortality from whooping-cough, or a full of only 10 per cent. in lung diseases. In the period under consideration the greatest gain of life has been under fever, phthisis, scarlet fever, diarrheeal diseases, and ““ all other causes ” ; under all these headings there still remains a considerable margin for profit, although this is getting narrowed in the case of fever. Scarlet fever is practically confined to the young; phthisis has special incidence on youth and early working life, and is of all diseases that which is most destructive to the most valuable lives; fever attacks those of all ages, but it is more fatal at the early ages than phthisis; “all other causes” account for many deaths of both young and old. With the exception of the last, a continued diminution of deaths from any of these causes would tend to increase the duration of active life; in the case of “all other causes,” a great portion of the saving would mean only some prolongation of old age, which, as it would almost certainly be accompanied with increased vigour of the survivors, would be a benefit only in a degree less. On the other side of the account we have diseases of che lungs, heart, and kidneys, and cancer. The three last exact most victims in middle and late life, the lives cut off by them are therefore less valuable than those that fall to phthisis ; lung diseases are extremely fatal to the young as well as to the old, so that they have a point of resemblance to the zymotic diseases. If all these diseases should continue to increase in fatality in like manner, the result on the duration of life would be in the case of the first three to cut short its later years, in the case of diseases of the lungs to cut it short at both ends. On the whole, balancing these results, we may safely say that if the individual causes of death alter in fatality XIV THE DECLINE IN THE DEATH-RATE 245 in the future as they have done in the recent past, for a time the duration of life may continue to increase, but ultimately the tendency would be for more people to survive the dangers of childhood and adolescence, but fewer to reach old age. Useful life would not be greatly shortened, and as the proportion of persons of the reproductive age would be increased, the birth-rate would also, presumably, be increased. (2.) The Report of the Census of 1881 (vol. iv. pp. 22 and 23) calls especial attention to the importance of taking into consideration the constitution of a population as regards sex and age, before comparing its death-rate with that of another population. In illustration, the sex and age constitution of the urban and rural sanitary districts are contrasted, and it is remarked, “if we take the mean (1871-80) death-rates in England and Wales at each ageperiod as a standard, the death-rate in an urban population, as constituted above,! would be 20-40 per 1,000, while the death-rate in the rural population would be 22-83. Such would be their respective death-rates on the hypothesis that the urban districts and the rural districts were equally healthy. We know, however, as a matter of fact, that urban death-rates, instead of being lower than rural death-rates, are much higher. The difference of healthiness, therefore, between the two is much greater than the difference between their death-rates.” This is a most important passage, and expresses a truth not generally known, and when known not sufficiently dwelt upon. As the work of calculating thus the ‘“ normal” death-rate of a population is somewhat laborious, it would, I think. be of oreat service to medical officers of health ! This is in reference to a table (Census, 1881, vol. iv. p. 22) showing the age and sex constitution (1) of London and the Urban Sanitary Districts in England and Wales, (2) of the remainder of the eountry. 246 . STUDIES IN STATISTICS CHAP. and others, if the Registrar-General’s Department would publish a table of factors by which the death-rates of the principal towns ascertained in the ordinary way could be reduced to the normal (e.g. in the circumstances given, the rural factor would be 93176, the urban factor -04275).> The age and sex constitution of the population of Fngland and Wales has changed, though not to a very great extent. It would be interesting to know what influence these small changes have in themselves had upon the death-rate and birth-rate. Any one approaching the subject on a prior: lines might well maintain that the agglomeration of human beings in towns affects a constantly increasing fraction of the people, and therefore since. life in towns is necessarily less healthy than life in the country, the tendency would be for the death-rate to rise. The first proposition is undoubtedly true: the town population in 1851 was but 51 per cent. of the whole, in 1881 it was 59 per cent. ; the increase has been and appears likely to be continuous. The second premise has appeared to be true hitherto, but nevertheless the conclusion has not followed. The explanation would seem to be that the increased attention paid to hygienic matters by the individual, the corporation, and the State, has more than counterbalanced the evils attendant upon density of population. This must not, however, be pushed too far; the ‘sanitary ” surroundings of many town populations are now superior to those of sountry populations, but yet the death-rates of these populations remain higher than those prevalent in the country ; whereas the above extract from the last census clearly shows that owing to peculiarities of age and sex constitu-1 The Registrar-General did publish such factors within a few weeks ifter the reading of this paper. See Annual Summary for 1883, p. v. 2 Calculated from a table in Census of England and Wales, 1881, rol. iv. ». 9. Xv THE DECLINE IN THE DEATH-RATE 247 tion they should be lower. The increased mortality from cancer and several of the “local diseases,” the increase in lunacy (though this is not so great as it appears),' the deterioration of eyesight, the unquestioned premature decay of the teeth of the rising generation, and to a much smaller extent the prevalence of premature baldness, convince me that the a priori argument 1s a sound one. There are conditions attendant on life in towns which no hygienic knowledge, no improved administration, not even a new municipal government for London, can wholly overcome. Some amount of degeneration of race appears inevitable, and in the United States of America, where all sociological phenomena are seen in more rapid evolution than here, it is said to be even already very evident. This seems to be a fitting place to meet an objection that will be brought forward by many ; they will say—* One thing is plain, no sanitary improvements, no amelioration of habits, will enable men and women to live for ever. The increased mortality from local diseases towards the end of life is the inevitable correlative to the diminished mortality in infaney.” There is a fallacy of time inherent in this objection. The total period in review from the first year to the last extends over only twenty years, while the average interval, from the middle year of the first period to the middle year of the second, is only twelve and a half years. The recent fall in the death-rate of children under 10 did not begin until the year 1867,” so that but few of the extra survivors could have attained their majority by the middle of the second period, far less could they have added some thousands annually to the deaths of persons aged 65 and upwards. Our records do not go L Mr. Noel A. Humphreys has recently shown that it is very doubtful whether lunacy is increasing at all in this country. Journal of the Royal Statistical Society, vol. liil, p. 201. 2 Forty fourth Annual Report of Reqistrar-General, Tables 24, 25. 248 back further than 1838. We know that during the five years 1841-45' low mortality prevailed at ages over 25, and during the first five years of life. The younger adults of those days, if now surviving, would be old men to-day ; the children then under 5 would by our second period be aged 85—40. The recent census shows, it is true, a slight excess of females at ages 35—45, but males are a trifle in lefect. But all this would prove nothing unless we wssume, as is perhaps reasonable, that the extra survivors are more delicate than the average, for the death-rates per thousand living hove increased at the later ages. Any way this loss is small compared with what we have yained at other ages, so that the balance in our favour remains large. STUDIES IN STATISTICS CHAP. Table VII? shows for childhood, working life, and old age the relative incidence of the various causes of leath upon males and females. This table, read together with Table VI, shows which sex gains most by each change in the death-rate. Thus, in childhood and youth, young girls suffer especially (as compared with boys) from phthisis, whooping-cough, diphtheria, fever, and heart disease, and all these have diminished in fatality. On the other hand, boys suffer especially (as compared with girls) from violence, kidney diseases, brain diseases, “all other causes,” lung diseases, and to a less degree from liarrheeal diseases, small-pox, measles, and scarlatina. Of these, lung and kidney diseases are increasing in fatality, the rest are declining, the net results being that virls gain more than boys in the proportion of 7 to 6. During working life, from cancer alone do women lose more than men, but here the loss is more than double, and cancer is increasing rapidly, though the rate of v Forty-fourth Annual Report of Registrar-General, Tables 24, 25. 2 Calenlated from tables IVA and IVn: X1vV THE DECLINE IN THE DEATH-RATE 249 increase is more rapid in the case of males than in the case of females; men, on the other hand, suffer more especially from violence (5 to 1), kidney diseases, lung diseases, phthisis, brain diseases, and fever. Of these, diseases of the kidneys, lungs, and brain are all increasing, the others are diminishing, but the net result shows that females gain more than males in the proportion of 7 to 3. In old age women again suffer from cancer more than men, and cancer is increasing in fatality, but on the other hand old men suffer excessively from kidney diseases (3 to 1), violence, phthisis, small-pox, and fever; and of these, the first two are increasing. In the net result the mortality of both old men and old women is increasing, but the death-rate of old men three times as fast as that of old women. It is quite plain that the recent fall in the death-rate favours the accumulation of surplus women, if one may be allowed such an expression, and should the change of mortality go on in the same direction the rate of accumulation will be increased. Several other factors are working to the same end. (a.) Forty years ago for every 100 female children, 105°3 male children were born, but now the number is only 103:9.' (b.) Emigration always removes an excess of males; thus, during the six years 1877-82, no less than 349,666 adult Englishmen left this country, but they were accompanied by only 179,869 adult Englishwomen, so that the men were 169,797 in excess, or nearly 2 to 1.2 TEinigration is, however, to a great extent balanced by immigration. In the ten years between the two last censuses the net result of migration amounted to a loss to this country of 123,467 males and 10.840 females, the excess of males being 82.627.° LU Forty-fourth Annual Report of Registrar-General, Table 13. 2 Board of Trade Return of Emigration and Immigration for 1882. 3 Report on Census, 1881, vol. iv, p. 15. 50 (c.) Independently of recent changes women live longer than men. The combined result of the operation of all these factors is less than might have been expected. In 1851 here were in this country 104-2 females to every 100 males. This proportion has risen gradually till it now amounts to 1055 to 100.2 As this excess of women is a common topic of conversation, I have drawn up, in the interest of the ladies, the following table, based on a longer table in the Census Report (vol. iil. p. v)— TABLE B.—-SHOWING EXCESS OF MALES OR FEMALES AT DIFFERENT GROUPS OF AGES, TOGETHER WITH THEIR CONJUGAL CONDITIONS. STUDIES IN STATISTICS CHAP. Age and Condition. Males. Numbers. ' Excess. 1,728,466! —-Females. —— Numbers. Excess. 47401251 11,6509 Children under 15... Unmarried 15—35 ... 35—45 ... over 45 2,704,062 195,427 200 255 | 72,629 0 9,631,433 233,820 292.151 38,393 01,306 Married 15—35 over 35 1,469,040 2.907.858 248,133] 1,778,237 2 659.725 309,197 Widowed, all ages ... Total ... a 134 TO 099 .(M A 12.629,902 320.762 694,635 13.334. 364,252 r.n15,397 Balance. excess of females The discrepancy between married men and women, some 60,000 more married women than men being returned, admits of ready explanation. Many men at sea, in the army in different parts of the world, in the zolonies, or travelling, had left wives at home. Some few men who were really married no doubt found it convenient to return themselves as unmarried; probably a \ English Life Tulle, No. 3, pp. 36—39. Y Revort on Census. 1881. vol. iv. p. 15. XIV THE DECLINE IN THE DEATH-RATE 251 considerably larger number of women did the reverse. Taking no account of these, the table shows that the total excess of females, amounting to close upon 700,000, comprises 564,252 widows, a simply appalling number whether viewed from a sentimental or an economic standpoint. But this is only the excess of widows over widowers. The total number of widows is very little short of a million! (999,046). The “really old maids” number only 292,151, or 91,896 more than the old bachelors; there remain 233,820 who may be termed rather old maids.” The 11,659 female children in excess correspond to some extent to the 72,629 young men in excess, since Women marry sarlier than men. (3.) There remains for consideration the effect of the present or future decline of the death-rate upon the increase of the population. Dut little need be said on this head. Tt is obvious that if fewer die, more will survive, and pro tanto the population will be increased: But there is more in it than this. Clearly, everything depends upon the ages at death. It is possible to conceive two populations with equal general death-rates, but increasing ab very different rates. If the mortality fall chiefly on the earlier ages, while those who reach maturity live on to a ripe old age, a slowly increasing population will result. On the other hand, if there be a high birth-rate, accompanied with high death-rates in middle and late life, the population will increase much more rapidly. The birthrate must mainly depend upon the proportion of the population living between the ages of 20 and 40. Now what all the figures show is this :—death-rates in early life falling, in late life rising; so that the population is increased in two ways, directly, by the diminution of deaths, indirectly, by the proportionate increase of births; or perhaps it would be more correct to say “ lreeding STUDIES IN STATISTICS CHAP. capacity,” since the birth-rate is largely influenced by the marriage-rate, which in its turn is affected by commerce. [t should, however, be borne in mind that emigration, just now very active, by removing chiefly those of the breeding age, and to a less degree by increasing the disproportion of the sexes, tends to counteract this rise in the birth-rate, so that emigration checks population in two wavs. The chief inferences which I think may be fairly drawn from the facts under review, are— (1.) The decline of the death-rate in recent years is shiefly to be attributed to a diminished number of deaths assigned to fever and phthisis, and, to a somewhat less degree, to scarlet fever, diarrhceal diseases, small-pox, liphtheria, and measles. (2.) There is also a great diminution under the heading “all other causes,” but this is largely to be explained by mproved diagnosis and greater precision in filling in certificates of the cause of death. (8.) There has been coincidently with a fall in the yeneral death-rate a considerable increase under certain sauses of death, notably diseases of the lungs. heart, and kidneys, and cancer. (4.) The figures prove that only a portion of the increase noted under the above headings can be explained by improvement in the returns. There has been a real, not merely an apparent, increase. More especially, the increase under lung diseases will not explain the fall under phthisis. (5.) The gain under phthisis is more important than the loss under cancer, kidney diseases, and some others, since the lives saved are more valuable than those lost; also in the case of phthisis parents beget children when far advanced in disease, a fact greatly to be regretted. X1v THE DECLINE IN THE DEATH-RATE 253 (6.) Preventive medicine should be especially directed against those diseases which kill the largest numbers, such as “local diseases” generally, the ill-defined diseases of infancy, and phthisis ; while among the zymotic diseases, diarrheea and scarlet fever, whooping-cough and measles require especial attention at the present time. (7.) Since the falling causes of death can only fall to zero, but the rising causes may rise indefinitely, the present changes may conceivably in course of time lead to a rise in the general death-rate. (8.) The tendency appears to be for useful working life to be increased, but for old age to be slowly shortened. (9.) The increased mortality in the latter years of life, and such phenomena as the increase of defective sight and rapid deterioration of the teeth, probably result from the increasing proportion of the population living under the artificial conditions of town life. (10.) The present increased mortality at the higher ages cannot possibly be explained by the recent diminished mortality at the beginning of life, even on the supposition that the survivors are more delicate. (11.) The changes that are taking place in the incidence of mortality, almost all aid the other causes which contribute to the great superabundance of women in England and Wales. (12.) They also tend to increase population both by diminution of deaths, and increase, positively and relatively, of births. In conclusion, if I have made mistakes in my figures, or if I have been over-ready in connecting cause and effect, [ must ask you to make allowance for the complexity of a problem which is bound up with so many of the infinite relations of civilized life, an equation in which the number of unknown quantities is itself unknown, 134 STUDIES IN STATISTICS CHAP. TasLe 1a.— England and Wales. Manes. Deaths from Various Causes at Eleven Groups of Ages. Averages of the Five Years 1876-80, compared with Averages of the Ten Years 1861-70. All Ages. 861-70. 1876-50. 1861-70. hsre-so. "61-70. 76-50 ¥61-70.| 76-60. soz | 1,128" 956 306). 189| 110 %] 69 4,784 1 4,950 4,434 4,5800 =2g0| 209 30 32 10,537 8720} 6,860 | 5,765 2,799] 2,815 534 382 1,847 uy) rog4 | 785] 427| 402} 122 yg 9,330 4,764 1,779 853 r,x12! 582 709 427 507: 1 6,027 4,906 5,883 | 55 181 6 $ ¥1.937 I 11,246 9,220 9,513 202! KC Q-20— 61-70./76.50.167-70. 76-80. 93] 83% 164 | 185 10 7 6 7 152 112: 72 | 51 59 33) 32 18 80s) 509 746] 504 3 2 T - sal 1a 2, og 3mall-pox ... Measles ... a Scarlet fever po Diphtheria ... oF Fever ee on Whooping-cough ... Diarrheeal diseases Phthisis ... ie _-— Lung diseases ... yish Brain ,, wo wen Heart diseases and drops; | Kidney diseases ... wh Sancer aw won y Violence ... ws we A1] other causes ... ...] 71,553 | 74,928 All causes ... wi ...} 245,049 1 269.871 26,061 51,283 86,867 18,340 6,766 4,118 12,244 13,257 5,431 1,3041 544] 520° Sqoi B35 4235] 1,014 4,319] 3,008 17,366 28,085 608' 85] 227 63 g20| 8951 474 5aT 16,466 16,947 754! 086 382. 462 559] 433[ 4or| 408 576 477 I 295| 248} 287] 804! 319) 383) 3r2| 383 68 334" zo8| 191] 7a wy r02| 147) 124] 181 19 ort 1 12 9 12 20 25 25 33 ara| 22510 asl veel ssid raed mes! sro] ger] ess 38,1361 | 40,822 ’ 1,042! 2,049] 1,004 | 1,080] 8:8 a1] 780( 811 NN 118.741 bro,2831 9,555] 5,086} 4,655] 6,292 5,556 | 7,659 7,152 : 1 ’ oo 75— 61-70. *76-80. “61-70. 76-50 F61-70.}76-80.F61-70. 76-80. ¥61070, 76-0. 116 15 62 29! to of 4 4 5 5° : . a 27 29 4 24 ° 21 20 1 l a40 | 488 817' 382 698 355 I a | — — 243 111 0 oc | 159 qin) 289 LL —-— Grego. | "76-80. 203 r 55 29 533 1 : " | dmell-pox ... oe Measles ... ee “ Scarlet fever re - Diphtheria ... J ” fever — ee v Whooping-cough ... tn Diarrhoeal diseases ww Phthisis ... .. ..! S109 6,164 Lung diseases... ..0 r,283 1,580 Brain ,, wre 4 ros0 L171 Jeart diseases and dropsyl ggz 1.176 Kidney diseases ... ..| 353 532 | Jancer ... .. .. 91 127 Tiolence ... .. | 5617 | 1,788 All other causes ... “1 ox000 | 1,793 All causes .., - 14,772 15.982 3 4 190 61 aie 1 563! 517 4940 ' 5,509 3,400] 8,906 1,087 2,107" 84' 88 2,042 2,607 5164] 4,001 4,572| 5,870 3.3261 4,652 1,591 1,940 2,029] 2,476] 2,811] 3,808 2,529 | 8,465 1,463 | 2,170 b2.000 2,703] 2,760] 8,981 630] 2,811 481 | 787 | s95| 933] 970 L200] 650] 911 244 352 4871 758] 727] 1,213 01 472 nsst | 1,698 140s] 1,608] r.ras] 1.382! qu7| 494 2,302 | 2,436 oe a 200] san) oan 6,201! 11,337 {11,474 15,970 | 18.254 17,327 (20,117 [10,882 24,487 21,082 20,611 o,ga6 24,381 Xv THE DECLINE IN THE DEATH-RATE 255 Tapre I8.—ZEngland and Wales. FrmaLrs. Deaths from Various Causes at Kleven Groups of Ages. Averages of the Five Years 1876-80, compared with Averages of the Ten Years 1861-70. All Ages. 1861-70.) 1876-80 980 4,709 1 8,321 | 1,698 | 4,812 + 7,265 to.152 | 25,030 } 44,768 32,123 19,907 | 3,751 8.875 bo 3525 | 8,567 —_— 47501 1,667! ft} 71,154 b 31,757 | 38,242 | 1,661 | 1,760! 233,501 | 251,242 9x,361 | 96,574 | 9,800 | 8,902 * 10— 20— 1-70. 76-80 5170.176-80.F 61-70. 5-50. 63 TI 83] 84f orl 110 36 81 2 11 10 9 bor| 874 162). 106] 107 | 3 86] 141 63. 47 42° 27 Bo7| 0361 043 | 5608 730! 498 12 12 3 2 2 no on 62! 28 40 3mall-pox ... Measles... vor Scarlet fever er Diphtheria ... ow Fever ws oe Whooping-congh ... Diarrhoeal diseases Phthisis ... wo wf Lung diseases ... wr Brain ” Pe . Heart diseases and drops) Kidney discases ... ws lancer ee ve ve. Childbirth and childbed fever ne i Violence ... .. All other causes ... \ All causes . 1,175 | 1,156! 247 | 272 3581 444 322 386 49 92 ] 12 3,220 | 2,802 3171 831 404 | 484 3531 419. 66! 121 18"! 23 4,010 388 411 394 103 33 3,445 410 429 174 179 93 167" 164} 152] 151 gor] 996] e20! sm 858 | $37 5,042 | 4,701 1 6,855 | 6,167 | 8,046 | 7,202 640| 660 30 Iam AR "61-70. | "76-80. 116 125 15 15 120 | 98 47 33 nob | 621 LI 2 281 107 "61-70. "76-80. 51-70.176-50 38 £ 7 1 pa 85— 65— 75 63-70.1"76-80.1'61-70. 76-80 Fo 1-70. 176-50. 19 3! 3 ” , Small-pox ... o a Measles ... ps sos Scarlet fever ee . Diphtheria ... ree be Fever yn we or Whooping-cough ... we Diarrheeal diseases and 64 84 7 39 26 409 1 199 6 i 42 2¢ 303 I 2QI 3! iz 21 7561 12 239 1 DID 626 440! f 154 108 2 rnd mee | 1 146°? ai 6rz! A476! mox | £505 Phthisis ... wen wt Lung diseases ... - Brain 25 so _— Heart diseases and dropsy Kidney diseases ... i Cancer ne oe wad “hildbirth and ehildbed fever os wi Violence ... wie yo d All other causes ... ..| 1.897 | 1,930 All causes ... vos ..] 16,033 | 15,385 7,248 6,409 ' 4,087 1,824 1,014 | 1,170 1,445 1,867 911 978 1,189 1,470 1044 | 1,270 | L323 2,005 260 441 | 311 544 270 333 860 ' 1,226 1,538 | b 224 275 24% 2,746 2,640 2,242 2,718 1,750 2,821. 2,006 2,702 goq4l 863 1.482 | 2.062 1,348 | 1,353 3,836 | 4,966 ° 2,621 | 8,688 3,110 | 4,202 332 | 666) 1.eoz | 2.270 + 4731 417 5,008 6,762 3,558 4,821 3,602 | 4,905 2761 542 1,072 | 1.873 74 3779 2,857 1,980 120° AGE 85 5,678 3,796 LEGO 246 04 1,156 8 "1 826 © yn ae bal ago) —_ 885 wi) 480 2,357 | 2,479 * 2,923 | 8,129 4,046 | 4,293 7,121 | 7,023 “5,380 | 16,685 hr4,086(17,058)18,136(22,82 1k22, 457/27, 28° 840[14,559 4, 06120 L071 5,3 25, 256 STUDIES IN STATISTICS CHAP. TapLe Ila.—England and Wales. MALES. Death-Rates per 100,000 Living at each Age, from Various Causes, at Eleven Groups of Ages. Averages of the Five Years 1876-80. compared with Awveraaes of the Ten Years 1861-70. All Ages D— Bu 1861-70.) 1876-80 1861-70. 1876-80 61-70. ]776-50 * Small-pox ... . Measles .. as Scarlet fever ow - Diphtheria ... on go Fever ws ws ve Whooping-cough ... . Diarrheeal diseases oo 18 46 01 8 » 49 118 9 40 7! 12 39 49 g99 | 66 307 475 76 123 33% £28 18 270 339 46 50 848 AHO | 15 23 222 34 88 [2 +6 8 20 153 27 28 2 n Phthisis ... oe “. Lung diseases Fr Yar Brain - we wt Heart diseases and dropsy’ Kidney diseases ... “ Cancer ie ve a Violence ... .e ”, All other causes ... A All causes ... es .. 247 369 307 131 42 24 8 213 419 301 150 55 34 108 148 99 1,201 14,139 40 12 I wr 1.406 996 28 20 1 1283 43 4 36 62 16 18 55 6o 686 2,361 612 2,204 2,672 2,295 7,316 6,681 vr 5 f15 ' 631 2H "61-70. | "76-80. boss 70. | 7680. 12 | 8 a 15 oon wor r80 -80 . 3mall-pox ... ‘ Measles... we Scarlet fever wee Diphtheria ... - Fever wr oo Whooping-cough ... Diarrheeal diseases Phthisis .. .. .. 417 102 | 386! Lung diseases .. ..I 172 197 1 3501 Brain ”» — at 134 142 224 Heart diseases and dropsy] 123 158 - 219 Kidney diseases ... wa 41 | 54 66 Cancer - ve pod 21 26 54 Violence .. .. A 13% | 126 All other causes ... | 115 | 102 194 | 176 All causes ,.. er aqo rT 1.246 1,381 30 ig ass} 308 F 244 | 266 02 75 158 328 | 306 1,917 | 1,981 20— 61-0. 176-80 Fér-0."76-80.F 61-70. "76: 0. 8 | 13 1 1 2 3 47 1 An 3 S 8 4 2 17 Q + {9 4 32 x y 9 2 61 20 7 9 219 22 57 32 35 1 2 288 52 44 5 4 3 288 52 38 as 17 3 102 88 CS] 845 | 067 nA 34 ne 17 1T oo ~ Q £46 348 €t6 £R9 LT. 75 61-70. 76-80. 61-70.]'76-80. 2 3 61-70. ]"76-80 3 9 r 43 1141 47 11) | 160 3 149 443! 1 43 1 S01 £0 330 318 1 202 | 192 66 62 750 837 I1,519 | 1,686 2,617 3.252 467 542 1,095! 1,265" 1,990 2,422 458" G60] 923]1,116 1,283 616 28! 180 | 290 875 | sit 637 121 | 1731 183] 2851 230 330 18¢ | wr 202 221 | 249 206 655 | al 1,926 | 1,628 8,019 | £,020 3,300 | 8.481 6.660 | 6.928 1:6,464]17.044 XIV THE DECLINE IN THE DEATHRATE 257 TasLE IIn—England and Wales. Fimares. Death-Rates per 100,000 Living at each Age, from Various Causes, at Eleven Groups of Ages. Averages of the Five Years 1876-80, compared with Averanes of the Ten Years 1861-70. All Aces. w0— | - 0 20— "61-70. |'76-80.¥61-70. '76-50.¥61 -70. [76-80 F 612-70. |76-50. 71 10 | 1 7 861-70. [1576-80 861-70. 1876-80. Small-pox ... Measles ... wa Scarlet fever . Diphtheria ... we Fever ee vis Whooping-cough ... Diarrhiceal diseases 15 42 93 ‘9 88 57 or 7 86 64 13 37 36 a 65 17 293 | 251 150 319 78 48 | r27 49 | 414 | 410 gz | ase | 7 25 | 28 215 | 143 4 | 38 o7 | 4 18 ‘8 +6 ” 5 2 28 | 10 Nn I IT L 3 2 36 3 7 3 5 34 i 6 i 72 Io 5 5 Phthisis ... wh 4 Lung diseases we wi Brain, foe . Heart disenses and dropsy Kidney diseases ... aw Cancer... . Childbirth and chitdbed | fever ... aw in Violence ... vg - All other causes ... ....} 630 | 631 All causes .., wo | 2,128 | 1,946 194 347 249 154 29 Ba 95 1,007 391 35 3 73 4,157 761 24 id 2 48 55 54 27 19 56 58 7 108 0B “11 227 27 36 34 | 10 9 397 38 4 39 0 3 293 35 87 40 15 3 22 20 32 2Q 33 °9 3x 39 4 6 2 53 56 x6 106 2 12 130 15 77 ' 73} 80 | 70 § 8s 78 76 1% N4q8 | 548 | 662 | 500 | 621 2,204 | 1,048 6,343 | 5,661 ra. 10 75— '61-70. "76-0. "61-70. | "76-80. *61-70. | "76-80. '61-70.]"76-80. S1-70.|"76-80 61-70. "76-80. Phthisis ... ses . Lung diseases .., wx Brain, ce a Henrt diseases and dropsy Kidney disenses ... a Cancer vie “ee _— Childbirth and childbed fever wm we Violence ... ws ir All other causes ... il All causes ... . : Small-pox ... Measles Scarlet fever Diphtheria .., wi Fever ee war Whooping-cough ... Diarrliceal diseases 7 4 3 2 32 3 5 8 1 3 c © 2 £ 3 2 7 27 64 5 2 Ie a 30) 34 120 426 4 2G ny 0 u 1 9 6s | 84 * 161 106 324 138 AL 334 51 31 36 23 17 390 | 113 93 119 24 67 327 | 283 126 233 L00 182 136 218 37 | 32 88 184° 287 207 244 586 208 402 248 176 51 51 185 | 230 } 5! 178 124 ' 106 634 31x 1,505 164 g3r 1,078 536. 943,001 85 | 72 | 121 2000 28: ! 872°" 45 2,296 1,736 1,203 73 283 46 3,045 2,035 1,638 132 97g 5 i3 6 6 a2 80 8a 8 f 14 227 285 101 184 | 166 | 303] 280) 620 | 548 1,866 |1,568 | 01s | 7,823 £00 | 1,203 | 1,123 rss6 [1081 2,777 | 2,849 |5,880 | 6,060 15,428 15,067 0 1 135 06g 58 STUDIES IN STATISTICS CHAP. TaBiE IIA — England and Wales. Mares. Deaths from Various Causes during Childhood and Youth, Working-Life, and Old Age. Averaues of the Five Years 1876-80. compared with Averages of the Ten Years 1861-70. Apres . Small-pox ... wo Measles ... ohh Scarlet fever ee Diphtheria . Fever we a Whooping-cough ... Diarrheeal diseases Phthisis ... sa Lung diseases ... Brain 55 - Heart diseases, &e. Kidney diseases ... Cancer ve ee Violence ... ce All other causes ... All causes ... 0—20, 1861-70] 1876-80 1,302 4,764 0,345 1,702 4,404 5,068 0.268 4,900 18,613 17,981 1,477 452 58 4,631 42.16% 4,378 25,444 18,777 1,412 781 76 4,629 44.902 133 83" 127,420 20—865. 1861-70.) 1876-80 575 536 20 23 189 | 145 130 88 4,252 2,168 3 2 1.248 R40 20,045 I, 335 7,882 7,504 2,325 1,574 6,632 11,697 20,869 14,763 9,808 (0,363 3,643 2,483 7,361 12,846 75.611 KF 949 65—. 1861-70.| 1876-80. 16 | 13 3 15 674 1.121 1 9 225 1 953 751 819 8,331 '1,076 6,138 8,287 4,672 6,565 1,605 2,342 910 1,559 981 1,267 17,691 17,675 12.008 | 50.799 TABLE ITIB.— England and Wales. Femavrrs. Deaths Jrom Various Causes during Childhood and Youth, Working-Life, and Old Age. Averages of the Five Years 1876-80. commnared with Averages of the Ten Years 1861-70 Acrng Small-pox ... “s Measles... - Scarlet fever po Diphtheria . Fever we ee Whooping-cough ... Diarrheeal diseases Phthisis ... - Lung diseases ... Brain 5 vee Heart diseases, &c. Kidney diseases ... Cancer ee sos Childbirth, &e. ... wo Violence ... wr . All other causes ... . All causes © 0—20. 861.70.) 1R78-RO 1,254 4,589 9,960 1,936 4,885 6,206 Q 28% 545 1,672 8,100 1,508 2,605 7,258 Q ard 6,361 5,797 15,769 21,197 14,277 14,675 1,443 1,480 288 570 54 74 167 167 2,368 2,453 35,114] 36.869 r12.0:c8 | 116 344 20—65. 1861-70. 1876-80 323 36 Te 28 Lae 376 36 220 110 1,999 7 a0Y 20,339 8,925 6,882 8,167 1,310 1 147 3,358 1.124 18,671 11,131 8,831 10,653 2,393 5.094 3,400 1,473 12.673 yon 12,081 mo qo | 65—. 1861-70.) 1876.80. Oo 9 3 11 638 1.216 1 10 208 1.081 547 8,787 6,415 5,582 396 1,520 562 12,440 8,617 7,774 788 2.377 648 824 21.661 21,612 47.822 | 56.8304 XIV THE DECLINE IN THE DEATH-RATE 259 TABLE IVA. —England and Wales. MALES. Death-rates per 100,000 Living at each Age, from Various Causes during Childhood and Youth, Working-Life, and Old Age. Averages of the we Years 1876-80, compared with Averages of the Ten Years 1861-70. Small-pox ... i Measles ... i. Scarlet fever ee Diphtheria es Fever i. won Whooping-cough ... Diarrheeal diseases Phthisis ... ee lung diseases ... Brain ,, ve. Heart diseases, &e. Kidney diseases ... Caneer oe a Violence ... ve All other causes ... All causes ... -_ 0—20 20—65. 1861-70.] 1876-80.) 1861-70.| 1876-80, 27 98 213 35 90 104 197 0 &5 148 23 41 104 1868 II 4 3 84 25 9 3 2 87 11 101 182 169 30 9 76 439 324 24 14 1 394 227 155 147 46 31 352 249 166 75 62 42 124 80 774 2.617 ' 2.300 130 228 207 1,488 | 1,439 85—. 1861-70.] 1876-80. 4 3 2 43 182 L 3 148 245 | 156 2,113 1,581 1,252 447 297 242 3,874 | 8,370 9,305 | 9,688 TaBLE 1IVB.—England and Wales. FEMALES. Death-rates per 100,000 Living at each Age, from Various Causes during Childhood and Youth, Working-Life,and Old Age. Averages of the Five Years 1876-80, compared with Averages of the Ten Years 1861-70. \ofBs au Small-pox ... sa Measles ... cen Scarlet fever A. Diphtheria “ Fever ve. ce Whooping-cough... Diarrhoeal diseases Phthisis ... wy Lung diseases ... Brain ,, —_ Heart diseases, &e. Kidney diseases ... Cancer e- _ Childbirth, &e. ... . Violence ... ne . All other causes ... . All causes ... .. 0—20 1861-70.1 1876-80 26 94 205 40 100 i28 173 9 80 139 26 45 125 144 (31 324 204 0 100 365 253 26 10 L 3 19 42 634 2122¢ | 2.002 20—65. 65—, 1861-70.] 1876-80. 1861-70.1 1876-80. 5 I 5 3 73 25 3 i 3 2 31 11 2 2 13 241 2 33 170 289 100 172 1,608 187 1,174 165 1,021 37 73 92 | 281 53 | 23 b 119 193 1 1,306 | 1,215 » 8 vzc | 8.854 260 STUDIES IN STATISTICS CHAP. TABLE Va.—England and Wales. MALES. Percentage Fall or Rise in the Death-Rates from Various Causes at Eleven Groups of Ages. Awerages of the Five Years 1876-80. compared with Averaaes of the Ten Years 1881-70. All Ages 0—15— lo bgt las tags. | A. x “Ao dF lA 3 maz] =12) 14 — 8-33 cL cf eee | ee ee a 15 29|- 31) ~40-38|~40{-34/— of ... |... |... | ... oy — 34h 39] - 22-28) — 53) — 51{ — 50| - 25| - 36| — 37| - 39 - 75 J Lda U2 he ee ee ee | Ld—2ol—19 asl — 77 _ 74] -73l 711 —60l — 53i~ 411 39] 18 Small-pox ud Measles ... Scarlet fever ... Diphtheria ve. Fever ... a Whooping-cough Diarrheeal diseases Phthisis lb Lung diseases ... ow Brain ,, I, Bo Heart diseasesand drov: Kidney diseases Cancer ... wo Violence... i” All other causes 4-23) 20-23 — 28|— 26{~ 15|- 4| ... |— B|~ 5-7 71+ 2) - 3|+ 3|— 1+ 5|+14|+14[+10/+11|+24 3+ 4|+ 1|- 5|-14|- 6|+ 5|+ 9|+16|+16/+22 01—80|=11| ... [+ 8| ... {+28|+22[+22(+21|+26 ol + 48] + 25| + 20| + 28| + 27| + 38| + 40| + 40| + 30| + 25 Sha | en [4-181 +251 +881 + 431 + 52) + 44 orl—17l-1al age ols oo +19 ol .16/-10 4 ol— gl_10 oa Tod Toa mal — 121 1+ 8l+ 5+ 4l. All eauses . .. TABLE Vs.—ZFngland and Wales. Femarus. Percentage Fall or Rise in the Death-Eates from Various Causes at Eleven Groups of Ages. Averages sf the Five Years 1876-80. comnared with Averages of the Ten Years 1861-70. All : an 0— | 5— lots lho—toslzs—_lss—lss_las Irs |= sof ~ 7451) — 5|—15{= 1|= 7[w14l+ 6I= Fe = ta] - 14) = 13|- 28-31) —80| = 11)... [pen | on] | -31f-29|- 38|— 48/45] 41] —s0|-21| | || oo] = 341 39] - 26] 37| — 89] — 44| ~ 39| ~ 18] Bo| - 32 14 ho |= 370-61] - 55| - 50] — 491 — Rol — 40] _51| - 65 — 691 — 70| — 76 f= I= 1-12) LL vin | -. o. eae | “rs | _ Zo Cal Z ssl Jal Jol Zee] Jarl Jad ZEe Daal Zeal gg Small-pox on Measles ... i Scarlet fever ... Diphtheria ... Fever ... po Whooping-cough Diarrheeal diseases Phthisis -~ _ Lung diseases ... wo Brain ,, wy - Heart diseasesand dropsy! Kidney diseases .. Cancer ... RN ol Childbirth and child- | i aid. | 11-1102 —- | wee | eee Violence i. 17 ols ol +1641 +211 +19[+ 8 All other causes c= I -12)- 11 - : i 4: 8|—-12(-16{~-13 All canses..., ...|- a Coded 17 7" at salt 8+ 1 -99|-19|-18/-27| -26|-24|-16|-17|—-16|—14]+ 2 +15(+ 1j- 9[-12(- 9| ... |+12{+ 5{+ 8/+15/+33 -15|~ 1|+ 3|- 8 -10/— 8(+ 7|+15|+16|+15/+17 —~31|-17| ... | ... |+ 8|+ B|+14|+12{+13|+16|+28 S701 1.601 4+ 50! + 56) +49] + 46( + 51) + 60 + 68| + 67{ + 81 ve ee eee I 793] 4 920] +2681 8214-84 XIV THE DECLINE IN THE DEATH-RATE 201 TasLE VI— England and Wales. Percentage Fall or Rise of the Death Rates from Various Causes, at Three Groups of Ages, representing Childhood and Youth, Working-Lafe, and Old Age respectively. Averages of the Five Years 1876-80, compared with Averages of the Ten Years 1861-70. Small-pox ... a Measles ... ee Scarlet fever ve Diphtheria rou Fever aan 4 Whooping-cough ... Diarrheeal diseases Chthisis ... vay - Lung diseases ... — Brain ,, ws a Heart diseases and dropsy Kidney diseases ... shi Cancer aw oo wi Childbirth and childbed ; fever ... - Violence ... ve All other causes ... All causes ... ok 0—20. f3 13 -30 -35 -55 1s - 25 +15 -12 -19 +45 — 1 4 *Xales. 20—65.! 65—. 0—20. -20 ~32 ~ 42 -56 ss — -65 1B i" -32 wel —- 85 71 1-55 . 2 “96 Fb _1% -11 +10 +7 +19 +35 36 -5 rl +18 +22 +27 449 —24 +13 -14 -13 ~ 67 - 9 +13 -18 -192 2 Females. 20—65. | 65—. J40 _33 _58 272 “29 hr -21 +7 +10 +12 + 54 199 —12 +22 +15 +20 +70 L888 12 +15 -10 + 9 -15 1 TasLe VIL—England and Wales. Death-Rates of Females, compared to Death-rates of Males taken as 100, from Various Causes, during Childhood and Youth, Working-Life, and Old Age. Averages of Five Years 1876-80. Small-pox ... . Measles ree . Scarlet fever... we Diphtheria ... . Fever... wow wa Whooping-cough ... Diarrhceal diseases ... Phthisis oo _ _ Lung diseases a vie Brain ,, ... _ _ Heart diseases and dropsy ... Kidney diseases ... -_ Cancer oy o_—. 1h Childbirth and childbed fever Violence .. oo 2 All other causes... . All causes ven a 0—20 | 20—865. 90 94 94 [18 ‘10 20 Sr 87 100 100 84 100 132 83 78 108 71 100 32 50 5. 61 219 3 8H — 37 100 77 93 56 28 88 98 28 26 U1 262 STUDIES IN STATISTICS CHAP. TasLe VIIIA. England and Wales. Malus. Diminution or Increase of Deaths from Various Causes at Eleven Groups of Ages. Averages of Five Years 1876-80, compared with Averages of Ten Years 1861-70. Showing Annual Gain (—) or Loss (4) of Lives under each heading. Allowance is made for Increase of Population at each Group of Ages. 20— - 820l- 108 - 7 - 26-59 - 632 ~ 49 — 41 - 51- 1 - 2,312 —- 1,044 |- 2531 — 70 |— 34 - 504,—- 110|- 41 - 38 |- 20 - 1,242 |- 7520-417 - 453'- 379 + 54 [~ 5 win eee -1.8451- 183'- 77 - BS 71 Phthisis ... .. ..]- 3,966 — 381|- 133|— 186|- 761|- 1,071 Lung diseases ... «|+ 6,490 + 8,486 + 13 | - 7+ 16/~ 4 Brain ,, i. c= 465 ~ 2,443 + 30 | + 71- 210 —~ 67 Heart diseases and dropsy |+ 2,565,— 201 |- 106]- 37 wes + 14 Kidney diseases ... vee F nyo 135 |+ 61+ 21 (+ 7+ 34 Cancer... ces wd E1871 1 4 i = 1 + 1t+ 3 Violence ... ce . 1,063 - 211 148" 282! 178. 145 All other causes ... ..!- 8,425 - tal, 280! 116 74 112 All causes ... — ~ 17,185] -10,808!- 2,784 '— 1,398 — 1,645!- 1,912 25— 25 Small-pox ... ee oo 381-20 Measles ... ce _— ~ 1 Scarlet fever cen = 29 — 2 Diphtheria - Es 21 - 7 Fever wn pr, w= 6091-649 Whooping-cough ... w+ 1 sue Diarrhoeal diseases w= 174-70 Phthisis ... or = 1,058 -— Lung diseases ... wl 73 : + Brain ,, en J 69 + Heart diseases and dropsy | + 4 + Kidney diseases ... J: 118 + Cancer vr avy ceed 19 Violence ... wo - 172 All other causes ... = 216 All causes ... Ld= 2171 48 — E5~ - 65-og - 9 12 + 1. - 9 - 3 - 8 I 8 - 5801-558 14 | 108 - or - q- 1 —— 5 a. 4 - wn 153 e+ 1 208 117 F + + + 14 | ~ 487 + 197 | + 483 1° 265 211 118'|— 552 | + 533 | + 716 '+ 362 + 2/4 40 | ~ 7 635+ 909 648 [+ 619 787+ 476 326'+ 179 371 + 143 67 1,282 a 21¢ 1.149 L 6531+ 1.279] + ogv Kiv THE DECLINE IN THE DEATH-RATE 263 Tasre VIIIs.—England and Wales. FEMALES. Diminution or Increase of Deaths from Various Causes at. Eleven Groups of Ages. Averages of Five Years 1876-80, compared with Averages of Ten Years 1861-70. Showing Annual Gain (—) or Loss (+) of Lives under each heading. Allowance is made for Increase of Population at each Group of Ayes. Al Ages. | 0— 5— 10— I s— | 20— Small-pox ... wen Measles ... - Scarlet fever - Diphtheria wn Fever re p_ Whooping-cough ... Diarrheeal diseases |” 045 = ...1— 3,879 ..|— 870 ...i— 6,510 sil — 93 l= 2,919 - 817 - 109 ~ ql - - 43)- 12)- . ~ 9,247 — 1,084|- 348|- 8 - s13!- 17] 82|- 30 - 1,326 — 808]— b4li- 554 - 8lj- 33|- 2] 2 - 1.817'- 141l- 56" - 52 2 - 2 - 51 - 22 - 420 + 1 - 79 Phthisis ... - i“ 6,806 Lung diseases wee + 5,404 Brain ” — 38 313 Heart diseases and dropsy + 2,121 Kidney diseases ... .. + 1,417: Cancer we soe c+ 1,661 Childbirth and child- | i ih bed-fever a, | 523 Violence . ... —. w 124 All other causes ... ..— 9,837 All causes ... ..l=22,993 - 864|—~ 1391-255 - 1,034 + 2,857 |+ 1210-25-47 -2,227{—- 10 + 14.- 87 - 186|- B54 - _— 1 + 106 [+ 44 + 32 + 42 In 4 = 2 - 1,216 - 41 - 49 + 16 + 58 5 35 34 171|~ 130/}- 34 3 _ 4,8720- 241|- 42 1151-159 ~11.645]— 2.912|- 1,352 | = 1,999 |= 2,060 25— 5— 415— 5H — BH— 75— Small-pox ... we Measles ... was Scarlet fever cs Diphtheria “ Fever —. -_ Whooping-cough ... Diarrhoeal diseases ary 2 . al - 1 . . . 41|- 10§- 7 . i+ 1 21|~ 6|- 12|- 5|- 1,- 1 609|— 634|- lf 519-364-170 1+ 104 1+ 1. 9181-915 1801-217 246 ls 7 - - 801+ 1 Phtl eee awl 1,083) 937) 534 - 265 5 Lung diseases we a 4+ 197+ 126 + 3623+ ip 4 139 Brain ,, TUT sal+ 95|+ 208|+ 491|+ 635+ 2 Heart diseases and dropsy + 61+ 246\+ 279+ 4711+ gos + 110 Kidney diseases ... LE 1410 18514 2124 269 |+ 177 Cancer ... ove We 20 2331+ 3501+" 4671+ dlll+ ! Childbirth and childbed)! fever ... bon Violence ... ne we All other causes ... i" All causes ... .. 15 o64l- 243'- 254'- 561l- 1,357]- 2,229 TTT ToT rls mele C328'- 1aoal_ 268ls s7l+ 8090+ gg 264 STUDIES IN STATISTICS CHAP. TasLe IX.—ZEngland and Wales. Diminution or Increase of Deaths from Various Causes at Three Groups of Ages. Averages of Five Years 1876-80, compared with Averages of Ten Years 1861-70. Showing Annual Gain (=) or Loss (+) of Lives during Childhood and Ye outh, Working- Life, and Old Age respectively, after Allowance is made for Increase of Population at eae ae. Males. Females. Aces 0—20. | 20—65. | 65—. - 961|- 138 - 7 - 690|- 1, .. - 8,679|- 77-9 - 693|- 64l- 9 - 2,864 |- 2,775|- 549 I ~1.6301- wgrl— a9r 0—20. 1 20—65. | 65—. —- 946+ 2 - 1 - T70i- 6|— 1 — 8,768{— 113] ... - 802|- 66|- 2 ~ 8,229 |- 2,747 (- 534 - 98+ 50 .. - 1.566 918 [— 435 Small-pox ... . Measles ... er Scarlet fever ve Diphtheria woe Fever ce en Whooping-cough ... Diarrheeal diseases Lh Phthisis ... ver I 1,461 |~ 2,458] ~ 471- 1,792 - 4,935 | 79 Lung diseases ... w+ 3,502 14 1,444 (+ 1,5441+ 2.497 |+ 640+ 2,267 Brain ,, wo oo= 2,427 4+ 695 |+ 1,267) 2,260|+ 753 |+ 1.194 Heart diseases and dropsy ~ 844+ 1,696|+ 1,213|- 242|+ 1,073 + 1,200 Kidney diseases ... at 24214 9551+ 5051+ 225|+ 865|+ 827 Cancer ve pa. al b 614+ 6671+ B14 8'+ 1,06561+ 588 Childbirth and childbed fever ... _— Violence ... .. .- 869] 836l« All other causes ... Lhe gl 1,178 1 - 2,431 ~All causes ... ..0—-16,6351— 2,366'+ 1.81% ls ll 78 = 4,770]- 1.481|~ 3,586 _17.908— 6.1981 + 1106 Posrscripr, June 1890. Since the foregoing was read to the Statistical Society, the death-rate has continued extremely low. In 1881 it fell to 189, the lowest point reached up to that date : after rising to 19-6 in the next year it fell again, till in 1887 it reached 188, and in the following year dropped to 17-8. From 1881 to 1888 (the last return published) the death-rates for both males and females have in every year been lower than wn any preceding year since the commencement of the record in 1838. The death-rate for both sexes has, in the thirteen years ending with 1888, averaged 19-7 per 1000, in the last eight years 19-1 per 1000.) This depression in the death-rate has now 1 Fifty-first Annual Report of the Reqistrar-General, p. xxiv. X1v THE DECLINE IN THE DEATH-RATE (POSTSCRIPT) 26; lasted so many years that it will no longer compare with the period of low death-rate in 1841-45, when the rate was 214. The following table, in which the death-rates from all causes at various groups of ages during the five years 1881-85 are compared with the like figures for the period 1876-80, shows that the mortality has decreased at every age-period in each sex. The fall has been but slight in males aged 65-75, and females aged 45-55. In both sexes the diminution in the mortality has been greatest at the two extremes of life. ENcLAND AND WaLEs.— Death-rates of males and of females from all causes per 1000 living at various groups of ages. The quinquennium 1881-85 compared with the quinouennium 1876-801 Walea 1876-80 wi ‘881-85 ror Difference ... on Fall per cent. . Females. 1876-80... 1881-85 ee ee Difference ... ee Yall per cent. wa AU) om Vs otis sofas sl as] 55 65 75 os — mmr —— — _—————_ siren 634 ro [re] 672 id ies 69°40 Jens jen 90-42350-60 | 5778 [316 14-56] 6-00 8:18 1274 119-42 [35-64 [48-75 ag-6 l2o64 v7} | o'56 | bw o'72 log] 64) oso] ras] 062 | 7'6 35'2 +s wiles |s1leeloviesies [ze ls loo | 50! 08 19-54 foe 5°88 yo I 624 funy Jon pesos Js jure 18-24 150-48 | 5:62 [3:30 [4-72] 5:94 [7°90 (10°90 115-24 |27°82 |59°46 [129°4 12678 30) 6°34 | o'26 - bse] 0°30 Joe] a) o'12| 0°78 | 6x 31°2 67 bial 44 | 57 63] 48 | 15] 30 08 | 27 [21 | 45 104 I | . It will be interesting to go a step further and make a like comparison between the figures for 1886-88, the three years in which the death-rate averaged only 186, with the quinquennium 1881-85 in which it averaged 19-3. A glance at the following table shows that for all ages below 65 the fall has continued, being proportionally 1 Compiled from the Forty-ninth Annual Reportof the Registrar-General, pp. x1-xli. The figures for 1876-80 differ slightly from those given in the tables in the body of the paper. This is in consequence of the Registrar-General's death-rates having been corrected for the true population as ascertained at the census of 1881. The discrepancies are inconsiderable. 266 greatest in both sexes between the ages of 5 and 15. Between 45 and 65 the fall is but very slight, whereas, above the age of 65 there is a notable rise; indeed for the age 65-75 the death-rate is now higher than in the quinquennium 1876-80, the slight fall in the intervening quinquennium having been more than made up by the rise since. I am now convinced that, in laying so much stress on the increased mortality at the later ages, I did not sufficiently allow for the fact that the period 1876-80 comprised two very severe winters which followed seven consecutive mild seasons. There is no doubt that these peculiar meteorological conditions tended to the survival of an unusual number of persons of advanced life, followed by a correspondingly heavy mortality on the advent of cold weather. With this partial exception I see nothing to modify in the conclusions arrived at in 1884. STUDIES IN STATISTICS CHAP, ENGLAND AND WALES. —Death-rates of males and of females from ail causes per 1000 living at various groups of ages. The three years 1886-88 compared with the cuinouennium 1881-85.1 N ———— ——————————— eet Witt] ne mi Au o— | = | so [ss] a0— | 25— 35 as—s5— | 65 | 75— | 85 —_— yy bak: on 316 6:00 | 818 1274 19-42 ou fom | 1446 2064 19:63 |56°93 | 4671 280 | 417 548 | 7740 [1187 [1913 135-28 | 7213 | 14676 | 519-4 Difference on 0°79 vty) _ 36 | 030] _— 078] o't7 vag) orl +8384 204-160 a5 lige 1alaalos los | gs ! 15 12 4 als vals ga 1881-85 we 586-88 ae Females. (881-85 ue pe 21 f50-48] 569 ners oo 790 i053 ro s040 1201 srs 886-88... N7-gafemes eve 3-00 [4°30] 5433 | T07 [10-10 14-90 l27-50 | 63-03 | 138-9 | 256g 1 Difference ve : ayo] gz “ oii) «Ro! ron [+357] 4 "J sgl 108] 105] 73 1ive 486) 70 feb As regards individual causes of death the later reports ! The figures are calculated from those given in the Fifty-first Report of the Requstrar-General, pp. xlii-xliii. X1v THE DECLINE IN THE DEATH-RATE (POSTSCRIPT) 267 of the Registrar-General show what changes have occurred since 1881.1 (n the deaths from the following causes there has been a further fall— Scarlet fever Diarrheea ... Fever ves Dropsy 2 ... 2 wns Phthisis ... - - Convulsions ves ee oe 1 » Hydrocephalus (Tubercular Meningitis) ... considerable fall. Small-pox vee vee I - slight 28 Cholera ... - 0 vis ve sus ’s Developmental diseases ... ve very considerable fall. i 9 On the other hand the fall in the mortality from measles noted in the paper has been reversed, the deathrate has risen so considerably as to have passed the point of 1850-54. As regards the diseases which were stated to have increased in fatality, in the following the increase has been maintained— Cancer ... co - very considerable increase. Diphtheria rar ‘58 considerable » Diseases of the kidneys ... ’ » Diseases of the heart vee slight v But in the case of several causes of death the course of events has changed, so that they have been less fatal of late years. Thus croup after a further rise has exhibited a marked fall. Diseases of the lungs, diseases of the brain, and rheumatism have all fallen slightly, but diseases of the liver, whooping-cough, and tabes mesenterica have all 1 See table A, p. 230 supra. 2 An explanation of the decrease under the heading “ Dropsy ” is given on pp. 228, 229. This explanation applies with still greater force to more recent years, owing to the practice of sending enquiries from the Registrar-General’s department in cases where the cause of death is ill-defined. See Fifty-first Report, p. xvii. 263 STUDIES IN STATISTICS CHAP. fallen considerably, even below the point of the period 1850-54. The continual fall of “fever” (chiefly enteric) and of scarlet fever deserves to be noticed, also the slow increase of diphtheria. The very satisfactory decrease of phthisis is to some extent counteracted by the alarming increase of cancer, the number of deaths now attributed to this fell disease being double that of forty years ago. As the question of the alleged rise in the death-rate at the later periods of life is one of great interest, and seems left in some doubt by the figures as exhibited above, I will recur to the subject, and look at it in a slightly different way. The tables in the Fifty-first Report of the Regustrar- General (pp. xlii-xliii) admit of the comparison of five periods of ten years each. I have accordingly constructed the following table, in which heavy type indicates the decade in which the maximum death-rate at each age was reached, also the mean for the whole period of half a century, and french type the comparison between this mean and the rate for the decade of specially low mortality, viz. 1879-88. A careful examination of this table shows that (with one trifling exception) the improvement has been con. tinuous from the first for both sexes at the ages .5-25, For infants under 5 the maximum was reached in the third period; for ages 25-85 in the second period ; for ages 55-65 in the fourth period; for ages 65-75 in the last period ; at the other ages there were discrepancies between the males and females. Both sexes, however, concur in showing in the last decennium for the ages 55-75 death-rates slightly above the average of the half century. It seems therefore safe to assert that the great fall in the general death-rate of recent years is almost confined to persons below the age of 45, or above the age of 85. xiv THE DECLINE IN THE DEATH-RATE (POSTSCRIPT) 269 ENGLAND AND WALES.—Death-raies of males and of females from all cases per 1000 Living at various groups of ages, I've decennia, the mean of fifty years, and comparison of the last decennium with the mean, showing rise or fall per cent. ~ Periods. All Lges 2305 a3 34 23°17" 20°48 09-Pr pe. ¥ | gg! Sel Bg | y=) Be 1839-48 ... 1840-58... 1559-68... 1869-78... 1879-88 .. 50 vears ... 7 72-09 9-0n rn L265 | 17°53|31°43/66-80 1475/3131 4 20 | 18'38|31°80{65°88|146°R|306°2 15-92 | 18-0032 62]6640|146-3}413-4 1397 | 20°01|34-53|60 201149 5|322°7 12-51 | 19-40!33'31/70-24147 7309-5 942 1303 | Ramm iens 3 Lat Te | 4:£1 Last decade above or below }| —g7}—13's5|- 22's 290 |—26'3 I :7°9 fess] —4'0 +2 +34 = torul—11 mean, per cend. 3 ! £ ew { TREMAT, RIS 1839-48 ... 1839-68 ... 1859-68 oo 1869-78 ... 1879-88 ... 50 vears ... 4st 6l52 96° 48 0° 19°58 '15°6397-04(60°02|185°1|293-3 1663 68:04 3°71 26 (6% ET ‘35 12°68 [15°80 [2774159°29|184°5(286°1 2127 | 63:20 783 460 67° 811 978 12°08 115°5127°65|3895(184°0(286"7 2047 {59°91 668° 894 5.77 714 894 | 1179 [15°71 |28-49{60°50{135°0{292°8 18°30 | 40°83 | 6401 3°22 | 4°62, 56 7°63 | 10°67 [15°16 {28°07|61°12|132'8{280°3 20-63 | 59-70 | 7-52 4-49 655 | 777! 9-40 | 11-96 [15-56 |27-98|59-981134-3l287-8 Last decade | | | nbove or below rss —28'7'—28'31—29°5 I —188 [—10'8|—26|+03|+1°g|—11|—2'6 mean, per cent, 27% STUDIES IN STATISTICS CHAP. CHAPTER XV. THE CAUSATION OF SUMMER DIARRHEA. $1. Some Facts bearing on the Atiology of Summer Diarvheeal THE subject of the causes that produce the annually recurrent epidemics of Diarrhcea, so fatal to the infant population of our large towns, has engaged the attention of many observers, and is by no means a new one to this Society. Many conflicting theories have been ably expounded, and as ably criticised, with the unsatisfactory result that medical officers of health in their endeavours to incite their several sanitary authorities to bestir themselves and cope with this gigantic evil, have, as yet, no mass of consistent professional opinion to appeal to. It is an undoubted fact that sanitary authorities, in many cases, think less of the deaths of ten or twenty children from Diarrheea than they do of a single case of small-pox. My purpose this evening is not to review successively all the conceivable explanations of the fatality of Diarrheea in any place, Leicester for example, but rather to consider broadly certain large groups of facts, more or less wellknown, and in this manner to endeavour to define certain limitations within which observers with special opportunities may construct a complete theory that will explain L Read before the Society of Medical Officers of Health, February 1880. LV CAUSATION OF SUMMER DIARRHEA 271 all the phenomena of the disease. If I merely succeed in placing before you three of these general facts in such a striking manner as to make them easy of remembrance, my object will have been attained. In the first place all are, I suppose, prepared to admit that many circumstances may predispose to Diarrhea, as to other diseases; such we may conceive to be— Artificial feeding of infants. Eating unripe or overripe fruit. Foul air—Bad water. High temperature—Overcrowding. Syphilis—Struma. In fact, any cause of depression of the general health may be regarded as a possible, or even probable, predisposing cause of Diarrheea. But secondly, it is at least not improbable that there exists some special or specific cause—at present unknown, or at best the subject of conjecture—which must be superadded to one or more of the predisposing causes before an outbreak occurs, or which may possibly produce an outbreak independently of any other cause. This specific cause may be a meteorological condition, a peculiar form of pollution of air or water, some animal or vegetable matter in a state of change, or even a living germ. Thirdly, the method of inquiry employed by me necessitates this limitation. the importance of which I fully admit, viz. :— When in the course of the paper Diarrhea” (or any other disease) is mentioned, it must be clearly understood that fatal Diarrhea (or other disease) is alone meant. This limitation is forced upon me, because the arguments to be brought forward are all derived from the statistics of the causcs of death in this country, so laboriously and 272 STUDIES IN STATISTICS CHAP. carefully collected by Dr. Farr and his assistants in the General Register Office. These statistics are without doubt liable to numerous, indeed serious, errors, due in the main to mistakes of diagnosis, and carelessness in filling up the registers by, it is to be feared, too many practitioners throughout the country ; while on the other hand, the classification adopted by Dr. Farr needs careful consideraation by any one who may desire to draw conclusions from the facts recorded. It is nevertheless but poor logic and unworthy criticism to say that these imperfections alone condemn the system, and make its results valueless. The sequel of this paper, not to speak of other work, should, I think, convince the most sceptical that when perfectly independent inquirers, working with entirely different methods, in one case on the London statistics alone, in the other on those of all England, arrive at the same results, there must be “something in ” the figures. The conclusion is strengthened when, by an appeal to the statistics of New York, most of these conclusions are confirmed. In the disease under discussion, the simple fact that many deaths are recorded as having been preceded by Diarrhecea—even as a mere symptom—is sufficient, with the very large numbers at our disposal, to establish certain laws of the disease as a fatal malady. From what has been said above, it follows that the attacks of sharp purging, of a few hours’ duration, which so often follow the ingestion of improper food or an undue quantity of food, exposure to cold, or violent exercise after a meal, are left out of account entirely. It is obvious that we shall be far on the road to the solution of a given difficult stiological point when we can, with respect to any disease under investigation, give satisfactory replies to the three following questions :— 1 See especially Chapter XVII infra. CAUSATION OF SUMMER DIARRHEA 273 L.—Where does it kill ? 2.—When does it kill ? 3.— Whom does it kill ? We will sce what the Blue-books answer to these questions in the case of Diarrhcea. Since 1870, Dr. Farr’s quarterly reports have given the death-rates from Diarrheea during the three months—July, August, and September, for— a.—The whole of England and Wales. b.—The twenty largest English towns. c.—Fifty other large town districts. d.—All that portion of England and Wales which is not included in the seventy large towns, and which tor shortness is termed, ¢“ The country districts.” [ have calculated the average death-rates in the summer quarters of the ten years ending with 1879, for the whole of the seventy towns. The small diagram shows the more important of the results in a striking manner. The vertical line on the left shows the scale in deaths per 1000 of the estimated population. The names are printed as nearly as may be opposite to the figure on the scale corresponding to the death-rate of the place—only the towns with very high or very low rates (and London) are indicated—mno less than thirty-five of the towns have rates between 2 and 4 per 1000.1 The average death-rate from Summer Diarrheea (or, more correctly speaking, from Diarrhea in the summer quarters) of the whole of England and Wales is 21 per 1000. ' It must be understood that «a death-rate of 4 per 1000 in the summer quarter” does not mean that of 1000 persons living 4 died in that quarter, but that if the mortality had remained the same for a whole vear 4 would have died 274 STUDIES IN STATISTICS CHAP. That of the twenty large towns is as high as 3%, while that of the fifty other towns is 8 per 1000. The death-rate of all the rest of the country, that is of all the registration districts not included in any of these seventy towns, and which comprise besides districts properly called rural, a large number of smaller towns, is 11 per 1000. This makes it quite evident that, in the summer, Diarrhea is twice as fatal in the fifty towns, and 2% times as fatal in the twenty towns as it is in the country districts. Therefore the answer to our first question, “ Where does Diarrheea kill 2” is—* Mainly in the large towns.” It is very probable that if we could separate from the “ country districts” a hundred of the small towns therein comprised, the result would come out still more strongly in favour of the country. Incidentally, it may be remarked, that diphtheria shows almost an opposite distribution, being essentially a rural disease.! When the death-rates of the several towns are investigated, very striking differences come to light. Thus the diagram shows that the death-rate in Leicester is over 8 per 1000, or 5% times the rate prevailing in the “ country districts.” Wigan and Preston follow next, with rates about 7. Salford, Leeds, Hull, Yarmouth, Bolton, Liverpool, Birmingham, Manchester, and Northampton follow, all with very high rates. The London rate is almost exactly the same as that of the fifty towns. Of the whole seventy towns, no less than fifty have rates higher than the mean of England and Wales, while but four—Dover, Halifax, Swansea, and Merthyr Tydfil— 1 See Chanter X VIL 2nfra. CAUSATION OF SUMMER DIARRHEA 275 enjoy lower rates than the average of the country districts.” DIARRHEA DEATH-RATES. Summer Quarter Average, 1870-79. RATE PER 1000 [LEICESTER WIGAN PRESTON SALFORD LEEDS HULL YARMOUTH BOLTON BIRMINGHAM LIVERPOOL MANCHESTER NORTHAMPTON I'WENTY TOWNS FIFTY TOWNS LONDON ENGLAND AND WALES BATH SHREWSBURY SOUTHAMPTON COUNTRY DISTRICTS DOVER HALIFAX SWANSEA MERTHYR The case of Merthyr Tydfil, a mountain town of collieries and ironworks, is very remarkable—its rate being 276 STUDIES IN STATISTICS CHAP. actually less than half the average °‘country districts” rate. The general death-rate of Merthyr for the ten years 1870-9 was 22'3, or a little below the mean rate of the fifty towns; and its zymotic rate was 89, or almost the same as that of the fifty towns. Merthyr in past times has suffered severely from cholera and from typhus fever, but it has now for many years been in the fortunate possession of a most excellent medical officer. of health. On the other hand, no efforts seem able to reduce the extraordinarily high mortality of Leicester from this disease. Again, why is Halifax so remarkably conspicuous among the northern towns for its low Diarrhcea-rate ? When the seventy towns are arranged in the order of their average Summer Diarrhea death-rate, the twenty towns at the top of the list are found to have an average rate of 5-2, while the twenty towns at the other end have a rate of only 1'8. What factor in common have the twenty towns comprising one group, which, being absent in the twenty towns comprising the other group, causes the latter to have a Diarrhea death-rate of little more than one-third that of the former ? Density or proximity of population is not the common factor, although the size of the towns does seem to have something to do with it, since out of the twenty worst towns no less than ten belong to the group of the “ twenty large towns,” whereas but one of these latter (Bristol) appears in the healthy class. The geological system upon which a town stands appears to exert no influence, although it is not improbable that porosity or impermeability of soil (independent of its geologic age) may be of great importance. Elevation above the sea does not make much difference, but this 1s not necessarily a measure of the fall of the XV CAUSATION OF SUMMER DIARRHEA 277 sewers—a point meriting investigation. Proximity to the sea is without effect. Accordingly, in the absence of any other clue to a solution of the problem, it is very desirable that the whole of the seventy towns should be classed according to— 1. System of excreta disposal. 2. Fall and ventilation of sewers. 3. Porosity of soil, depth of ground water, and movements of same. 4. Character of water supply. 5. Prevalent trades, and especially the employment of women. 6. Consumption of alcoholic liquors per head. 7. Existence of the habit of opium-eating, and the like. Meanwhile the answer to the question, “Where does Diarrheea kill ¢” may be thus modified—* Mainly in towns, but with some remarkable exceptions.” We now come to the second point of the inquiry— “When does Diarrheea kill 2” This has received great attention, and is fairly worked out. Messrs. Buchan and Mitchell in their admirable paper On the influence of weather on mortality from different diseases and at different ages, have answered a part of the question. Their diagrams, constructed from the London mortality tables for the thirty years 1845-74, show with great clearness the average fluctuation per cent. from the mean in each week of the year for a great number of diseases. (See Plate XVI.) In some cases—notably Diarrhosa—the weekly death-rate fluctuates so regularly as to form a simple curve. Thus, taking one year with L Journal of the Scottish Meteorological Society, July 1875, p. 187, William Blackwood. 278 STUDIES IN STATISTICS | CHAP. another, the mortality from Diarrhea is below the mean from the beginning of October to the end of June, while. it very greatly exceeds the mean during July, August, and September. From the beginning of December to the end of May the curve may be regarded as a straight line. It rises gradually at the beginning of June, the maximum being attained in the first week in August, whence it falls less rapidly than it rose, until the end of November, following the law that is usually observed in epidemics, the rate of decline being probably some such function of the communicability of the disease that, as the communicability decreases, the mortality decreases, but at a slower rate, It is not my object in this paper to enter very fully into the relationship of Diarrhea fatality to meteorological conditions, but I will merely remark that the commencement of the rise in June and the end of the fall in November are very approximately three weeks after the rise of the mean temperature of the air above 50° and its fall below the same point. Whereas the portion of the curve that is above the mean corresponds closely with the portion of the mean temperature curve that is above 60°, but is about fourteen days later. The population of London is so large, the number of gears (1845-74) so long, and the average weekly number of deaths (minimum 18, mean 52, maximum 213) is so great ; moreover the curve is so smooth, so nearly a true curve in the ordinary acceptation of the term, that i may be regarded as the correct expression of a natural (or rather wmnatural) law, at least as regards London. Messrs. Buchan and Mitchell have shown that the Diarrheea curve for New York is almost the same as that for London, the discrepancies being explicable by the climatic differences of the two places. LONDON WEEKLY DEATHS FROM DIARRHOEA & CERTAIN OTHER DISEASES Excess or Defect per cent. compared with the Mean. AVERAGE OF 30 YEARS, 1845-74. Mean Temperature of the Air at GREENWICH GF hC JAN. rT 5 JUNE lJuLY] AUG.l £-n.'0OCT.|NOV.I DEC. - - 1 30 50 a0 4 MEANS - Epi et ZNCLISH PHGLE®R A YSENTERY THRUSH TABYS EERE AUR BE, 3 9) VeWT CF BYE wR Lg] JAUMDICE ENTERITIS ATROPHY DEBILITY Srale nfo ee - LoL Lt JAN.| FEB.IMARCH | APR.| M AY JUNE JULY|AUG.| SEP. OCT. Nov. | DEC. _ - ’ STANEORDS GEOG ESIABT LONDON "he curve of “Want of Breast Milk” is deduced from the statistics of 77 years only. ’ "he small divisions correspond to ten per cent of excess or defect from the Mean. ‘he temperature is the average of 20 years, taken from the weekly Returns of the Redistrar General for 1872 ~~ : S Xv CAUSATION OF SUMMER DIARRHEA 279 The second question, “When does Diarrheea kill 2” may then be answered as follows :— During six months of the year (December 1st—May 31st) the deaths from Diarrhea in London are practically constant, averaging between 14 and 15 per week. Three weeks after the mean temperature of the air rises above 50° F. (i. e. about June 1st), the deaths begin to get more numerous, and a fortnight after the temperature exceeds 60° F. (z. e. about July 1st), the deaths begin to exceed the mean (in London 52 per week). The greatest number of deaths occurs in the first week in August (on the average), whence the numbers fall as they rose—mutatis mutandis. The number of deaths thus appears to be some function of the temperature. The coincidences here pointed out, it must be clearly remembered, are merely between meteorological conditions and deaths registered. In searching for their causal relations, allowance must be made for— 1. Interval between registration and death, usually under a week, more often only two or three days. 2. Duration of illness. 3. Period of incubation (if any). The corrections for these would tend to make the temperature curve and disease curve correspond still more closely, unless, indeed, Summer Diarrhoea has a very long period of incubation, which we have no reason to suppose. The constancy of the deaths during a large portion of the year suggested to me the desirability of investigating the geographical distribution of winter Diarrhoea. This was a work of more labour, since the figures were not ready to hand, and some enumeration and calculation were required. The first quarter appeared the most favourable for 280 STUDIES IN STATISTICS investigation, since it is in no way affected by the summer rise. Taking again the average for the ten years 1870-79, it is found that in the first or winter quarter, the death-rate for the country districts” is identical with that for England and Wales, viz. 0°83 per 1000. That of the twenty ;owns is somewhat below, that of the fifty towns somewhat wbove, the mean. I examined more minutely the twelve towns at the top as well as the twelve at the bottom of my first list, and found that in place of the unhealthy towns having a rate nearly four times that of the healthiest towns, it was barely twice as great. Wigan showed the highest rate, L‘0 per 1000 ; then Preston ; Leicester on the other hand was amongst the best, 0'2 per 1000; while Shrewsbury, seven from the bottom in summer, came out third from the top. There were many other cases showing a very different state of affairs in winter from that which exists in summer, and strongly confirmatory of the theory that Summer Diarrheea is a disease sui generis. Messrs. Buchan and Mitchell's diagrams show that the curves for the following eight causes of death— Simple, or English, Cholera, Dysentery, Want of Breast Milk, Tabes Mesenterica, Thrush, Jaundice, Enteritis, Atrophy and Debility, nll, to a greater or less extent, resemble the curve for Diarrheea—but in the case of the last seven, they exhibit less ample fluctuations. By the kindness of these gentlemen CHAP, ENGLAND AND WALES. DEATH — RATES FROM DIARRHOEA AND CERTAIN OTHER DISEASES WITH RAIN AND HOT DAYS AT GREENWICH 84748 49 50 51 '52 53 54 '55 '56 '57 58 '59 '60 '61 '62 63 ’64 65 66 '67 '68 '69 '70 ‘71 72 73 ‘74 '75 ’'76 77 78 '79 ) " (| 1 TF 1 - " 7 1 I TT TI 7 — I vr] — i ; om ma oe FAITRAT|[GREENWICH ~~ ~~ o oo RE CC — RAINJATIGREENWH , _ SE ES NS SS _ 1 _ wpe TT rT - ferme Jar” -. NE NN LT — 5 N rr TT _ 1 —_— - - oc = a odie Number of Days qn wh fe os p me emperathice bf the a TE : CL wi 82°Fio whe EA —00 VERN(77 ond 50 — T a — me ra fay i Ta 4 Hale, ap gem aeks 1 TTI oT J apes a N py M poly = S The Jemperature curves indicate Numbers of Days, all the other curves indicate excess or defect per cent from the mean srrans ceoey sme ‘aeh divison of the vertical scale represents ten per cont "he Meteorological Means are for the 30 years 1850 — 1879. fhe Death-Rate Means are for the 3Z years 1847 = 1878, Except Atrophy & Debility Mean which is for the 2/ yesrs 1858 -1878 A A. Asistic Cholers fpidemic in these years consequently figures of English Cholera valueless he Death-Rates are all calculated per million persons living. XV CAUSATION OF SUMMER DIARRHEA 28; I have been enabled to reproduce (with certain modifications as to colour, &c.) some of their diagrams in Plate XVI. Simple, or English, Cholera may be provisionally regarded as a term applied to the more severe cases of Summer Diarrheea—and, as might be expected on that theory, its curve is simply the Diarrhea curve intensified —almost disappearing in winter, and rising to a proportionally higher point in summer. Some of the winter cases are probably due to arsenical poisoning. To attack the question from another point of view— The diagram (Plate XVII.) represents the rise and fall of the annual death-rate from Diarrhea in the whole of England and Wales. In this case the vertical columns represent years, and the angular coloured lines (euphemistically termed curves) represent the fluctuations of the death-rate from year to year—the horizontal portions indicating the height to which the rate rises, the sloping line being put in to give continuity to the whole. The brown line represents the Diarrhcea rate, the dotted line the mean for 82 years. It will be at once noticed that the Diarrheea curve reaches its highest point in the year 1868, with its hot and dry summer, and its lowest in the years 1860 and 1879, when the summers were remarkably cold and wet. In fact, in general terms it may be stated that Diarrheea is most fatal in years with hot summers, especially if they be at the same time dry ; and is least fatal in years with cold summers, especially if they be at the same time wet. It must, however, be admitted that although this rule is strictly correct for extreme years, there are nevertheless many exceptions, and we are as yet far from knowing what are the precise conditions which determine the fatality or otherwise of the disease. The number of rainy days appears to be at least as 282 important in its pathological relations as the amount of rain that falls. I have endeavoured to represent on the chart the wetness or otherwise of the three summer months by combining these two meteorological data into one, and representing this by the pale blue area; the quantities (being the algebraic sums of the percentage excess or defect of the rainfall and rainy days) being measured from the top line downwards, so as to show that the relation between Diarrhoea mortality and wet seasons is inverse. This method we owe to Dr. Buchanan. By a careful detailed examination of the deaths from Diarrhea in London during the summers of 1875 and 1878, and the meteorological conditions at Greenwich for the same periods, I came to the same conclusion as Dr. Farr, namely, that the temperature of the water of the Thames has a closer relation to the mortality from Diarrheea than the temperature of the air, but the temperature of 62° Fahr. appeared to me to be more critical than that of 60° Fahr., which he mentions in some of his reports. It therefore occurred to me that the number of days on which the mean temperature of the Thames rose to 62° Fahr. or higher, would form a good measure of the « pathological heat” of the summer, if I may be allowed the use of such a term. This enumeration, a somewhat laborious one, was carefully made for me by Mr. F. A. Dixey, and the result is shown in the diagram by the pale crimson area, the boundary line of which will be seen to correspond fairly with the Diarrheea curve. Above this is a pale brown area representing the number of days in each year on which the mean value of the temperature of the air equalled or exceeded 60 Fahr.; this will be seen also to correspond fairly closely with the Diarrhea curve. It is my intention to construct other curves cor STUDIES IN STATISTICS CHAP. XV CAUSATION OF SUMMER DIARRHEA 283 responding to various air temperatures; but so far as I have gone, and I believe the same is true of the labours of Dr. Buchanan in this direction, all attempts to express the Diarrhoea mortality in a given place as a function of the temperature only, have failed. The reason is probably a simple one, viz.—that Summer Diarrhea is a disease very greatly influenced by temperature, but not caused by heat alone ; it is rather a communicable zymotic disease which thrives best during hot weather. This is to some extent borne out by the observations of Dr. W. Johnston on Summer Diarrhoea at Leicester in 1878; that gentleman remarked that during a prolonged microscopical investigation of the dejecta of Diarrhea patients, he three times had to desist, owing to severe outhreaks of the disease in his own person. The curves on the chart, below that of Diarrheea, viz.— English cholera, thrush, tabes mesenterica, want of breast milk, jaundice, enteritis, atrophy and debility, and dysentery, all exhibit, when compared in what I may for shortness term my method, much the same relation to Diarrhea that is indicated in Messrs. Buchan and Mitchell's diagrams; 1 fact that, to my mind, is alone sufficient to prove that Dr. Farr’s fioures have at least some value. We now come to the third question. “ Whom does Diarrheea kill 2” In the next diagram (Plate XVIII) the annual fluctuations in the Diarrhoea death-rates at different ages are represented in a similar manner. Owing to the very small number of deaths from Diarrhoea at certain periods of life, I have (to avoid fallacies as far as possible) added together some of the groups of {Summer Diarrhea, its Nature, Cause, and Treatment. Lancet, 1878. STUDIES IN STATISTICS CHAP. ages in the annual reports, so as to reduce these from seventeen to seven. My “seven ages of man” are as follows— [. 0-1 Infants under 1 year. Period during which, normally, all subsistence should be drawn from the mother’s breast. II. 1-2 Infants between 1 and 2 years. Period of transition to a mixed diet. IIT. 2-5 Young children between 2 and 5 years. Mixed diet established. IV. 5-15 Boyhood and girlhood. To puberty. V. 15-25 Youth. Puberty to manhood. VI. 25-55 Prime of life. VIL Over 55 Period of decline. The vertical columns again indicate years; the seven coloured lines indicate the fluctuations of the death-rate at the seven ages, from year to year. As before, first the death-rates per million persons estimated to be living in each year at each age were calculated, and from these a mean rate was obtained for cach age. Then, to insure uniformity of scale, the curves were so constructed that in every case each one of the smaller divisions of the vertical scale corresponds to a deviation of 10 per cent. above or below the mean. The actual value of the mean is indicated in figures on the right hand, from which it appears that in the fifth period (youth) the death-rate from Diarrheea is an almost vanishing quantity, only 0°04 per thousand (or 40 per million) living at that age; during the preceding period the mortality is little greater, and during the prime of life only 0-1 per 1000. During the third period (2-5 years) it is *6 per 1000, but during the period of decline it rises to 1'2 per 1000, during the second year of life to 5°1 per thousand, and ENGLAND AND WALES. DEATH — RATES FROM DIARRHOEA AT SEVEN DIFFERENT AGES Dotted lines indicate mean Death-Rates for thirty two years 1847-78 BOTTI SEXES. AGES _ oo . aT 48149150 1s 22 15a Taal en Pe Tar Toa Teo Tea r OJ TO 9 (J UT 0 v FO v 0 oe 6416566 67 68 'a9 170 sr 12 [ge oes lige Pra hag Dog T o oh \ N50 L__ DES _E ~838 SD — NZ =X tte I —— A=STANFORD'S GEOG: ESTABT S CAUSATION OF SUMMER DIARRHEA 285 during the first year to no less than 17-2 per 1000 living at that age. Tt will be at once apparent that the variously coloured curves in the diagram deal with very different numbers. This difficulty is obviated by the method above described, which was suggested to me by Messrs. Buchan and Mitchell's paper, viz. calculating the fluctuations from the mean, and plotting out these instead of the death-rates themselves ; e.g. the top (red) line in 1868 rises 85 per cent. above the mean, and in 1860 it falls 50 per cent. below 7t. In the same years the yellow line rises 24 per cent. and falls 20 per cent. If now the relations between the curves be analysed. it will be observed that— Firstly. The top (red) curve shows a general upward tendency during the thirty-two years, as does also the second (pale red), but to a less degree; the third curve (dark green) shows, if anything, a tendency to fall ; all the other four curves are decidedly declining, the fall being most marked in what, for shortness, we may term the curve of youth (dark blue), and least so in that of old age (yellow). Secondly. The three great cholera epidemics of 1849, 1854, and 1866 affect the curves very differently. All the curves from two years up to fifty-five are profoundly influenced by the presence of cholera, the old age curve to a much less extent. The infantile curves are hardly at all affected. Thirdly. All the curves rise to local maxima in 1857, 1861, 1868, and 1878. Fourthly. All the curves fall to local minima or are very low in 1850, 1855, 1856, 1860, 1862, 1869, and 1877. Fifthly. Those fluctuations which are independent of the visitations of Asiatic cholera are less marked in all the 56 STUDIES IN STATISTICS CHAP. ages above five years than they are in the three earlier periods of life. Siwthly. The second curve is almost identical with the first, a fact having an important bearing on the breastmilk theory. Here it is necessary to make reference to some deductions from data derived from Messrs. Buchan and Mitchell's paper. They have analysed the average weekly numbers of deaths from Diarrhoea in London at six groups of ages for the six years 1869—1874, and have shown that at every age the number of deaths rises during the summer months, though in very different degrees. To avoid fallacies arising from small numbers, I have thrown the six periods into three. The average number of deaths during those six months of the year—December to May—during which the morbality from Diarrheea in London is known to be practically constant, is taken as unity ; then the average mortality at cach age during the three months of high mortality is compared with this. The following are the resulting ratios :— Ages 0— 5 .. ... ... 1:216 5—60 ... ... ... 1: 858 60 and upwards ... 1 : 98 Summer Diarrhoea, then, kills persons of every age, but young infants in by far the largest proportions. It must here be observed that the figures upon which both these diagrams depend, are the deaths from Diarrhcea during the whole year. (Plates XVII. and XVIIL) But on the average, in the whole of England and Wales, twice as many deaths from Diarrhea occur during the three summer months as during the remaining nine months of the year, so that two-thirds of the numbers XV CAUSATION OF SUMMER DIARRHEA dealt with are to be referred to summer Diarrhoea, though as the deaths in June and October are evidently closely connected with the summer outbreak, the proportion is really still larger. In the last diagram (Plate XVIIL) we may, judging from the London data, refer about six-sevenths of the deaths from which the first three lines are deduced to summer Diarrheea, but only one-half of the deaths on which the other four lines are based. This is no doubt pro tanto a source of fallacy. 287 We may, I think, legitimately draw the following conclusions from the facts that I have laid before you: — A.—A variety of causes produce throughout the year deaths which are registered as due to Diarrhea. These causes affect persons at all ages, and are activein all places in such a way that the Diarrhcea death-rate in the first quarter of the year is pretty nearly the same in town and in country, and in one town as in another town. B.—There exists, in addition, a specific form of Diarrhea which makes its appearance in the summer months, disappearing again with the advent of cold weather ; this disease, which may be termed Summer Diarrhea, kills persons at all periods of life, but in much larger numbers at the extremes, and in by far the greatest proportion during the first two years after birth. This disease is greatly affected by the warmth or otherwise of the season, and to a less extent and in an inverse manner by its wetness; but this relationship is such as to make it probable that high temperature and drought are only secondary causes in its production, a fact which is strongly confirmed by its remarkable geographical distribution, showing, as it does, a marked partiality of diarrhea for certain large towns. 27 STUDIES IN STATISTICS CHAP, If these general conclusions be admitted as proved, we may dismiss as untenable certain theories as to the causation of the disease which have been from time to time put forth, viz. :— 1. Bad water and filthy privies. These are as common or commoner in the country than in the large towns, and are therefore not likely to be the cause of a town disease. 2. Fruit, ripe or unripe. This is not extensively saten by children under 1 year of age, and therefore could only affect them through the maternal organism, which is possible, though not in a high degree probable. 8. Artificial feeding. This, which, during hot weather, would often imply feeding with substances in a state of incipient fermentation, would not account for the deaths of adults or old people, nor of children between the ages of 2 and 5. Moreover it is stated, on good authority, that many children have suffered from Summer Diarrheea who have been entirely fed from the breast. 4. There remains then bad air. This may be either— a. A locally-bred miasm from the soil, or bh. Sewer air. As a working hypothesis, I agree with Dr. W. Johnston's theory, and suggest that the exciting cause of Summer Diarrhoea is intimately connected with the process of putrefaction—however and wherever arising,—but that in a large proportion of the cases the infective material has its sources in the public sewers, and is introduced into the system through the lungs. Further observations are much needed on the following points :— I. What proportion of children under 1 year of age, fed solely from the mother’s breast, suffer from the disease, compared with those fed artificially ? Lists of such cases compiled with the utmost precautions are very desirable. CAUSATION OF SUMMER DIARRHEA 289 II. In such cases did the mother, or any members of the family, suffer from Diarrhcea ? IIT. What is the usual sequence of cases in families ? If medical officers of health in whose districts the disease is prevalent, would systematically compile records of the kind here hinted at, we ought, in a few years, to know more of the wtiology of Summer Diarrheea, and therefore be in a better position to prevent it. KV S 2. Summer Diarrhea in the large Towns of England tn 1880.1 In February of the past year I made a communication to this Society on the subject of Summer Diarrhcea. No pretence of extensive clinical acquaintance with the disease was alleged, the materials for the paper being entirely derived from the returns of the Registrar-General, and the inquiry was, therefore, limited to fatal cases. An attempt was made to establish certain general facts, which any complete theory of the disease must account for, viz.:— (@) A variety of causes produce throughout the year deaths which are registered as due to Diarrhcea. These causes affect persons at all ages, and are active in all places in such a way that the Diarrhea death-rate in the first quarter of the year is nearly the same in town and country, and not remarkably excessive in any individual town. (b) There exists in addition a specific form of Diarrhoea which makes its appearance in the summer months, disappearing again with the advent of cold weather; this disease, which may be termed Summer Diarrhea, kills persons of all periods of life, but in much larger numbers at the extremes of life, and in by far the greatest proportion during the first two years. It is greatly affected by L Read before the Society of Medical Officers of Health, 1881. uF STUDIES IN STATISTICS CHAP. the warmth or coolness of the season, and to a less extent by its dryness or wetness; but this relationship is such as to make it probable that high temperature and drought are only secondary causes in its production, a probability which is greatly strengthened by its remarkable geographical distribution. Thus, as a general rule, it is a disease of towns; but different towns are very differently affected — Hull, Leeds, Salford, Preston, Wigan, and notoriously Leicester, suffer in a remarkable degree ; whereas Dover, Halifax, Swansea, and especially Merthyr Tydfil, °njoy a no less remarkable immunity. (c) As a provisional hypothesis it was suggested that some products or concomitants of putrefaction are the exciting cause of Summer Diarrhoea. That these are probably chiefly taken into the system through the lungs, and that the distribution of the disease is connected with the axistence in certain towns, and parts of towns, of sewers vith but little fall and insufficiently flushed. I propose this evening to consider shortly the prevalence of fatal Diarrhoea in England and Wales during the past summer. The summer of 1880 was a warm one ; in August and September the mean temperature was above the average— remarkably so in the latter month. The highest temperature of the year was recorded in most places as late as September 4th. The rainfall of July and September was excessive ; but from August 8th to September 6th scarcely any rain fell. The death-rate from Diarrhea in England and Wales for the three months, July, August, and September, 1880, vas 332'—a little over 21,000 persons, mostly infants, having perished from this cause alone during the above period. This death-rate exceeds the average of the pre-1 See note on p. 273. Xv CAUSATION OF SUMMER DIARRHEA _ . 291 vious ten years by nearly 50.per cent., and since 1847 has only been exceeded once, viz. in 1868. Some counties enjoyed comparative immunity, viz. Wilts, Westmorland, Monmouth, and North and South Wales, whereas others suffered severely, viz. Warwick, Lancashire, East Riding of Yorkshire, and Leicestershire. Certain points in the distribution of the disease during the past year call for special comment, and for convenience I may state that the following remarks are confined to the facts of the eleven years 1870-80. The death-rate from diarrhcea in the ‘twenty large towns,” which had hitherto averaged over 66 per cent. in excess of the death-rate of England and Wales, was in 1880 but 33 per cent. above the latter, the smallest excess in any previous year being 50 per cent. in 1878. On the other hand, the death-rate of the “fifty other towns” was somewhat above the average, so that for the first time it slightly exceeded that of the “twenty large towns.” The mortality in the country districts was, in 1880, 24 per cent. below that of England and Wales, or proportionately slightly greater than in any previous year, so that on the whole Diarrhea seems to have been more generally distributed than usual in the past summer. Perhaps this is what we should expect, since small towns are growing up in the so-called country districts, whereas the sanitary administration of the large towns ls improving. } When, however, the separate towns are examined, we find Preston heading the list with a death-rate of 18:6 (for the three months), which is greater than any that has prevailed during the past eleven years in any of the seventy towns, the nearest approach to such a high rate being 12-4 in Wigan in1870. Preston has had the distinetion of heading the list for the last two vears: it has also 202 STUDIES IN STATISTICS CHAP, on three occasions stood second, and for the ten years has averaged third. Second in the past year comes Leicester, with the very high rate of 10'6. Leicester has seven times headed the list and been twice second—in the average of the ten years it has stood facile princeps. Third comes Stockport, with the high rate of 10-0. Stockport has never enjoyed a low death-rate from Diarrhoea, but has never before been so undesirably distinguished (average 16th). York comes out fourth with a rate of 8:6; it has never suffered so much before, and indeed there is reason to fear that Diarrheea is on the increase in that ancient city, its average place being 29th. Fifth comes Salford, one place below its average. Sixth stands Coventry— average 28th—but very variable throughout, having ranged from 11th to 56th. Seventh comes Ipswich— average 24th, but ranging from 6th (1878) to 64th (1870). Diarrhoea seems to be increasing in this town. Kighth stands Hull, two places below its average. The eight towns at the other extremity of the list were in 1880, Merthyr, Swansea, Halifax, Newport (Mon.), Bath, Huddersfield, Rochdale, and Bristol. Of these Merthyr, Swansea, and Halifax occupied their usual places, but it is remarkable that the very small mortality from this disease in Halifax and Merthyr continues to decrease. In the summer of 1880 sixteen towns stood much higher in the list, 7. e. suffered much more from Diarrheea than usual, viz. Stockport, York, Coventry, Ipswich, Blackburn, Walsall, Stoke-upon-Trent, Cambridge, Reading, Maidstone, Plymouth, Southampton, Cheltenham, Dover. Hastings, and Chatham. On the other hand, eighteen towns stood much lower in the list than usual, that is, enjoyed greater immunity -— Wigan, Bolton, Birmingham, Leeds, Yarmouth, Liverpool, Manchester, Bradford, South Shields, Newcastle, London XV CAUSATION OF SUMMER DIARRHEA 203 (never so low before), Barrow, Colchester, Oxford, Dudley, Bristol, Rochdale, and Huddersfield. To take, in the case of any individual town, its average death-rate for a period of ten or more years, and compare this with the like average for other towns, does not give all the information that the figures are capable of giving. Again, to quote in the case of the same town the deathrates in a series of successive summers, is most deceptive, since it may lead to the inference that Diarrheea is on the decrease, or the reverse, in the district. It is necessary to have a standard for each year. The best available standard is the death-rate of the whole country during the same period. I classify the towns in each year in seven classes, according to the ratio that their death-rate from Summer Diarrhoea bears to that of England and Wales. In this way any one town is readily compared with any other, and we obtain such curious results as this, viz. :—that whereas Salford, with certain fluctuations, maintains its average place in the lists thus formed, Manchester, on the other hand, is slowly improving. The Sanitary Authorities of those important centres of population are striving, in a most praiseworthy manner, to remedy the filth nuisances for which Lancashire is notorious in sanitary literature: if their Medical Officers of Health, with all their local knowledge, were to investigate thoroughly the causes of the different behaviour of Summer Diarrhoea of late in Salford and Manchester, the results could scarcely fail to throw much light on the comparative atility of ‘the various methods of sanitary reform employed. In like manner Newcastle and Gateshead appear to be sometimes very differently affected by what must be very similar, if not identical, meteorological conditions. Why is this? In several years the death-rates of Plymouth 204 STUDIES IN STATISTICS CHAP, and Devonport differed materially, while during the six years that East Stonehouse was ranked among the fifty towns,” its death-rate was distinct from either. On the other hand, during the same six years, the Gosport and Portsmouth figures were more in harmony. The case of Manchester and Salford may be also looked at in another way. For the four years 1870-73, the average rate in both towns was 7-2; in the seven years 1874-80, the rate in Manchester averaged only 38, whereas that of Salford was 54. The figures, however, as thus stated, might lead us to imagine that the Salford rate was diminishing, but this deceptive appearance is mainly owing to the very low mortality from Diarrheea throughout the whole country during the cold and wet summers of 1877 and 1879. When the past eleven years are compared together, after making all possible allowance for the great fluctuations which the death-rates of many of the towns exhibit, it will be found that in the following fifteen towns there is a tendency to a decrease in the death-rate from Diarrheea in the summer quarter, viz. Wigan, Bolton, Manchester, Sheffield, amongst the worst towns; South Shields, Bradford, Bury, Colchester, Gateshead, Middlesborough, Oldham, amongst the medium towns; Rochdale, Hudders-Geld, Chester, Chatham, amongst the healthier towns. On the other hand, the Summer Diarrheea death-rate appears to be on the increase in eight towns of the medium class, viz. St. Helens, Lincoln, Ipswich, York, Worcester, Cambridge, Plymouth, Maidstone, and also in one of the healthier class, viz. Dover. In the remaining forty-five towns the mortality, though subject to great fluctuations in some cases, does not show any decided tendency either to increase or decrease. Unfortunately in only eleven out of the twenty large XV CAUSATION OF SUMMER DIARRHEA 295 towns is it possible to compare the Diarrhcea mortality in the past summer with the meteorological conditions prevailing locally at the time. In the case of these eleven towns such a comparison fails to connect the fatality from the disease as closely with the temperature as might be expected—though the connection is sufficiently close to prove (as indeed all admit) that high temperature is a necessary factor in the causation of the disease. The mean temperature of the air was 60° or higher, in London for no less than ten weeks during the four summer months, whereas in Nottingham such a mean temperature was only reached in six weeks; nevertheless Nottingham suffered proportionately more than London to the extent of 60 per cent. The highest weekly mean temperatures were reached in London and Sunderland, both in the first week of September. In Sunderland the Diarrhoea deaths reached their highest point in that week, but in London the deaths were only one-fourth as numerous in proportion to population, and the highest point of the epidemic had been reached a month earlier. Mean temperatures of 60° or higher occurred during June in seven out of the eleven towns, but whereas common experience teaches that these early high temperatures are most effective in causing lassitude and exhaustion, in no case were they accompanied or immediately followed by numerous deaths from Diarrhea: a fact long ago noted, and which is easy of explanation on the theory that the disease is excited not by Ze direct action of high temperature upon the nervous system, but ‘by the products of putrefaction which are dependent upon prolonged high temperature. In nine out of the eleven towns, the highest mean temperature was registered in the first week of September. The highest point of the epidemic was before that week in four of these nine towns. 29€ STUDIES IN STATISTICS CHAP. It is a remarkable fact that although August was hotter and more free from rain than July in London and Brighton, in both towns Diarrhea deaths diminished in frequency as the month approached its end, in spite of rising thermometer and rainless sky. The incidence of Summer Diarrheea from week to week in the past summer no more admits of explanation by meteorological conditions alone, than does its distribution in different parts of the country. To explain the course of the annually recurring epidemic in any particular year in any one district, it is requisite that besides meteorological conditions we should also be acquainted with other factors, such among others as variations in quantity and quality of water-supply, flushing and disinfecting of sewers, street watering and scavenging, removal of house refuse, &e. In the last week of July the Diarrhoea rate in the “queen of watering-places,” Brighton, rose to 102 per shousand, a rate double that of any other of the great owns in the same week. What did the frightened inhabitants, or the sanitary officers, do in Brighton, that made she death-rate fall from that date, while in all the other great towns it was rising? Or what had they permitted to be done previously to the end of July, that made such a state of things possible? The epidemic in London reached its height as usual in the first week of August. Why was it almost entirely unaffected by the high temperature in September ? Bristol as usual suffered but little, and that little chiefly in September; the maximum being in the third week of that month, or eight weeks later than in Brighton. The incidence of rainfall in the eleven towns was not sufficiently different to account for these discrepancies. The events of 1880 I regard as confirmatory, so far Xv CAUSATION OF SUMMER DIARRHEA 207 as they go, of the theory which regards the exciting cause of Diarrhoea as an organism, or other concomitant of decomposition of organic matter, which can only exist, or become virulent in its properties, after prolonged high temperature. One writer on the subject has recently attributed much of the mortality from Summer Diarrhoea to decomposing food, especially fish, and more especially mackerel. This may no doubt kill a certain number of individuals; but if it be an important factor in the mortality of infants in the first two years of life, it must act through the maternal organism. Is it or is it not the case that the mothers of infants dying from Summer Diarrhoea usually themselves have the disease, and if so, in which generation does it show itself first ? Dr. A. McCook Weir, of Nottingham,! admits one of my points, viz. that the disease affects persons at all ages of life, but he denies the influence of sanitary surroundings. He says a fatal issue results only where there is a previous atrophic condition. This is very probable ; but as he also admits that persons not in an atrophic state of system take the disease readily, it merely proves that atrophy is a part of the cause of death, not of the disease. Is it not more than probable that the crippled child of three years old, who had been “crawling on the pavement under a broiling sun,” and afterwards died of English cholera, eight hours after the first symptoms, had been playing, as children often do, immediately over a sewer grating? Again, Dr. McCook Weir says, “ Dr. Johnson is well aware that infantile mortality must to a great extent, other things being equal, depend on the number of infants.” and again, in the same paper, “given, therefore, 1 Sanitary Record, Jan. 1881, p. 276. STUDIES IN STATISTICS CHAT, a uniform Dirth-rate, an equal degree of infant susceptibility, whether congenital or acquired, or both, and au equal degree of heat, and I venture to submit that there would be an equal prevalence of infantile Summer Diarrhoea throughout the country.”* To test the importance of birth-rate in the question, I have for each of the seventy towns calculated tke ratio per cent. of the Diarrhea death-rate in the summer quarter fo the mean birth-rate of the two preceding quarters, whereby all errors of estimation of population are avoided, this being in fact but another way of stating the percentage of Diarrhea deaths to births registered in an equal period of time. Now it is evident that if a high birth-rate were an important factor in the question, these ratios should be approximately equal for all the towns. What is the case, however ? Preston gives the highest ratio, 33 per cent. The nine towns with the highest Diarrhea rates give an average ratio of 24 per cent. Whereas the nine towns with lowest Diarrhoea rates give an average ratio of 5 per cent, Swansea and Merthyr having a ratio of 1 per cent. only. In the accompanying table all the towns having higher birth-rates than Leicester are in italics. An inspection of the table proves conclusively that a high birth-rate has little to do with the matter. I cannot but feel that induction from an insufficient basis of facts (but one or two towns being considered by each inquirer) has been the chief source of error in inquiries hitherto made into the causation of infantile Summer Diarrhoea \ Sanitary Record. Sept. 1879, p. 91. 93 [UNHEALTHY TOWNS. Preston res res Leicester ces ree Stockport oe ve York ... es Jo Salford oe es Coventry awn yor ‘pswich I _ Tull ... pa ns Northampton - St. Helens ... a. Wigan pa we 4shton-under-Lyne ... Norwich ce ee Blackburn... ve Bolton ve re Walsall ee ve. Sheffield ce. we Birmingham ... vee Leeds ... - wy Yarmouth ... ws a Stoke-upon-Trent ... Lincoln er 0 Liverpool we po Averages vee .. i538 YEE Bua | SERRE ELZ2| 450% 2S oR Sa 5e EEOLIERSEA agg |Z al IA wu: SR os Ra I¥ 5 BiB 36 106 00 2.8 417 33 384 | 28 41-2 24 346 25 10-2 ® 35-3 356 39:5 38-3 189 18°56 41-6 35-8 385 36:4 405 36-0 11-2 36:7 | ° 33:1 : 381 1 401 15 39-7 15 o" 3" 3-8 3 3c 30 1 31 51 31 51 5.0. | 75 “301 5 | 30 a 25 Mepiom Towns. Vottingham ... — ‘underiand ... - fanchester... sxe Vorcester ... we ‘ambridge ... Po 'ynemouth vk “Tolverhampton ee {eading ee . derby fs Le Bury... — ve Jrighton wun wo radford sin I. {aidstone ... wae Yateshead _ re “iddlesborough as ‘ortsmouth ... son lymouth ... 2 Southampton ... wi Cheltenham ... -— Jouth Shields se Carlisle . nad Newecastle-upon-Tyne ixeter cs ce Cardiff ves ee Dover ... vee an 225 32% 18,4 2354585. ERE |gneg EEIES neP E558 |8a28 Si £557] gE I= 2 MEE Sig Ss 258 Ee 0 3 3 5-0 1-9 3 380 14 397 13 875 “2 30-2 395 28% 28-2 27-R 94 1 i” 95, Ta-4 2R-9 "56 97 25 337 26°6 353 41-8 i 366 Fer 1 4 3 , 2 RQ “7 ‘6 6 36 pg 10 3 10 10 © oer RD ~ - Gt » BEE lang, tay 92 Heavray TOWNS, |Z FEF RwEgl 208 fego|2wEs| 88° |EES=|E5E%| 55s ag [gm HHS sondon ee ee Jdham er es [astings - - Yarrow-in- Furness ... “hatham ne ow Jolchester ... av Jevonport ... or Oxford - was West Bromwich ves dacclesfield... vi hester en Li Mrkenhead ... 5 Dudley ee ae Shrewsbury ... — 3ristol ee — 2ochdale en ee Tuddersfield... .. 3ath ... ae _— Newport (Mon.) ne Halifax os gus Swanseq oh won Merthyr Tydfil wo 3-3 38 3-2 Qe 3-5 12 1 1. 374 34-7 29-4 12°8 316 329 294 °52 58 4-0 34 379 28-9 33:8 353 299 3049 26-0 37-8 29-6 40-4 304 9 10 11 | 8 Le, 0 e 9 %8 25 9-4 70 211 21 i 4 - Averages 9:6 | asx INGLAND AND WALES Towns in Ztalics have higher birth.rates than Le’ 300 STUDIES IN STATISTICS CHAP. § 8. Summer Diarrhea in England and Wales. 1880-87. In the third quarters of these eight years the deathrate from Diarrhea averaged only 1-94 per 1000, as agaist a rate of 2:25 in the ten years 1870-79, being a fall of 14 per cent! This diminution in the Diarrhcea mortality may have been caused by either (1) meteorological differences between the two periods, or (2) sanitary improvements. ~ Without denying that meteorological conditions had a share in the result, I am disposed to attribute more weight to the latter cause. In the first place it is an undoubted fact that sanitary reforms have been in active progress during the whole period under review, and in the second place, prolonged high temperature and drought in the summer of 1884 did not produce such an increase in Diarrhea mortality as micht have been expected. In the long table (pp. 302, 8083), on the left hand side the seventy chief towns of England and Wales are seén arranged in the order of their average mortality from Diarrhoea in the third quarters of the ten years 1870-79; on the right hand the same towns are arranged in the corresponding order for the eight years 1880-87 ; in each case the town with the highest mortality is placed at the top of the list. In the middle column is shown the vise or fall, per cent, in the death-rate of the second period compared with the first. An examination of the table reveals, among many changes, certain constant features that are very striking. Of the 20 towns at the top of the list in the first period, no less than 16 were among the first 20 in the second period ; while Leicester, Preston, Salford, Leeds, and Hull were among the worst six towns in each period. 1 The rate varied from 1°14 in 1885 to 3-29 in 1880. xv CAUSATION OF SUMMER DIARRHEA 301 Again, of the 20 towns at the bottom of the first list, no fewer than 16 were also among the last 20 in the second period; nay more, Merthyr Tydfil, Swansea, and Halifax, as the healthiest towns, take the three places at the foot of both lists. This persistent selection and equally persistent avoidance of certain towns by Diarrhoea should give some clue to its causation. It will be noticed that while the death-rate fell more or less in 59 out of the 70 towns, it rose in nine; and in two, Merthyr Tydfil and West Bromwich, there was no change. Not only did the rate fall in the 28 most unhealthy towns, but even in 16 out of the 22 healthiest, especially in the case of Swansea, Bath, Huddersfield, Bristol, and Rochdale. The improvement should also be noticed in the case of Carlisle, Exeter, Oxford, Barrow-in-Furness, Tynemouth, Colchester, South Shields, Ashton-under-Lyne, Northampton, Manchester, Liverpool, Leeds, and more particularly Wigan. This last town was formerly a close rival of Leicester, whereas its mortality, having fallen one-half, from 7:2 to 3'6 per 1000, is now exceeded by that of thirteen towns. It occurred to me to write to the medical officers of health of those towns in which any marked change in the Diarrheea mortality had taken place, and ask them to name 1 Since the table was calculated the returns for 1888 and 1889 have been published, but as the Registrar-General no longer publishes the death-rates for the “fifty towns” (sce supra, Chapter XIIL p. 204), the results could not be incorporated without undue labour. The death-rate for England and Wales was, in the cold and wet summer of 1888, only 0:90; in 1889 it was 1-67. As regards the twenty-eight towns, for which the rates are still published, the averages for the two years showed Preston and Leicester still at the top (Salford and Leeds following closely), and the same four towns as before at the bottom of the list. Sheffield, Norwich, Sunderland, Portsmouth, and Birkenhead showed a marked relative deterioration, but Hull and Nottingham an equally marked improvement relatively to the other towns. 302 STUDIES IN STATISTICS CHAP. Average Death-rates from Diarrhoea in Seventy Large Towns of England and Wales tn the Swmmer Quarters of t'e Ten Years 1870-79, compared with the Bioht Years 1880-87. Ten years 1870-79. dose or fell . ! en 1 T60 GOL, | Right years 1880-87. i Leicester ... -19 Wigan - -50 Preston ... -1 Salford as -19 Leeds . -31 Hull — 7 Yarmouth ... -25 Bolton wn 25 , Liverpool ... rr | 31 to Birmingham wep ol -24 1 Manchester... ” -37 :z Northampton .. b0 -36 3 Sheffield ... .. 49-22 t4 Ashton-under-Lync 45 -47 tg St. Helens ... i. } 4d] 5 t6 Stockport ... ee ~-18 t7 Nottingham oe 43-7 18 South Shields .. 42" -36 tg Norwich ... oor 7 zc Lincoln ... an } 3.91 -15 21 Tynemouth ver -41 22 Newecastle-upon-Tyne -30 23. Sunderland ve 0 37 11 24./Ipswich ... av 8 25. Bradford ... _ } 36 -28 26. {Bury cn “ -25 27.|Wolverthampton ... 35 - 8 28.1Coventry ... . 34 -12 2g.|York - +21 30.{Colchester ... -33 31.|Blackburn ... - 8 32. Derby a vi ves -31 33.|Gateshead ... ves f 392 +13 34. Worcester ... “ee -31 35. Stoke-upon-Trent ... 3-1 3 81 72 6:9 ! 58" 5g Freston .e Leicester wih Salford . St. Helens ... Hull ... _ Leeds oe. Nottingham ... York ... vee Yarmouth ... Birmingham ... Bolton ss Norwich ... Sheffield Wigan " Gateshead ... Stockport ... Liverpool ... Ipswich _— Sunderland ... ss Lincoln er er | 33 Wolverhampton - Manchester ... van Northampton .- | 3-9 Walsall ave ‘es Blackburn ... oo 91 Stoke-upon-Trent ... Portsmouth ... po | 3-0 Coventry ... ow Cardiff ee 28 South Shields ves : } 2-7 Bury... win . Newcastle-upon-Tyne Bradford... ee | 2:6 London ad "id Brighton ... .. 925 6'8 66 4-7 42 | , 3:9 38 36 35 34 1 The towns that have gained improved places in the list (i.e. have gone down) are printed in #alics, those that have relatively lost ground as regards diarrhea (4.c. risen in the list) in elarendon tvpe. Xv CAUSATION OF SUMMER DIARRHEA 303 Average Death-rates from Diarrhea in Seventy Large Towns of England and Wales in the Summer Quarters of the Ten Years 1870-79, compared with the Eight Years 1880-87. R : fel] Ten years 1870-79. | a ont. 36. London ... aw 37 Brighton ... o } 30] 38 Portsmouth oy 28] 39 Barrow-in-Furness ... | 10 Cambridge ... cr | 27 11 Walsall ... fs 12 Oxford we. 13 Dudley ... 14 Middlesborough 45 Exeter — 46. Carlisle ... 47 Oldham ... 18 Cardiff ve 49. Plymouth ... | 9:4 50. Maidstone ... 51 Rochdale ... wn 52. Birkenhead - | 22 53. Bristol ve ce 54. Reading ... ver | 91" 55. Huddersfield 56. Macclesfield . ! 90. 57.|Chester ... 53. Devonport ... vee 59.| Cheltenham i | 19 6o. Hastings ... oun J 61.| West Bromwich ... 62./Chatham ... _— { 1-8 63. Newport (Monmouth) J 64.| Bath we wee IY 65. [Shrewsbury an 66.| Southampton . J 16 67.| Dover i -_ 14 68.| Halifax .., _. 69. Swansea... ve } 13 vo. Merthyr Tydfil ... 0-7 -13 -17 + 7 -48 19 +19 16 15 15 -38 -44 24 +12 -17 -17 - 55 5 33 29 -38 20 30 26 16 11 0 +28 +11 -41 -81 +13 43 23 54 Eight years 1880-87.1 Ashton-under-Lyne ... 24 Tynemouth ... er | 9-3 Chatham .. 2 Colchester Derby Worcester .. Jambridge ... Dudley wo Middlesborough Birkenhead ... Maidstone ... su Newport (Monmouth) Dover .. Lop 20 Plymouth ... ve DUdham ove see 1-9 West Bromwich ... 1 8 Southampton ve Hastings ... 17 Exeter tes Cheltenham Macclesfield Reading ww Devonport ... Chester sor Barrow-in-Furness Oxford _ Darlisle ces Bristol ws Huddersfield Shrewsbury ... Rochdale ... Bath ... . Halifax es . Merthyr Tydfil ... 07 Swansen. . N 06 1-3 11 1-0 ! The towns that have gained improved places in the list (i.e. have gone down) are printed in #alics, those that have relatively lost ground as regards diarrhea (i.¢. risen in the list) in elarendon type. 304 STUDIES IN STATISTICS CHAP. any changes of circumstances in their sanitary districts to which they would attribute the decrease (or increase) in Diarrhea. No fewer than 25 medical officers were kind enough to reply, some of them in great detail, to my queries, and I take this opportunity of once more thanking those gentlemen for their great kindness. The replies, as might have been expected, were very varied ; a number of circumstances were detailed as possible causes of the diminution in Diarrhea mortality; in a few cases the cause was assigned with some confidence, but more usually with hesitation. I have endeavoured to group the answers as far as the facts allowed :— « Qubstitution of water-closets for privies”:—Clesier, Liverpool, Worcester. ““ Abolition of privies” :— Derby, Oxford. « Substitution of dry-ash closets for privies” :—South Shields. « Substitution of the pail-system for privies”:— Askfonunder- Lyne, Rockdale, Tynemouth, Wigan. [N.B. These are four towns in which the improvement as to Diarrhoea was very marked.] « Substitution of ash-tubs for middens” :— Hudders-Jicld. “ Improved drainage, or sewerage ’—DBristol, Carlisle, Derby, Beeler (application of the “smoke-test” in all houses in which it was suspected that a case of illness might possibly be connected with the drains), Huddersfield, Ozford, Rochdale, South Shields, Swansea, Tynemouth. « Severance of connection with the sewers ”:— Derby, Tynemouth. “ Improvement in the removal of nuisances ”:— Ashton under- Lyne, Bristol, Carlisle, Northampton (VY, Ozford. South Shields, Swansea. « Weekly removal of house refuse ”:— Derby. xv CAUSATION OF SUMMER DIARRHEA 305 ‘“ More frequent emptying of ash-pits ”:— Ashton-under-Lyne. “ Emptying ash-pits in June, and then letting them alone during July and August” (to avoid disturbance of the night-soil during hot weather) :— Liverpool. “ Ventilation of house drains, and ventilation and flushing of sewers :—Clester. “ Flushing local drains during the summer”: — Bath (with disinfectants), Huddersfield (Dr. Cameron lays great stress on this), Liverpool. “ Watering the streets with disinfectants ”:— Erefer. “Paving back streets ”:— Wigan. *“ Cleaning and paving courts ”:— Eweter, Worcester. “Cleaning and whitewashing tenement houses” :— Laeter, Worcester. “ Improved houses for working people ":— Ozford (and zonsequent improved habits), Rockdale, Tynemouth. “ Various sanitary operations ”:— Liverpool. “ Purification of the river ”:— Ozford, Wigan. * Abolition of surface wells ”:— Derby. “ Extension or improvement of water supply ":— Bristol, Chester, Derby, Huddersfield, Northampton, Ozford, Tynemouth (said to be still defective). “Better feeding of infants” :—dshton-under-Lyne, Lzeler, South Shields (in these three towns circulars on the feeding and management of infants are given to the mother on the registration of each birth), Rockdale. “ Possibly increased hygienic knowledge” :— Carlisi, South Shields. “The good work done by eréches ”:— Bat/. “Improvement in milk ”— Derby, Tynemouth (inspection). “ Improvement in food ":— Derby, Rochdale. “ Food no longer kept in cellars ”:— Rochdale. 306 STUDIES IN STATISTICS CHAP. ¢ Improved habits and mode of feeding of the people”: — South Shields. “ Checking the consumption of unwholesome fruit, bad fish,” &e. :— Baik. “ Women perhaps somewhat less employed in factories. The people more regular in their habits of living. Have not so much to spend. Have not so many children, and therefore take more care of what they have. Give their infants more milk and less bread and water ”:— Rockdale. From Newcastle the report was that no important change had taken place to explain the diminished mortality. So much for the towns in which Diarrheea mortality has diminished. In only nine towns was the Diarrhoea mortality greater in the second period than in the first; but if we bear in mind that during the time under consideration the mortality in England and Wales fell 14 per cent., we shall see that the rise in the death-rate of these nine towns has perhaps more interest and significance than the fall in fifty-nine. Here, one might be justified in saying, we should expect to find some clue to the cause of the disease. In the following table are given details as to the nine towns, and also as to five others, in which the death-rate was stationary, or nearly so. It will be observed that the increased mortality was in no case absolutely great, although in Dover, Chatham, York, and Walsall it amounted to a considerable percentage increase on their mortality in the first period. It was indeed sufficient to alter notably the position of those towns in the list, so that, e. g., Dover has lost its proud place of fourth in point of health, while York is now amongst the conspicuously unhealthy towns. Whereas formerly Chatham, Dover, and Newport all had rates con- Xv CAUSATION OF SUMMER DIARRHEA 307 DEATH-RATE FROM SUMMER DIARRHEA IN CERTAIN TOWNS. York ... Dover -... Walsall ... Chatham ne Gateshead ce Cardiff ... ie Portsmouth abn Newport (Mon.) Southampton .. Merthyr Tydfil ... West Bromwich Birkenhead cee Stoke-upon-Trent Preston. ... es 1870-79. 3:3 1-4 27 1-8 3-2 2D PR 1-8 [-6 no ~ 25 6:0 1880-87. 0 2:0 3-2 2:3 36 2-8 3:0 2:0 1:8 2 n-Increase. Actual. | Per cent. +21 +43 +19 +28 +13 +12 +7 +11 +13 +07 +06 +05 +o5 +04 +03, 402 +02 +02 0*0 2'0 a 3 1 siderably below that of England and Wales, now the Diarrhoea mortality of each of them exceeds that of the whole country. From Gateshead and Walsall 1 have no replies. York has no changes to report, except in boundaries, and those do not affect my comparisons. The medical officer of health cannot assign any cause for the increase of Diarrheea in Newport, and the same applies to Dover. In Southampton improved drainage and water supply, as well as house to house inspections, are reported ; a “ Destructor ” furnace was erected three years ago, to which all house refuse is removed weekly ; the milk is frequently analysed and often found to be adulterated, but no explanation of the increased mortality is forthcoming. The Chatham medical officer reports the removal of nuisances ; a system of surface drainage; the closure of wells. He says that a very large number of women are 308 employed in clothing factories ; also that when the women go “hopping,” their children are left at home to be fed out of foul-smelling bottles. He adds, however, that he cannot indicate any new circumstance as likely to have caused “the increased mortality. So Dr. Walford suggests that the increased fatality of Diarrheea in Cardiff may probably be explained by the exceptionally rapid increase of population (possibly underestimated) and the excessive birth-rate, which is the highest in the kingdom. The population has outgrown the water supply, so that in the summer-time the sewers are imperfectly flushed, and there is consequent atmospheric pollution. Dr. Mumby reports that in Porfsmouth women are more employed in factories than formerly. He is convinced that in 1886 much fatal Diarrhea resulted from repairing and cleansing the sewers during very hot weather. At the same time he asserts that sanitary defects are now rarely found on inspecting houses in which deaths from Diarrhoea have occurred. STUDIES IN STATISTICS CHAP: The following table is designed to show the incidence of fatal Diarrheea in England and Wales on the two sexes at the several periods of life during 20 years. At first sight it appears that although the disease differs greatly in fatality at different ages, the proportionate incidence on the two sexes is very similar. The figures in the right-hand column, however, prove there are some decided differences. Thus, in the first year of life, six boys die from Diarrheea for every five girls. Again, at the ages from 20 to 45, women suffer decidedly more than men. This is probably to be explained by the fact that women being more at home are more affected by any insanitary conditions of the house. Moreover, they are exposed to infection (if CAUSATION OF SUMMER DIARRHEA 309 any) in nursing the sick. The great excess of male deaths in the first year of life would appear to be connected with the great excess of male deaths from all causes. We saw in the second chapter (p. 6), that for every 4 girls that die in the first year of life from all causes, 5 boys die. DEATH-RATES AT VARIOUS AGES OF MALES AND OF FEMALES PER MILLION LIVING AT EACH AGE. AVERAGE OF THE TWENTY YEARS 1861-80.1 Age fy ) |) 5 20 — 25 — 35 — 45 — , 956 — ’ 65 — ’” 5 — All azes Males. 20,963 5,799 1,049 370 T0Q 71 25 92 167 tan Th 2,8 x - 1.c06 Females. 17,595 5,641 1.075 365 198 78 925 27 47 79 07 160 420 .,259 3,704 _77 ixcess per cent, Males 19 » ’ Females 2 Males 1 Females 1° Males ¢ Females ¢& ~7 re Males LL Males 15 In conclusion, I would merely add that the experience of recent years has confirmed my opinion that Summer Diarrhcea is caused by the pollution of air, water, and food with some product or products of the decomposition of organic matter during very hot weather. ! Calculated from the figures given in the Supplements to the 35th and 45th Annual Reports of the Registrar-General. j10 STUDIES IN STATISTICS CHAP. CHAPTER XVI ON SOME STATISTICAL INDICATIONS OF A RELATIONSHIP BETWEEN SCARLATINA, ERYSIPELAS, PUERPERAL FEVER, AND CERTAIN OTHER DISEASES! ABOUT a year and a half ago? the idea occurred to me that by plotting out by the graphic method the annually fluctuating death-rates from various causes in England and Wales, as given by Dr. Farr, in the Registrar-General's Annual Reports, interesting results might be obtained, more especially as the resulting curves would allow of easy comparison. It must be well known to all present that results which can only be arrived at with much labour, and some uncertainty, by the study of long columns of figures, become at once obvious when the same figures are reduced to diagrams. The object of my investigations was, not so much to ascertain the Huctuations in the fatality of the several diseases, interesting as these may be in themselves, but rather, by the careful comparison of the curves, to see whether any, and if so what, relations subsist between diseases believed to be distinct, with a view to correcting, if need be, the methods of classification now in use. but chiefly to search for etiological clues. [ accordingly traced eighty-nine curves. representing ! Read before the Epidemiological Society of London, April 7th, 1880. 2 72 1n 1878 XVI SCARLATINA AND PUERPERAL FEVER 311 the death-rates per million in England and Wales from as many “ alleged causes”; the figures being obtained from tables following the “letters of Dr. Farr to the Registrar-General” published in the 35th and 41st annual reports. By a simple application of the law of combinations, it will be found that to compare all these eighty-nine curves, two and two together, would involve 8,916 operations. Of these I have as yet actually made only 1,425, leaving still no fewer than 2,491 to be performed. However, as the zymotic diseases have been compared with nearly all the others, there is reason to believe that the most important results have been obtained already, and that the remainder of the work will be less fruitful. I should remark that several of the causes of death have not been examined at all, since the numbers are so very small that they could not, by any possibility, yield trustworthy results. Since 1855, the Registrar-General has given, in the form of a supplementary table, the deaths in England and Wales in each year, classified under very many more heads than are shown in the other tables. To insure the greatest accuracy attainable, I have abandoned my first charts, and am now working from the © supplementary tables,” the labour of calculating the rates being reduced to a minimum by the aid of the arithmometer of M. Thomas—a most useful machine. I soon found it impracticable to draw all the curves to one vertical scale, and finally decided on the method employed by Messrs. Buchan and Mitchell! viz. of finding the mean for each disease, and plotting out the annual death-rate at so much per cent. in excess or defect of this mean. This entails, of course, additional calculations; but the results obtained are strictly comparable, and enable us to note with case such facts as 1 Journal of the Scottish Meteorological Society, 1874, p. 187. STUDIES IN STATISTICS CHAP. that two diseases are affected in the same manner by hot summers, but the one threefold as much as the other: so that it is, perhaps, allowable to infer that, in the former, summer heat is a main cause of the disease; while, in the latter, it is but one of many contributory causes. By smploying in all cases death-rates per thousand living in sach year (or per million, as is more convenient when the death-rate from a single disease is under consideration), all fallacies due to increase of population are avoided, or rather reduced within the limits of error of the estimates »f population. It is my intention this evening to put before you, very shortly, certain of the indications of the above 1,425 comparisons. A variety of circumstances have interfered with my work ; so that I must ask you to consider this as a preliminary paper, which may or may not be followed by others, according on the one hand as you may be disposed to hear more from me on the subject, and on the other as further work may lead to results of importance, or the reverse. At a very early stage the close relationships of the curves of diphtheria and croup caught my attention, and soon after the practical identity of the curves of puerperal fever and erysipelas. More recently I observed that the disease curves may be divided into four groups at least, viz. :—1. The dearrhoea! group, characterized by having the oreatest mortality during the summer quarter, and by being most fatal in years with hot and dry summers. 11. The bronchitic group, characterized by having the greatest mortality during the winter months, and being most fatal in years with very cold winters. mr The scarlatinal group, under consideration to-night. 1v. A provisional group, comprising all the other diseases which do not appear, so far as we know at present, to be greatly DEATH - RATES FROM ENGLAND AND WALES. ERYSIPELAS, SCARLATINA AND CERTAIN OTHER DISEASES WITH RAIN AT GREENWICH. ALL AGES AND BOTH SEXES. RA~ AA RT AR 'A0 RO AL 'A2°A3 64: '65 'R6 "67 68 69170 71 "72 73 7a '75 7] 77 IR TQ R() Mean leath- Rates ar million living RAIN AT GREENWICH PUERPERAL FEVER ERYSIPELAS 1-63 - 92 IHEUMATISM OF HEART — SCARLATIN _ARYNGITIS 01 —. ~57 ATTN SY io 2A) 1 1 = — DIPHTHERIA =| GA = CYNANCHE MALIGNA 3} opunny 17 '% 8 5 5 8555615758 591 sol *61T 626 seat ssl eel ert 6st eat rol Tri 72 731 14 15 Pre rr brs tra tao STANENRN'Q Ft TART ‘he dotted lines indicate means for the twenty four years 1855-78. : gree es 'he curves express the fluctuation per cent above or below the Mean. Each division of the vertical scale corresponding to 10 he figuras at the side give the gctua! values of the Means of the Death-Rates per million persons estimated tn be liviné XVI SCARLATINA AND PUERPERAL FEVER 313 influenced, or at all events mainly regulated, by any meteorological conditions. Among these may be mentioned small-pox, measles, whooping-cough, phthisis. and (probably) enteric fever. Remember, that this classification is provisional, and is only drawn up to facilitate the examination of the deathreturns of England and Wales. Thus, for instance, 1 know full well that whooping-cough is most fatal in the early months of the year, and that if whooping-cough breaks out during cold weather, it is more likely to be fatal than under the reverse conditions; but what I do assert is, that taking a long term of years, and the whole kingdom into account, the mortality from whooping-cough does not vary directly with the coldness of the winters, as that of bronchitis or pneumonia does. The larger diagram (Plate XIX.) accompanying this paper shows certain curves constructed in the manner above described, and grouped together in order to bring out as clearly as may be the points to which I wish to draw your attention. The vertical columns divide the curves into twenty-four portions, corresponding to the twenty-four years, 1855-78." The horizontal portions of the coloured lines mark the distances above or below the mean of the death-rate in question in each year—on one scale throughout, each small division corresponding to 10 per cent. The horizontal lines are connected by sloping lines, to give continuity to the curve, and. to facilitate comparisons. The mean death-rate from each disease per million persons living in England and Wales is indicated by a ! In revising this paper I have carried forward the curves (without altering the means) to 1880. As some changes in classification were introduced by the Registrar-General in 1881, I have thought it well to stop the diagram at that point. The figures which form the basis of the diagram are given in the fable at the end of this chapter (pp. 832-34). 314 dotted line of corresponding colour, and by the figure at the right hand, while the names of the diseases are given on the left hand in corresponding colours, viz. pysmia, puerperal fever, erysipelas, rheumatism, with disease of heart or pericardinm,” scarlatina, laryngitis, quinsy, croup, liphtheria, and cynanche maligna. It will be scen that the curve of scarlatina presents four marked elevations, or periods of epidemic fatality, separated by three equally well marked depressions, or periods of minimum fatality. In the most fatal year 71863), the scarlatina death-rate exceeded the mean by 73 per cent.; whereas, in the least fatal year (1861), it tell to 47 per cent. below the mean. These four periods of ‘epidemic fatality are, to some sreater or less extent, apparent in all the ten curves, the Auctuations varying more in degree than in direction. [n the curves of puerperal fever, erysipelas, pywmnia, and “rheumatism accompanied with heart affection,” the fourth or last epidemic period, 1874-5, is the most marked ; whereas, in the curves of cynanche maligna, diphtheria, and quinsy, the first epidemic period is far the most prominent. The second epidemic period is most distinct in scarlatina, croup, and diphtheria; the third in scarlatina and “rheumatism with heart affection,” being but faintly indicated in most of the other curves. Cynanche maligna and quinsy would not be sufficiently important causes of death (averaging only 17 and 14 per 1,000,000 respectively), were it not for their very interesting relations to diphtheria. Previously to 1855, but few deaths from diphtheria were registered in England and Wales, and they were not kept separate in Dr. Farr's analysis of the Registrar-General's returns. STUDIES IN STATISTICS CHAP. XVI SCARLATINA AND PUERPERAL FEVER 315 In 1856 the total number of deaths registered in England and Wales, as due to diphtheria, increased from 186 to 229—not by any means a striking rise—but in the same year the deaths attributed to quinsy rose from 3871 to 418 ; whereas those assigned to cynanche maligna rose from 199 to 874, a very notable increase. In 1857 the diphtheria deaths rose further, from 229 to 310; those attributed to quinsy from 413 to 485; but those attributed to cynanche maligna increased suddenly from 374 bo 1,273, such a leap up as must have alarmed the few persons who were cognizant of the fact. This may be taken as probably the first indication of the coming outburst of what was long spoken of as ‘“the new disease, diphtheria,” though probably, in reality, but an old foe under a new name. In 1858, practitioners began to be more alive to the real facts of the case; and although the deaths attributed to quinsy and cynanche maligna showed still further increments, viz. from 485 to 623 in one case, and from 1,273 to 1,770 in the other, the deaths more correctly assigned to diphtheria rose with alarming suddenness from 310 to 4,836. By 1859 the great stir created by letters in the press had drawn the attention of all observers to “the new disease,” and, accordingly, we see that in that year the curves of quinsy and cynanche naligna began to fall more rapidly even than they rose; and, with the exception of the period of the recrudescence of the diphtheria epidemic in 1863, have continued to fall, till those diseases now have a very insignificant place in the list, and accounted for only 216 and 179 deaths respectively, on the average of the five vears 1874-78.1 ! Quinsy caused, in 1878-79-80 respectively, 294, 299, and 353 deaths 3 after 1880 it was classed with sore throat, and the two together reached » maximum in 1884. corresponding to a diphtheria maximum. 316 STUDIES IN STATISTICS . CHAP, But, meanwhile, diphtheria rose, in 1859, from 4,836 to 9,587. The epidemic may be said to have lasted, with diminished but varying fatality, till 1865. During the five years 1874-78, the deaths averaged 3,092, or about one-sixth of the number slain by scarlatina.l It must be borne in mind, however, that the disease is far more prevalent in the United States, in Germany, and, more sspecially just now,” in Russia, than it is with us. The croup curve seems to have been affected in like manner, but in the earlier years to a less extent, by the epidemic of diphtheria. Since 1861 the curves of croup and of diphtheria have been so very similar, as to suggest that at the very least there is great confusion in the liagnosis of the two diseases, and even to be a weighty argument in favour of those pathologists who maintain their identity. Pyzemia has no place in the Registrar-General’s reports prior to 1862, and there can be little doubt that the steady rise from that year must be attributed to changes of nomenclature ; for if not, what is to be said for our boasted improvements in operative surgery and hospital hygiene ? In spite of this steady rise, its resemblance to the curves of erysipelas and puerperal fever is strikingly suggestive. You have, probably, already observed that the curves of erysipelas and puerperal fever are practically identical, with the exception that the increased fatality of puerperal fever in 1874-5, was about twice as great as that of 1 From 1878 to 1886 the death-rate from scarlatina fell steadily from 753 to 215 per million, and during the five years ending with 1888 averaged only 269. On the other and there has been a recrudescence of diphtheria, the rate (including cynanche maligna) having risen from 109 (in 1880) to 185 (in 1884); the average in the five years 1884-88 has Leen 164, or three-fifths of the scarlatinal average for the same years. 2 7 e In 1880. v1 SCARLATINA AND PUERPERAL FEVER 317 erysipelas. I confess that I find it difficult to avoid the conclusion that they are both due to one poison. Though the puerperal fever curve has great points of resemblance to the scarlatina curve, it is far more like that of erysipelas; it is also extremely like that of “rheumatism, with disease of heart or pericardium.” So that we may infer some close relationship to rheumatic fever—a point that I must reserve for further investigation, hut may merely remark’ that the erysipelas curve is still more like the rheumatism curve. The small diagram on the next page attempts to represent the mutual relationships of the various diseases in the large diagram-—the distances between the circles being supposed proportional to the closeness of relationship. Dr. Thos. C. Minor, of Cincinnati, U.S.A., writing, in 1874, on “ Erysipelas and Child-bed Fever,” * said (according to an abstract in the Practitioner), in reference to the deaths from these two diseases in the United States in 1870: “1. Erysipelas and child-bed fever seemed to prevail together throughout all the States. 2. Any marked increase, in any one locality, of one disease, was apparently accompanied Dy a corresponding increase of the other. 3. Where histories of past epidemics, of either disease, were obtainable from any of the States, the apparent connection of the two diseases was noticed by physicians at the time of such epidemics, and remarked on. 4. This relationship indicates that there is an intimate connection existing between child-bed fever and erysipelas, and justifies the inference that in any place where erysipelas is found, there will be found child-bed fever.” Again, referring to an epidemic of child-bed fever in Cincinnati, Dr. Minor says— Lp and Child-bed Fever, Cincinnati, 1874. Robert Clarke nd Co. 318 STUDIES IN STATISTICS CHAP. “1. The two diseases, child-bed fever and erysipelas, prevailed at the same time, in the same localities. DIAGRAM SHOWING SUGGESTED RELATIONSHIP OF THE SEVERAL DISEASES. [Pome | Pyamia ) i | Scarlatina Erysipelas ) Rheumatism of Heart f Cynanche Maligna ! Diphtheria ( Laryngitis Croup 1 I \ Quinsy T 3 » 2. Where an isolated death from child-bed fever was noted outside of the infected districts, a corresponding death LONDON WEEKLY DEATHS FROM SCARLATINA & CERTAIN OTHER DISEASES Excess or Defect per cent. compared with the Mean. AVERAGE OF 30 YEARS, 1845-74. v \ MAP « E auc.! sep. locT. 'NOV. DEC. MEAN AVERAGE Mo of weekly Deaths = SCARLATINA ERYSIPELAS PUERPERAL FEVER CER 10: da PIE 19.6 7 4 DAE a Eh TR Lm Ay ~~ ad wi dc Bebe BE DIPHTHERIA 1861 = 1874 CROUP 10 + & O-| DW uw JAN. JUNE [JuLyi auc. ! St NOV - DEC, The numbers have been adjusted by "Bloxam's Method previous to the construction of the curves. Scale. Fach small division of the vertical scale corresponds to ten per cent of excess or defect from the Mean XVI SCARIATINA AND PUERPERAL FEVER 319 from erysipelas was noted in the same locality. = This was almost invariably the case. 3. Infants die of erysipelas shortly after or before their mothers die of child-bed fever. 4. A few physicians attending child-bed fever cases and erysipelas cases at the same time, as exhibited by the death-register, were the most unfortunate in their practice. 5. Physicians having large obstetric practices, but who were known to be believers in the close connection of child-bed fever and erysipelas, returned few death certificates from either causes.” Again, on examining the alleged connection of typhus fever and scarlatina with child-bed fever. Dr. Minor found that— “1. Epidemic typhus is not always associated with an outbreak of epidemic child-bed fever. 2. Epidemic scarlet fever is very seldom associated with an outbreak of epidemic child-bed fever, or vice versd. 8. Epidemic erysipelas is invariably associated with an outbreak of child-bed fever, or wice versd.” I will now examine the subject from a different point of view. The smaller diagram (Plate XX.) is reproduced with modifications of detail from an admirable paper by Messrs. Buchan and Mitchell, in the Journal of the Scottish Meteorological Society, by the kind permission of the authors. These gentlemen analysed the weekly mortality returns for London for the thirty years 1845-74, and constructed curves showing the average numbers of deaths from a number of causes in each successive week of the year. The curves are on such a scale that in every case the same percentage deviation from the mean is represented by the same distance above or below the dotted line indicating the mean average weekly number of deaths. A little reflection will show that the increase of population does not greatly affect the result, except that great epidemics 320 STUDIES IN STATISTICS CHAP. in recent years must disturb the curve more than epidemics of like relative extent in the earlier years. Diphtheria was not separated from scarlatina in the London returns until 1859, but as there is reason to suppose that but comparatively few deaths from this cause occurred in London in 1855-58, this is of little consequence. The coloured lines in this diagram correspond to one year—that is, one average year, which is divided by the vertical black lines into 52 wecks. The months are indicated by figures at the top and bottom. For the sake of distinctness the areas representing the excess of mortality above the mean are coloured in. The coloured figures to the right hand indicate the mean average number of deaths from each cause in every week in London, and the distance of the continuous coloured line above or below the mean indicates the percentage deviation from that number, one small division of the vertical scale corresponding to ten per cent. It must be premised that laryngitis” in this diagram comprises laryngismus stridulus and cedema glottidis (the latter of small account), which have been excluded in the construction of the larger chart; also, “rheumatism” includes all cases returned under that heading—presumably many of them chronic joint affections, whereas these cases, along with many others, are excluded in the larger diagram. “Rheumatism, with disease of heart or pericardium,” may be fairly regarded as comprising few cases other than acute rheumatism, though of course not including by any means all the deaths due to rheumatic fever. The colours on the two charts otherwise correspond. Now, in certain respects, all the eight curves resemble one another. All the curves are above the mean from the last week in November to the first week in January inclusive, and they are all below the mean from the first week in June to the second week in August inclusive. XVI SCARLATINA AND PUERPERAL FEVER 32I There are grounds for believing that if the deaths from laryngismus stridulus could be separated from laryngitis deaths the curve would be somewhat modified, so as to approach more nearly that of croup. As it is, this curve resembles more closely that of bronchitis than my examination of the annually fluctuating death-rates from these two causes would have led me to expect. Messrs. Buchan and Mitchell state that the sudden depression of the curve in January is almost constant from year to year. May not this be caused by the deaths from laryngismus stridulus commencing to rise to their maximum at this point, as the curves of epilepsy and convulsions do ?* Omitting for a moment scarlatina from consideration, and dismissing laryngitis as unsatisfactory, we have remaining six curves bearing a very striking resemblance one to another, which is more especially marked in the case of erysipelas and puerperal fever. The scarlatina curve, if shifted about five weeks further on, would be not unlike the other curves. Tt is conceivable that similar causes may tend to the production of scarlatina first, and later on to the other affections—somewhat as a change of type is frequently observed during the course of an epidemie, notably in cholera. It should be added, that all of these curves have been modified from those published in the memoir referred to by the use of what is known as “ Bloxam’s method,” whereby, instead of taking the average number of deaths in each ! Dr. B. A. Whitelegge has called my attention to the fact that the croup curve has two maxima, one in March, the other in December. He believes that this is due to confusion in diagnosis, the spring maximum corresponding to cases which’ should have been returned as laryngitis, whereas the winter maximum corresponds to cases that should more correctly have been attributed to diphtheria. In short, he believes that shere is no such disease as eroup, and I am very much disposed to agree with him. 322 STUDIES IN STATISTICS CHAP. particular week, the mean of the average of that week, the week preceding, and the week following is taken. The method is very useful when the number of deaths or the number of years is too small to give a smooth curve; inequalities produced in accordance with the law of probabilities are removed, and a close approximation is obtained to the curve that would result from the plotting out of a large number of deaths for a longer term of years. You will naturally ask, “Is there any peculiarity of scason or otherwise which determines, or has a share in determining, the fatality of this group of diseases in certain years?” As the curves as a rule rise in the autumn and remain high throughout the winter, one is naturally led to suspect that a very hot summer, or a very hard winter, is the determining cause. So far as I have been able to investigate the point, it does not appear that the heat of summer or the cold of winter varies either in a direct or inverse relation to the fatality of these diseascs. But when the amount of rain is examined from this point of view, it appears that there is some kind of inverse relationship between the two. The same result is obtained whether the amount of the fall or the number of days on which it falls is considered, the connection being somewhat closer in the latter case. I have calculated the percentage variation from the mean of both the rainfall and the number of rainy days, and added the two together. The result is shown at the top of the larger diagram, by the blue area; the ordinates being measured from the top downwards, so that the curve formed by the line limiting the blue area is placed upside down, and the inverse relation appears as a direct relation, and is thus much more easily realized. (Dr. Buchanan’s method.) This relationship is nearly as close as that which sub- XV1 SCARLATINA AND PUERPERAL FEVER 323 sists between diarrhea and hot summers on the one hand, or bronchitis and cold winters on the other; and it is strikingly suggestive. How the rainfall influences the fatality of these diseases (we do not know how far it may be related to their prevalence) is another question. Assuming the connection to be real, it seems most natural to suppose that the frequent fall of rain purifies the air, and removes from it the particles of contagious matter; that it acts, to some extent, as the process of inunction in scarlatina is supposed to act, by making the poison less diffusible. The average distribution of the rainfall throu ghout the year does not throw any light upon the question ; most rain falls in the five months June—Oectober, during which period the scarlatina curve is continually rising. There are many indications, in recent medical writings, of a growing opinion that the process of analysis has been carried too far in our classification of disease ; and there an be no doubt that, more especially considered from an wtiological standpoint, such a result is to be deprecated. [f great care be not taken, such confusion may be created as “ species making ” has caused in zoology and botany. Increased accuracy of diagnosis is all important in this connection ; without a clear diagnosis the death certificate cannot be filled up in a lucid manner, and consequently the death cannot be placed in the appropriate pigeon-hole by the abstractor. If we get our sub-groups correct it is 2asy by lumping them together to diminish their number. or to alter the larger groups and classes. It 1s, moreover, important not to lose sight of ‘the possible influence of the individual on the disease that attacks him. There may in fact be a question of soil ; we know indeed already that the germs of some diseases grow with facility in the tissues of certain patients, whereas they find difficulty in effecting a lodement at all oh STUDIES IN STATISTICS CHAP, in others. May it not be that the process of development differs among the individuals that succumb to attack? Or even granted that the germ breeds true,” may not differences in the constitution of the host lead to greatly different symptoms—even to what we now call a different disease ? Further observations on this point are greatly needed. APPENDIX : August 1890. THE returns of ten more years are now available. As regards rainfall, one of these years (1879) was exceptionally wet ; three (1880, 1882, and 1888) were decidedly wet; two (1884 and 1887) were dry, whereas the other four were about the average. In the year 1881 extensive alterations in the classifica tion of the alleged causes of death were introduced by Dr. Ogle. Although there can be no manner of doubt that these changes were in almost every case a great improvement, yet, in a series of statistics, changes of any sort, howbeit imperative, are of necessity an evil per se. The particular changes referred to so far affect the causes of death now under consideration that the curves of “rheumatism of the heart,” “quinsy,” and “cynanche maligna” cannot be carried beyond 1880. As regards the last there is no reason to regret its absorption into diphtheria ; quinsy, again, is of little moment, but it is a matter of regret that there should be a break in the statistics of “rheumatism of the heart,” since they are important, and promised interesting results. It is a good set-off against this inconvenience that the new group “rheumatic fever, rheumatism of heart.” is better than the old. The pyemaia curve has in recent years exhibited no V1 SCARLATINA AND PUERPERAL FEVER distinct relationship to any of the others; it has been generally higher than the average of 1855-78, notably in 1879-80-81. The puerperal fever curve was above the average in nine out of the ten years, and the disease was exceptionally fatal in 1882 and 1883. In order to compare the years following 1880 with that year and those preceding it, allowance must be made for the following circumstance :— In 1881, with a view to making the returns more complete, the Registrar-General, in certain cases in which the certificates of death were not satisfactorily filled up, wrote to the medical men who had signed them for further particulars. Thus a letter was sent in every case (among others) in which a woman of child-bearing age was stated to have died of either peritonitis, septicemia, or pyemia.! As a result of these special inquiries, which were continued in subsequent years, numerous deaths have been added to the heading puerperal fever, so that in 1881 and 1882 the alleged mortality from that cause was increased by 13 per cent., and in 1884 by 7 per cent. If, however, full allowance be made for this, the years 1881-85 will remain distinguished by a puerperal fever mortality only exceeded by that of 1874 and 1875. This increased mortality is quite contrary to what one might have expected; but it is probably to some extent due to increased accuracy in the returns, as medical men gradually learn the scientific importance of the certificate of death. The mortality from erysipelas has, unlike that of puerperal fever, been somewhat low in the last ten years, having reached the average only in 1882; nevertheless a close correspondence in the directions of the fluctuations of the two curves has been maintained. - Forty-fourth Report, pp. XX, XXL. CHAP. The scarlet fever death-rate fell in 1879, and in every succeeding year till 1886, so that it then amounted to but 25 per cent. of the average. There is good reason to believe that the increased use of fever hospitals for the isolation of the sick® has materially contributed to this result. In consequence of this continued fall there is little correspondence to be observed between this and other curves, including that of rain. The curve of laryngitts shows a fair amount of correspondence with those of croup and diphtheria, which again have continued to run similar courses. Diphtheria was more fatal in 1884 than it had been for 19 years; it has indeed been exceptionally fatal since 1881, so that it now approaches scarlet fever in importance. It should be mentioned that from 1881 onwards the recorded mortality has been raised by about five per cent. by the inclusion of cynanche maligna deaths under the same heading.? It cannot be said that the close inverse rclationship between several of these diseases and the rainfall observed in the earlier period has been maintained, but in the very wet year 1879, erysipelas, puerperal fever, and rheumatism of the heart were all at a minimum, whereas in the dry year 1884 diphtheria was at a maximum and puerperal fever very high. The exact figures from 1855 to 1888 will be found in a table at the end of the chapter. Dr. Astley Gresswell® has independently observed the inverse relationship between rainfall and scarlet STUDIES IN STATISTICS 1 See below, p. 327. 2 This applies to the rates shown in the table on p. 334, but not to those published by the R.gistrar-General. In the latter, eynanche maligna is throughout included with diphtheria. 8 A Contribution to the Natural History of Scarlatine, by Dr. D Astley Gresswell. MLA.. M.D.. Clarendon Press, Oxford. 1890, p. 191. Xv SCARLATINA AND PUERPERAL FEVER 327 fever, and he calls attention to the fact that the effect seems to persist, that is, that a second wet season following a depression of the scarlet fever curve depresses it still further, and he suggests that possibly the long-continued heavy (summer) rainfall during the period 1878-82 “may have had to do with the want of increase of scarlatinal mortality in the period 1888-87.” He adds—* It may be cemarked further that the mortality did rise in the last of these five years, in which the total rainfall was below the average, and again that it fell in 1888, when the rainfall was again largely increased.” Dr. Gresswell considers that deficient sunshine, an ordinary accompaniment of heavy rainfall, may have something to do with the matter ; and he adds some very suggestive remarks® as to the way in which meteorological conditions may affect human scarlatina through its influence on some lower animal, such as the cow, which may serve as an intermediate “host” to the organism which we may suppose to be the cause of the disease. In 1876 the late Dr. Matthews Duncan read an important paper On the alleged occasional epidemic prevalence of puerperal pyemio or puerperal fever, and of erysipelas,® which I unfortunately did not see till after my paper appeared. In it Dr. Matthews Duncan set himself to prove from the London statistics that erysipelas and puerperal fever are not epidemic diseases, or at all events that they have never appeared in the epidemic form in London since the commencement of the Registrar-General’s records. It has always seemed to me that the very word “epidemic” is so vague that discussions as to the epidemic or non-epidemic character of diseases are apt to be as fruitless as many of the speculations of metaphysicians. The increased precision of medical knowledge naturally L Op. cit. pp. 195-197. 2 Bdinluroh Medical Journal. March 1876. 328 STUDIES IN STATISTICS CHAP. thrusts into the background theories as to mysterious influences in the air, but brings more and more into prominence facts tending to prove the propagation of disease from patient to patient. The really important matter is not whether epidemics of puerperal fever occur, but whether the disease is propagated from case to case, and if so, what conditions favour its propagation? No one would, I suppose, deny that puerperal fever is a “septic” disease, 7. e. that it is connected in some way with processes of putrefaction or fermentation, and few would deny that the poison or ferment may be conveyed from one case to another. In regard to the first of these propositions [ would refer to Dr. Boxall’s recently published paper. Personally I much question whether any disease is ever really “epidemic” in the sense given to those words in Dr, Matthews Duncan’s paper (quoting Dr. Priestley), 7. e. “dependent on general atmospheric or other influences which pervade a community, as distinctive from the miasms of unhealthy hospitals, or the propagation of the disease from one patient to another by inoculation.” The communicable diseases are (certainly in the enormous majority of cases) spread by infection, mediate or direct, from case to case. Atmospheric conditions, doubtless in many instances, favour or tend to oppose such spread. As regards its relationship to scarlet fever, I am at one with Dr. Matthews Duncan, but cannot at all agree with his conclusion that puerperal fever is not very closely related to erysipelas. In the first place I find two flaws in his argument. In spite of what he says, his great diagram is deceptive owing to his comparison in a diagram on the same vertical scale of a common ailment like scarlet fever, with a comparatively uncommon ailment like erysipelas 1 Fever tn Child-bed, by Robert Boxall, M.B., B.C., Cantab, M.R.C.P. XVI SCARLATINA AND PUERPERAL FEVER 329 or puerperal fever. That the force of mere numbers greatly deceives the eye may be seen by plotting out the puerperal fever deaths in the eventful (I will not say “ epidemic”) year 1874 on a scale four times larger. The other point is this: Scarlet fever and puerperal fever differ entirely in their pabulum—the one attacks mainly children, who are a large proportion of the community, and by constantly moving about, and mixing in schools and other places, afford great facilities for the dissemination of any communicable disease that may appear among them ; the other, ex hypothesi, attacks only puerperal women, a very small fraction of the community, who are by the very nature of their case isolated from one another to a remarkable extent. Hence one would not expect a puerperal disease, however infectious, to occur in such extensive outbreaks as are common in the case of scarlet fever. In the second place, I entirely deny that there can be said to be two mean mortalities from scarlatina, one including, one excluding epidemics. It is very singular that Dr. Matthews Duncan should have lighted upon rheumatism as a typical non-epidemic disease ; he states that he selected it on account of the similarity of its numbers to erysipelas and puerperal fever. Now, in my case, after comparing together the curves of scores of diseases, I was surprised to find that the curve of rheumatism bore a curious resemblance to those of erysipelas and puerperal fever, a resemblance that so far as I know is not equally marked in the case of any other disease. The fact that the curve of rheumatism of the heart” is still more like those under discussion suggests strongly that if we could obtain the true curve of “rheumatic fever” or “acute rheumatism,” we should find a still stronger resemblance. 1 Tt should be explained that until 1880 the Registrar-General’s STUDIES IN STATISTICS CHAP. To pursue more into detail the apparent relationship between these three diseases I have, as the result of no small labour, constructed a diagram (Plate XXI.), showing for each of the eleven Registration Divisions of England and Wales the deviation from the mean per cent. of the death-rates from erysipelas, puerperal fever, and rheumatism during the thirty-four years 1855-88. Unfortunately the data for “rheumatism of the heart” are not available for Registration Divisions except for a few years. At the bottom of the diagram are given the corresponding curves for the whole country. In the following table (p- 381) will be found the population in millions, and the average death-rate from each of these diseases in each Registration Division ; also the ratio of the maximum to the minimum death-rate, which Dr. Matthews Duncan takes as a measure of the epidemic character of a disease. The diagram (Plate XXI.) affords a good illustration of the value of large numbers in statistical inquiries. The curves for England and Wales exhibit smaller fluctuations than those for sections of the country, and the correspondences between them are in nearly all cases much closer. Nevertheless, if the diagram be looked at as a whole, or analysed in numerical detail, it will be found that the correspondences in nearly every case are far closer than mere chance coincidence would account for. The great outbreak (if the word epidemic is inadmissible) of both erysipelas and puerperal fever in 1874-5 is very marked throughout, but more especially in the West Midland, Northern, and Welsh divisions, The curves of erysipelas bear a closer resemblance to Returns gave “rheumatism” as a cause of death, but also in certain tables subdivided this into “rheumatism of the heart” and other “rheumatism.” In subsequent reports “ rheumatic fever and rheumatism of heart,” and “rheumatism ” have been shown separately. REGISTRATION DIVISIONS OF ENGLAND AND WALES. DEATH-RATES PER MILLION LIVING, rroM ERYSIPELAS, PUERPERAL FEVER & RHEUMATIC) SHOWING THE PERCENTAGE DEVIATION FROM THE MEAN OF EACH DIVISION. 1855-1888. 1855 57 59 61 63 66 67 69 71 73 75 - L To a r 50 = 3 ar "8 I. ‘85 ‘87 — 100 + A. [1 _ = 111 5 — Le) — - VV - VI Vil Vil IX AN X] 1 E&W. 1856 '58 60 '62 64 66 '68 70 72 74 78 78 '80 ‘82 ‘84 '86 ‘88 | London. Stansord’s Geog’ Estaly J 3 = : = = Se XVI SCARLATINA AND PUERPERAL FEVER 331 REGISTRATION DIVISIONS of ENGLAND and Wares. POPULATION in MiLrioNs and AVERAGE DEATH-RATES per MILLION from ERYSIPELAS, PUERPERAL FEVER, and BHEUMATISM. Also Ratio of Maximum to MINIMUM D EATH-RATE wn each case. Thirty-four years 1855-88. Registration Division. gi “w, 22 23 Ya J} London ... _-II South-Eastern ...0 [IT South Midland we [V Eastern ... sword V South-Western ...| VI West Midland ... VII North Midland ... VIIT North-Western ... IX Yorkshire... o. ¥ Northern ... oe {Monmouth & Wales 33 2'2 1'4 12 I'9 2 1°4 34 25 14 1'4 \ England & Wales | e28 Average death-rate per 1,000,000. Byape- Pyespe Riga: Er [Buen | py 112 7b 134 21 19 88 54 116 26 18 92 58 115 36 23 86 55 109 3g 28 72 48 113 34 22 84 66 118 39 Ig 81 62 104 55 270 94 , 76 122 5+ 0 21 84 . 84 92 33 i: 28 57 | 78 198 y3¢ 46 23 87 | 60 | 121 [261371 19 Ratio of Maximumn to Minimum. those of puerperal fever than either does to rheumatism, but ““ rheumatism” is too vague a term for our purpose, and another ten years must pass before it will be possible to come to any sound conclusion as to “rheumatic fever.” As regards erysipelas and puerperal fever it must, I think, be admitted that a relationship which holds good in the several parts of the country, as well as in the entire area, cannot be fortuitous. [ would ask the reader to compare the curves of erysipelas and puerperal fever in England and Wales with those of bronchitis and pneumonia on the one hand, and those of measles and whooping-cough on the other (see Plate XXIIL p. 371). He will observe that though the puerperal fever curve docs not usually fluctuate as much as those of the latter discases, it does fluctuate much more than the former. I have no wish to dogmatize, but prefer to leave the facts for others to interpret. 332 STUDIES IN STATISTICS CHAP. ENGLAND AND WALES. DEATHS and DEATH-RATES from ceriain diseases ; also the percentage fluctuation of the DEATH-RATE from the MEAN of 24 years 1855—1878.1 Pyemia,. Puerperal Fever, | oi | Rheumatism of Erysipclas. the Heart. r £ 5 = 518p . flog 28 Pd 4 sgl FA 3 sl = HS SR Brat ¢ E08 AE RE 8 EEE 2 =| 8 |SE|38| § SE 8+ Dod a Q =| 8. a 2 Sey = © Sof Be S| B 28° aga SE|&° 3 1855 i856 1857 1858 (859 1860 1861 1862 1863 1864 (865 1866 1867 :868 .869 1870 i871 i872 i873 1874 £875 1876 877 1378 0791 51121021 11213%] 740 067 39/1998 105 a): 381 836 89/1478 77| 8° 198 068! 88/2000 'to3 11% 567 238 011924 ¢8[107 784 987 jo! 80/1662) 341 9% 753 386 14 0 1535 76 83 361 21 010 46 13/1520! 75| 8¢ 704 47 1155] 561 89/1906 yz id 370 39114841 71 1138/2091 -0o 10: 101 0 13331 63 01/1953 gz id 95] i8:1197 36| 89(1665 78! 8 346 701066 is) 78/1445 67 T¢ 793 41196 34] 86/1942 35) 3 071 08 1181 52 851853! 33] 91 160 (a1 1492 66 10512120" 92 103 230 101° (464 64 102{2216 97 LOG .203 98 [400] Gr 97/1771 77: 84 049 10711740) 0118/2027 87 951034 810 1312093358 142 155 1363 fa 2505 [105 1673017 26 138 1817 S046 ne R397 ~% 05 1591 ii gg 94/2077 93' 1504 415 571 91/1870 8211426 39 36 26 A C 75 5. > 1¢ “ 69 159 202 249 273 246! 337 355 or 375 366! LOS 374 345 749 - 84 | 5TT . il ¥ ’ ’ tl 04 - 3 8 10 IT 13 [x 15 16 18 16 16 17 20 £7 { f 0d 15 19 «14 99 « 96 5. 125 76 164 66 142 61 133 57 124 02110011017] 46/100 23 Lean ofl3507 169 100/ 1376 | 63 2100/1994. 1879 1880 881 1882 883 R84 QQ, 3431 45 44. 25 1561464 154 |1659 "| 2287 719564 a | 58 64 nn 921824 1052014 1402314 5512435 1192166 2145 496 ra 72! 78 Rey 78 85 a7 1356] 53,116 1902| 74160 2429) 93] 2590| 98| & 2357) 88 2 "27140100 B® ¥12912 rof| g = 12530 o1| 4 RN “A ; aC YA Hh ~a 092:2697 J4 2 Mean of 17 years. ! The death-rates in this Table from 1855 to 1878 were calculated sefore the census of 1881 was taken: the rates given for the years XVI SCARLATINA AND PUERPERAL FEVER 333 1855 856 857 1858 1859 (860 1861 1862 1863 18604 865 1866 1867 (868 (869 “870 1871 1872 873 374 873 1876 (877 1848 Mean of 24 years 1879 1880 1881 1882 1883 884 885 1886 (887 88 Scarlet Fever. - Quinsy. ' @< Q ts DQ " Lo B= id ow Ss 8 . Lee . ‘ +5 = Se o gS wg 4 S28 Bel 6» B.E| Pey £2 | £4 a la) oo = Hele al = Hr 2 ow ~ Teed = eZ dg Er SZ 8€l 8 a2 88 - 947238 2 ca gf T EL EEL E13 gE EHS A 3 So | 2% ~ Sul BY Owl BY 3, Bu 3 c S 4 a 88x a gle a glam ALL Croup. 6,836) 894 105 3,463] 707, 83 2,558] 6G52| 76 23,600!1,212|142 ‘9,221 976/114 9,681 486 57 9,077 451] 53 14.834] 728! 85 30,475 (1,478 173 29,700 (1,422 167 L7,700| 837] 9¢ 11,685 a 54 2,300| 567| 66 21,912] 998 117 27,641 11,244 146. 39,543 1,446 169. 18,567) 8x35! 96 11,922 517| 61 13,144) 563: 66 24,922 1,054 123 20,469] 855 100 16,893 697 | 82. 14,456| 589’ 69 18,842] 758 B9 915 1,068 1,171 1,202 1,112 952 1,023 ,,203 1,229 1,252 1,046 960 921 9781 1,149! [,142| 57 [L149] 50 1,061 46 1,254 | 54 1,397] 59 1,494" 62 1,224 3 70 40 1.014! 41 93 371 108 113 117 481 119,623 ‘09 426 92)310. 981342" 1141323 15 334 ‘15 378 95 319 36.271 83)201 361208 N78 98 134 97} 144 7 881110 R5 103' 158 ~a 114 173 3 120,226 9 97 156 © ¢% 76,3 ol 70 781294! 17 87 1,419 2353 3,207 273 3,279 274 6,220 319 5,636 286 4,380 220 4,397 219 5,667 278 6,957 337 B,777 325 5,921'280 5,168 1241 4,387 | 202 4,491 | 205 4,478 202 4,302 | 191 4,116 [181 3,640 158" 4,282 183" 5,010 | 212 4,542 | 190 4,204 173 3,910|159 4,061 163 103 119 120 139 125 a6 96 1-1 1.49 <0 122 {0A T) 7, 69 v1 18,435] 854 100 1,120] 52 100}2841 14 [100 £ 894] 220[100 7,613! on 4,975 3,739 2,649 0,863 5,355 3,986 7.859 5.378 694 675 548 520 472 399 231 215 27% gra 81 79 64 51 55 47 7 FE ar 26 L077) (,022] ,160| 1,444 499, 650 [,597 L440: 805) 1.590 43 40 45 35 39 82 77 8¢ 105 "08 7% 19 a9 g 2 2 12 | 87 353] 14 101 3,574 140 3,571: 138 2594 138 6091174 091 171 1,748 175 4,235 154 3,685 |132 3,979 141 36321 4 62 Bo on = So 24 5 Ta so 5 = £2 5 Soy r 50 54 [872-78 are consequently a trifle too high, but the error being under one per cent. it was not thought necessary to incur the expense of re-emgraving the diagram. The rates from 1879 are based on the revised estimates of population. STUDIES IN STATISTICS CHAP, 1855 1856 :857 858 i859 1860 861 862 1863 i864 1865 “866 1867 1868 1869 i870 1871 1872 1873 1874 1875 1876 1877 1848 Mean of 24 vears 1879 1880 1881 1882 1883 :884 (885 1886 1887 QR] Diphtheria. L86 229 1: 310] 106 ,836/ 248 9,687] 487 1,836|243 4,214 209 1,562| 224 6,123) 297 5,098! 244 3,952 187 2,774{ 130] 2,437 112 2,847 130! 2,411 108 2.566114 2,376 104 2,051| 89g. 3,384) 102 3,366 142 3,236|135 3,000 124 | 2,669| 1053 3.2901 132 8 10. 161 316 158 136 “45 93 58 i21 84 73 34 70 4 57 58 56 2 2 3 69 36 3,302 154] 100] 2,870(113 2,625) 102 3,153 121 3,992 151 £,218/ 158 3,020 185 4711163 398| ran ,443 Is, 1,815! 168 72 6é °f E uf *0%5 o 2 98 Cynanche Maligna. | Rainfall | Rain ; py | Rain. ~ + - Q oe — — = a0 = 23 aE 3g yy & <=2} a —_— Vo 4 25 ow < LE As n at, A = = q31s8=23 |S wo = J SESH ~5 | x & ’ 199! 3741 1,273 1770 397) 376 303 341] 384 366 193 226 163 “66! 195 gt 149 201 147 194 179| » 151] 6 162 - 2081 82)21-1| 84 115]215| 89, 389|21-4| 88 535[16:8| 6g: 178 259107 111 316{130 89 20-8 861 97 265 109) 110.20:0] 82 103 17-2| #1 54°290|119-62130°5!126 4’ 286118 45 253104 521240| 99 35[185| 76 39223] g2 261300) 124" 371234] 96 48120:0| 82 44 28-2|116 37124:2| 100 39 269|rr1 49 29:2| 120 144] 95 144] 95 129] 8s 112} 74 1481 97 192] 126 146: 96 179 18 131 86 112] 74 136° 9o 184 121 L653) 207 143" 94 147 97 128 3s 164 10¢ 197 r30 144 95 139 or 149 ¢8 168 r11 183 120 167 110 91 92 87 72 102 128 91 113 BZ 72 05 128 l& 99 Bs 80 100 127 96 87 107 105 116 115 34s] 1711001243] 100152] 100 | 100 2 7| 49 185{ 7 | 42 313 129] 29-8123 2521104 2521 104 21-91 gol 18-1) 74 24-0 99 24:2|: oo 19:9] 82: M-Blira 1791 118 1521 100 154 101 179 118 173 114 50 99 RR 109 o7 93 OO 124 112 103 111 102 87 104 1108 88 112 n== oe = = = =e = oP Ea 2.3 Po ~-Wi * Calculated on the sum of the means of Diphtheria and Cynanche Maliona for the 24 years. XVII THE DISTRIBUTION OF DIPHTHERIA 335 CHAPTER XVII. THE GEOGRAPHICAL DISTRIBUTION OF DIPHTHERIA IN ENGLAND AND WALES. In spite of the immense labour and great ability that have been for a long time devoted to the subject, the mtiology of diphtheria is still very obscure. A faint hope of lighting upon some clue has induced me to undertake a tedious investigation into the geographical distribution of 89,603 deaths from the disease, recorded in England and Wales during the 26 years from 1855, when it first appears in the returns of the Registrar-General, to the close of 1880. The influence of age and sex on diphtheria, so far at cast as it is a fatal malady, is peculiar. TABLE 1.2 Deaths from Diphtheria in England and Wales. 1855-60 1861-70 1871-80 Males. 9,844 18,467 13,815 Females. 10,879 20,987 15,611 Persons. 20,723 39,454 29 426 Total - + 49.126 | 47477 | 89.603 ! Reprinted from the Supplementary Report of the Medical Officer of the Local Government Board Jor 1887. 2 Compiled from the Supplements to the Registrar-General’s 25th, 130 STUDIES IN STATISTICS CHAP. TABLE II. 1855-60 1861-70 1871-80 Averave - Deaths from Diphtheria. Females to 1000 Males. 105 - aga Persons Living. Females to 1000 Males. 1,047 1,053 1.054 1.051 The facts as to sex are set forth in Tables I. and II, from which we learn that more females die than males by about 12 per cent., whereas the number of females living exceeds that of males by 5 per cent., so that an excess smounting to 7 per cent. remains, which cannot be accounted for by the disparity of the sexes in the general population. Table IIL! gives for each sex the number of deaths at several groups of ages, and from it we learn that 57 per cent. of the deaths of males and 51 per cent. of the deaths of females occur in the first five years of life ; in the next five years 25 and 29 per cent. respectively, while after the forty-fifth year is reached only 2% per cent. of the deaths 5f males and 2 per cent. of the deaths of females take place. This table also shows that the excessive number of female deaths occur mainly between the fifth and fifteenth years of age; indeed, after the age of 45, slightly more males than females die of diphtheria. In Table IV. the death-rates per million living at each age are given, and from the figures we see that not only 35¢h, and 45th Reports, The figures for 1855-60 are slightly less than those shown for the same years in later reports; probably cynanche maligna was not included with diphtheria in the earlier publication. \ Tables III. to XV. are printed at the end of this paper. XVII THE DISTRIBUTION OF DIPHTHERIA 337 absolutely, but relatively to the numbers of each sex living, the excess of deaths of females is mainly between the ages of five and fifteen. This excess, however, does not seem to me to be sufficiently marked for any irregularities in age or sex distribution seriously to affect the figures as to geographical distribution that I am about to lay before you. A few words as to the death-rates in the first five years of life generally. For the first period the materials are not given iu the returns. In the remaining 20 years several considerable discrepancies between males and females are apparent, but I do not care to lay stress on these since the ages given in the census are notoriously uncertain for these early years. The mortality is, however, without doubt least in the first year, thus offering a marked contrast to summer diarrheea ; for the next four years it is, roughly speaking, constant, and during these years (second—fifth) no less than 47 per cent. of the deaths of males and 443 per cent. of the deaths of females take place. | should say that nearly all the figures quoted are obtained from the supplements to the 25th, 35th, and 45th Annual Reports of the Registrar-General, but in a few cases it has been necessary to refer to Annual Reports themselves or to the census, also to two Parliamentary Papers known as Mr. Lowe's Returns, 1864 and 1873. The populations given are the means of successive censuses, and are in most cases taken direct from the decennial supplements. As the diphtheria deaths are in the first period only for the six years 1855-60, whereas the populations are the means of the two censuses for 1851 and 1861, it follows that the latter are too small, and the deathrates calculated from them consequently too large; the error affects all districts more or less, but mainly those in 7 338 which the population is increasing very rapidly. The mean error in the population thus arising amounts to 22 in every 1000. Thus for England and Wales I have given the death-rate for the six years as 182 ; it should be only 178. In some few instances this might affect the argument slightly, but the error, such as it is, must remain, since the labour of recalculating the populations of the six hundred odd Registration Districts of England and Wales is more than I am inclined to face. Moreover, the assumption of the decennial supplement, that the population may be taken as the arithmetical mean of two successive censuses, itself overstates the population of rapidly-growing districts, since the growth is in a geometrical and not an arithmetical progression. To avoid overloading the paper with tables, I have thought it best to omit areas, populations, and deaths, giving only densities of population and death-ratcs. STUDIES IN STATISTICS CHAP. In examining the geographical distribution of the deaths from diphtheria, I shall begin with the eleven registration divisions into which England and Wales are livided. The registration divisions are large areas, most of them comprising several counties. In the second period under review, the smallest of them, the Kastern Division, made up of the counties of Essex, Suffolk, and Norfolk, contained a population of 1,180,655 persons, of whom 8,498 died of diphtheria; while the largest of them, the North-Western, comprising the counties of Cheshire and Lancashire, contained a mean population of 8,159,406, of whom 5,442 died of diphtheria in the decennium. It is therefore obvious shat the argument concerning the fallaciousness of small numbers does not apply here at any rate. In Table V. the registration divisions are arranged in XVII THE DISTRIBUTION OF DIPHTHERIA 339 the order of their diphtheria death-rates per 1,000,000 persons living for each of the three periods under review. In Table VI. the whole period of twenty-six years is taken, and the registration divisions are arranged, (1) in the order of their mean annual death-rates from diphtheria; (2) in order of their density of population; and (3) in order of their mean annual death-rates from all causes. In columns (1) and (3) the death-rate of England and Wales is taken as 100, so that the excess of each division is read off at once in percentages of the mean. An examination of these tables shows that the South-Western, South-Midland, and Northern Divisions have always experienced a low diphtheria mortality. Again, the North-Western, and to a less marked degree the London Division, have suffered little on the whole, but in the third period each of them exhibited a slightly higher death-rate than England and Wales. It will be noted that the London and North-Western Divisions (this last comprising Cheshire and Lancashire) are by far the most densely populated, and they also have the highest general deathrates, the North-Western Division enjoying a most unenviable notoriety in this respect. On the other hand, diphtheria is seen to have been above the average of England and Wales in each period in the West-Midland and South-Eastern Divisions; also in two out of the three periods it has been above the average in the Welsh and Eastern Divisions, and has in no case fallen much below the average in those Divisions. The North-Midland Division, although on the whole it has suffered most of all, has been affected very irregularly. Now all these five divisions have general death-rates below the average. In point of density of population they vary considerably, The Eastern and Welsh Divisions have populations mainly rural ; that of the South-Hastern 340 STUDIES IN STATISTICS CHAP. Division is about half and half ; that of the North-Midland Division is more rural than urban, but in the West Midland Division urban conditions predominate. So far as we have gone it would appear that rural life is not unfavourable to fatal diphtheria, while on the other hand the disease shows no special proclivity for densely-populated districts. A comparison of columns (2) and (3) of Table VI. shows that, as is well known, it is far otherwise when deaths from all causes are considered. This seemingly undue incidence of diphtheria on sparse populations I shall proceed to investigate in some detail. By breaking up the divisions into their component registration counties, we shall get much more specialized populations to deal with, but at the same time, as the numbers fall the results deduced may be less trustworthy. With a view to partially obviating fallacies due to small numbers, I have made the following adjustments :— Huntingdonshire (population 58,591) has been added to Cambridgeshire. Rutlandshire (population 23,432) has been added to Leicestershire. Yorkshire has been divided into its three Ridings, as sheir characteristics are so different. Westmorland (population 63,038) has been added to Cumberland, together with the Ulverston district of Lancashire (including Barrow-in-Furness), the whole making a natural group which I have termed “the Lake District.” Wales I have divided into— 1. Glamorganshire (population 366,026). 2. The remainder of South Wales. 3. North Wales. [t will be noted that the object has been to group These are the mean populations of the middle period (1861-70). LVI THE DISTRIBUTION OF DIPHTHERIA 341 together districts having similar characteristics. As thus adjusted, the smallest county (Hereford) has a population of 107,726, while five other counties range up to Hertford with 186,082. At the other extreme, Lancashire has 2,657,312. In Table VIL the registration counties (modified as above) are arranged, for the three periods severally, in the order of their diphtheria death-rates; England and Wales being taken as having a death-rate of 100. In Table VIII. the same counties are arranged in parallel columns in the order of (1) Diphtheria death-rates for 26 years; (2) Mean density of population, 30 years; (8) General death-rates, 30 years. Where the diphtheria rate of a county has been above (or below) the average of England and Wales in each of the three periods it is printed in italics. It will be seen from these tables that of the 12 counties which have uniformly suffered less from diphtheria than the average of England and Wales, four are counties with dense populations, including indeed Metropolitan Middlesex and Lancashire. On the other hand, of the seven counties which have uniformly suffered more than the average from diphtheria, only two (both suburban) have somewhat dense populations, while one (South Wales less Glamorganshire) is the most thinly populated of all. Looking at the matter in another way, there are seven especially densely populated counties, three of which, Lancashire, West York, and Metropolitan Middlesex, suffer from the highest general death-rates ; Metropolitan Surrey and Warwick also have high death-rates, Metropolitan Kent near the average, and extra-Metropolitan Middlesex, alone of the seven, a low death-rate. But as regards diphtheria, three of these, viz. Metropolitan Middlesex, West York, and Lancashire (the last particularly) have enjoyed com- 42 STUDIES IN STATISTICS CHAP. parative immunity from the disease in all three periods; two of them, Metropolitan Surrey and extra-Metropolitan Middlesex, have enjoyed like immunity in two periods out of three, while two only have twice suffered more than the average of England and Wales, namely, Metropolitan Kent and Warwick, the last very severely. Of 12 counties standing at the other end of the list as districts of especially sparse populations, we find :— Three counties above the average for diphtheria in each period, viz. Shropshire, Lincolnshire, and South Wales (less Glamorgan) ; five counties above the average twice out of three times, viz. Norfolk, North Wales, North York, Hereford (all of which suffered very severely in one or sther period), and Northumberland. On the other hand, Wiltshire, Dorsetshire, and Cambridge (with Huntingden) suffered more than the average in but one period ; Cambridge very severely. The Lake District, one of the least dense “counties,” had a mortality somewhat below the average throughout. In the map (Plate XXII) the counties are tinted so that the depth of colour is proportional to the mortality from diphtheria. The contrasts between Lancashire and the West Riding, when compared with the sparsely populated adjoining counties; London and the home counties ; (lamorganshire and the rest of South Wales; as also the remarkable prevalence of diphtheria in the Eastern Counties and Wales, are very strikingly indicated. At this point we may safely say that density of population is not a factor of prime importance in the production of fatal diphtheria; indeed, on the contrary, I hope to be able to show in the sequel that this disease has a strong predilection for scattered rural populations, and even for districts reputed to be exceptionally healthy. Having got so far, it is obviously desirable to compare, Plate xx 6 wr ND WALES Scale of Enflish Miles Ww 60 W 20 #0 40 50 — ey 4-79 32-88 32-109 ie —14l 49 -191 52 1 Average DIPHTHERIA Death-Rates Twenty-six years 1856-80 ‘ENGLAND & WALES taken as 100) ZN iE Cr London. Stanfords Geoa’' T° London: Edward Stanford X¥I1 THE DISTRIBUTION OF DIPHTHERIA 343 if it be possible, the relative mortality from diphtheria of town and country districts within the same county. The easiest way of doing this is to lay down definitions of what are to be considered densely populated, medium, and sparsely populated districts respectively, and then examine the several counties and the districts of which they are made up. The criteria that have seemed to me best are the following, viz.:— 1. Dense Districts.—Those in which there is less than one acre to each person living. Practically these comprise all the very large towns, together with a few small ones. 2. Medium Districts.—Those in which there is one acre to each person living, but less than two acres. This class comprises a large number of unions containing mediumsized towns. 3. Sparse Districts.—Those in which there are upwards of two acres to cach person. In these there are but few towns of any size, and even where considerable country towns are included, their population bears but a small proportion. to that living outside under strictly rural conditions. As the populations are very different in the different periods, and the areas themselves are by no means fixed, it is only possible to compare districts for each decennial period separately. Where changes of area have been made during the decennial period, I have made such adjustments as seemed possible, and, although the results for the districts involved are necessarily inaccurate, I feel pretty sure that the discrepancies are not sufficiently numerous or important materially to affect the argument. It is fair to state that the following comparisons involved very considerable labour in the calculations of areas, populations, and rates. 344 STUDIES IN STATISTICS CHAP, In Table IX. are given for the six years 1855-60, for each registration county, the population of its dense, medium, and sparse districts, together with their diphtheria deathrates per million persons living. It will be seen that, as might have been expected, there are several counties which do not contain registration districts falling under each of the three above defined descriptions—for instance, all the registration districts within the metropolis come under sur definition of “dense districts,” whereas those in Herefordshire are all “sparse districts.” There are, however, 25 counties in which we can sompare dense districts with medium districts, and in no less than 20 of these the diphtheria mortality was higher in the medium than in the dense districts. Again, there are 32 counties in which we can compare medium with sparsely populated districts, and it will be seen that in 28 of these the sparse districts suffered more severely from diphtheria than the others. Lastly, there are 29 counties in which it is possible to ompare dense districts with sparse, with the result that in no less than 26 the sparse districts suffered most. There are, moreover, three counties having only dense districts, all with but moderate diphtheria mortality, and four with only sparse districts, which all suffered very severely. The figures for the whole of England and Wales for this period are : Dense districts 128, medium districts 182, sparse districts 249 diphtheria deaths per million persons living. It is therefore plain that so far as the first period is concerned, diphtheria showed itself to be twice as fatal a malady in rural districts as in large towns, while districts with medium density of population occupied a middle position. XVII THE DISTRIBUTION OF DIPHTHERIA 345 In the second period (Table X.) the results are less striking : thus, we see that for all England and Wales the death-rates for dense, medium, and sparse districts are 163, 164, and 223 per million living respectively. Out of 28 counties in which it is possible to compare dense with medium districts, in 21 instances the medium districts have the higher diphtheria mortality. Out of 83 counties in which medium districts can be sompared with sparse, there are 24 in which the mortality in the sparse districts is greater. Out of 81 counties in which dense districts can be compared with sparse, there are 25 cases in which the mortality in the sparse districts is greater. Lastly, in the case of seven counties in which the foregoing comparisons are not possible, and for which the method of comparison with the rate of the whole country is made use of, the influence of density and sparseness was not so uniform as in the first period; in two of the three dense (metropolitan) districts the mortality was above the average of the whole country, and in one of the counties with only sparse districts, the mortality was below that average, though the other three sparse counties suffered very severely. In the third period (Table XI.) the results are similar to those in the second period, but still less marked. The death-rates for England and Wales are 114, 125, and 132, for the dense, medium, and sparse districts respectively. Out of 30 counties in which dense and medium districts can be compared, the medium districts have the higher leath-rate in 15 cases only. Out of 34 counties in which medium and sparse districts can be compared, the sparse districts have the higher death-rate in 21 cases. 346 STUDIES IN STATISTICS CHAP. Out of 32 counties in which dense districts can be compared with sparse, the sparse districts have the higher death-rate in 23 cases. Lastly, as regards the five cases in which no comparisons are possible, the dense and sparse districts suffer about equally. If we assume the mortality of the dense districts as 1000, and take the average of the three periods, we get a sufficiently striking result : Mortality from diphtheria in dense districts - 1000 ” medium ,, - - 1178 , ’e sparse ,, - ~- 1507 [t is, however, not easy to conjecture why in each succeeding decade the towns should have, relatively to the rural districts, suffered more and more. If we grant for a moment the exciting cause of the disease to have its origin in the country, it is just possible that the constantly increasing communication between town and country, by affording additional opportunities of importing the disease, might account for its increased prevalence in towns. Although the greater proximity of people in towns would at first sight seem to increase greatly the chances of infection, it is by no means certain that the individuals of a town community come so much into personal contact as the dwellers in a lonely hamlet. There may be but few opportunities for the introduction of the poison into an isolated village, but once introduced there are great facilities for its spread ; in a village every one knows his neighbours, whereas in a large town dwellers in the same street are often complete strangers to each other. But facility of communication, or the reverse, cannot be the true explabation in this case, since other communicable diseases, such as measles, scarlatina, and whooping-cough are most fatal in towns. It would, of course, greatly XVII THE DISTRIBUTION OF DIPHTHERIA 347 facilitate the investigation if there were statistics of cases of disease available, whereas it is only possible to discuss deaths. The Inféctious Diseases (Notification) Act of 1889 will, it is to be hoped, in course of time provide us with such case statistics. [ have been anxious to avoid the charge of dealing with numbers too small to give trustworthy results ; hence, instead of dealing with individual registration districts, I have grouped them together as far as possible, and then have drawn my inferences from the comparisons of numbers of these groups. It would obviously be most unsafe to attempt to argue from single instances in which some single special circumstance might altogether outweigh general considerations. But in order to be able to form a more definite idea of the sort of localities which are either specially liable to diphtheria, or enjoy an unusual immunity therefrom, I have compared the diphtheria death-rate of every registration district, in each of the three periods, with that of England and Wales, and divided the districts into the four following classes: — Class A.—Diphtheria mortality 50 per cent. and upwards above that of England and Wales. Class B.—Diphtheria mortality above that of England and Wales, but not 50 per cent. above it. Class C.—Diplitheria mortality below that of England and Wales, but not 50 per cent. below it. Class D.—Diphtheria mortality at least 50 per cent. below that of England and Wales. For my purpose it is obviously of little importance to take note of occasional prevalences of, or occasional exemptions from, diphtheria. Therefore, I give, in Table XIL, those registration districts which were in all three periods greater sufferers from diphtheria than the average of the whole country, indicating the severity of the disease STUDIES IN STATISTICS CHAP. more precisely according as they fell under Class A., in one, two, or all three periods, or not at all. Similarly, the districts whose mortality was always below the average are given in Table XIII, classed according to the number of times that they came under Class D. In the tables dense Jistricts are printed in SMALL CAPITALS and medium districts in ttalics. It will be noted that there are 80 districts in which the liphtheria mortality has been uniformly above that of England and Wales. Classed according to density of population, only three of these are dense (one a metropolitan istrict and two in the suburbs of Birmingham), six are medium, but as many as 71 are sparse districts. In Table XIII. are shown the 140 districts in which the Jiphtheria mortality was uniformly below that of England and Wales. Classed according to density of population, 10 fewer than 51 of these were dense districts and 26 medium, whereas only 63 were sparse. It can scarcely be denied that these two tables show in a remarkable way that very many of the largest towns of the country have suffered comparatively little from diphtheria. Indeed, that will be found to be the case with districts comprising the whole or parts of one-half the great towns of England ; among them, towns with such avil sanitary reputations as Hull,! Leeds, Bolton, Newcastle-on-Tyne, Leicester, Manchester, Liverpool, and East London. On the other hand, in parts only of three of the large towns—London (Islington), Birmingham, and Brighton —has there been continuous excessive diphtheria mortality. [ am not prepared to point out any special characteristics which the places that have suffered most from diphtheria have in common, further than that the great 1 Hull has greatly improved in recent years. 248 XVII THE DISTRIBUTION OF DIPHTHERIA 349 majority of them are poor-law unions whose populations live mainly in seattered villages or in country towns of small size. Of the 80 districts, 12 are in Sussex, 7 in extra-Metropolitan Kent, 7 in Lincolnshire, 7 in South Wales, 5 in extra-Metropolitan Surrey, and 4 each in Shropshire, West York, and East York, leaving only 30 districts for all the rest of the country. To check these results, and with a view to partially eliminating the effects of communication of the disease by infection, I have prepared Table XIV., in which are given the registration districts whose diphtheria mortality has always exceeded that of ‘the county in which they are situated ; also Table XV. in like manner, mutatis mutandss, for districts of low mortality. There were 82 which persistently suffered from diphtheria beyond the average of their counties, of which 10 were dense, 8 were medium, and 64 sparse. On the other hand there were 120 districts which persistently suffered less than the average of their counties; of these 48 were dense, 17 medium, and 53 sparse. These tables, then, tell the same tale as Tables XIL and XIIL, only somewhat less emphatically. As might have been expected, very many districts appear in both lists. Thus, out of the 80 districts that suffered severely in comparison with the whole country, 47 also suffered severely when compared with their own counties, Again, of the 140 districts that, as compared with the whole country, enjoyed comparative immunity, as many as 86 enjoyed a like immunity when compared with their respective counties. Taking both standards together, the districts’ which suffered most from diphtheria during the 26 years were :— ! Districts of medium population in ¢talics ; of dense population in SMALL CAPITALS. 350 STUDIES IN STATISTICS CHAP. (1) Tenterden, Chelmsford, Ellesmere, Horncastle, Howden, Driffield, Aberayron. All of these are very sparsely populated unions. (2) The following suffered only slightly less: Peterborough, Smallburgh, Gainsborough, Rothbury, and Haverfordwest. All of which again are sparsely populated. (3) Next after these came Bridge, Hailsham, Petersfield, Witham, Blything, Sturminster, Ledbury, Penkridge, Aston, Solihull, East Retford, Selby, Patrington, Morpeth, (Cardigan, Lampeter, and Crickhowell. Of these Aston is a dense district, but all the others are sparse, and, moreover, with the exception of Crickhowell, they all, during the 30 years 1851-80, enjoyed general death-rates below that of England and Wales. Per contra those districts which have enjoyed the greatest immunity, both when compared with their counties and with England and Wales, are— (1) SournamrroN, Borrox, and Machynlleth. (2) Only slightly less favoured—Hoo, Uxbridge, Bicester, Halstead, Chard. Yeovil. Nuneaton, Blaby, WIGAN, and HALIFAX. (3) Distinctly worse than the preceding, but still notably healthy as regards diphtheria, are—Farnham, Fordingbridge, Andover, St. Albans, Amersham, Newport Pagnell, Brixworth, Helston, LeicH, Settle, Krieniry, HouvcrroN-LE- SPRING, Chester-le-Street. Cockermouth, Llanelly, and Pembroke. As regards density of population, out of the above 29 districts, seven take rank as densc, and six as medium, while 11 had general death-rates above that of England and Wales, the three Lancashire towns, Wigan, Bolton, and Leigh, being the worst. It may be remarked incidentally that there is no registration district, however small, which has escaped from XVII THE DISTRIBUTION OF DIPHTHERIA 251 fatal diphtheria for longer than one of the three periods. The smallest actual number of deaths registered in the 26 years was four in Hoo, a Kentish district with a mean population of 3,043 inhabitants, which give a death-rate of 51 per million living. The 12 districts in column 4, Table XIIL, bad mean diphtheria death-rates ranging from 42 to 69 per million, against 159 for England and Wales. To sum up, whatever units of area are taken in the examination of the geographical distribution of diphtheria in England and Wales, the same general results are arrived at. The distribution of diphtheria, is apparently sus generis, the mortality from the disease clearly is not regulated by the same causes as influence the general mortality, certainly density of population does not favour it. This disease appears rather to have a preference for certain special districts, the great majority of which may be described as rural, and which have but a small proportion of their people living in towns; indeed, several of these districts—notably, Rothbury, Solihull, Hailsham, Petersfield, Ledbury, and Tenterden—have such very low general death-rates for the 30 years 1851-80 (varying from 156 to 17°6 per thousand), that they may fairly claim to be among the healthiest areas in England. On the other hand, whilst among those which have to a great degree escaped the ravages of diphtheria, there are some notably healthy areas, such as Farnham and Andover (with deathrates of 164 and 17-6), there are also a number of towns of medium or large size, some of which, such as Leigh, Bolton, and Wigan, have high general death-rates (ranging from 24°6 to 28'3). Some one will object, “ You have yourself stated that the incidence of diphtheria mortality according to age and sex is peculiar, and yet in your argument you make no 352 STUDIES IN STATISTICS CHAP. allowance for differences in the age and sex constitution in the populations of the districts.” This is quite true, but it should also be remembered that four-fifths of the whole of the deaths from diphtheria occur among children under 10 years of age, and an extremely small proportion among elderly persons. Now it is a well-known fact that in densely populated districts the proportion of children is larger and the proportion of elderly persons notably less than in rural districts. If, therefore, it were worth while to undertake the immense labour of correcting all my calculations for differences of “age and sex constitution,” the striking contrast that I have endeavoured to bring out between the liability of rural and urban populations to fatal Jiphtheria would be still more marked. Again, the geographical distribution of the other zymotic Jiseases is totally different to that of diphtheria. That summer diarrhea is especially a disease of towns is a familiar fact, to which attention has repeatedly been called by the Registrar-General. Measles is especially fatal in London! Scarlet fever is most common in the mining and manufacturing counties—Durham, Yorkshire, Northumberland, Staffordshire, Warwickshire, Cheshire, Lancashire, Monmouthshire, and South Wales.? The practical deduction suggested by these facts is, that the cause, or causes, of diphtheria should not be sought for primarily in any high development of civilization, such as sewers, but rather in some condition associated with a more primitive mode of life. Again, privies and ash-pits con hardly be important agents in breeding or disseminating the disease, or we should expect to find diphtheria sxceptionally prevalent in those northern towns where such ~ Forty-seventl, Annual Report of Registrar-General, p. Xiv. ? Jpid. p. xiii. See also Note ab the end of this Paper. XVII THE DISTRIBUTION OF DIPHTHERIA 353 nuisances reach their worst, whereas the contrary is the case. On the other hand, low vegetable organisms developed in damp dwellings would perhaps fit in with the facts that I have brought forward; or again, some evil special to shallow wells or other primitive sources of water supply. The line of investigation, however, which seems to me most promising lies in comparative pathology. The peasantry live on terms of great intimacy with domestic animals, more particularly cows, sheep, pigs, and poultry (including pigeons). Some little known disease of some one or other of these creatures may be capable of inducing in men or women brought into frequent and close contact with them a trivial “membranous sore throat”; then, under suitable conditions of recipient and environment, the more generally recognized form of the disease “true diphtheria,” or, as it used to be called, cynanche maligna,” may result. The poison, perhaps gathering intensity and infectiveness, may then be conveyed into towns by farm produce such as milk, cream, or poultry, conceivably by eggs, meat, or vegetables; and lastly, through the persons first infected, the sources of water supply or the public sewers may get poisoned, and so indirectly aid -in the spread of the disease. But these speculations are taking me beyond my facts, and I must be content to leave others to derive what indications they can from the data I have set forth. [See Postscript, p. 855.] Note.—In a previous examination into this subject some years ago, the results of which were never published, [ compared the geographical distribution of diphtheria, scarlet fever, and diarrhea for the ten years 1861-70. The registration districts were divided into “ dense towns,” “open towns,” and “country,” in a somewhat different manner to that adopted in this paper, viz.:— 354 STUDIES IN STATISTICS CHAP. Districts with 10 persons and upwards to an acre were called Dense Town Districts. Districts with one-tenth of an acre, but less than two acres to each person, were called Open Town Districts. Districts with two acres and upwards to each person were called Country Districts. The death-rates from each disease were compared, not with the death-rate from that disease in England and Wales, but with the death-rate in the same registration division. Thus — Number of Districts. Dense| Open pens Town | Country. DiraTHERIA. 25 per cent. below mean of division and lower | 15 Within 24 per cent. above or below mean - 30 25 per cent. above mean of division and higher| 5 50 SCARLATINA. 25 per cent. below mean of division and lower Within 24 per cent. above or below mean 25 per cent. above mean of division and hicher 3 36 11 DIARRHEA. 25 per cent. below mean of division and lower Within 24 per cent. above or below mean 25 per cent. above mean of division and higher 4 22 24 on) 84 | 116 57 | 132 93 | 155 215 219 188 169 | 403-1 goo 26 | 210 93 | 159 50 ‘34 239 288 95 50 169 | 403 629 351 262 301 88 121 | 231 46 20 920 89 103 | 699 These figures, which were derived from the Supplement to the Registrar-General's 35th Annual Report, and from Mr. Lowe's “ Return of the average annual proportion of deaths from specified causes, de., dc., during 1861-70,” show’ clearly that, in its relation to density of population. avin THE DISTRIBUTION OF DIPHTHERIA 355 fatal scarlatina holds a place midway between those of fatal diphtheria and fatal diarrhoea. It is, of course, possible that diphtheria and scarlatina may be, as was once generally believed, two forms of one snd the same disease, whereof the one form is more frequently developed under rural conditions, the other ander urban; but it is doubtful whether any one would now venture to maintain this position. POSTSCRIPT. Dr. J. T. R. Davidson has quite recently published some interesting and suggestive notes on diphtheria in Buenos Aires, where he found the disease very frequently associated with a like ailment in hens, the conditions being such as to bring children a good deal in contact with the birds. See British Medical Journal, Oct. 25th, 1890, p. 954. t Dr. Edgar G. Barnes has published maps showing the distribution »f diphtheria and scarlatina in England and Wales in the six years (880-86. They show that while the diphtheria mortality was above the average of the whole country in every one of those years in North Wales, Essex, Hertford, Middlesex, London, Kent, Surrey, Sussex, and Hants, and below the average every year in Northumberland, Durham, Cumberland, Westmorland, Lancashire, Yorkshire, Derby, Stafford, Leicester, Worcester, Oxford, Gloucester, and Wiltshire ; on the other hand, the scarlatina mortality was above the average in every one of those years in Durham, West York, Leicester, and Monmouth, and below the. average in every one of those years in Westmorland, Rutland, Huntingdon, Suffolk, Essex, Hertford, Buckingham, extra-Metropolitan Middlesex, Kent and Surrey, Sussex, Hants, Berks, ‘Wilts, Dorset, and Devon. The contrast in the distribution of the two diseases is very marked. On the other hand, Dr. Barnes’ results agree very fairly with mine, We both make Essex, Kent, Surrey, Sussex, and North Wales very liable to diphtheria, and we both show an exemption from that disease in Durham, Lancashire, Leicestershire, West York, Oxford, Gloucester, and Wilts. The principal discrepancies, or perhaps it may be developments, are as regards London, Hertford, and Hants on the one hand, and Noxth York and East York on the other. See The Etiology of Diphtheria, by T. G. Barnes, M.D., British Medical Journal, July 28, 1888. } + pn TABLE IIL.—Deaths from Diphtheria in England and Wales, at Groups of Ages, Males and Females separately. All Ages. 0-— First Five Years. 0-5 10— 15-20— | er eB AR BI ar | mE MALES. 1855-60 sug 1861-70 oor po. 1871-80 ” woe Total Percentage of the whole FEMALES, 1855-60 i. 1361-70 oe 1871-80 Total Percentage of the whole Bveess of female over male deaths 9,844 18.467 13,815 42,126 100 10,879 20,987 15.611 v7 ATT 100 5.351 1,076 1,236 2,124 2,635 1.153 + 1,670 4,353 1 5,541 1,072 ' 1,086 2, #5 2.968 1,781 **"-9 984 | 5,454 2,530] 860 | 358 1,795 wo i = 591 “856 7,769} 38,7301 9421 326 24,170 pose I 3,021 | 1,275 nr ————— re ———n— | 57.4) 25°01 102 174 149 315 1 364 202 ' 244 124 239 216 71 197 165 59 1 41 182 | 107 129 73 438 | 370) 221] 8a FA 2:6 4 R818 AR 3 691 | 7571 m70 790 ' 1,104 '.596 2,531 912 1,456 3.208 | 5,091 15,174 1,123 11,236 1 1,163 2,430 ' 2,421 2,225 1,621 ° 2,005 1,807 5,416 8,100|1,287 | (1,203 5,660 1,863 ' 7.801" 4,77311,364 407 627 448 171 “g 223 459 359 135 284 265 a4 207 129 51 143 1923 37 18 80 | 33 77 ber ~ 5,662 5.195 24 490 113.533 t 11,482 K&R 1.044 | 68 45 | e112 194 -— om 12°5 2-1 2501 3.00811,443 | 207 | 147 | 287 | | —10 | —53 | —27 —6 Mm ment mene Excess of Male Deaths. % ~ ~~ ay “ h 3 - TABLE IV.— Average Annual Death-rates from Diphtheria in England and Wales per Million Persons living at several Groups of Ages, Males and Females separately. : MALES. 1855-60 vis 1861-70 _ 1871-80 ; Averaces of 96 years ... TEMALES. 1855-60 ve 1861-70 es . 1871-80 Averages of 26 years ... All Ages. 177 | 177 | 16 | 187 101 | 125 First Five Yrs 0— Nn Bi 10 15 | 90— | 95 35— ! ar | BH— | 65-— | T5— 760 1 655 474 | 539 | 523 57) 4901 | 724 | 617 | 667 | 589) 649 "ag 718 | 380 - 142 65 87 7s7 | 338 | 107 52 35 aro] os61 721 28 + 20 ' 20 | ; 1€ 30 19 23 32 20 | a5 Po i 17 29 1 18 325 | 107 ———— ‘ 9291 50 | 50 20 | 10 Vis | 02 | 2501 21 717 467 778 | 448 4731 325 207 166 105 78 61 28 32 25 17 22 a1 16 1 18 | so Vr | 18 | 19! 1 20 500 <99 845 870 818 570 20 GEA 455 491 621 9% 15 168 | 377 | 674 668 | 746 | got 1 656 | 413 | 159 _.-656 | 298 RY 5s 1s 3 358 STUDIES IN STATISTICS. CHAP, TABLE V.—REecIsTRATION DIvIsions 4 EnGLAND and WALES arranged 2 the Order of their DIPHTHERIA DEATH-RATES per Million Persons VIR. 1st Period, 1855-60. North-Western South Midland London wo Northern ... South-Western Welsh na England & Wales West Midland Yorkshire ... Eastern a South-Fastern North Midland 102 127 130 156 65 167 182 202 220 230 250 24R 2nd Period, 1861-70. South-Western Northern .... South Midland North-Western London es Yorkshire ... England & Wales West Midland South-Eastern North Midland Welsh ... Eastern .. 119 164 1 172 176 180 18/4 195 200 202 204 296K 3rd Period, 1871-810). Yorkshire ... North Midland South-Western Northern ... South Midland Eastern ee England & Wales 95 98 103 104 109 113 127 London ver North-Western West Midland South-Eastern Welsh vs 122 125 141 143 163 TABLE VI.—REGISTRATION DIVISIONS of ENGLAND and WALES arranged en Order of (1) DipaTHERIA DEATH-RATES, (2) DENsITY oF Popu-LATION, and (3) DEATH-RATES FROM ALL CAUSES. Adwerages of 26 years 1855-80). 1) Diphtheria Death-rates. (England = 100.) South-Western North-Western South Midland Northern ... London oa 80 82 34 87 RE Enoland & Wales 100 Yorkshire ... Welsh ‘es West Midland South-Eastern Eastern ws North Midland 09 (2) Density?! of Population. ‘Acres to 1 person living.) Welsh cee Northern ... South-Western Kastern . North Midland South Midland South-Eastern 3-79 2-81 270 2:67 2-59 2-33 2.092 England & Wales Yorkshire ... West Midland North-Western “ondon , 1°74 1-67 1:52 0-63 702 (3) 3 Death-rates from all Causes. (England =100.) South-Eastern South-Western South Midland Fastern v-North Midland Welsh wi West Midland 85 90 an "9 3 97 99 England & Wale: Northern ... Yorkshire ... London ve. North-Western 100) 103 105 106 115 1 The use of the word “ density” as applied to a set of figures in which the greater the density the smaller the figure is somewhat awkward hut has the sanction of usace. XVID THE DISTRIBUTION OF DIPHTHERIA 359 TABLE VIL—REcisTRATION COUNTIES arranged in the Order of their DIPHTHERIA DEATH-RATES, in each of THREE PERIODS; the Rate for ENGLAND and WALES being taken as 100. 1st Period, 1855-60. ancashire ... North Wales .. Bedford... ve Northumberland... Buckingham ... Oxford ... ee 3merset ... wn Metropol. Middlesex Hertford ... A. Northampton ... Devon... po Metropolitan Kent Jambridge and Hunts Lake District ... Flamorgan .. Metropolitan Surrey Hants - vee Extra-Met. Middlesex Suffolk... —— Leicester and Rutland Floucester ee “heshire ... wa Warwick ... . West York oi England and Wales Cornwall ... wr Worcester hia Durham ... cen Monmouth _— Stafford ... . Fssex wi fo Extra-Met. Surrey Berks wi a Wilts ... ves Dorset ... ee South Wales es Nottingham sg Extra-Metropol. Kent" Shropshire -_ Norfolk ... -— Derby ... ‘ee Hereford ... - Bast York ws Sussex... ee North York .. Lincoln ... . 9nd Period, 1861-70. Somerset ... gon Buckingham os Jertford ... ve Jornwall ... voy donmouth wie Wilts us “. Glamorgan. ol Devon ... oe "eicester and Rutland Durham ... i Txtra-Met. Middlesex Jxford ... _— Rloucester a Jerks - ... cee ereford . . - Jedford ... . Jants ae 3 Dorset ... ce derby ... a Worcester... A. Lancashire ... Nottingham — West York es detropolitan Surrey Metropol. Middlesex Stafford ... “e “ake District ... Northampton ... England and Wales 5 5: 5 r 1 i 30 33 56 37 59 59 71 r 3 3 3 32 36 36 i] 30 90 02 94 or 20 J) 25 95 96 oy 100] 100 02! 11 11 Al 13 17 21 23 23 31 38 29 7 3 35 65 183 186 208 o32 Extra-Met. Surrey Northumberland... Extra-Met. Kent... Metropolitan Kent Tssex es ve Fast York ee South Wales ai Shropshire .. Cheshire ... North York a Suffolk... is Warwick ... .. Jussex ... _ North Wales .. Cambridge and Huntx 7% Lincoln ... ad LTE Norfolk ... “ |219 04 113 14 19 19 23 24 2 2€ 3? 20 51 59 3rd Period, 1871-80. Buckingham Jorthampton Yottingham ‘vo derby ... si Wilts - ee Floucester a West York ... Somerset ... ves Cambridge snd Hunts Durham ... ve. Norfolk ... J Jertford ... ~ Leicester and Rutland Juftolk ... wh Dorset ... yo Devon... ei East York ve Huuts ain wn Stafford ... I. Metropol, Middlesex Berks ve I. Lancashire ai Lake District ... North York -. England and Wales Glamorgan vor Oxford ... ver Northumberland... Lincoln ... a Metropolitan Surrey Essex ee . Cornwall ... es Metropolitan Kent Hereford ... ake Wxtra-Met. Kent... Monmouth a Txtra-Met. Middlesex Worcester... as gxtra-Met. Surrey 3edford ... i Sussex... ws Cheshire ... ce Jouth Wales es Shropshire es Warwick ... .. North Wales ae 9 36 "1 3 a0" 5 23 08 09 100 01 07 7 3 0 40 15 37 60 138 STUDIES IN STATISTICS CHAP TABLE VIIL—REGicTRATION CoUNTIES of ENGLAND and WALES arranged in order of (1) DipururriA DeaTH-RaTES, (2) DEnsiTY oF PorPuraTiON. and (3) DEATHS FROM ALL CAUSES. 1) Diphtheria Death-Rates {England = 100) 26 Yenrs, (4) | Density! of Population {Acres to 1 Person living) 20 Years, 8) Death-Rates from All Causes (England = 100) 2 30 Years. Extra-Metrop. Surrey ' © Sussex _ - £ Dorset “yy ce 582 Jertford ... .. 338 Lincoln ... . 61 Tssex I~ ... 6} Tants . ee £'1¢ Extra-Metrop. Kent “% 3erks wo i. 35( Ixtra-Met. Middlesex 354 Wilts .. —-38% Jereford ... “. 3:24 Juffolk ... I. 3:1€ Shropshire — *0¢ Worcester ... ve "7€ North York I © 7! Somerset ... .. “71 Cambridge and Hunts 2°69 South Wales ve 2:63 Buckingham sa 2:61 Devon ... ain 250 Oxford ... po 40 Cornwall ... sa 39 Northampton ... 32 Bedford ... Su ‘81 North Wales oo 426 Norfolk ... wh 223 Gloucester... . 2°16 Lake District ... 2:12 Metropolitan Kent 2:09 Derby - - 1:98 Leicester and Rutlanc L180 Nottingham wi 74" Cheshire .... oe. (74! England and Wales Hast York... _ Monmouth ave Warwick ... vou Northumberland ... Flamorgan sas Durham ... . letropolitan Surrey stafford... - Metropol. Middlesex West York I ‘ancashire.. oo South Wales a North York w+ Lake District ved North Wales - Tereford ... .. inecoln ... “ Torthumberland ... hropshire... - Dorset woe vou Wilts en . Cambridge and Hunts Norfolk ... vee Devon ve en Oxford ... “ Suffolk... i. 3uckingham “ Northampton ... Berks _ ve Cornwall ... - Bast York... - 3ussex po yas Hertford ... »e lssex i - Somerset ... — Leicester and Rutlan lants ™ ed Bedford ... Sor Monmouth ee Stafford ... cae Derby veh A Nottingham sor 100 England and Wales Buckingham? .. Somerset... as Hertford ... - Devon —_ ww Northampton on Leicester ond Rutland Floucester ... .. Lancashire vee Oxford ... ee Wilts ve i. Jurham ... IN Flamorgan “. West York wae Hants - sn Cornwall ... vou Bedford ... _ Wetropolitan Middlesex Extra-Met. Middlesex Lake District qu Monmouth oi Metropolitan Surrey Nottingham po Berks po o Dorset _ » Northumberland ... Stafford ... aie England and Wales Derby a hoe Worcester ... i. Metropolitan Kent Suffolk... ee Bxtra-Metrop. Surrey Essex va von Cambridge and Hunts Hereford ... ” Cheshire ... Sat Bxtra-Metrop. Kent Kast York... i. North Wales ee South Wales ee Warwick ... .. Shropshire... ig North York ws Sussex a av Norfolk ... 3 Tinenln t See footnote to Table Vi, ® These rates are the arithmetic means of rates for three decades given in the Supplement to he Registrar-General’s 45th dwnual Report, Table I. 3 The name of a county is printed in italics when its diphtheria death-rate was above (or below) he averace of England and Wales in cach of the three periods XVII THE DISTRIBUTION OF DIPHTHERIA 361 TABLE IX.—Figst Prriop, Siz Years, 1855-60. POPULATION and DipuTHERIA DEATE-RATES in Dense, Medium, and Sparse Districts of Registration Counties, Metropol. Middlesex Metropolitan Surrey Metropolitan Kent Extra-Met. Surrey Extra-Met. Kent Sussex... vee Hants pp vee Berks ce wos Extra-Met, Middlesex Hertford ... J. Buckingham ~~ ... Oxford ... ee Northampton ... Bedford ... ves Cambridge & Hunts Essex won ver Suffolk... wale Norfolk ... _ Wilts vee sos Dorset ... ee Devon... it Cornwall ... Ves Somerset .. ve Gloucester _-Hereford ... vee Shropshire =... Stafford ... ve Worcester es Warwick ... ve Leicester & Rutland Lincoln ... . Nottingham ‘es Derby ... a Cheshire ... .. Lancashire .. West York whe Rast York ve North York ws Durham ... . Northumberland Lake District . Monmouth .. Glamorgan . South Wales a North Wales BN England & Wales Population living in Medium Sparse Danse Districts | phi Districts (1 Acre, bub (upwards of {less than less than Fre) 1 As £3 2 Acres tO |g, 0} Person) sach Person)! 175 Coy 888,207 531,091 163,813 88,166 112,208 95,554 142,013 94,758 08 999 20,725 192,889 15,178 80,812 20,346 70,743 64,429 21,922 16.478 129,002 210,049 249,488 206,509 157.326 111,278 123,509 134,157 184,954 107,243 219,389 267,646 064,738 312,049 200,448 124,799 198 873 193,917 244,167 147,856 102,958 179.197 198,087 88,245 128,498 131,281 102,190 113,888 95,895 70,718 168,228 312,726 103,731 202,257 59,708 152,133 216,862 94,753 77,445 370,414 204 609 20,105 37 507 27,088 68,485 48,929 118,541 97 008 27.899 | 25,786 92.106 11,041 54 845 148,153 55,401 69,091 229. 684 112,193 149,567 50.984 24,444 179,264 131,470 316,201 64.416 51,316 22,691 57,127 75,546 57.069 95,719 47,366 92,283 ),058,415 991,164 102.115 99,474 134,023 282,654 39,690 133,154 57. 146 195,372 71.163 291.821 43,189 92,301 205.730 84.521 3.766.416] 2.626.288 7.604213 Diphtheria Deathrates per Million. Dense| Medium | Sparse Dis- Dis- | Districts.] tricts. | tricts. 195 148 137 164 116 | 181 83 108 138 269 294 468 222 255 161 327 198 87 132 178 78 137 60 25 8 531 110 105 65 112 178 152 ‘10 4? 43 07 142 50 175 370 260 262 57 230 wh 195 333 350 292 263 70 J7 563 367 165 283 85 262 518 122 102 168 155 245 155 251 a0) 226 120 45 188 85 90 77 138 167 120 177 75 70 | 88 78 na 202 220 140 108 215 183 272 73 88 143 170 190 107 163 58 175 308 7 5a 593 70 87 145 | 83 123 | 182 | 249 its, fo Gi Mean 182 362 TABLE X.—Skcoxp Pzrriop, 1861-70. PoruraTiON and DipHTHERIA DEATH-RATES in Dense, Medium, and Sparse Districts of Registration Counties. STUDIES IN STATISTICS CHAP, Population livine in Diphtheria Death-rates ver Million, Dense Districts Jess than Acre to each Person). Medium Districts 1 Acre, but less than 2 Acres to ach Perean Sparse Districts ‘upwards of 2 Acres to aach Person) Dense Disricts, Medium Sparse Dis- | Districts. | tricts. Metropol. Middlesex Metropolitan Surrey Metropolitan Kent Extra-Met. Surrey Extra-Met. Kent Sussex... _ Hants... ves Berks 0 ws Extra-Met. Middlesex Hertford ... .... Buckingham ~~ ... Oxford ... ve Northampton ... Bedford ... ee “ambridge & Hunts Essex va oe Suffolk... ves Norfolk ... ves Wilts ve A. Dorset... ss Devon ... cee Cornwall ,.. wa Somerset ... ne Sloucester vee Hereford ... oe Shropshire ni Stafford ... i. Worcester - Warwick ... 44 Leicester & Rutland Lincoln ... _ Nottingham dis, Derby ... wow Cheshire ... Firs (.ancashire a West York ver East York au North York er Durham ... _ Northumberland Lake District ... Monmouth _— lamorgan ve South Wales _— North Wales LL. 2,158,691 660,951 209,482 133,842 202,821 114,989 201,467 53,623 133.309 175 186 219 130 37 "23 133 ‘17 31 175 202 290 299 265 317 305 155 156 74,824 231,059 63,107 [8 468 116,632 153.318 215,772 215,438 1692 299 92,781 71,732 23,641 19.436 11,674 39.490 118 100 35 177 .65 342 245 5 195 120 148 148 126 82 55 47 275 73 50 97 ‘51 318 "24 178 318 162 231 374 243 156 309 175 149 113 236 288 7.500.396! 163 | 164 (223 ee — Mean 184 43 2 7% 29 1 20,526 15 059 98,220 04,311 54,483 27,307 26.950 08.977 07707 10,561 60,290 63,632 113,327 57,717 97 009 164,322 55,338 68,064 265 EOS 26,517 598,421 159,824 386,367 R1.70R 54,912 46,846 60,233 77,810 70.530 112,791 56,692 166,477 2,418,072 1,261,342 129.651 112,672 154,180 267,336 95,123 121,058 62,408 313.795 289,944 207 355 45,723 88 492 108,647 93.021 10.926.808/2.962,041 England & Wales XVII THE DISTRIBUTION OF DIPHTHERIA 363 TABLE XL—TaIRD PErIoD, 1871-80. PoruraTioN and DIPHTHERIA DEATH-RATES en Dense, Medium, and Sparse Dzstricts of Registration Counties. Population living in Medium Dense Districts Districts (1 Acre, but Jess than Jess than Acre to each | 5 a uroc'to Person). ia1ch Person). Sparse Districts (upwards of 2 Acres to ach Person). Metropol. Middlesex Metropolitan Surrey Metropolitan Kent Extra-Met. Surrey Extra-Met. Kent... Sussex ... oe Hants oe wa Berks ce ee Extra-Met. Middlesex Hertford ... ws Buckingham Oxford ... wii Northampton ... Bedford ... sa Oambridge & Hunts Essex -— — Suffolk... -_ Norfolk ... — Wilts I. ver Dorset ... ws Devon ey . Cornwall ... _ Somerset ... 4 Gloucester ce Hereford ... yon Shropshire I. Stafford ... wan Worcester... ih Warwick ... wa Leicester & Rutland Lincoln ... -— Nottingham ve Derby ~~... oe Cheshire ... wis Lancashire ves West York ee Fast York ee North York ve Durham ... ce Northumberland ... Lake District i. Monmouth .. Glamorgan " South Wales ee North Wales . 2,418,565 361,340 255,472 197,895 314,190 145,823 230,180 67,775 274.303 168,930 138,357 85,999 01.230 15,909 48,531 80,689 26,324 24,820 14,944 16.884 46,342 16,280 295,730 219,867 153.845 I 21,459 57 403 17,804 29,114 33,285 .60,197 .16,015 214,489 85,144 252,857 295,733 177,337 138,683 364,550 182,092 043,844 137,742 119,257 182,908 205,116 100,922 "39,054 96 350 145,568 16,041 a7 075 32,720 77,410 62,100 38,441 97.995 34,020 | 35 491 11,470 18,303 71,599 | 110,458 12.037 | 63 456 171,102 49,815 130,746 205.765 98,861 373,728 298 263 159,650 108.798 54,673 63,253 30,458 81,797 18.954 240,586 70,482 198,878 2,863,753 1,599,649 157,274 43,090 197,020 245,117 18,332 53,801 274.809 40,396 217,250 381,860 89,146 87,520 70,801 ma mos | 41,628 138,916 .06,403 140,515 38,937 :65,249 151,726 72,707 114,576 329 405 251 060 103,126 100,512 72,707 39,674 56. 958 England and Wales 13.701.980/3.512.210:7.120.163 Diphtheria Denth-rates per Million. Dense Diskricts. Medium Sparse Dis- | Districts., tricts. 15 (35 As 5 =3 91 04 03 166 139 149 186 148 138 58 92 91 193 60 103 165 162 ‘98 04 {1R 228 42 101 52 01 nw 52 101 18 85 59 109 116 139 143 77 75 51 118 70 100 80 131 0 Hd 02 a7 7 21 38 8 104 95 138 217 106 132 179 141 | 84 73 45 82 169 200 110 73 90 0 33 28 90 53 71 {68 109 34 16 57 "2 90 245 54 85 i154 05 57 "30 9 al 32 55 77 187 180 138 138 69 81 lr 114 | 125 132 Read Mean 121 364 TABLE XIIL—REcistraTioN DisTriors dn which the DIPHTHERIA DEATH-RATE was above that of ENGLAND and WALES in each of the three Decennial Periods. STUDIES IN STATISTICS CHAP. 1. Always 50 per cent. ahove. 11. Twice 50 per cent. above. 11. Once 50 per cent. above. IV. Never 50 per cent. above. Tenterden. Bridge. Hailsham. Ticehurst. Fast Grinstead. Horsham. Petworth, Chelmsford. Smallburgh. Ledbury. Ross. Ellesmere. Wem, ASTON. Solihull. Horncastle. Gainsborough. Howden. Driffield. Cardigan. Aheravron. Chertsey. Guildford. Bromley. Tunbridge. Cranbrook. Faversham. Battle. Uckfield. Cuckfield. Steyning. Midhurst. Petersfield. Peterborough. Witham. Blything. Depwade. Penkridge. Kings Norrox. Holbeach. Grantham. Caistor. Glanford Brigg. East Retford. Selby. Patrington. Thirsk. Bedale. Morpeth. Rothbury. Haverfordwest. Newecastle-in-Emlyn., Lampeter. Crickhowell, Knighton. (34 districts.) N.B.— Medium districts in #falics ; dense districts in ECL, CAPITALS. Reigate. Godstone. Hollingbourn., Eastbourne. Chichester. Wokingham, Lexden. Docking. Sturminster. Bridgnorth. Oswestry. Cheadle. Burton-on-Trent Bourn. Worksop. Great Quseburn Hemsworth. Tadcaster. Bridlington. Guisborough. Montoomery, IsLiNGTON. Dorking. Camelford. Foleshill, (4 districts.) (21 districts.) XVII THE DISTRIBUTION OF DIPHTHERIA 365 TABLE XIII. — REGISTRATION DISTRICTS 7% which the DIPHTHERIA DEATH-RATE was below that of ENGLAND and WALES nn each of the three Decennial Periods. Ao Never 50 per cent. below. II. Once 50 per cent. below. III. Twice 50 per cent. below. iv. Always 50 per cent. below. Sr, GEORGE, HAR-OVER SQUARE, Sr. GILES. BETHNAL GREEN, Sr, GEORGE-IN-THE-EAST, POPLAR. Ringwood. Wallingford. READING, BRENTFORD, WesT Ham, YARMOUTH, Honiton, EXETER. Newton Abbot. Totnes. Taunton. JLIFTON, Stroud. CHELTENHAM, Tewkesbury. Stone. WEST BroMwiICcH, DUDLEY. STOURBRIDGE, Stamford. Macclesfield. Houir. Slendale. Pontypool. 99 districts.) WESTMINSTER. VHITECHAPEL, Ii. Exp, THANET, fareham. 3asingstoke. Taringdon, Tton. Vycombe. ylesbury. .enley. ‘rackley. Jardingstone, lettering, - Jobum. JAMBRIDGE, Noodbridge. Jalne, Nestbury. Jorchester. Sherborne, Bridport. Crediton. Penzance. Scilly Islands. Bridgwater. Bedminster. Westbury-on-Severn. Cirencester. SHREWSBURY, ‘WORCESTER. Warwick. LEICESTER. 1ADFORD. LIVERPOOL. 3uny(Lancashire). BARTON-UPON~-IrRWELL. MANCHESTER. JLDHAM. ROCHDALE. 3URNLEY. Lancaster. Ripon, BRADFORD. LEEDS. Durham. NEWCASTLE-ON-TyYXE. TYNEMOUTH. Pontypridd. Neath. Tarmarthen. Rhayader. Bangor. {53 districts.) N 1 —Medium districts in italics ; dense districts in SMALL CAPITALS ‘arnham. oo. ‘ordingbridge. wndover. -Vare. loyston. Titchin., \ St. Albans. Temel Hempstead. Jerkhampstead. ~.mersham. 3uckingham. icester. Brixworth. Nellingborough. Jundle. Laton. Halstead. Sudbury. Bury St. EDMUNDS. Jevizes, Bradford-on-Avon. SALISBURY. 3T0KE DAMEREL. Telston. Langport. frome, Keynsham, “hipping Sodbury. Thornbury. Nuneaton. Hinckley. LEIGH. ASHTON-UNDER-Lyne. BLACKBURN. Settle, Skipton. KEIGHLEY. Todmorden. SADDLEWORTH. HouGHTON-LE-SPRING, Dhester-le-Street. JATESHEAD. Jockermouth, Llanelly. Pembroke. 46 districts.) SOUTHAMPTON. Uxbridge. Newport Pagnell, Chard. Yeouil. Wells. BATH. Blaby. WIGAN. 30LTON. JALIFAX. Machynileth. (12 districts.) 366 TABLE XIV.—ReastrATION DISTRICTS 00 which the DipPHTHERIA DEATH-RATE was in each of the THREE PERIODS above that of the COUNTY wn which they are situated. STUDIES IN STATISTICS CHAP. 4. Always 50 per cent, above Ll. Twice 50 per cent. above. IIL Once 50 per cent. shove. Iv. Never 50 J.er cent. above. Tenterdett. Peterborough. Chelmsford. Sturminster. Axbridge. Ellesmere. Horncastle. Prescot. Howden. Driffield. Rothbury. Haverfordwest. Aberavron. Petersfield. Woodstock. Witham, Blything. Smallburgh. Mere. Tavistock. Launceston. Northleach. Stow-on-the- Wold Penkridge. Market Harborough. Gainshorough. Fast Retford. Worksop. Newark. Runcorn. West DERBY. Great Ouseburn. Hemsworth. Wortley. Doncaster. Selby. Tadcaster. Patrington. Morpeth. Lampeter. Crickhowell. Presteigne. Ruthin. HAMPSTEAD. [SLINGTON, Bridge. Faversham. Hailsham. Romsey. ‘Wokingham. Hendon. Aylsham. Blandford. Poole. Camelford. Williton. Clutton. Ledbury. Cheadle. Burton-on-Trent. Kins NORTON. ASTON, Solihull. Ormskirk. Pontefract. EOCLESALL BigRLOW. Cardiff. Cardigan. Knighton, Qt. Asaph. KENSINGTON. Sr. PaNcRras. (GREENWICH. Cranbrook. East Grinstead. Marlborough. Stafford. Caistor. Southwell. Bridlington. Thirsk. Guisborough. (12 districts.) (13 districts.) (27 districts.) (30 districts.) rr ————— een eet reset en ee See TR, PTE = 4 Medium districts in italics ; dense districts in SMALL CAPITALS. XVII THE DISTRIBUTION OF DIPHTHERIA 367 TABLE XV.—REGISTRATION DISTRICTS in which the DIPHTHERIA DEATH-RATE on each of the THREE PERIODS was BELOW that of the COUNTY in which they are situated. Never 50 per cent. below. _ WESTMINSTER. 31. GILES. BETHNAL GREEN. MiLE EXD. PopLaR. KINGSTON. Rye. Eastbourne. Ringwood. Hungerford. Kettering. WesT HAM. Epping. Tendring. Sudbury. Henstead. Calne. EXETER. S10RE DAMEREL. Keynsham. CLIFTON. DUDLEY. STOURBRIDGE, Boston. Lincoln. DERBY. Bury (Lancashire). BARTON-UPON-IR-WELL. Lancaster. York. Durham. TYNEMOUTH, Once ou cent, below. WHITECEAPEL. RICHMOND, (GRAVESEND. THANET. 3RIGHTON. Fareham. 3asingstoke. Faringdon. READING. BRENTFORD. Berkhampstead. Newport Pagnell. Brackley. Wellingborough. Luton. - Dunmow. Bury St. Ep-MUNDS. Stow (Suffolk). Woodbridge. NORWICH. Kivaes Lynn. Swaftham, SALISBURY. Shaftesbury. Sherborne. Penzance. frome. Thornbury. Madeley. Wellington (Salop). Newport (Salop). WORCESTER. Stratford-on-Avon. Hinckley. Stamford. Mansfield. NOTTINGHAM. ASHTON-UNDER-LYNE. BURNLEY. Skipton. SADDLEWORTH. LEEDS. HowL. GATESHEAD, NEWCASTLE-ON-TYNE. Neath. Rhayader. Newtown, Dolgelly. Carnarvon. (50 districts.) ——— N.B.—Medium distriets in italics : dense distriets in SMALL CAPITALS. | 111. - Twice 50 per cent. below. Farnham. MEDWAY. Fordingbridge. Andover. Uxbridge. St. Albans. Amersham. Chipping Norton, 3rixworth. CAMBRIDGE. YARMOUTH. Thetford. Jelston. Chard. Yeovil. JHREWSEURY. Atherstone. Warwick. Blaby. STOCKPORT, WIGAN, LiEigH. Settle, KEIGHLEY. HALIFAX. HoUGHTON-LE-SPRING. Chester-le-Street. Cockermouth, Llanelly. Pembroke. Bangor, 31 districts.) 1v. Always 50 per. cent. below. tloo. SOUTHAMPTON, Bicester. Halstead. Nuneaton. Borrow. Machynlleth. (7 districts.) (32 districts.) 36° STUDIES IN STATISTICS CHAP. CHAPTER XVIII PHTHISIS, BRONCHITIS, AND PNEUMONIA: ARE THEY RPIDEMIC DISEASES ?! SoME explanation is perhaps needed for this evening calling your attention to diseases which are not usually admitted to be epidemic in character. My reasons for doing so are briefly twofold. First, to fully understand epidemic diseases, it is well to contrast their behaviour with that of diseases of the non-epidemic class; and in these days, when great epidemics are few and far between, we may well turn our attention to the principal causes of death that are always active in our midst. Such are phthisis, bronchitis, and pneumonia, to which are attributed almost exactly one-fourth of all the deaths in this country. Secondly, Koch's discovery of the tubercle bacillus has set all men a-thinking, while the proceedings of that most excellent institution, the Committee for the Collective Investigation of Disease, remind us that the idea that pneumonia may be an epidemic disease is gaining ground. The question naturally arises, Is bronchitis wholly above suspicion ? As in previous papers, I shall to-night deal with the statistics of mortality in England and Wales, as furnished 1 Read at a Meeting of the Epidemiological Society of London, April 1th. 1883. xviii ~~ PHTHISIS, BRONCHITIS, AND PNEUMONIA 369 by the Registrar-General’s Department, and with those statistics alone. The numbers of deaths are taken from the Table of Causes of Death in each year’s annual report, the “Supplementary Table” being used in every case to correct the classification as far as possible ; and the deathrates have all been “ground out” again by Thomas's machine, the corrected population values having been interpolated between the successive census numbers.’ The results obtained by these processes may, it is believed, be relied upon, though in such a mass of figures small errors may have been overlooked. A comparison of the ‘curves that result from plotting out the death-rates from the majority of alleged causes, both with one another and with curves representing various meteorological elements, has led me to classify diseases into four groups. These groups may shortly be called hot diseases, cold diseases, dry diseases, and indifferent diseases. The indifferent diseases show no marked relation to meteorological fluctuations. This is a residual class of doubtful extent. As Messrs. Buchan and Mitchell have shown, most diseases exhibit a seasonal or week-to-week fluctuation which bears some more or less direct relation to weather; but in the case of some, which, like small-pox, measles, and whooping-cough, show very marked and constant fluctuations of this kind, no similar fluctuation is observed from year to year. For instance, measles is most fatal in summer and in winter, but we do not find it especially fatal in hot summers or in cold winters. By “ hot diseases” I mean those which are most fatal in summer, and as a rule especially fatal in very hot summers. Examples of this class are diarrheea, want of breast milk, atrophy and debility, tabes mesenterica, thrush, enteritis. dysentery. ! In many of the Registrar-General’s Tables this has not been done, but the death-rates calculated on old estimates have been reprinted. BRR 370 STUDIES IN STATISTICS CHAP. By “cold diseases” I mean those which are most fatal in winter, and more especially in severe winters; this class includes most affections of the respiratory organs, apoplexy, paralysis, mortification, heart disease. By “dry diseases” I mean those that are greatly checked by rainfall, especially frequent rainfall—such are scarlatina, puerperal fever, erysipclas, acute rheumatism, laryngitis; this class and “hot diseases” overlap to some extent. It must not for one moment be supposed that it is alleged that heat, cold, or drought is the cause of either of the above-named diseases—what is meant is merely that fluctuations in the various meteorological elements correspond more or less closely with fluctuations in the various death-rates. They are to be regarded rather as favouring or predisposing conditions. Thus, for instance, on the average the mortality from scarlet fever will be greater in a dry than in a wet season. More will die of bronchitis in a cold winter than in a mild one ; more of diarrhoea in a hot summer than in a chilly one. Of what use, it will be asked, are researches which lead to conclusions that are familiar to all ? Parturiunt montes, nascetur ridiculus mus. The answer is twofold :—1. This classification groups together diseases that were supposed to have no @tiological connection, and conversely separates discases that were supposed to be allied. 2. The connection between the meteorological elements and the death-rate is much less close than might have been expected. This is markedly the case with bronchitis and diarrhea. There are, then, indications of other, more obscure, causes at work ; the history of the progress of the natural sciences abounds in instances of important discoveries that have originated in the study of such “residual phenomena,” as they are termed by logicians. ENGLAND avo WALES DEATH - RATES FROM PHTHISIS, BRONCHITIS, PNEUMONIA AND CERTAIN OTHER DISEASES CALENDAR YEARS. 850 '51 52 53 54 55 56 57 58 59 60 '6l '62 '63 '64 65 ©6 67 '68 69 70 '7I '72 73 "74 75 76 ‘77 '78 79 '80 '81'82 ee 2 2 =x Mean Death Rates Cg per million living — SURAS _ hedge leSHNTERIS “THER ij AEN ep — we, -- 5Hi8 P84 —= R44 -_ 2347 aa 865 __ 5 i — TT ant -— [37 1 2 3 RHEUMATIG A= HEART - [YIN] ) MN VIAVIPAIR VA VAR) : ~E , ed Ir 1 1 1] CARS| COMMENCING OCTOL — 2697 “E Ave LTE ma 72300 -_ {322 7 also Psrlmstastsa ion sass 50 en ater eae as reste tea testo: AR =o bre Kear ’! - ‘he curves express the fluctuation per cent. above or below the Mean. Each division of the Vertical Scale corresponding to ten per cent. he figuras at the side give the actual values of the Means of the Death Rates per million persons estimated to be living. The curves nclude all ages & both sexes. The lowest black curve shows the number of days in each year on which the Mean Daily Value of the Temperature if the Air at Greenwich was 40°F lower KVIIL PHTHISIS, BRONCHITIS, AND PNEUMONIA 371 I claim, then, for my method of research that it brings into play all the four inductive methods; but that it does this imperfectly, and that only inductions of but slight orobability can be drawn from it, no one is so well aware as myself. In all the charts laid before you, please to note that the ups and downs indicate percentage fluctuations from the mean. Hence all the curves may be said to be strictly comparable, or drawn to one scale. The actual mean number of deaths is indicated by figures on the righthand side. One of my first objects in this investigation was to compare together all those alleged causes of death, such as measles and whooping-cough, in which the respiratory organs are more or less involved, or in which the fatal issue is commonly due to pulmonary complications. Another was to compare the various causes which are, or have been supposed to be, connected with tubercle. For the first of these purposes the following causes of death have been carefully studied :—Laryngitis, cedema olottidis, croup, measles, whooping-cough, influenza, bronchitis, asthma, emphysema, pneumonia, pulmonary congestion, pulmonary apoplexy, pleurisy, empyema, phthisis, haemoptysis. For the second :—Phthisis, heemoptysis, tubercular peritonitis, tabes mesenterica, tubercular meningitis (called hydrocephalus in the returns), scrofula, psoas abscess, joint and bone disease, rickets. As might have been expected, most of this labour was thrown away, and from the majority of the above curves little was to be lecarned-—chietly owing to the smallness of the numbers dealt with, but partly to difficulties of zlassification. I will now call your attention to Plate XXIII., and to 372 STUDIES IN STATISTICS CHAP. the upper portion of it. The facts apply to England and Wales, and the curves show the annual fluctuations of the death-rates from phthisis, bronchitis, and pneumonia, and some more or less allied diseases throughout a period of twenty-five years. At the top of the diagram are two curves, coloured gray and black, which represent respectively the death-rates of two typical epidemic diseases, measles and whooping-cough, which have one character in common, viz. that when they kill the patient, the fatal issue is commonly the result of pulmonary complications. Two points will be noted with respect to the curves of measles and whooping-cough. 1. The fluctuations are greater than those of the other curves on the diagram. 2. The curves either bear no resemblance to one another nor to any of the others, or at least the points of difference are as striking as the coincidences. From these facts we are justified in drawing two inferences :—(a) That whatever influence inclement seasons may have on the mortality from whooping-cough and measles, that influence is quite out-balanced by other considerations, viz. the prevalence or otherwise of the infective poison, its varying potency, and the number of susceptible individuals. (B) Since the curves of phthisis, bronchitis, and pneumonia do not exhibit such wide fluctuations, there is a presumption pro tanto that these latter are not epidemic diseases, at all events in the sense that measles and whooping-cough are. Now look at the green line representing the mortality from phthisis. Two points catch the eye at once. 1. It has a tendency almost uniformly downward. 2. It does not deviate much from a straicht line. ! The Plate was originally prepared for the twenty-five years 1855-79. but in revising it IT have carried forward the curves to 1882, giving a period of twenty-eight vears. xvi PHTHISIS, BRONCHITIS, AND PNEUMONIA 373 As to its tendency downward, it must be borne in mind that the death-rate from respiratory diseases and phthisis taken together has risen over 5 per cent. during the last :wenty years, while that from phthisis alone has fallen 20 per cent. It might very naturally have been supposed that this apparent decline in the mortality from phthisis was to be attributed mainly to points of diagnosis and nomenclature, but that this is to a very small extent the case has been proved in Chapter XIV. (p. 234). The figures there quoted show that the average mortality from phthisis in the five years 1876-80 was at the ages 5—35 very considerably lower than during the ten years 1861-70, whereas at all those ages the mortality from ‘‘ diseases of the lungs” remained comparatively stationary. The approximation of the curve to a straight line is remarkable. No other cause of death gives a straighter curve ; but those of cancer, apoplexy, paralysis, convulsions, and fractures are about equally straight. Messrs. Buchan and Mitchell have shown that in London the maximum death-rate from phthisis occurs in the eleventh week of the year, the rate being then 13 per cent. above the mean, whereas the minimum death-rate occurs in the thirtyseventh, thirty-eighth, and thirty-ninth weeks, the rate being then 12 per cent. below the mean, or a total range of 25 per cent., a smaller range than in any other disease, except cancer and premature birth. (See Plate XXV.) This straightness of both the week-to-week and the year-to-year curves of phthisis is, I think, a very remarkable fact. Any one would naturally have supposed that severe weather would influence the phthisis death-rate to a much greater extent than is actually the case; but it would appear that the inexorable diseases, phthisis and cancer, keep their victims alive, so to say, till they have finished with them, jealously resenting any interference. 374 STUDIES IN STATISTICS CHAP. Of course the very chronic character of these diseases to a great extent explains this. Those who believe in the communicability of phthisis would probably explain by this same chronic character the fact that it exhibits none of the characteristics of an epidemic disease. The curves of tubercular meningitis and tabes mesenterica are inserted in Plate XXIII. to show that in several respects they resemble those of phthisis. Such a resemblance was not observed between these curves and that of scrofula. The two red curves are drawn to a common mean, and exhibit the phenomenon of two almost identical curves, differing mainly in the fact that one has an upward, the other a downward tendency. The explanation is easy, for it is well known that many deaths are now returned as due to chronic bronchitis, which in former times were attributed to asthma. This is probably a mere yuestion of nomenclature. But although fewer deaths are now attributed to asthma than formerly, the transfer of these to bronchitis will not account for the great rise in that curve, since the joint mortality of the two diseases has risen 62 per cent. in 20 years. Average Death-rates in England and Wales per million. 5 “-ars 18F: Bronchitis ... es Asthma Asthma and Bronchitis 3: ©” Increase . “vg, Rise or Fall per cent, 31 +62 Neither will the decrease in phthisis and pneumonia altogether account for the increase in bronchitis, although it will to a great degree, as the following table shows :— XVIII PHTHISIS, BRONCHITIS, AND PNEUMONIA 375 Average Death-rates in England and Wales per million. . Phthisis ... Bronchitis ... Pneumonia... Other lung diseases... y Years, 1855-59. ~ ny TR by ~~ R Vears, LuTh 0 Rise or Fall per cent. 20 1 "d Total ... ... D805 Increase... vee It must then be admitted that more persons dic now from diseases of the respiratory organs than did twenty years ago. Probably the diminution in the mortality of infants from exanthemata will account for this. Delicate persons (indeed, all persons must die of something) and weaklings who escape a pestilence, may be expected to fall victims to so-called inflammatory diseases. Reverting to the diagram, the dark-blue curve, oneumonia, shows a downward tendency, though not so marked as in the case of phthisis. In amplitude its fluctuations are midway between those of phthisis and bronchitis; but they correspond in direction very closely with the latter. The pleurisy curve is in some points allied to the pheumonia curve ; but it resembles far more closely that of “rheumatism of the heart,” which in turn, as I showed in a former paper,! has a very remarkable resemblance to the erysipelas curve. Messrs. Buchan and Mitchell called attention to the fact that the week-to-week fluctuations of pleurisy resemble more closely the corresponding fluctuations of rheumatism, than they do those of other respiratory diseases. May we not conjecture from these two facts that the majority of deaths assigned to pleurisy have their origin in 1 See Chapter XVI. p. 317. 376 STUDIES IN STATISTICS CHAP rheumatic fever, rather than in inflammation induced by cold? The numbers are, however, too small to draw any very positive conclusions. Turn now to the lower portion of the diagram. It seems more natural, in studying a winter disease, to adopt a method of dividing the year which shall avoid the inconvenience of cutting every winter in half, The plan that commended itself to me as most convenient for this purpose was to make the year commence with the present fourth quarter, or October 1st, a period that very nearly corresponds with the first rise of the curves of winter diseases from the summer minimum. This method is applicable to the statistics of London only. The death-rates from phthisis, bronchitis, and pneumonia are given in the same colours as above, and resulting curves will be found to be very similar to those for England and Wales. The black curve indicates the number of cold days at Greenwich in each year—calling a cold day one on which the mean daily value of the temperature of the air was 40 deg. Fahr., or lower. This curve resembles those for bronchitis and pneumonia in many respects, but not so closely as might have been expected. Several measures of the coldness of the winters have been tried, but the one adopted corresponds more closely than any other with the mortality due to the two great inflammatory diseases of the breathing organs. The next diagram (Plate XXIV.) shows the weekly fluctuations of the deaths from bronchitis and pneumonia in London during the past five winters—three of exceptional severity and two very mild. The death-rates have not been calculated, but the curves all express fluctuations from the mean weekly number of deaths for the five years ; each small division of the vertical scale in this diagram corresponds to twenty per cent. of excess or defect from the mean, so as to economize space. The shaded squares LONDON WEEKLY DEATHS FROM BRONCHITIS & PNEUMONIA. Xcess or Defect per cent. compared with the mean of the 5 years [878-82 ALSO COLD AT GREENWICH BRONCHITIS ! veekly average 219 1878 ocT./Nov.] DF = TI [MARCH ar. MAY | JUNE [JULY ! AUG.| SEP, PNEUMONIA veekly average 86 - 1879 10° 0" } MLTR Eo STE RE ELTA irr ilaenl wiser mma LD ROLLIPY VESTAS (ENS 1879 - 1880 10°F: S0°F-1880 ~ - 1881 10° F 50° F - i188] - - 1882 20° f 50° 1 1882 - 1883 40°F 50°F = B.|MARCH | APR. | MAY | JUNE July] AUG. | SEP. | ’ - STANFORD3 GEODG* ESTABT LONDON ach small division of the vertical scale corresponds to 20 per cent. of excess or defect from the Mean. he grey shaded squares indicate mean weekly temperature at Greenwich below 50°F. he black squares in like manner indicate mean temperature below 40°F. Each division of vertical scale corresponding to 5°F. Thus the mean temperature of the first week of November 1878 was 41; that of the third week of December was 29°F. LONDON AVERAGE WEEKLY DEATHS FROM BRONCHITIS , PNEUMONIA, PHTHISIS & CERTAIN OTHER DISEASES. Excess or Defect. per cent.compared with the mean AVERAGE OF 30 YEARS 1845-74. MEASLES J! cr Thea N AVERAG “ARCH | APR.| MAY | JUNE JULY | AUG.1 «EP. [0CT.INOV.| DEC, oot actly Be asic — -— 8) AL = WHOOPING COUGH Spal A Te ym ta "UBERCULAR MENINGITIS PHTHISIS BRONCHITIS RTH MA PNEUMONIA SLEEP SY RHEUMATISM JAN.| FEB.|MARCH|APR.I MAY JUNF UG. — 44.G — + pw —148.! 13-0 123 5 7 1 1 —7 DB 3 - a on A. OCT. (NOV. | DEC. LONDON, STANFORD'S GEOG: ESTAGT tach small division of the vertical scale corresponds to 10 per cent of excess of defect from the Mean xviit ~~ PHTHISIS, BRONCHITIS, AND PNEUMONIA 377 indicate the mean weekly “cold” (if one may be allowed the expression); the pale shading corresponding to temperatures between 40 deg. and 50 deg. Fahr., or what is usually called “raw” weather; the black to temperatures between 20 deg. and 40 deg., or “ frosty ” weather. It will be noticed (1) that the pneumonia curves (black) correspond generally with the bronchitis curves (red), but deviations from the mean are much smaller in the case of the former. (2) While the pneumonia curves rise in the autumn as rapidly as those of bronchitis, they subside more slowly in the spring (8) Long-continued cold raises the mortality, from both causes, more than lower temperatures lasting a shorter time. (4) In the spring of 1879 there was prolonged high mortality from both bronchitis and pneumonia out of all proportion to the severity of the cold. (5) In the winter 1878-9 the correspondence between the two curves is much less than in the others. It should be explained that the extraordinary mortality in February 1880 and 1882 followed intense frost that was accompanied by dense smoky fogs. But I have not been able to prove that fog per se has the effect that might be expected. Lastly, the two dotted lines are the phthisis curves during the two severe winters 1878-9 and 1879-80. It will be scen at once that the severe cold had an insignificant effect upon the mortality. The curves are, however, much more like the pneumonia curves than those of bronchitis. The next diagram (Plate XXV.) is occupied with curves traced from Messrs. Buchan and Mitchell’s figures, and applies to averages of the thirty years 1845-74, Measles and whooping-cough are seen to have very characteristic curves, quite dissimilar to each other, and I may add all others. Tubercular meningitis is again seen to be allied to phthisis, but tabes mesenterica is a 378 summer disease mainly. Asthma and bronchitis are closely allied. Pneumonia looks like a compound curve having a secondary spring maximum.! Pleurisy and rheumatism are very similar, but, owing to the smallness of the numbers, both are very jagged. Pneumonia is one of those diseases in which the question of sex has to be considered when discussing its stiology. In equal number living, for every 1000 females that died in 1880 :— From all causes .. From phthisis . From bronchitis ... “ry . From pneumonia, no less than ... STUDIES IN STATISTICS CHAP. “.131 males died. RH Lis , 1.460 .. . This great difference points to occupation and habits. [ntemperance is no doubt a factor, but the difference between bronchitis and pneumonia is very striking. The returns of the Registrar-General's Department necessarily deal only with futal cases. Bearing this in mind, the following table, based as it is on more than twenty-four thousand cases, has a value which admitted ‘naccuracies in the returns cannot destroy. The figures are obtained from the Annual Reports for the several years, and from the census of 1881, vol. iii., and show the following facts for males and females separately :— (1) The number of deaths from pneumonia at all ages and at various groups of ages for each year from 1871 to 1880, both inclusive. (2) Do. do. do. for the decade 1871-18R0. 1 Several other curves exhibit a deep notch in the month of February, the mortality rising again afterwards. This to my mind corresponds with the common experience that the raw weather and parching east winds of March are more trying to the constitutions of many people than the lower average temperature of February. Ages 0 Deaths per I dillon living Shi 10.000 - 3.000 8.000 ah 7 O00 1 23.000 5.000 4 000 ir Tr 3.000 I 2 000 - | | O00 1 p 10 TTT NGL YT WALES. *VERAGE DEATH RATES TROM PNEUMONIA. DECAY 1871 '88¢. 40 50 60 0 20 TTT rT 80 20 Years = Deaths per Mitlion living *0.000 93.000 ] 3.000 bt 5 7.000 5.000 = 5.000 E 4.000 Males 3.000 IHFemales 2.000 000 Los CITY ; . Ty. roo Yoo I 1 London. Stanford. Geog b Estab® 0 : «Jales. 1871 .., ee I. 1872 ... es ve 1873 ... ee we 1874 ... cr soe 1875 ... wos a 1876 ... - i. 1877 ... - _ 1878 ... =... viv 1879 ... . cee 1880 ... wg ve Total ve Mean annual rate per 2 1,000,000 living at each age § All Ages. 12,776 1,631 3,233 14,999 15,968 14,085 13,672 14,131 15,721 14.606 140,822 1.188 Under 1 | 1— 4,213 3,699 4,181 3,873 4,032 8,779 3,535 3,848 3,780 3 995 2,016 1,861 1,914 2,062 | 2,056 1,925 2,050 2,106 2,182 2.055 38,165 20,221 10,5771 6.334 Deaths from Pneumonia in England and Wales fe under 5 9_ 756 837 213y7,535° 877 126 672 315 196 | 6,743 286 85 694 ' 349 190,7,328 303 119 841 362 | 248 ! 7,386 | 424 152 731 866| 221,7,406| 8921 153 712 344 | 230 6,990 | 349 140 710 360 | 219 6,874 836| 131 7591 420) 219 7.361 400| 127 822 | 409' 252 7,445' 460] 160 707" 417 217 6.621 484! 134 189 i51 197 266 271 233 | 199 | 297 272 org 232 255 288 385 473 359 340 310 <0] 494 ane 150m 2,205 224911149' 696 1,689, 3,791 4353 | 260 1,327 101 2,28¢ 194 «,458 335 25— 85 45 — 55. 65— 673 741 951 639 865! 907 750 1,016 1,097 933 1,255 1,448 1,101 1,464 1,565 970 1,194 1,240 881 ' 1,210 1,265 801 1,127 1,216 969! 1,418 1,53¢ 948 1.391 L.45¢ 891 761 1,004 1,397 544 ,302 206 1,220 473 {444 732 642 792 978 127 906 888 917 1,092 1 020 8,665 11,681 TR a A ' 5081 882'1270|1.796 2.441 75-— | 85 286 262 301 325 407 349 289 364 458 301 ‘1 - | 35 °8 a on 41 3,432 9.731 476 1 3.265 4 3 : , Females. 1871 ... Ho . 1872 ... . J. 1873 ... o. oo 1874 ... we we 1875 ... wre Io 1876 ... won ane 1377 ... why Po 1878 ... wos Po 1879 ... vue 35% 1880 ... sus a. Total Le ve Mean annual rate per ) 1,000,000 living at each age 9,992 8,651 9,671 10,998 11,193 10,407 10,024 10,632 1,453 10.560 103,511 828 3,221 2,743 2,959 2,965 2,849 2,794 2,664 2,786 2,634 9 407 1,914 ,675 1,782 1,816 1.815 JT :,830 1,920 1,919 1.8”R0 7421 546 647 791 714 729 704 809 752 2 369 347 305 434 374 350 350 384 424 209 232 184 186 246 208 238 02 253 «B55 9255 6,478 5,595 5,879 6,252 5,960 5,588 5,741 6,152 5,987 5 644 370 317 317 383 391 344 369 | 378 441 449 130 97 138 169 43 | 61 43 143 133 165 181 208 221 168 167 189 229 284 27 389 369 465 578 730 R29 417 378 492 643 699 460 ann 132 372 508 61° ie “6! 566 483 603 758 300 Je 581 48] 591 732 312 40 nf 142 152 eo TA T Neds "0 LOS 794 7 a0 28,022 | nas Sa | Bees S78] alee 1,420 Sh aa 6,001 sod 0004 1,08 [302 7,800 | 5,723 | 2,201 11551 71113613] 256! 109 | 1501 192! 286! 419) 516] 921 1.622 (2.088 280 226 280 340 380 315 322 387 379 203 38 a3 44 ‘0 1 9 7 0 528 2 962 IN . . 1 2 ‘te,—This Table and Plate XX VI. appeared in the Collective Investigation Re" ~ 380 STUDIES IN STATISTICS CHAP. (8) The average annual death-rate per million living at each age (calculated on the mean of the two censuses, 1871 and 1881). It should be noted that males suffer more than females, almost in the proportion of 3:2. But the disparity is slight at ages 2—15, and is most marked at ages 35—65, when males suffer more than females in the proportion of 2:1. The greater mortality of male infants can only be dependent upon that difference in the male constitution which causes the well-known excessive mortality of males under 2 years of age. The greater mortality in middle life may conceivably be due to sexual differences, or more ikely to occupation or drinking habits. In both males and females, deaths from pneumonia are proportionately least frequent at ages 10—15, increasing on either side of that point to the extremes of life— but the mortality is three times as great, according to the return, in the first year of life, as in old age. The facts are more clearly shown in the diagram (Plate XXVI.), in which the continuous curve represents the male death-rate, the dotted curve the female deathrate. For years 3—15 they are practically coincident. In noting the varying mortality from year to year it should be borne in mind that the winters 1870-1, 1874-5, L878-9, 1879-80, 1880-1 were severe, The Influence of Geographical Distribution. If the mortality in each registration county, from sronchitis and pneumonia respectively, be compared with he like mortality throughout the country, it will be found chat there are ten registration counties in which pneumonia is comparatively much more fatal than bronchitis :— 1 More than 15 per cent. XVIII PHTHISIS, BRONCHITIS, AND PNEUMONIA 381 Mean Death-rates tn Counties per million living, 1876-80, and percentages of the Rates in England and Wales. BRONCHITIS. Sounty. Lancashire ... _ Metrop. Surrey ... Metrop. Middlesex West York - Warwick ... ses Monmouth ... oh Nottingham ve Stafford ... ves Cheshire ... m Cumberland res Metrop. Kent vor Derby wr rr Leicester .. wo North York ov Wilts 5 ie Durham ... aos Somerset ... ve Devon ve ve North Wales “e Worcester ... ee East York ... _ Northampton ary Lincoln wn ce. Dorset oe ve Middlesex, Extra-Met. Hssex - _ South Wales - Hereford ... re Northumberland ... Huntingdon A. Oxford ce ce Shropshire ... on Buckingham “e Hants io _ Berks a oi Rutland ... — Hertford ... oy Bedford ... wn Westmorland on Kent, Extra-Met, ... Suffolk -_ vow Cornwall ... oe Sussex a oh Norfolk ... en Surrey, Extra-Met. Cambridge ... i, Gloncester ... . England and Wales ee Rate. yi 3 4 oo DA Ay ¢ 3460 3241 2997 2913 2891 2718 2662 2550 2506 2463 2423 2375 2196 1182 2177 2177 2128 2057 2053 2038 1994 953 936 928 918 L918 906 901 891 369 862 858 840 | 790 771 768 ~64 07 700 668 664 627 € 1620 68 i578 6 542 65 1462 61 1263 53 9388 | 100 145 132 12¢ 12¢ 121 114 111 107 105 102 103 90 a n a" 2 PNEUMONIA, County. Lancashire ... we Monmouth ... i South Wales vo West York... au Stafford ... vay Metrop. Middlesex Metrop. Surrey ... Rutland ... ore Metrop. Kent Ai Leicester ... ce Derby wo vou North Wales or Devon -_ on Warwick ... og Tuntingdon a Worcester ... po durham... _— Northampton ao Sloucester ... oe Nottingham wih East York ... sod Bedford ... wn “heshire ... wo Jornwall ... po Surrey, Extra-Met. Middlesex, Extra-Met. Kent, Extra-Met. ... Berks wr -— Jertford ... oe Shropshire ... wt Cumberland a Essex aes 5 Dorset i. “5 Cambridge ... Zon Suffolk .. ih Northumberland ... Lincoln ... po Norfolk ... on Somerset: ... fh Tereford ... ren Sussex es vee Wilts ve er Tants _ _ Suckingham we Jsford _ oe North York web Westmorland ve England and Wales ses sen von cea see see van aes 45 Rate, £8 1895 1330 254 1237 292 197 131 078 69 048 987 959 958 754 335 93¢ p27 92% bal 31° NT 309 "97 72 “54 340 329 314 306 303 303 796 "81 78 78 6f AF "3r “98 of 580 869 868 350 522 207 41 139 132 125 123 122 119 “12 of °4 LC Cc” ow c To 87 Cn < ups | ¢ 67 €7 65 62 59 h4 100 1004 Cr )-STUDIES IN STATISTICS CHAP, Gloucester, South Wales, Rutland, Extra-Metropolitan Surrey, Cornwall, Bedford, Cambridge, Huntingdon, Extra-Metropolitan Kent, and Monmouth. There are ten registration counties in which bronchitis is comparatively much more fatal’ than pneumonia :— North York, Wilts, Westmorland, Cumberland, Warwick, Oxford, Somerset, Nottingham, Metropolitan "Surrey, Cheshire. = [ am by ne means anxious to force conclusions from the meagre facts thus laid before you; but, nevertheless, [ venture to propound the following propositions for liscussion :—1. The mortality statistics of England and Wales throw no light upon the question of the communicability of phthisis,” but the evidence, so far as it goes, is negative, 2. Cold weather has but very little influence on the plithisis' death-rate. 3. The bronchitis death-rate is greatly affected by temperature, that of pneumonia also, hut not quite so markedly. 4. It is difficult to explain, by meteorological phenomena only, all the fluctuations of the mortality from bronchitis and pneumonia. 5. The diagrams do mot give any support to the theory of pneumonia being a disease allied to typhoid fever. 6. Although the facts do not exclude the possibility of a sommunicable pneumonia having no connection with cold, they show that such a theory can account for but a very small proportion of the deaths attributed to pneumonia.’ 7. While the mortality from pneumonia, and more 2specially from bronchitis, appears to be mainly regulated hy intense cold, by fogs, and the cold winds of early spring, the occasional want of correspondence between the assumed causes and the alleged effects suggests that 1 More than 15 per cent. 2 The question of the communicability of phthisis is treated in considerable detail, from another point of view, in the next chapter. 8 See Appendix, p. 443. xvil ~~ PHTHISIS, BRONCHITIS, AND PNEUMONIA 383 possibly bronchitis and pneumonia, or many of the cases so designated, may be communicable, or they may be complications of a communicable disease, more or less fatal according to the prevailing meteorological conditions. 8. Many domestic incidents suggest that common catarrh, or some form of it, is more or less. communicable. Is this communicable catarrh, assuming it to exist, a mild form of influenza ? and are many cases of bronchitis and pneumonia to be regarded as serious complications or sequelae of that neglected but most troublesome ailment ? 9. The fact that whereas, in equal numbers living, 1,103 males die from bronchitis to every 1000 females, but as many as 1,460 males die from pneumonia to every 1000 females, suggests that the two diseases have different causes. 10. The curves suggest a common factor in the two diseases phthisis and tubercular meningitis. This inquiry has been less productive of positive results than I at first expected, but if it contribute, either to clear away false notions, or to indicate lines of future research, it may be labour not wholly lost. 384 STUDIES IN STATISTICS CHAP. CHAPTER XIX. A CALCULATION OF THE PROBABILITY OF THE ACCIDENTAL AND FATAL INCIDENCE OF PHTHISIS UPON BOTH HUSBAND AND WIFE. A FEW years ago the attention of medical men was salled to the frequency with which such an observation 1s the following is made :—A. B., a healthy man, to all appearance free from any phthisical taint, marries a consumptive wife, C. D.; after a year or two C. D. dies; A. B. is found to be phthisical at the time of his wifes Jeath, and dies of the disease after a short interval. Several collections of such cases, many of them very striking, were published, and the inference was drawn that they only admitted of explanation on the hypothesis of infection of the man by his wife, or vice versd. On the other hand, it may be maintained that phthisis is a very common ailment, and especially fatal between the ages of twenty and fifty, and that consequently, as » mere matter of chance, such cases must occur more or ess frequently. As the matter is one the importance of which cannot oe exaggerated, it seemed worth while to attempt to ascertain from the materials at hand how frequently such ! The substance of this paper was written at the request of the Secretary to the Collective Investigation Committee of the British Medical Association, and appeared in the Collective Investigation Record, July 1883, XIX COMMUNICARBILITY OF PHTHISIS 385 coincidences might be expected to occur as a pure matter of chance, on the hypothesis that phthisis is not a communicable disease. The following calculations have been verified by three competent mathematicians, one of them an actuary of experience. If the results are anywhere near the truth, they prove that a far greater number of coincidences of the deaths of both husband and wife within a short interval from phthisis, than has as yet been brought forward, would be required to prove that the ome had contracted the disease from the other by infection. The available dato suggested the following form for the problem— Required—To find in how many cases in England and Wales during the decade 1871-80 both husband and wife would die of phthisis, assuming the chances for married and single to be equal in all respects. Let the mean number of married males living in the period 1871-80 be M. Then from the census we get— M = 4,130,131. These numbers are also given in the census reports for seven aces. Viz. — At age 15— 20— 25— 35— 45— bb— BRR. 5,906 = py, 239,332 = pu, 140,048 = yu, 093,663 =p, 830,395 = us 528,442 = yy 200 345 =p. Sothat M = wu, + ws Similarly, let the mean number of married females living in the same period =F. He 1 NY 386 STUDIES IN STATISTICS . CHAF. Then— F — 4,193,245 and— F—M = 63,114 showing a considerable excess of wives over husbands, which is probably to be explained by («) husbands absent on or beyond the seas, (b) deserted wives, the men returning themselves as single, (c) single women, for various reasons returning themselves as married. To find the age distribution of the M wives corresponding to the M husbands, the census numbers giving the age-distribution of all the married women in the country must be adjusted by multiplying by the factor . i. = « = ‘0849049. That is to say, it is assumed that the errors arising from the various causes given above affect the census numbers at all ages equally. Thus :— 15— 20-35- 381,668 £63,439 09a.’ ofp 202. ~ 32,991 = ¢ 375,924 = ¢, ,244,422 = os 1,076,919 = ¢, 764,360 = ¢; 435,279 = ¢s 200,236 = o, Qothat M = « F = ¢ + Now we get from the Registrar-General’s Supplementary Report for the ten years 1871-80, the mean numbers living and the deaths from phthisis in the period. (See Table on next page.) These figures include the single, married, and widowed : by our hypothesis the single and married are (as regards phthisis) equal in all respects. If there are m males living during a certain period, and XIX 15— 20— 25 35— 45— | 65 — COMMUNICABILITY OF PHTHISIS 387 MALES. FEMALES. Mean No. living. Dying of Phthisis. Mean No. living. Dying of Phthisis. Say 1,176,491 my, 1,082,135 my 1,705,568 ' my 1,324,405 ' m, 997,999 ! my, 681,568 | me 512.991 | m, 19,723 31,927 63,070 54,549 38,532 21,774 8.015 Ser dy dy d-¢- A. 1,187,331 1,134,358 1,861,449 1,431,548 1,087,587 755,902 618.702 Sav I J I 98,459 35,617 | 65,973 18,663 26,786 13,436 5.511 Say 5 . 3 rf during the same period d males die of phthisis, the prooe : . d bability of any one of the m males dying of it — ~ But we are dealing with M males, of 7 ages, Viz. 15—20 = p;, 20—25 = p,, &c., the probabilities of dying oi . dd of phthisis at these ages being =, 2, &e. my me .*. Of these M married males the number who will probably die of phthisis within the assigned period is— _ d 1 a 2 ad = Rt te + po +... + ey d But 2 = ‘01676426 Multiplying by p, = LL = 03093297 my As = +03697888 ns ds = 04118755 m 4 2s = -0386092¢ ms ds — = "03194692 mg ar = 01562406 YY) Total = X = 148,109 Le. Married males who will die of Phthisis during the decade. 388 STUDIES IN STATISTICS CHAP. Corresponding to these X husbands who die of phthisis sre X wives. The age-distribution of the X husbands Joes not differ materially from that of the total married males. We may therefore safely assume the same agedistribution for their wives as for all wives. The number of wives alive at the respective ages may then be assumed to be proportional to ¢,, 9. . . . . #5 and to form a series Xs Xz « «00 XT nn... Tia 2 = F MN X X x - = 3p Pr des de xr = 3F $v + X + And x, + 3 Fo. +0 CM {¢: + ¢2 Cee @r) "M =X Buy X _ 148,109 WNT T 4130131 = ‘0358606 = A Ad = X1 = 1,183 Ap, = x» = 13,481 Ags = xs = 44,626 Ng = -. = 38,619 Apr = 27,410 875 ~ 15,609 \ 0 + 7,181 Tetal = 148,109 = X We are now dealing with X females of 7 ages, the proba-Mad . id 0,0, bilities of dying of phthisis at those ages being = rx &e. The number of these X females (wives of X husbands sho die of phthisis) who will probably die of phthisis within the assioned period will be in like manner— 31 Os 37 Y = _— 4 _— se. — X17 X27 + xr 7 XIX COMMUNICABILITY OF PHTHISIS 389 dy But — I = 02396888 Multiplying by x1 = 28:36 3, 03139838 Xz 492398 = 03544174 Xs 158162 - 03399327 xa = 131279 a — 02462883 Xr 67508 6 IA — 01777479 Xe o7'7-45b 85 _ 00890736 fq X71 = Total = V = 4.363 63-96 The closest approximation to the number required that the data admit of. The above calculations admit of being stated in the form of tables which, as Mr. Waters has suggested, may appear less repulsive to non-mathematical readers. Thus :— ENGLAND AND WALES.— Males. Ten Years, 1871-80. 10¢, 15— 20— 25— 35 — 15— 55— BE Mean number living. 1,176,491 1,032,135 1,705,568 (,324,405 997,999 681,568 519.991 Dying of Phthisis in decade. 19,723 31,927 63,070 54,549 38,532 91,774 8.015 Number per million dying of Phthisis in decade. 16,764 30,033 36,979 11,188 38,609 31,947 15.624 Then assuming the same phthisis death-rates for married males we get :— 390 STUDIES IN STATISTICS CHAP. ENGLAND AND WALES.—Married Males. Ten Years 1871-80. Age, Mean number living. 15— 20— 25 — 35 — t5— 25 — BH— 5,906 232,332 1,140,048 1,093,663 830,395 528,442 999.345 Total | 4,130,131 Dying of phthisis in decade (at rates in table above). 99 7,187 £2,158 15,045 39,061 16,882 4.677 148.109 ENGLAND AND WALES.— Females. Ten Years 1871-80. Age. 15— 20— 25— 35— 45 55-— AR. Mean number living. 1,187,331 1,134,358 1,861,449 1,431,548 1,087,587 755,902 618.709 Dying of Phthisis in decade. 28,459 35,617 85,973 18,663 26,786 13,436 5.511 Number per million dying of Phthisis in decade. 23,969 31,398 35,442 33,993 24,629 17,775 8 907 Then since the age-distribution of the 148,109 husbands who die of phthisis is not very different from that of the total 4,130,181 husbands, we may fairly assume that the age-distribution of their 148,109 wives is not very different from that of the total 4,193,245 wives. ‘See Table next page.) So that, in accordance with the data and the various necessary assumptions, we may state that, during the ten years 1871-80, sn England and Wales, assuming marriage to have no influence upon phthisis, by selection, infection, or otherwise, it would happen 4,363 times ‘that both husband and wife would die of that disease. X1X COMMUNICABILITY OF PHTHISIS 391 ENGLAND AND WaLEs.—Married Females. Ten Years 1871-80. Age, Age-distribution of 4,193,245 wives (mean of Censuses 1871 and 1881). Age-distribution of 148,109 wives whose husbands jie of Phthisis in decade. Wives in column II. dying of Phthisis in ‘ecadeat rates in previous tahla 11 (5— 33,495 20— 381,668 25— 1,263,439 35 1,003,376 5 776,040 s5— | 441931 65— 203.296 Tot~1 ; 10% 1,183 28 13,481 423 44,626 1582 38,619 1313 27,410 675 15.609 | 218 7181 64 TAQ QO 4363 Next as regards the interval between the deaths of husband and wife :— Tt may be assumed that the deaths from phthisis were spread evenly over the ten years; and to simplify the calculations, we will assume for a moment that the double event we are concerned with occurred 1000 times in the decade, ¢.e. that 100 husbands and 100 wives died of phthisis in each year. Now 100 husbands died in the first year ; of the wives of these husbands 10 (ex hypothesi) died in each of the 10 years. Those wives who died in the first year, and half of those who died in the second year (10+ 5==15), died within one year of their husbands. The other half of those who died in the second year, and half of those who died in the third year (5 + 5==10), died within an interval of between ome year and two years of their husbands ; and so on. Finally, half of those who died in the tenth year died between mine vears and ten vears of their husbands. Again, 100 husbands died in the second year. All 392 STUDIES IN STATISTICS CHAP. the wives who died in the same year, and half of those who died in the first year, as well as half of those who lied in the third year (104+5+5=20), died within one year of their husbands. The other halves of those who died in the first and third years, and half of those who died in the fourth year (5+5+5=15), died within an interval of between one year and two years of their husbands; and so on. And finally, half of those who lied in the tenth year died between eight years and nine years of their husbands. By similar reasoning as to the 100 husbands who died in each of the remaining years, the following table is -onstructed,! in which it will be noted that the figures in the last line but one form a series 10{20 —(2n—1)}. Wives who die tn each Interval from Husbands Deaths. INTERVAL. Year of decade. 1. Huspands | vho die] in year. 100; 1s 100; 20 100 20 Loo} 20 100] 20 100] 20 100] 20 nol an nab Within 1 year | 2 years | 8 years 4 years | 5 years | 6 years 7 years | 8 years | 9 years | year. | and and and and and and and and and Year. | under 2 | under 3] under 4| under 5 | under 6 | under 7| under 8 | under 9 funder 10 l 5 TA 3 In 10 3 101 5 5 ot 2 20 20 20 20 20 20 J = ) ) an 10 10 TU i) 29 20 20 a 0 lv 15 ir n 5 4 \ } — _— 10 | 10 . h —_— il 110 10 101 5 — YU J WJ 10 5 Nl Vv 1101 90] 701 501 30] 10 742 | 654 | 567 | 480 | 393 | 305 | 218 | 131 | 44 9 t The above method of stating the argument as to the interval between the deaths (as well as the following table) is due to Mr. Waters. XIX COMMUNICABILITY OF PHTHISIS 393 The last line of the table gives the actual number of wives dying in each interval, obtained by substituting for 1000, the number found in the previous process. viz. 4,363. Exactly the same results would of course be obtained by calculating from the wives to the husbands. If there be N. medical practitioners in England and Wales, and n of these send in returns, then (assuming no infection, and assuming that those who send in returns have an average share of practice) they should have observed during the ten years a number of cases equal to 7 of the above numbers, though presumably they would only record a part of the cases observed. According to the Medical Directory for 1883, N = 15,462. But within the decade both husband and wife would die of phthisis, within a twelvemonth of each other, in 829 instances. If, therefore, 500 medical practitioners sent in returns as to their experience, they should have observed collectively about 27 such cases. This is the * probable ” number that would have been observed, but quite apart from infection the number actually observed might be greater, since the number of practitioners and number of years are not very large, and the coincidence is a somewhat uncommon one. Again, marriage may have some effect on phthisical mortality otherwise than through infection (e. g. sexual selection might be imagined to have a notable effect). It is plain, therefore, that to show any substantial argument for the existence of infection it would require a much larger collection of cases than has vet been published. 1 This number does not include army and navy medical officers, but on the other hand it does include many retired medieal men. 394 STUDIES IN STATISTICS CHAP. CHAPTER XX. I'HE SEASONAL PREVALENCE OF CONTINUED FEVERS IN LONDON.! Bur fifty years ago the various species of continued fevers were commonly confounded together under the name of typhus, and, as a natural result, different writers pave most conflicting accounts of the symptoms and course of the disease, and likewise of the morbid appearances found after death. The honour of disentangling this sonfusion is mainly due to Gerhard and Pennock (1837), Stewart (1840), and Jenner (1849). In 1854, Murchison entered the arena, and made sarious contributions to the voluminous literature »f the subject; his classical Treatise on the Continued Fevers of Great Britain was published in 1862, and a second edition appeared in 1873. In this work, and in his lectures, Murchison laid great stress upon the etiological distinctions between the different fevers, and exhibited with this view, in a series of tables and diagrams, the weighty statistics. of the London Fever Hospital, which showed, among other points, that as a rule typhus and enteric fever prevail at different seasons of the year. It was, however, not till 1869 that the various forms of L Originally read before the Epidemiological Society of London, April Rth, 1885, CONTINUED FEVERS IN LONDON 395 continued fever, previously classed together under the head of “typhus” or fever,” were distinguished in the returns of the Registrar-General. In Table I are shown the deaths in London from « foyer” during a period of forty-eight years, taken from the TasLe L— Deaths from Continued Fevers in London ; also Death-rates per 1000 living, from the same diseases. in each of the forty-eight years 1837-1884. Year, | Deaths. 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1859 19331 1078 1819 1262 1151 184 2094 1721 1324 1838 3297 3685 2564 2032 2374 2183 boi, Ja) vears 134,539 | Mean 2159 Rate | per 1000. 2121 2°31 1°01 | 69 062 062 1°08 0'83 0°64 0°87 1°47 | 1°65 113 088 1°00 0'a0 1.10 | Year. Deaths. Rate per 1000. 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 L864 1865 L866 1867 (RAK 2617 2816 2460 2717 2195 1919 1840 1476 1848 R673 2871 3782 3217 2688 2184 SARK 1°06 1°12 097 105 | 083 o'72 068 0°53 0'66 128 ©°'99 1°28 10% 088 071 0°79 ‘otal, 16) | years 40,771 Mean ora8 | 0-91 Vear. 1869 [870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 otal, 16' vears Mean | Deaths Rate per 1000. 2400 2018 1691 1303 1510 1528 1217 1130 1252 1381 1080 910 1197 1123 1120 1035 076 063 0°52 ©'39 0°45 0°45 0°35 0'32 0°34 0°37 0°29 0°25 0°30 028 028 Anh 21 895 1368 | 0°39 Annual Summary for 1885 ; also the death-rates per 1000 estimated to be living in London in each year. It is at once seen that, in spite of the enormous increase of population, the absolute loss of life from * fever” 1s less now than it was ten years ago ; much less than it was twenty, thirty, or even forty years ago. When allowance is made 1 Qix months only. 396 for the increase of population, it appears that the mortality from this cause has, during the last sixteen years, been but 35 per cent. of what it was during the sixteen years ending with 1852. The decrease has been almost uninterrupted since 1864. In Table IL are shown similar facts from the same STUDIES IN STATISTICS CHAP. asLE IL.— Deaths from Typhus, Simple Continued, and Enteric Fevers in London, and also Death-rates per 1000 living, from the same diseases, in each of the sixteen years 1869-1884. Year. Typhus. Deaths. . Rate. 869 716 [870 472 1871 384 1872 174 1873 R77 L874 312 1875 128 1876 159 L877 157 1878 151 L879 71 .880 "4 1881 92 1882 | 63 [883 55 1884 32 023 o'15 012 0°05 0°08 209 0°04 | 0°04 0°04 0°04 2°02 2°02 0'02 o'oT o'0I o'01 Total ...| 3,307 Mean ... | 0:06 | Simple continued. Deaths. Enteric. Deaths, Rate. Rate. 815 570 136 3922 325 337 272 202 194 197 160 134 134 95 102 | 78 °'19 0°18 0°13 | 0'10 010 2°10 0'08 n*06 2°05 0°05 2°04 0'04 2°03 0'02 0'03 0'02 1,069 976 871 807 908 879 817 769 801 1,033 849 702 p71 B75 p63 095 0°34 0°30 o'27 0°24 0°29 0°26 0°23 0°22 0°25 0°28 0°23 °'19 0°25 025 024 0°23 1173 | 14.415 961 | 008 | 900 095 sources, but under the three headings first adopted by the Registrar-General in 1869, viz. Typhus, Simple Continued Fever, and Enteric Fever. The first point to notice is, that during the sixteen gears typhus was made accountable for 3,307 deaths; simple continued fever for a considerably larger number, XX CONTINUED FEVERS IN LONDON 397 4,173 ; but enteric fever for nearly twice as many deaths as the other two diseases put together, viz. 14,415. Secondly, whereas in the earlier years typhus was accredited with about the same number of deaths as simple continued fever, and enteric fever with numbers only half as large again, latterly the proportions have been very different : typhus now numbers only half as many victims as simple continued fever, whereas enteric fever claims ten times as many. In other words, the deathrates per 1000 show that whereas the relative mortality from simple continued fever is but one-tenth of what it was, and that from typhus only one-twentieth, the fall in the enteric fever rate amounts to only 30 per cent. The great diminution in the fever mortality of London is, then, mainly due to the diminished prevalence of typhus and simple continued fever, while the mortality from enteric fever has remained for ten years comparatively stationary.’ Typhus is a definite and well-known disease ; it is far otherwise with simple continued fever. Murchison says, ¢ Simple continued fever, when uncomplicated, is rarely, if ever, fatal in this country. The numerous deaths from simple fever recorded weekly by the Registrar-General are due for the most part to enteric fever with latent abdominal symptoms.” Again, “ Many cases are designated Simple Fever or Febricula, which are in reality mild or abortive cases of typhus or enteric fever, or relapsing fever without the 1 More accurate diagnosis and increased ‘precision in the use of terms have doubtless to some extent increased the recorded mortality from enteric fever at the expense of typhus. The seasonal prevalence of the two diseases as registered shows that this cause cannot have heen so efficient as might have been supposed. 3 Continued Fevers (2nd ed.), p. 682. 398 relapse, or catarrh with an unusual amount of febrile disturbance.” ! STUDIES IN STATISTICS CHAP. Again, ‘“ Medical men often decline to call a fever ‘enteric,’ in which, as often happens, there are no enteric symptoms, and hence the intestinal lesion is apt to be overlooked until it unexpectedly becomes a source of danger. This mainly accounts for the circumstance that in the returns of the Registrar-General deaths are weekly ascribed to ‘Simple Continued Fever,” a disease which in swenty-five years has not once been fatal in the London Fever Hospital.” ? Again, he states that the cases of Febricula admitted ‘nto the. London Fever Hospital during the twenty-three vears 1848-70, amounted to 2,232.° On the one hand we have the emphatic and reiterated statement of a deservedly respected authority, founded upon vast experience; on the other the opinion of the large body of medical men in London, who certified that 4,173 deaths in sixteen years were caused by this disease— simple continued fever. It is now fifteen years since Messrs. Buchan and Mitchell, in the Journal of the Scottish Meteorological Society, compared the seasonal prevalence in London of the three forms of continued fever. The curves they obtained were based on but six years’ results. Concerning typhus they remark, “It is, however, a curve which calls for clearer definition from future observations, though it seems probable that it has two maxima, the larger in the early months of the year, and the smaller in the height of summer.” Of simple continued fever they say, This curve . . . is not very satisfactory, and further observations are required to define it with greater precision.” L Op. cit., p. 678. 2 Ibid. p. 420. 8 Ibid. p. 684, table. XX CONTINUED FEVERS IN LONDON 399 More facts are now available, viz..— (1) Monthly admissions to the London Fever Hospital for the twenty-three years 1848-70.) These are available for typhus and enteric fever only. See Table IIL (2) Weekly admissions to the hospitals of the Metropolitan Asylums Board for the ten years 1875-84. From the returns supplied to the Registrar-General, and published in the weekly reports. These also do not give simple continued fever. See Table IV. (3) Weekly deaths in London during the six years 1869-742 See Table V. Tapre [IL.— Admissions to the London Fever Hospital in each month during the twenty-three years 1848-1870. January February March ... April ... May ... June ... July ... August... September October November December Total ... . Mean ... Jo Twentyhree years. 1,976 1,621 1,906 1,642 [,525 1,296 1,251 1,183 1,162 ,429 1,667 L610 18.268 1.599 Typhus. Average. Percentage of Mean Twenty-;hree years, 85°9 70°5 82'9 71°4 66-3 56°3 54°4 51°4 50°53 621 725 40'0 30 106 125 108 100 85 B2 78 76 94 110 108 433 306 318 209 232 335 434 721 303 839 819 5309 fhm oo 499 5.088 Enteric Fever. Average. Percentage of Mean 188 13'3 13-8 2 [0°I . 146 189 31'3 349 365 356. 22°4 7 1 ra Ae 18 RB? 7 44 161 168 164 108 217 100 1 Murchison, Treatise on the Continued Fevers of Great Britain (2nd ed.), pp. 66, 446. 2 Buchan and Mitchell, Journal of the Scottish Meteorological Society, Feb. 1875, p. 256. 400 STUDIES IN STATISTICS CHAP. TABLE IV.— Average weekly admissions to the Hospitals of the Metropolitan Asylums Board during the ten years 1875-1884. Wool st wy ind ©... 3rd... th... th... 8th... th... 3th... th... [Oth ,.. [1th ... 12th... 13th... 4th... 5th... 6th... 17th ... 8th... Oth... 20th... 21st... 22nd ... 23rd... 24th... 25th... 26th... 2th... 28th... 29th... 30th... 31st... 32nd ... 33rd... 34th ... 35th... 36th... 37th ... 38th ... 39th ... 10th ... st... 12nd... 43rd ... 14th ... 5th... 16th... 17th ... 18th ... 19th... 50th ,.. 5st... 32nd... Mean ... Month. January yy ees ” [I i. February ” oe og Ee 5 eee March ... yo eee ww Bees 13 pe April... 9 es 33 or » “re May ... 3p eee ww EE Be wow June ... so anh 2 ees 5p. seh i July ... 1g eee 3 ees Ww A August ... yo ees wo EI September 5 ees pu yee »”, ore Netober w te Been » “ts November EEA » Les » re December nee — 5 we yy ees Typhus. | Enteric Fever. Scarlatina. Percent-Average. age of Mean. Average. Percent ! Percent age of Average. age of Mean, | Mean. 16 . io | Liss | 2°0 3 '1 L150 | 36 87 77 3 55 7 2 8 ‘0 7 106 108 95 78 £9 L 2 24°5 239 22°3 216 259 19° 156 ils 192 172 “6g 69 ro - 2 19°4 7°2 7'9 77 189 18°5 16'9 235 220 232 148 237 256 25'4 “74 :69 280 297 289 34 62 43 Ly wr 81 79 287 {07 340 42'0 336 36'0 387 300 322 343 27°3 24'0 85 3c ° [@ 8 ‘ g 58 68 95 | 95 | 5 4 75 ‘2 A°4 4°0 3 33 7 9 17 ’ . 90 )} ) tl 46 i 5° 4% 41 9 ) 0 ‘6 60 3 48 5-8 56 8-0 68 70 77 90 06 | 38 27 168 71 182 19°3 181 15°1 36 26 270 qr i 70 29) vy _ 7 ~ ’ } J R) 1 95 1 31 71 57 208 212 225 239 224 187 168 i156 £0 45 LA ’ 1 . 85 J ) “5 27 ‘3 | 95 L140 3 50 "8 27 48 115 123 131 104 92 100 133 2-4 1 Kf 20 | 100 | 81 100 | 262 CONTINUED FEVERS IN LONDON 401 TABLE V.—Average weekly deaths in London, and percentage of the Mean. - SIX YEARS. 18e- 4 ¢ TEN YEARS, 1875-1884. Enteric | Simple Fever. continued. + 155 Fol ES 4 $y 2 £5 fw Ist + Jan. ° 2nd, 3rd +, ith 5th Feb. 3th » ith ” 3th 9th Mar. 10th ’s 1th vs L2th » 13th ' L4th April 15th |, 16th vs {7th i 18th May 19th |, 20th + 21st ’s 22nd June 28rd. ,, 24th | 25th ' 26th |, 27th = July 28th ’ 20th | goth | 31st Aug, 32nd | 33rd |» g4th | 35th Sept. 386th | 57th | 1) 38th or 39th 40th Oct. 41st n 12nd » t3rd 4 14th Nov. 15th | 16th » 47th ’ 48th Me: 19th 50th 51st 29nd 95 125' 180 107 141 78 8c 112 :7¢ 97 128 17° 77 101 173 83 109 175 '0°3 136 2172 97 128 160 ol 92 150 yy 88 7 3 117 1502 90 (19 180 78 108 160 ol 92 155 2108 14'5 ny 186 1570 “7116 1472 i 99 (3-2 )'105 130 er! 83 135 v2 94 11-8 ol 92 145 57 88 142 58 89 14% 52 68 28 58 86 12°3 7°, 99 128 8 180 138 7 75 157 C112 42 92 170 95 i4'r 95 20°C £5 192 13 223 83 16°35 79 193 £3 19'5 82 20°1 128 207 116 20°3 01 19°35 72 2775 99 232 0% 220 24°% 24°C a 8 .02 101 99 99 98 99 120 91 85 101 86 102 91 88 8¢ 85 81 75 1007 122 tro 125 £3 10 7-7] 88 97 110 27 110 82 100 132 150 83 04 93 106 05 119 9'0 102 2°0 102 88 .00 77 ]e ny Pl By go er 97 93 108 111 00 91 7€ 83! 82 “05 93 §9 89 70 83 76 105 122 100 03 91 1 74 38 02 £ 00 ‘0 91 9's 108 81111 ~ ey 7 26 19 20 2X 2°0 2°1 "3 [ad bi12 ! I -102 18:8 110" 17-2 101 162 95 176 108 14:8] 87 55. 91 66 97 41 83 32 1. ‘50 88 ‘64 96 159 9% 65 © 141 8? 6:3 95 3 SE 31 Ge. 23 72 08 BR 1'4 67 1373 78 ‘1'2 66 21 71 rz'0' 70 ‘0's 62 ‘°C 68 1'8 6¢ 32 7 23 72 46 86 1'3 66 52 89 53 90 i4'5 8b (5's 91 169 99 20°1 118 22°4 131 21'9 "128 246 ' 144 25-3 '148 26'8 ' 157 26°3 154 272 160 27'0 ' 158 "54 14° 250 147 2577 11 Lh ’ 10 41 32 35 20 23 8 | pu - 092 2% i) [7 2'1 ad 20 26 2'0 2 a) 10 27 ve Uy 20 2'0 2'0 I'1 7 rg 1 12 1 2'0 3 20 9 1-8 12 18 21 (7 8 2'f 2c 27 2 1°€ "107 ‘0 9 10 3K 102 ' 112 ‘4 4°0 4 40 2g 3'5 2°3 38 ~8 >eg 32 2g 28 2 ba 41 2g + 27 38 6 2°} 35 ve "3 29 +8 3'5 37 9 6 "2 I 2°I 108 ” 67 £0 31 3? 73 70 5 r=39 81 97 82, L1€ 109 (26 94 110 in 118 118 115 11 151 132 125 141 36 7 - 08 ~108 102 . 92 UN) 102 70 . 98 - OR . RA 107 L105 128 L 95 ) 10%. 105 Mean re t Pe 9. 00 1 Messrs. Buchan and Mitchell, Journal of the Scottish Mctooroloyicsl Society, vol. iv, pp. 250. 256. 3) STUDIES IN STATISTICS CHAP. (4) Weekly deaths in London during the ten years 1875-84. From the weekly returns of the Registrar-General. See Table V. All these facts are set forth in the accompanying liagram (Plate XXVIL), with the limitation that in the sage of typhus and simple continued fever, I have considered it best, as the numbers are small, to deal with months in preference to weeks. As (with the exception of the London Fever Hospital statistics) the available Ggures are given for weeks only, the number taken to represent January is the sum of the first four weeks ot the year, February the second four weeks of the year, and so on ; but for March, June, September, and December the number taken is four-fifths of the sum of the five weeks :aken to correspond to these months. For the purposes in hand it is necessary to consider the months as of equal length; but as there are fifty-two weeks in the year, it is necessary to drop out four weeks. This may seem a complicated arrangement, but it is really simple, and does not materially affect the accuracy of any conclusions Irawn. Throughout the diagram the blue curves relate to ;yphus, red to simple continued fever, and yellow to :nteric fever. The curves show the average deviation per cent. from the mean, the dark colour indicating excess, the light solour defect. They are all strictly comparable, being on one uniform scale, each small division corresponding to ten per cent, The actual mean weekly, or monthly, number of admissions or deaths is given in ficures on the right hand. je The first point to note is that the various curves lerived from different sets of figures, and relating to CTONTINUED FEVERS IN 1LCNDOCNM, oD ne 5 3 C oC Pc SE Ls > [a 8 © 0 2 TYPHUS FEVER. Monthly ADMISSIONS to the LONDON FEVER HOSPITAL Monthly ADMISSIONS to the Metropolitan Asylum Board HOSPITALS Monthly DEATHS n the METROPOLIS Monthly DEATHS nthe METROPOLIS SIMPLE CONTINUED FEVER. Monthly DEATHS in the METROPOLIS Monthly DEATHS In the METROPOLIS NTERIC FEVER Aonthiy ADKISSIORS to the LUNDON EVER HOSPITAL cooky ADMISSION ibe Metronsinan avium Boon HOSPITALS Veoh DEATHS fro NMOTROPOLIS Vs oy le dy E fd < VET ELY DEATH 3 he RTTROROLDY 5 JUNE my! aug. "P. OCT. INnr-328. JAN.IFEP NE JULY] AUG, sep. | oct. [Nov DEC. 1 Son drds Ging Etah? Gondor .he percentage deviation above or below the newn o. .aonthly or weekly admissions or deaths. Jne small division of the vertical scale corresponds to 0 per cent. The figures at the side sive the actual means AVERACE ¥0.Twenty three Years 1848 - 1870 AVERAGE Ten Years | 875 - 1884 ” Hw AVERACE Six Years 1869-18712 AVERAGE .Jen Years. {875-1884 - 83. AVERAGE Six Years 1869 - 1874 AVERACE Ten Years 1875 - 1884 AVERACE 2 went three Year jas - 1870 AVERACE Ten Years 1&7 5-185 AVERACE Six Years LSE - 147 Ay Ee RACE Ten Years 18785 = ]a4. T KX CONTINUED FEVERS IN LONDON “403 different periods, give in the main concordant results in the case of each disease. This general agreement—which is, indeed, in the case of enteric fever, very close, and in the case of typhus sufficiently marked—is in itself a proof that these diseases are, in some way or another, directly or indirectly influenced by season; and that they arc differently influenced is evident, for enteric fever, whether measured by hospital admissions or by deaths, is most prevalent in the autumn, whereas the other diseases show a preference for winter and spring. As the curves of enteric fever are most definite, and as that disease is now the most fatal form of continued fever, it will be convenient to consider it first. The curves of hospital admissions in the main agree. In both the curve is decidedly below the average from the beginning of March to the end of July, and as decidedly above the average in the last four months of the year. But in two respects they differ: in the older returns of the London Fever Hospital the admissions were much above the mean during August; in the later returns of the hospitals of the Metropolitan Asylums Board somewhat below. Again, the admissions to the former hospital were about equal for September, October, and November, whereas in the latter the October admissions were by far the most numerous. An explanation of one at least of these discrepancies is not far to seek. The Metropolitan Asylums Board has, actually or potentially, an indefinite amount of accommodation ; hence, in all probability, the number of admissions is a true measure of the prevalence of the disease, and the curve resulting is some measure of the real influence of season on the epidemic. On the other hand, in the case of the London Fever Hospital, it is far otherwise; the amount of available accommodation is strictly limited. In the carlier stages of 104 STUDIES IN STATISTICS CHAP. the epidemic all cases are admitted, but no doubt, as the cases become more numerous, a portion of them have to be refused admission. The hospital is constantly full during several weeks when the epidemic is at its height, and so the true maximum is not represented in the curve ; moreover, by the same process the horizon of the mean is lowered, and a glance at the second curve shows that if its mean were lowered a little, the July admissions would appear above the average, and so correspond with the surve above. As regards the two mortality curves for enteric fever, it may be noted that, as we might expect, the second curve, representing as it does the figures of a much longer period, is “smoother,” and probably truer than the first. But, in spite of minor differences, the general agreement of these curves with one another, and with the curves of admissions, Is very striking, and generally it may be stated, (1) that in the case of enteric fever the disease is least prevalent in late spring and early summer; (2) that the rise from the minimum may occur gradually as early as May, or suddenly as late as July ; (8) that it sometimes exceeds the mean of the year in August, always in September, and reaches a maximum about the end of October or beginning of November; (4) that from that point it declines more or less rapidly, usually reaching the mean again about the close of the year, and then declining with occasional small outbreaks throughout the first three months of the year 0 its minimum. It may be remarked that a comparison with the smaller diagram (Plate XXVIIL) shows the curves of enteric fever to approximate very closely to those of scarlet fever, although they are not identical with them. The most essential difference between the two curves is, I take it, that in the case of scarlatina, the rise from the minimum SCARLET FEVER, LONDON. Weekly Average 10 Years 1875-1884 Weekly ADMISSIONS to HOSPITALS of Metropolitan Asylums Board (Percentage of Mean) Weekly DEATHS in LONDON (Percentage of Mean JAN.: FEB. — MARCH APR.| MAY. JUNE.| JuLy.] auc] SEP. 1 OCT NOV. DFC. i —————r s——————_—— le irrala—— «Ta .. . London. Stanfords Geog* Estab One Division of the Vertical Scale Corresponds to 10 per Cent. Mean 26-19 - Mean 43-45 x gs $ 3 <3 iro CONTINUED FEVERS IN LONDON 405 TABLE VI.— Average weekly deaths from Scarlatina in London during the ten years 1875-1884. Week. L8t ... 2nd ... 3rd ... ith ... 5th ... 6th ... 7th ... 3th ... 9th -... Loth ... 11th ... ath ... 3th... 4th... 5th ... 16th ... 7th —w 18th ... 19th ... 20th ... 21st ... 29nd ... 23rd... 24th ... 25th... 26th ... Month. January ... ” -_ % ee ”» ee February... » J. —_ » wen March ... ”» » » bh April ” ” 1 May » »” » June ” " ” ry Average Deaths. 7 83 470 61 393 354 28 28 4 7 74 282 7 26-8 268 29'8 77 “6 2 al ie) > 3 3 Percen. age of Mean. .05 111 108 106 90 81 5. 63 65 7o 6 62 Bo Bs 57 2 ¢ Week. 27th ... 28th ... 29th ... 30th ... 3st ... 32nd ... 33rd ... 34th ... 35th... 36th ... 37th ... 38th .., 39th ... 40th ... Aalst ... 49nd ... 43rd ... 14th ... 45th ,.. 46th ... 47th ... 48th ... 49th ... 50th ... lst... 52nd ... dean... Month. July ... no ees "new ” oon August ,., 5 bee » » “ee September 9g eee "a yy ees a October... nae gy ae pees November 3 eee wa ys eee December 5 ies wees Average Deaths, 9'3 599 46's 40°8 462 41's 460 47 12°0 436 182 36°3 7'8 I1°5 660 675 48-3 ‘ey 54 94°8 593'0 74 JR Percentbr of ean. £ 2s 92 107 94 10° 0 10° 16 3 + 3a 36 "19 vil 100 to the maximum is longer and more gradual than in the case of enteric fever. See also Tables IV. and VI. This great resemblance between the curves should be kept in mind ; but although we can say that scarlatina and enteric fever are both, whether directly or indirectly, greatly influenced by season, and are moreover affected in a very similar manner, and even to precisely the same extent (so far at least as the mortality curves go, the maxima and minima in the two cases being identical), I do not consider that any further inference is justifiable. There does not appear to be any correspondence (as is found in the case of scarlet fever), either direct or inverse, 1 See Chapter XVI, STUDIES IN STATISTICS CHAP. 106 between the figures in Table 1I. and the rainfall at Greenwich. Next let us examine the typhus curves. The numbers being, with the exception of the admissions to the London Fever Hospital, very small, it seemed better to deal with months rather than weeks. The two sets of hospital statistics are scarcely liable to errors of careless or incompetent diagnosis, and, as might be expected, give better curves than the Registrar-General’s figures. The hospital returns agree in the main; they both show the prevalence of typhus to be below the average in the five months June to October inclusive, and above the average in the four months November to February inclusive. The earlier returns give a much greater number of cases for March, April, and May than the later. According to the first curve, the maximum is in January ; according to the second, in February. The two curves of deaths agree in representing the mortality as below the average in May, August, and September, and above the average in the first four months of the year, and in October. The most striking discrepancies are in July and November. If we attempt to combine the four curves, we may perhaps say :—The mortality from, and the admissions to hospitals for, typhus are at a minimum in July, August, or September; from this minimum they increase to an autumn maximum, commonly in November, but it may be in October ; they usually continue excessive throughout the winter, and another maximum is reached in January or February, which commonly exceeds the autumn maximum. The fall is irregular; it may be sudden in March, or more gradual and as late as May. Typhus is more infectious than enteric fever; its outoreaks are of less frequent occurrence, but may be much CONTINUED FEVERS IN LONDON 407 more severe. It is not nearly so much influenced by season as scarlatina or enteric fever, but so far as it is influenced, it may be described as a disease of late autumn, winter, and early spring. Lastly, as regards simple continued fever, the only facts to go by are the London deaths. The numbers are small and the curves very irregular, and not very concordant. The fluctuations from the mean are éven smaller than in the case of typhus. December, January, March, and, in the first curve, February, are months of high mortality ; June and August, months of low mortality. : The statements of Murchison quoted above would have led us to expect some indication at least, in the simple continued fever curves, of the well-marked autumnal maximum of enteric fever, but there is not a trace of this. On the other hand, there is some sort of resemblance between this curve and that of typhus mortality; and, indeed, if the two typhus mortality curves be combined, and also those of simple continued fever, this resemblance becomes very marked. In both the minimum isin August, the maximum in January; in both the mortality is decidedly low from June to September, and decidedly high in January, February, and March. It is not disputed that the deaths ascribed to simple continued fever are largely, probably mainly, due to something very different, and that, were it not for ignorance and carelessness in diagnosis, the deaths would have been registered in still smaller numbers than has been the case of late; but it does not, therefore, follow that these Jeaths have been mainly due to enteric fever. I contend that the diagram proves that but a very small proportion of them can be attributed to that cause. Among diseases the true nature of which may conceivably be overlooked, there might be suggested, in addition to typhus and XX STUDIES IN STATISTICS CHAP. enteric fever, the following :—General tuberculosis, septicemia, pneumonia, intermittent fever. It must be remembered that it is easier to diagnose the presence of fever than to ascertain its cause. Incidentally the figures obtained for the purpose of this paper throw light on the extent to which the hospitals of the Metropolitan Asylums Board are made use of. Mr. A. C. Waters has prepared for me a table (Table VIL.) TABLE VIL— Deaths from Enteric Fever, Typhus, and Scarlet Fever in London ; also those occurring in the Hospitals of the Metropolitan District Asylums Board, 1873-1889. 1873 1874 L875 1876 1877 578 (879 1880 1881 1889 Mean Nn years 83 god 1885 PEC Mean seven yes ™ £8 Ei oC 908 879 817 |B Total "> 2" XY 645 "R48 "7 ‘8 2 OM. J 1,630 3,839 4,622 £36 Cs 3 4 =] Deaths in Metropolifar District Asylumg Bee = Hospita’-{ , —-a i 258 284 17 lag v » i f . i « Percentage of deaths m se hospitals to total Teathe in London, £3 ti % 329 340 2's 176 22°9 { 2. 4 3 pa £ i 1 t ¢ 1 . 2 10% Po [17 re 74 105 141 i2-8 153 26-7 29-9 30-9 3 25 | 192 showing for each of the seventeen years 1873-89, the total deaths in the metropolis from enteric fever, typhus, and scarlet fever; also the deaths in the hospitals of the Metropolitan Asylums Board for the same years, and the percentages that the latter are of the former. This shows that, as regards enteric fever, proportions varying from 6 to 11 per cent. of the deaths have occurred in the hospitals ; as regards typhus, the proportions have varied from 8 to 51 per cent. (in 1882); and as recards scarlet xX CONTINUED FEVERS IN LONDON 409 fever, proportions steadily increasing from 1 per cent. in 1873 to 47 per cent. in 1889. In the case of the two former diseases the proportions treated in the hospitals of the Board have increased of late years, but in an irregular manner and not to a very great extent. In Table VIII. are given the ratios of the total deaths TABLE VIIT.— Enteric Fever, Typhus, and Scarlet Fever ; proportions borne by Deaths in all London to Admissions to Hospitals of the Metropolitan District Asylums Board. Enteric Fever. I Typhus. Scarlet Fever. 220 8B . Som ~ A 258. = Ro] IN = Qa z Engl © gc | aH = y J £ moe 253% Ea 2 388% Q og Hom a - 2 J g Es By LAg em = = T = Hag BOS = 252% 2 Re 838% am on 0 am B- x - et sw Td LEs 22% £3 | 595% EEE £8 | §82 og ~ sem aH a " g oH 873 1874 1875 1876 1877 1878 1879 1880 1881 LR82 381 | 435 299 | 288 372 484 8 908 879 817 769 901 1033 849 702 977 a1 AR 238 401 202 536 273 65 267 139 949 170 213 16¢ 221 48 283 28 234 219 189 1148 277 312 128 159 157 151 71 74 92 59 69 58 197 114 92 90 148 264 43 36 92 804 1,182 671 479 679 1,469 1,949 1,477 1.8c0 645 2,648 | 3,677 2,308 1,580 1,808 2,661 3,100 2,114 2 008 701 329 311 344 330 266 181 159 142 108 1 Mean LO vears, ago | 880 ' 230 192 | a7 ' wt 1.085] 2.955! 212 1883 1884 {885 1886 (887 1888 |RRO I ' 56 193 220 399 963 925 597 518 5H 198 88 71 $3] 0 15 0 55 32 28 | 12 122 110 63 130 54 300 65 1,020 2,006 5,845 1,430 1,353 | 722 na8 | 690 3,900 7 443 2,408 ule 4,518 78¢ 104 78 63 39 24 28 1% 53 Ir ro 200 F3R Mean 7 vears. | 388 | 707 | 10° 4. a 86 43.1031 184 38 in London to the admissions to the hospitals,—another way of measuring the work done by the latter. There was antil quite recently no means of ascertaining the total number of cases in London, and some time must elapse before we shall have a sufficient series of returns under the Infectious Diseases (Notification) Act. 1 STUDIES IN STATISTICS CHAP. [I have added to Tables VII. and VIIL the corresponding figures for the five years 1885-89. It will be noticed that as regards typhus and enteric fever, the proportion borne by the deaths in hospitals to the total deaths has not very materially altered, but as regards scarlet fever it is plain that the Metropolitan Asylums Board has dealt with a constantly increasing proportion of the London cases. In 1888, cases of diphtheria were admitted for the first time, ninety-nine cases coming under treatment in that year; the number increased to 722 in 1889. The proportions borne by these numbers to the total cases in London are indicated by the following table :— DIPHTHERIA IN LONDON, Admissions to Hospitals Deaths in T Proportion of deaths in of Metropolitan Asylums these Potal deaths Hospitals to 100 Deaths Board. Hospitals in London. in London. 99... LL 46 1311... .. 35 22 0... 215... 1588 ... ... 173 It is then evident that the new regulations as to the admission of cases of diphtheria have been largely made use of. Though scarcely coming within the scope of this chapter, I have appended a table (Table IX.) showing the mortality in the hospitals compared with the cases treated, or the case-mortality, as it is termed. It will be observed that scarlet fever was more severe in the earlier years of the series than in the later. It will also be ohserved that diphtheria is a much more mortal illness than typhus or enteric fever, a fortiori than scarlet fever. My conclusions may be shortly stated. (1) The mortality from enteric fever has not materially diminished in London during the last twelve years. (2) The mortality from both typhus and simple con- KZ CONTINUED FEVERS IN LONDON 411 TasLE IX.— Mortality per cent. of Patients treated in the Hospitals of the Metrovolitan District Asviums Board! 1873-1889. Enteric Fever. | Typhus. r 4 ra gre: +5 19 Ed 873 874 435 } 875 299 i 576 288 877 372 7 878 474 879 gh 380 2. € *81 82 83 £4 5 6 37 £888 QR] § 49 Wu 20°F ba ry 536 6s 139 170 | 168 or L 106 1¢ 28 86 pt 6 34 19-3 "3-1 "4-3 PS Searlet Fever. § Diphtheria. < 9z 804 1,182 671 479 6,9 ary nag - 160 an r ee 1 tp fe Total Mean 1 From a Table in the Report of the Metropolitan Asylums Board for 1889. The mortality of 108pital patients is calculated according to Dr. Farr's formula :— Mortality = Deaths. = 1 Admissions + # (Deaths + Discharges). tinued fever has decreased very greatly during the same period. (3) Enteric fever is by far most prevalent in the autumn ; the average weekly hospital-admissions, and the mortality in the metropolis, can be expressed as welldefined curves, having maxima at the end of October or beginning of November. (4) The curves of enteric fever strikingly resemble those of scarlatina, but differ in one important respect. The minimum in the case of the latter disease occurs in March or April, and the ascent from that time to the autumn maximum is very gradual; but in the case of enteric fever the minimum occurs in June or July, and the ascent to the maximum is comparatively sudden ; the converse being the case as regards the descent to the minimum. STUDIES IN STATISTICS CHAP. (5) The statistics of typhus and simple continued fever, so far as figures are available, do not yield welldefined curves of seasonal prevalence. The curves, such as they are, bear no resemblance to those of enteric fever. (6) Typhus is often prevalent in late autumn, but is rather a disease of winter and spring. In early autumn, when enteric fever is approaching its maximum, typhus is far below its average. (7) Simple continued fever is a term applied in a great many cases in error, since, though deaths are often attributed to it, it is but rarely if ever fatal in fever hospitals. (8) The statistics of the last sixteen years show that not more than an inconsiderable proportion of the deaths attributed to simple continued fever (however caused) can be due to enteric fever. (9) The hospitals of the Metropolitan Asylums Board deal with an increasing proportion of the cases of infectious disease in the metropolis, the increase being most marked in the case of scarlatina. (10) During the seven years 1883—1889, the deaths in the said hospitals have amounted, in the case of enteric fever, to nearly 9 per cent. of the total deaths from that cause in the metropolis, in the case of typhus to 21 per cent., and in the case of scarlatina to 25 per cent. (in 1888 it amounted to 41, and in 1889 to 47 per cent. ). It is hoped that the tables are free from errors. Some small apparent discrepancies are due to the fact that in calculating the percentages the figures were, in some cases, worked out by the calculating machine to several places of decimals, while the figures given in the tables are the nearest whole numbers. KXI HYDROPHOBIA STATISTICS 413 CHAPTER XXI. HYDROPHOBIA STATISTICS.! It is, I take it, an especial object of this Society to take a general view of the phenomena of the prevalence of epidemic diseases, or diseases supposed to be epidemic in their nature, whereby facts may be brought to light that might escape the clinical observer ; just as the geologist ascertains one set of facts by going into a quarry and examining the face of the rock, but ascertains another scarcely less important set of facts by observing the general contour of the hills as seen from the other side of the valley. The unprecedented number of deaths from hydrophobia in London during the year 1885, and the prominence given to the outbreak by the Police Order, as well as the great interest excited by M. Pastewr’s researches, point to the present as a suitable time to search the official records of this country, and review the facts therein buried. The arrangement of the Registrar-General’s reports makes it convenient to consider separately the prevalence of hydrophobia in England and Wales generally, and in the metropolis. The figures for the latter supply several facts not recorded for the whole country ; moreover, it will ! Originally read at a meeting of the Epidemiological Society of London, 1886. STUDIES IN STATISTICS CHAP. be noted in the Tables that the periods covered are not the same in all cases. TasLe L—Deaths from Hydrophobia in England and Wales and in London.— Forty years, 1846-1885. a 22 [ oe] n = < = = < — IL = BZ Eo & = 55 Bg = < = 2 2 - 1846 1847 1848 1847 1850 1851 852 1853 18654 1855 1 5... 71 7. 13 1 2... 15 1 16] 7 14 2 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 5 2 2 2 13 1¢ Total... 123] 13 El 13 Rate ) Rate eL 071d 056 per { Hon 1 lion \ Cotal.. ean... 5 2 023 [0-47 IIL 1866 1867 1868 .869 870 871 1872 373 1874 1875 lotal.. &%a vean...| 33-4! Re” per million Es ZF = ma c og = <Q ny ww. Ea wu = Rg a 2 = a 2 — RI 10 7 18 32 36 29 11 3 - : 1 6 1876 1877 1878 1879 880 1881 1882 1883 1884 1885 53 79 50 25 cn "6 16 | cA 28 34] 8 28 | 9 60 | 27 Total...l 430] 85 ean... 43.0 85 Rate | rod 0 Ul srl rod 2 Cel 22s { lion £ Total Deaths. Mean. Rates per million. England and Wales 39 years 944 ve 24:2 wi 1-11 London +... ... ... 40 years 145 we. 36 nt 1-19 ¥ No returns of ¢ Causes of death ? in this year, except for London. England and Wales.—During the thirty-nine years 1847-85 inclusive, no fewer than 944 persons are stated to have died from hydrophobia. (See Table I.) The deaths were very unequally distributed throughout the period. Thus, during the eight years 1856-63 only 26 deaths occurred, or an average of 3 per annum. The figures then rose to 36 in 1866, and fell to 7 in 1868, and rose again to 56 in 1871, from which time they have Auctuated from a minimum of 28 (in 1873, 1882, and KXI HYDROPHOBIA STATISTICS 1884) to a maximum of 79 (in 1877), making an average since 1870 of 43 deaths per annum. If the whole time be divided into four periods of ten years each, it will be seen that the mortality from hydrophobia has increased considerably, both absolutely and relatively, in the more recent years. In the second period the number of deaths was less than half that in the first, but in the third and fourth periods, even after allowing for increase of population, the mortality was double that in the first period, and five times that in the second period. Of the total 944 deaths from hydrophobia in England and Wales during the thirty-nine years for which the returns are given in Table I., no less than 413 occurred in the eight years 1871-78, during which period the Jeath-rate from this disease averaged 2°13 per million persons living. London.—As regards the metropolis the figures are available from 1845 to 1885 inclusive, a period of fortyone years, during which 147 deaths were attributed to this disease, or an average of 384 deaths per annum. In fourteen of the years no death occurred: indeed, there was complete immunity for the five years 1858-62; but, on the other hand, in 1866 there were 11 deaths, in 1877 as many as 16, and in 1885 no less than 27. Taking periods of ten years, as was done with the figures for England and Wales, similar results are obtained, but the correspondence is only general. Thus, the increase of the fourth period when compared with the third is much more marked in the case of London. Again, while 1874 and 1877 were periods of high mortality, both in the whole country and in London, on the other hand the greatest mortality in England and Wales during the first nineteen years in the Table was in 1851, when there STUDIES IN STATISTICS CHAP. 116 were 25 deaths registered, but there were none in the metropolis that year. "There are also marked discrepancies in 1854, 1865, 1866, and 1871. : There seems to be no doubt that the disease is getting commoner in London, and the figure of 1885, viz. 27, has not been approached before, so far at least as our records go. During the first seventeen weeks of 1886, as many as six deaths have been registered, as against four during the corresponding period of 1885.1 Taking the three periods covered by the decennial reports, and taking the mean populations given in these reports, we get the following average annual deatli-rates per million living. 1851—1860 1861—1870 1871—1880 Total England and Wales. deaths. Rates. o'5z } o'6y hs 1°96 3 S17 718 London. Deaths. Rates. Yo... 046 28 .. o'92 49 ... 139 K0 The Table shows that, relatively as well as absolutely, the disease is, both in London and the country as a whole, much more fatal now than it was formerly—some threefold. Seasonal Prevalence.—The data for determining the seasonal prevalence of hydrophobia are available only for the metropolis. In Table II. the 147 deaths that have been attributed to hydrophobia in London during the ! The prediction in the text has not been fulfilled ; the deaths from hydrophobia since the reading of this paper have been— England and Wales. In 1886 26 » 1887 29 y, 1888 we 14 ee ,, 1889 ... Not published London. 9 2 3 7 KXI forty-one years 1845-1885 are arranged according to the weeks in which they were registered. The weekly average for each month and quarter is also given. It will be noticed, as might have been expected, that in several weeks (seven) no death has occurred, whereas in two weeks—the third in July and the last in September—mno less than nine deaths have been registered. As regards months, the weekly average is lowest in March and February ; rises gradually to a maximum in September ; falls in October; rises to a second maximum in November, then falls rapidly to February. Happily, the disease is too rare to allow of any positive conclusions being drawn as to the way and extent in which it is influenced by meteorological conditions. The figures given in the Table show that it is by no means confined to the ‘dog-days,” though certainly deaths from hydrophobia are more frequent at about that time of the year. HYDROPHOBIA STATISTICS 417 TasLe IL.—Hydrophobia in London. The weeks in which 147 Deaths were registered during the forty-one years 1845-1885, Weekly Mean. Weekly Mean. Week]: Mean. Ist 2nd jrd | 4th 2 — | os ] Jan, 2:50 vdth 15th 16th 17th 4 > April 2:00 27th 1 3 28th + - 29th | 9 30th | 3 July 375 2 hth - 2 Bth + — 7th | = Rth | 3 Feb. 1:25 18th 19th 20th 21st 1 May 1°75 31st 32nd 33rd 34th 3 9 Aug. 3:75 2 2 2 Ist Qul21 1 162 2nd Qri29! 2:23 3rd Qri54 | 415 Ith LOth “1th ! [2th | [3th | 35th 36th 37th 38th 29th Weekly Mean. 40th 41st 49nd 43rd 3 3] Oct. 4' 275 1 Ath 15th 46th 17th { 3 2 6 Nov. 4:50 “8th ‘9th 50th Hlst 52nd 5 | Dec. 0 | 2-80 2 1th Qr43 3:31 1 EE 1. STUDIES IN STATISTICS CHAP, The exciting cause of hydrophobia is undoubtedly inoculation with the poison of rabies, and the manifestations of the symptoms are delayed by a period of incubation, usually ranging from four to twelve weeks. It should, nevertheless, be borne in mind that, conceivably, meteorological conditions may influence hydrophobia in several ways i—e. 9. (1) it may directly or indirectly favour the spread of rabies in the lower animals; (2) it may render human beings bitten by rabid animals more or less susceptible to inoculation; (3) it may hasten or retard the development of the characteristic symptoms. Anyway, the fact remains, that although from the small number of facts available it is not possible to draw a regular curve for hydrophobia, the curve, such as it is, shows considerable seasonal fluctuations from the mean. It may not be possible to say exactly in what way season affects hydrophobia, but it is pretty evident that in some direct or indirect manner the high temperature of summer has a decided influence in favouring it. This influence, as measured by the curve, is stronger than is observed in the case of scarlatina and enteric fever, but very much less marked than in the case of diarrhoea. Rabies. —The inquiry that naturally suggests itself is : What is definitely known about the seasonal prevalence of rabies in the lower animals ? It is, unfortunately, not possible to get any complete or extensive statistics of the prevalence of rabies among dogs in this country. But che figures to be laid before you, scanty and imperfect 1s they are, possess considerable interest. Through the kindness of Mr. Ernest Batt, Veterinary 1 Rabies and hydrophobia are of course one and the same disease ; for convenience, however, throughout this paper the disease as it manifests itself in man is termed Aydrophobia, the word rabies being reserved for the lower animals. Xi HYDROPHOBIA STATISTICS 419 Sargeon, and Mr. Victor Horsley, Professor Superintendent, [ am able to give the numbers of admissions for rabies to the Brown Institution from January 1880 to the end of July 1890, and the months in which they occurred. TaBLE 11L.— Brown Instiiution.— Admissions for Rabies. [pots 1850.| 1881.| 1882. 1853. |1884. 11885. 11886. [1 867. 1585. l1ss9. 10 (1890. Tears. ! January February March April May June Tuly August Septemr. Jetoher Novemr. Decemr. S { 3 ga g <A 2 2 Year These figures show, firstly, a considerable increase of rabies in the two years 1884-5; secondly, that rabies, so far as the small numbers admit of any inference, is least prevalent in the summer months, but is a disease of autumn and spring. Mr. J. C. Colam, the secretary to the Home for Tost and Starving Dogs at Battersea, and his successor, Mr. M. Colam, have been good enough to supply me with the numbers of cases of rabies detected there in each month from January 1883 to the end of July 1890, by Mr. Sewell, the veterinary surgeon to that establishment. These figures (Table IV.) show little difference in the totals for 1883 and 1884, but agree with those of Table IIL. in showing great prevalence of the disease in 1885, and a remarkable immunity in 1887. There is no clear 420 STUDIES IN STATISTICS CHAP. indication of influence of season, though fewest cases occurred in August and September. I attach great weight to the returns of rabies from these two institutions, as, though the figures may be small, they have great value, owing to the fact that they are certified by persons well qualified to diagnose the disease, and especially interested in the matter. The number of dogs brought under observation yearly at the Dogs’ Home is very considerable, so that we are able to form a fair idea of the proportion of these animals that suffer from rabies. The Brown Institution figures give low mortalities for January, February, June, July, and August; maxima in March and November. The Dogs’ Home figures give low mortalities for February, August, and September ; the maxima being in March and June. It would appear, then, that if influenced by the heat of summer, rabies follows it like scarlatina, rather than accompanies it like diarrheea. Rabies is least fatal to dogs in the summer, contrary to common report. Police regulations are of paramount importance in the prevention of hydrophobia, therefore it is interesting to observe carefully the results of recent measures. The unexampled prevalence of rabies and hydrophobia in London in 1885, especially in the latter half of the year, led in November to the issue, by the Chief Commissioner of the Metropolitan Police, of an order under § 18 of the Metropolitan Streets Act, 1867, requiring all dogs while in the streets of the metropolis, and not led by some person, to be muzzled. This order came into force on December 10th, 1885. On September 20th, 1886, it was supplemented by a further order of the Chief Commissioner, under the Dogs’ Act, 1871, for the seizure of all dogs found at large and not under the control of any person. Both orders expired on December 20th, 1886. XX1 HYDROPHOBRIA STATISTICS 421 The immediate result of the muzzling was remarkable. A case of rabies was admitted into the Brown Institution on November 22nd, but no other until the end of January 1886, and that was the case of a dog that was believed to have contracted the disease outside the Metropolitan area. At the Dogs’ Home at Battersea the case was very different, for ten cases were detected in December 1885 in that institution, being the largest TasLe IV.— Temporary Home for Lost and Starving Dogs, Battersea. Admissions for Rabies. January ... february ... March ... April . May ‘er. June ve July _ August .. September Dctober ... November December Year ...i 1883. | 1884. | 1885. | 1886. | 1887. J 3 ) ( 4 1 2 Ae 1888. | 1889. 2 3 2 1 Total pal | 1890. .3 R 16 “0 i | 1 1 1 10 5 | 11 11 | 12 5 oy £8 << 2 ° py ‘39 number that had occurred in any one month. The discrepancy is explicable if we remember that the Police Order, while it cleared the streets, filled the Home to overflowing. The total number of dogs brought in was, in 1883, 14,647; in 1884, 14,772; in 1885, 25,578. Mr. Colam attributes the extra 10,000 dogs admitted in the year 1885 entirely to the Police Order issued towards its close. Thus rabies greatly diminished in London during 1886, and in 1887 was almost absent from the metropolis. In the spring of 1889 it increased somewhat rapidly, and several deaths from hydrophobia occurred. In consequence, 422 STUDIES IN STATISTICS CHAP. partly of a representation from the London County Council, a Rabies Order, under the Contagious Diseases (Animals) Acts, applicable to the city of London and the Metropolitan Police District, was issued by the Board of Agriculture! The order came into operation on August 12th, 1889, and has been continued since by subsequent orders. The effect of the Rabies Order was as follows: —One rabid dog was admitted to the Brown Institution in October 1889, and two in November; but from that date until the end of July 1890 (my latest information), no case has been seen there. Again, at the Dogs’ Home, although fifteen cases had been detected in the first seven months of the year, there were but one case in August and four in October, and in 1890 single cases in February, March, and April, but none in May, June, or July, the atest month for which I have returns. In Table V. are shown, month by month for the last seven years, the cases of rabies in the two institutions, compared with the deaths from hydrophobia registered in London. The same facts are shown more clearly in the diagram (Plate XXIX.). The unusual number of cases of rabies in the Dogs’ Home in November and December 1885 may well have been largely due to the enormous number of dogs swept into the Home by the Police Order, and so brought under the notice of the surgeon (Mr. Sewell). Making allowance for this, it will be seen that the outbreaks of hydrophobia followed those of rabies, after an interval of from three to six months, which is considerably longer shan the average period of incubation ; but the agreement 1 The Board issued this order on July 31st, and the London County Council was the statutory authority to enforce it in the county of London. However, the majority of the Council, disappointed that the Local Government Act, 1888, hal not given them authority over the police, declared that they had no means of enforcing the order. The Government accordingly appointed Major Tennant to do this, '383 Sp LONDON | RABIES IN DOGS & HYDROPHOBIA IN MAN. 1884 1885 1886 1887 1888 1889 = TTT, oT - : ~ - TT F — 7 1890 — pis T 10 RABIES 10 Cases each month at, Brown Institution 5 & Dogs Home. 0 0 J HYDROPHOBIA Deaths each month in London 3 20) GREATER LONDON. RABID DOGS KILLED BY THE POLICE. Commencing July 1886. 5 0 3 3 Fos dem. Stanford (eogt Estab” XXI HYDROPHOBIA STATISTICS 423 Tare V.—Monthly Cases of Rabies at Brown Institution and Home for Lost Dogs, compared with Deaths from Hydrophobic in London. Seven years. 1883-1889 january... February March... April... May June ...1 July ...| August ... September Dctober ... November December Year ...| i853, 1884, : oo) 2 2 ¢ 2 29 2c | 1885 5 c 2 1 2 1 1846. u 2 <9 1587. - 2 «3 ~~ = a = > 2. =, pd 1838, a Total, 1889. | 7 years, re 2 = o z . a oi o : 2 <2 = Z E mE =. be . 3 2 Q 1+ 1G o-2. 3 1 I Cee ! 4 AL. f 10 19 | 98 16 9130267 ¢ fet _192% 65 is perhaps as close as could have been expected from such small numbers. As to the cause of rabies, or, rather, the influence of meteorological conditions on it, the fact that outbreaks have occurred both in spring and autumn makes it unlikely that either heat or cold has any direct influence on the disease. The winters of 1854-5, 1860-1, 1864-5, 1869-70, 1870-1, 1874-75, 1878-9, 1879-80, 1880-1, were all remarkably cold; in the years following these hard winters there was a notable increase of hydrophobia in two zases only, namely, in 1870, when 82 cases were registered as against 18 in the previous year, and in 1871, when the aumber rose still further to 56. On the other hand, the summers of 1857, 1859, 1865, 1868, and 1884 were remarkably hot. In 1865 the cases of hydrophobia were 19 as against 12 in the previous year, but they increased to 36 in the year following. In 1868, when summer diarrhoea reached its maximum, there were but 7 deaths from hydrophobia (less than in either of the previous STUDIES IN STATISTICS CHAP. four years, or in any subsequent year) against numbers, increasing to 18 in 1869, 82 in 1870, and 56 in 1871. In 1885 we know that in London there was a great increase of hydrophobia, as well as in the whole country. By the kindness of Dr. A. O. MacKellar, Surgeon-in-Chief of the Metropolitan Police, I have at the last moment been enabled to add to Plate XXIX. the Police returns of dogs found to be rabid in Greater London from July 1886 to the end of July 1890.) They indicate a considerable outbreak in 1889, which led to the issue of the muzzling order of that year. Some very interesting information as to Rabies is to be found in the Reports of the Veterinary Department of the Board of Agriculture, from which I extract the following paragraphs and tables: — “The following Table shows the number of cases of rabies in England during 1887-8-9, the three years it has been included as a disease undor the Contagious Diseases (Animals) Acts, also the number of each kind of animal attacked— Dogs Cattle Sheep Swine Horses Deer Kind of Animal. i 1887. 217 nh ur if - 1888, !' 1889. 160 ' 312 2 9 7 i1 176 340 “If the deer entered in the above Table for the years 1887-8, which were attacked with the disease in Richmond Park under very unusual 1 During this period the number of “mad dogs and dogs suspected of being mad ” killed in the streets by the police, within the area of Greater London, amounted to 947. Of these it is stated that 272 were actually rabid, the others not. In Plate XXIX., in addition to these 272 cases, are included -16 dogs that died of rabies. XX1 HYDROPHOBIA STATISTICS 425 sircumstances, be excluded, it will be seen that the animals of the farm are but very rarely affected with this disease as compared with the dog. ES «The following Table shows the number of dogs affected in each county, and thus indicates the districts in which the disease has maintained a rermanent existence— Cases of Rabies in Dogs in England in 1887-89. Counties. I 1887. 1888. 1889. Bedford se Cambridge ... Chester Derby Dorset Durham Essex Hants Herts Hunts Kent vs Lancaster ... Leicester ... : Lincoln (Kesteven) » (Lindsey) . London I. Middlesex ... Northampton Notts es Oxford . alop i Stafford ve. Suffolk .e Surrey yo Sussex (East) » (West) Warwick ... Westmorland Wilts ve. Worcester ... ee York (W.R.) i. Soke of Peterborough 1 — 33 | 20 6 6 3 ) 9 3 11 7 10 2 2 36 3 9 3 32 34 29 8 30 9 3 2 61 24 - 23 1 8 26 R > 15 | 67 Total os 917 | 160 | 319 126 STUDIES IN STATISTICS CHAP. “In the Report of the Department for 1888, reference was made to the fact that the counties in which rabies was most prevalent could be conveniently arranged in two groups, the first comprising Lancashire and adjoining counties, and the second London and the counties round it. It appears from the Report of the Registrar-General for 1888 that these are the very centres in which hydrophobia in man has been most prevalent during the past 20 years. * * * * * * * «The first effect of this Order was to clear the streets of London and the suburbs of a large number of useless stray and ownerless dogs. The Order did not ab first seem to have much effect in reducing the rabies returns, but taking into consideration the period of incubation in this disease, which often extends over several weeks, it was not to be expected that any material decrease could take place until the Order had been in force about three months. In the months of November and December there was a considerable reduction in the number of cases of rabies in the district to which this Order applied. “ Towards the end of the year, as the disease was not eradicated, it was deemed advisable not only to continue the Muzzling Order so far as the Metropolitan Police District was concerned, but to extend the area over which it should operate, so as to include not only the Metropolitan Police District, but also the whole of the counties of Essex, Herts, Kent, Middlesex, and Surrey, and the boroughs within them, from the lst January, 1890. « When affected with rabies the dog develops an irresistible tendency to wander from his home, and in the course of his travels he snaps and bites at any person or animal that happens to come in his way, and thus the disease is spread. The first and most essential thing to be done in the suppression of rabies in a district, therefore, is the seizure and destruction of all wandering dogs. The necessity for this is apparent from the rabies returns of the past year, which show that of the 312 rabid dogs, 121, or over 38 per cent, of them, were returned as stray dogs, which, if not exactly ownerless, were, when found diseased, so far from their owner's premises that the ownership could not he traced by the police.” ! Age and Sex.—As regards age, the figures of the forty years 1849-88 are available. Of the total number of deaths during this period, viz. 1001, nearly four-fifths were of males and rather more than one-fifth females. But of these 1001 deaths, 124 were of children under five + Board of Agriculture—dnnual Beport of the Vetsrinary Department, 1889, pp. 78.81. XX1 HYDROPHOBIA STATISTICS years of age, the males being more than twice as numerous as the females; 323 were aged between 5 and 15 years, the males being three times as numerous as the females; 528 were aged between 15 and 65 years, the male deaths at those ages being nearly five times as numerous as the female; over 65 years there were 26 deaths, the males being four times as numerous as the females. The exact figures are given in detail in Table VI. 427 Tanne VI.—Ages and Sexes of 1001 Deaths from Hydrophobia in England and Wales during the forty years 1849-88. All Ages. Males 798 females | 218 dersons ! 1001 I. “ le Total : | | “nder sho 15— Jonas tan sn—los—|rs— ss i 65 72 122 90 | 54 3a [10] 2 | 46] 8511/10 18[12|21 [17 | &| 1 .. ha 76 82 140{102| 75 | 53 | 23 3 [t will be observed that five deaths occurred under the age of one year (four males and a female) ; three deaths securred over the age of 75 (two males and one female) ; no death occurred of a person aged over 85 years. The next Table (Table VIL) shows the mortality from hydrophobia in relation to the number living at each age. As the calculation of the mean numbers living at the various ages throughout the period would have entailed considerable labour, I have made the Table apply to the thirty-five years 1854-88, and have assumed the population at each age enumerated in 1871 (the middle year of the period) to be the mean in the whole period. This method gives, I think, a sufficiently close approximation,—indeed, with such small figures to start with, it would be pedantic to attempt great accuracy. 428 STUDIES IN STATISTICS CHAP, TasLe VIL.—Ages and Sexes of 920 Deaths from Hydrophobia in England and Wales during the thirty-five years 1854-1888, together with the mean annual rates, per million living, at each age. All Ager 10-LL Males. Deaths ... Rate per million Females. Deaths ru Rate per million 722 | Hie 3 60 179 1°42 [22 |1°6 198 ' ) 43) 23 0'5 0°710°Q (27 Persons. Deaths ~~... 920 108[169): Rate per million) 12 + 170f1°8 |1°, 20— | 25— | 85-so-lss—|os—|rs—| 85— 66 112! 85 | 51) 32 19] 2°... 2'0 ©» o2'¢ 1'5i14|1°5l0°5 ... 21/17 410. 06lo7]0'3 oz ... 1 72149 23 3. i'oj10loglo3 ... 0s It would appear that the greatest relative as well as absolute mortality occurs between the ages of five and ten years, and next after this in the succeeding period of ten to fifteen years. The explanation of this may perhaps be, that of those bitten a greater number of children contract hydrophobia—there may be a greater predisposition at those ages. On the other hand, it should be borne in mind that young children are less completely clothed than adults, and their partially naked limbs are without protection against the saliva of the rabid dog; moreover, they are less able to defend themselves against attack, and to some extent the smallness of their limbs is a danger, the dog getting hold of a small limb more easily than a large one,—though perhaps this point is not so important here as in the case of snake-bites.! It is remarkable that at all ages, from one year to seventy-five, many more males than females die from hydrophobia. Under five, boys suffer twice as much as 1 Tn the discussion which followed the reading of the paper, Dr. Hime suggested that the greater liability of children to hydrophobia might be due to the small bulk of their bodies in relation to the amount of poison inserted by the dog’s bite. {X1 HYDROPHORIA STATISTICS 429 girls; boys from five to fifteen, three times as much as girls; youths from fifteen to twenty, five times as much as maidens of like age; while from twenty to forty-five, men suffer eight or nine times as much as women ; but after forty-five the disparity is not so great, men of fortyfive to sixty-five dying of the disease only twice as often (relatively) as women ; at ages over sixty-five the figures are too small for useful comparison. At the extremes of life the conditions as to occupation and habits are similar; in middle life men are brought much more into contact with animals by their occupations, and on the whole are not so well protected by their clothes —for it must be remembered that all authorities lay great stress on the efficacy of clothes in wiping the saliva off the teeth of a rabid animal. In young children it might have been imagined that girls were less protected than boys, but the latter are of course more venturesome. Dr. Ogle informs me that boys die from violence more than girls, and that the difference, which can only be accounted for by their greater daring and general wildness, shows itself at a very early age. Thus in the case of fractures and contusions the difference shows itself at three years of age, and in the case of burns, scalds. and drowning at one year. There is. however, one further consideration in reference to the far greater incidence of hydrophobia on males than females. We know that many more persons are bitten by rabid dogs than ever develop hydrophobia. May it not be conjectured that there is actually a greater predisposition in males to develop the disease ? Is the proportion of females bitten, but showing no after ill-effects, greater than shat of males? This is an interesting point for future inquiry. It may be well to summarize the facts in a short Table. STUDIES IN STATISTICS CHAP, Analysis of 1001 Deaths from Hydrophobia in England and Wales during the forty years 1849-1888. Children under 10 years or Persons aged 10, but under 55 years Persons aged 55 years and upwards Total . Males. 2.1 Females. Total, “ 305 617 79 1001 At the Dogs’ Home the rabid animals are distinguished eeording to sex, thu 1883 1884 18K" Te. Total. 13 15 30 &8 From this it would appear that six times as many dogs suffered from the disease as bitches. Unfortunately, no record is kept of the total number of dogs and bitches admitted, so that it is not possible to calculate their respective death-rates. Mr. Batt is, however, of opinion that dogs are more liable to the disease, and that the difference is not to be accounted for by a smaller number of bitches, nor by these being kept in greater seclusion than dogs. Prof. Coleman stated that when rabies finds its way into a kennel of foxhounds the mad dogs bite the logs but spare the bitches. Among 892 collected instances, 356 dogs and only 36 bitches were bitten! Hydrophobia is, in this country, in the vast majority of cases, communicated to man by rabid dogs; but of 246 cases that were registered in the five years 1876 to 1880, as many as seven were attributed to the bites of cats. On the continent of Europe hydrophobia is often due to the i 8ir T. Watson, Principles and Practice of Physic, 5th edit., vol. 1. n. 614. KXI HYDROPHOBIA STATISTICS 431 bites of rabid wolves,—a danger that we are happily exempt from in this country. TasrE VIIL— Occupations of 152 persons who died of Hydrophobia vn FEnaland and Wales in the five years 1876-80. Labourer ... ve Agricultural labourer Porter i. ee Collier ... .es Blacksmith “in Carpenter, joiner ... Mason ... es Cotton operative ... Weaver ... iy = Tailor, clothier ... ee Clerk gu ve I. Publican, beerseller, potman Groom, ostler, coachman ... Livery-stable keeper Pa Farmer, farm-bailiff a Gardener ... (Gamekeeper Carter ... Drover ... Shepherd ... Letter carrier as Errand boy a vee Hawker, Costermonger ... Veterinary surgeon 2 a = J gE @ = 22 | Eg — b> 15 3 15 3 4 | = 3 o [a 3 2 34 18 4 19 3 211189 Occupations.—For the five years 1876-80, the Registrar-General published the occupations of 246 persons —194 males and 52 females—whose deaths were attributed to hydrophobia. But an analysis of the occupations does not give any very interesting results. These 246 persons boo STUDIES IN STATISTICS CHAP, are described as having upwards of 100 different occupations. I will only call attention to the cases of the above 152 persons, which have been selected, either owing to the fact of several persons of that occupation having suffered, or to the fact that there is something suggestive about the occupation itself. (Table VIII.) Geographical Distribution.—The deaths from hydrophobia in England and Wales have been by no means equally distributed. Thus, taking the period of thirtytwo years, 1852-83 inclusive, out of 789 deaths there sceurred in each registration county the following numbers — TasLe IX. —Geographical Distribution of 789 wm 32 years, 1852-83. Lancaster... ... 225 West York ... 100 Met. Middlesex 71 Durham ... ... 34 Chester ... ... 33 Metrop. Surrey ~~ 31 Stafford ... ... 29 Extra-Met. Kent 1!® South Wales .... 7 Nottingham ... = Northumberland Extra-Met. Surrey © Derby ... ... iI Sussex... ... 1 Hants .. 0 Warwick ... ... 10 Fssex . .. 10 East York . 10 Shropshire ... 9 Extra-Met. Middlesex ... ... Leicester and Rutland ... ... Buckingham ... North York ... Monmouth Somerset ... Devon ... Hertford ... Wilts... Cambridge Oxford ... Cumberland North Wales r Deaths from Hydrophobia Gloucester Bedford ... Cornwall ... Berks cer een Suffolk ... ... Worcester vee Lincoln ... ... Dorset ... ... Metrop. Kent ... Norfolk ... ... Northampton .:. Hereford ... ... Huntingdon ... wv Westmorland ... 0 Total ... 789 4 3 By a rough method of approximation, based on the numbers enumerated in 1871 and 1881, the death-rates in the various counties may be obtained for the twenty years 1864-83. (See Table X.) Plate X¥F TNGLAND & WALES Scale of English Miles 050 MW 20 30 40 50 pee ed ———t Average death-rate from HYDROPHOBIA Twenty years 1864-83. {53 J) ATED tt — 3:6 per million living 27-2 br... 9-10 09-0 lpcececw, bY gas Wes se Yes ef : W.Gr London: Edward Stanford 0 TES London, Stanfords Geog® EBstab® XX1 HYDROPHOBIA STATISTICS 423 LasLe X.—Hydrophobia Deaths, and D:ath-rates per Million in each County of England, and wn North and South Wuales.—20 years, 1864-1883. comprising 707 deaths. Lancaster ... ... Chester on Bucks — West York ... Durham ... Nottingham ... Metrop. Surrey ; Northumberland ... East York ... » Derby ws - Extra-Metrep. Surrey Shropshire ... yor Stalford .e. ve. Extra-Met. Middlesex Metrop. Middlesex ... Leicester and Rutland Hertford i _ Extra-Met. Kent ... Sussex I. Oxford ve North York ... Cambridge... Wilts S40 [Peco 214’ 36 30 2 8 26 3 23 29 | 2°1 15 | 21 29 11g 14 18 0 16 1 16 12 8 26 9 (2 8 5 15 nl 3 14 13 13 3 12 12 I'2 12 ge) 0 Bedford —-— Hants ven Somerset ... South. Wales Dorset i. Devon cee Cornwall ... Cuniberland .. North Wales Berks een Suffolk cee Metrop. Kent Warwick .. Worcester ~~ .. Gloucester ... Norfolk vis Monmouth ... Lincoln ... Northampton Huntingdon ... Hereford — ... Westmorland Eneland and Wales 10 : 09 07 06 o'5 og o's oy 2’'5 24 04 04 04 0'3 22 o'2 o'2 OT nN D'o 707 1g N.B.—Of the 214 deaths in Lancaster, some occurred in every one of the years 1864-1883, with the single exception of 1868. Dut there was no death in that tounty in 1854. 1855, 1856. nor 1858-1862. inclusive. The preference shown for Lancashire is seen to be very remarkable ; Chester, Buckingham, and the West Riding of York coming next, but a long way behind. At the other extremity come the thinly-populated counties of Northampton, Huntingdon, Hereford, and Westmorland, in which an aggregate mean population of close upon half a million (498,223) kept quite free from hydrophobia for the twenty years under consideration.? ! Compare Plate V., p. 29, showing the density of population in the various counties of Inoland and Wales Tr. STUDIES IN STATISTICS CHAP, It will be said that the populations of many of the counties are too small to yield death-rates of any value from such a rare disease. There is less force in this objection than might be supposed. When the counties are coloured in a map with intensities proportional to the death-rates given in the Table, it will be seen that, with one or two exceptions, the patches of colour are grouped in a very definite manner (Plate XXX.). Thus the maximum mortality of Lancashire forms part of a patch of dark colour, including Cheshire, the West Riding, and Nottingham, while the home counties, with Cambridge, Oxford, and Buckingham, form another group of somewhat high mortality. The south-western counties and Wales form pale areas, connected, through Hereford, Gloucester, Warwick, Northampton, ‘and Huntingdon, with Lincoln, Norfolk, and Suffolk. There is another pale area in the north; but Durham suffers from an exceptional mortality. The slight exceptions of Wilts and Buckingham are of little importance, since the population of cach of these counties is small, being under a quarter of a million, and the rates are based on five and eight deaths respectively. The mortality from hydrophobia appears to be closely related to the density of the population, with the striking exception that London is comparatively exempt. Can this be explained by the efficacy of police regulations in the metropolis ? In Table XI is shown the number of deaths from hydrophobia in each of the registration divisions of England and Wales for the thirty-four years 1850-83. These arcas are not liable to the charge of being too small in population, since the smallest of them has more than one million of inhabitants. An examination of this Table would of itself be sufficient to prove, if such proof were needed. that the disease was a communicable one. TABLE XI.— Deaths from Hydrophobia in the Reaistration Divisions - elon d ged Wades dn each 0.” Fmtorfour years, 1850-1883 1 Divisions. | oboe $=3,3 @ | w — — | —-© Total, 34 Years, — _.—London ... es Sh eee] eee en fees HL I | 5 21 31 og, IL. —South-Eastern ... 3m. 11 i wey il cles 51 2[...1 2 6 TIL.—South Midland ... ATT ng ms 1 1]1] 2 | 1V.—Rastern ... ve .. “tem von ww Teun) ous Dove | wen] osm 2h] les 2 an ae] 1 V.—South-Western ... re Cenc geee ren feme fern food deere Tipe le fee beg ee 20 401 2... Vi, —West Midland ... 5 2]teet 20 Ti |etonigeeey 2 30] 0 8) 8] BE 2] 1 1 2 VY . North Midland... “ cde deen fen] Tho en feb I TL AL BY 4] 1 4 Cpa 2) 1 ¥ North-Western ... vad 2h] ee] 2]10] 4(14| 2]...| 6|12]22|16(15117{17| 8 19{22{16|10|15] 9|1 Jorkshire wih ... To] Lhd [en ee] oF BY TLI0(16Y 6) 2{15(10|1. A] 5B] 7| 6] 6] 8i.. Northern 4% .. cole tnbld er 1) 5E 84. 21 6) 6) 4b 4) 2] 1] 6 3] 2... tl] -Monmouth and Wales ... ait -] 1 3] af] 2ec|odin eee neon [nd 2] a LTB] ENGLAND AND WALES ... 113 25 18.0110 1 5| 3| 2 4| 3| 4] 1| 4|12|19|36{10| 7|18|32,56|39|28|61{47(53|79|50/35 20/54/283-Divisions affected... ... | 71 9,5 5 Zl ° Af 2 2 8) 8] 1| 1| 2| 2| 7| 7| 5; 3] 4 ‘ 8 0 6| 91 811 11h gy sl 8] 7 Counties affected ... ce ceelnn Pt 6] 8/11] 5| 2) 2 4; 3! 2} 1| 2{ 3} 9|11 7| 3| 5] 9/18[12(10{18(15{18|26/20|12|12|10{14|1 woo Ly ones 2.0. --g 106 59 or a 265 123 58 a8 827 +7 Thus it is seen that for the first fifteen years not a death occurred in the south-western division; then a solitary death occurred (in Devon); after this there was another interval of ten years, followed by a period of six years during which the annual deaths varied from one to nine. Take as a contrast the north-western division ; for the first four years there were annually from two to five deaths ; then a blank year, followed by one death ; another blank, followed by two deaths; then complete absence of the disease for five years, after which the disease, with the exception of 1868, was fatal every year, killing in some years as few as two persons, but more asually 10, 12, up to 17, 19, and even 22 in the year. The comparative immunity for a long series of years of isolated counties like Cornwall, suggests that more strenuous police measures against rabies, if carried out uniformly throughout the country, and for a sufficient length of time, would be successful in stamping out the disease. STUDIES IN STATISTICS CHAP. Summary. —Briefly, the points to which I would call your attention are these :— 1. Hydrophobia is stated to have killed 944 persons in England and Wales in the thirty-nine vears 1847— 1883. 2. During the latter half of this time it has been about five times as fatal, relatively to the numbers living, as it was during the first half. 3. Nearly half the above number of deaths occurred during the eight years ending with 1878. 4. As a rule, London has not suffered so much as other parts of the kingdom. 5. Hydrophobia in man is most fatal towards the end of summer and throughout the autumn, {XT HYDROPHOBIA STATISTICS 437 6. So far as has been ascertained, rabies in dogs is most fatal in late autumn and in the spring, and least fatal in the summer. 7. Directly or indirectly, hydrophobia appears to be more influenced by season than scarlatina or enteric fever, but much less so than diarrhea. But neither exceptionally hot summers nor cold winters appear to influence the diseasc. 8. The police order which came into force in December 1885 checked the spread of rabies, but brought a large number of cases into the Dogs’ Home. 9. Hydrophobia is relatively most fatal between the ages of five and fifteen. 10. At all ages, except the extremes of life, very many more males die than females. The sanve is believed to be the case with rabies among the lower animals. 11. Occupation seems to have comparatively little influence. 12. The geographical distribution of hydrophobia is peculiar. It is most prevalent in Lancashire; Cheshire and the West Riding of Yorkshire coming next. It is least prevalent in the Lake district, Wales, the southwestern counties, and the eastern counties, and in a strip of country stretching from Hereford to Huntingdon. 138 STUDIES IN STATISTICS CHAP. CHAPTER XXII. CONCLUSION. In the preceding pages my constant endeavour has been to lay before the reader, as clearly as possible, matters which in their very nature are prone to be both difficult and dull ; but in so doing there has been an almost inevitable tendency to be either too prolix and detailed, or unduly dogmatic. The general reader will therefore complain of unnecessary detail, while the severe statistician will say that I have thrown into the form of diagrams facts that should rather have been displayed in lengthy tables. To both forms of error I plead guilty. The objects that I had in view were mainly two : Firstly, to urge the economist and the politician to keep always in view the scarcely perceptible changes due to the constant growth of population—changes which they can indeed ill afford to neglect; secondly, to indicate to medical writers the vast mine of material existent in the voluminous returns of the Registrar-General—material of which the great majority of the medical profession scem to be still ignorant, although it is plainly desirable that every hypothesis put forward by searchers after the causes of disease should be tried by the touchstone of this mass of crystallized experience. The cardinal fact for the statesman and political economist 1s the marvellous growth of population that is XXI11 CONCLUSION 439 taking place in the present century ; if I have succeeded in any respect in making this more striking than it before appeared, I shall not have wasted my labour. Subordinate to the main fact is the local growth of population. At home this is gradually changing us from a country to a town people, and so greatly influencing our social, political, and sanitary conditions. Abroad, far beyond the seas, the growth of large populations in remote places, which were but the other day mere wildernesses, is producing unnoticed, though before our very eyes, new nations—nations which, while supplanting the picturesque, if filthy, customs of the savage, by the healthy but hopelessly unpicturesque civilization of Europe, are gradually but surely changing the centre of gravity of the political world. The innate conservatism of statesmen tends to make them think of little or no account movements that take place outside the family circle of European diplomacy. While I have, naturally enough, treated in the fullest detail the growth of population near home, and the migrations of our own people most familiar to us, I have also called attention to other, less familiar, migrations, such as that from the shores of the Mediterranean to the Pampas of the River Plate and the mines and forests of Brazil, and that from the rugged coasts of the Scandinavian peninsula to the vast prairies of the Western States of America. The racial characters of the new nations that, as a result of these movements, are springing into such vigorous life in the far West and South, I have treated of in considerable detail, since the subject appears to me to be one pregnant with interest to the historian and philolooist, though it has as yet been but little studied. One great omission will probably have been noticed, 440 STUDIES IN STATISTICS . -CITAP. viz. that I have almost ignored Russia, the most populous. of all the European nations. I have done so designedly, and for two reasons. Though Peter the Great, among his many reforms, did not neglect the numbering of his people, Russian statistics have of late years been much behind the age and scarcely comparable with those of other European countries. Again, Russia seems to me to be essentially an Asiatic Power, and for various reasons I have thought it better to let Asia alone; my primary concern was with Britons and with the new nations of the earth that have been founded, or are in the process of foundation, by Britons; as yet Russia has had but a trifling influence upon any of these. [ have called attention to the indisputable fact that the alleged “ depopulation of the rural districts,” and the movement of the people into towns, is not a phenomenon peculiar to this country, and is not to be explained as the result of a system of land tenure, but depends upon causes of much wider reach, embracing alike countries under the rule of military autocracies, and democratic republics—eauses which operate alike on these islands and on the continent; in the Old World and the New ; in the north and in the south : in the east and in the west. None of the data of statistics are mathematically exact ; thus it is well known that even a census, however skilfully planned, and however carefully carried out, fails to ascertain the exact number of the population of a country, but the discrepancy, even if it amount to thousands, is very small in proportion to the numbers dealt with, and may therefore be neglected in practice. Of course such numbers as those in the Registrar-General's Reports, indicating the alleged causes of death, XXII are subject to proportionally far greater and more serious arrors. There are, however, two considerations which tend to minimize the uncertainties that might be expected to arise from these errors and uncertainties. The first is the familiar principle of statistics, that large numbers are, other things being equal, more to be trusted than small. The facts for a village may be most fallacious ; for a large town (since the perverseness or ignorance of an individual therein can produce but a small result in the whole) the facts are more valuable than for the village ; a fortiori the returns for the whole country will be more trustworthy than those for any single town. Speaking generally, in a large collection of figures of any kind, there will be numerous errors, but these are not likely to be all in the same direction; in many cases they will tend to balance each other more or less, and the larger the aggregate the more likely is this to be the case. The second consideration is this: the facts for one year may with advantage be compared with similar facts, similarly collected, relating to the years that precede and follow. It will be recollected that in the great majority of cases I have dealt with large figures and have compared Like with like. The results have, to my mind, shown that the statistics of the alleged causes of death are more nearly correct than might have been supposed. Among the points to which I would wish to revert are the altered fatality of certain causes of death at certain ages ; the great saving of life among children ; the slightly increased mortality at the later ages; the increase under the heads of cancer, -Lieart disease, bronchitis, and other liscases of the lungs; the diminution of fever. scarlet fever, and of phthisis or consumption. In reference to the last point,—the decrease in the number of deaths attributed to consumption,—it is, I CONCLUSION 441 142 STUDIES IN STATISTICS think, a point of importance that an examination of the figures proves that the increase under the head of bronchitis is not due to a mere transference from phthisis, but that the one disease has really increased in severity, or in prevalence, while the other has diminished ; this is proved by the fact that the ages at death differ materially in the two diseases. The indications that rheumatic fever tends to be more prevalent when erysipelas is more prevalent, and vice versd, are very interesting, and may possibly lead to results of value. The facts brought to light as to the geographical distribution of diarrheea, diphtheria, and hydrophobia are curious, if difficult to explain. The chapter on the census may seem somewhat out of date, but I do not yet quite despair of moving the stony heart of the Treasury. My hope throughout has been that increased knowledge of facts, alike in matters political and matters medical, may tend to make legislative and administrative efforts more reasonable, and less empirical. Sure progress is more likely to be attained by the diligent and patient study of details than by the more showy and more attractive method of @ priort speculation, which is but too apt to lead to rash and possibly disastrous experiments. VOTE 14 3 NoTE oN AN EripeEMic oF INFECTIOUS PNEUMONIA IN THE EIGHTEENTH CENTURY. IN the parish registers of George Ham, in North Devon, the following remarkable note seems to record a severe epidemic of pneumonia in the vear 1746-7. «egg All these marked * died of a violent Pleurotic Fever or Peripneumene (sic), most of them in four or five days after the first seizure. ¢ Memord Jon WW." It is not evident who “John W.” was, but the note appears to be contemporaneous with the entries, if not actually in the same handwriting. The twenty-two names in the register marked with the *¥ are given below. West Down is an adjoining parish. My attention was called to the passage by the late Rector, the Rev. W. E. Cox (now Rector of Lynton); the present Rector, the Rev. W. H. Hole, has been good enough to check my transcript with the original. DEATHS. JE * Thomas Bennet Margaret Benne! * John Davies * Agnes Gorden * David Bennet * Mary Woollin * Joan Dyer * Toby Dyer Alarv Odann 1747. * Francis Adams —. —. * Mary Nutt . ee * Zacharias Dyer ... . Susanna Hooper ... *John Thomas, of West Down t Elizabeth Thomas, of West Down *Rleanor Knill a * Morrice Michael ... * William Passmore ... * John Howard . * John Skinner ... * Elizabeth Austen ... Anne Loelo March 20. 3 ER] March 22. March 24. + 5 ” I’ March 25. March 26. 23 3 March 27. March 31, April 5. April 6. April 8. April 9. May 7. May 8. May 24. May 27. Tune 5. ERRATA. page 11, second line from bottom, for reproducive read reproductive. «220, bottom line, read Mr. Charles Booth and Prof. Marshall INDEX $313 INDEX. AppoT, Mr, Census Bill of 1800. 196 Aborigines of Australia (Table), 145-147; of New Zealand (Table), 146-147; of Tasmania, 146; in United States (Table), 66-68 Acts dealing with Public Ifealth, 19. 847, 409 Age as a factor in death-rates, 6, 245 Age-constitution of population, afactor in birth- and death-rates, 9-11, 246, 251 Alberta. See Canada, provisional districts Amalgamation of races (Tables), 86-89, 92-95, 98-101 American cities, mortality statistics of (Table), 59; rapid growth of (Table), 159 Approximation to unknown terms of a series, methods of, 242, 243 Argentine Confederation, census of (Table), 132; development and prospects of, 132; migration to (Table), 51, 135, 136 ; nationalitics in (Table), 132 Arithmometer, saving of labour by, 5,227, 311 irsenical poisoning, a cause of winter diarrhoea, 281 \ssiniboia. See Canada, provisional districts Asthma, decreased mortality from (Table), 874 Xuckland, population of, 157 Australasia, birth-places of people (Table), 147; growth of population (Tables), 141-145 ; Irish in, 146 ; race-constitution (Tables), 145-150 Australia, aborigines, 145 Austro - Hungarians, emigration of (Table). 48. 106-109 BaLrour, Dr., on Census of house accommodation, 218 Baltic countries, See Scandinavia Barnes, Dr. Edgar G., on geographical distribution of diphtheria and scarlatina, 355 Barrow-in-Furness, rapid growth of (Table), 160, 162 Basutoland, 138 Bath, declining population of, 163 Batt, Mr. Ernest, statistics of rubies (Table), 418, 419, 430 Battersca, rapid growth of (Table), 161 Bechuanaland, 138 Belgium, migration from, 49, 106, 107, 110, 136 Birth-places, unsatisfactory tabulation of, 222,224 ; value of Tables of, 43 Birth-rate, curve of (Plate), 12; definition of, 5; dependent on ageconstitution of population, 11,251 ; affect of, on death-rae, 11 ; effect of marriage-rate on, 13-15, 252; effect of migration on, 252; illegitimate, 13; recent fall in, 14; variations of, since 1838, 12-14 Birth-rates, in European countries, variations of, 14 Jirth-register, defective at first, 12 3irths, in England and Wales, 12; illegitimate, 14; imperfect registration of, 42 ; proportion of mules to females, 249; registration of, made compulsory, 12; registration of, now fairly complete, 13 Bloxam, Mr., modification of curves in graphic method, 321, 322 Board of Agriculture, rabies order by, 422; report on rabies (Tables), 424-426 Booth, Mr. Charles, on census of emvlovment, 207. 220. 221 146 INDEX Boxall, Dr., on fever in child-bed, 328 Brain diseases, increased mortality from (Tables), 230, 239, 241, 254-264 ; in relation with convulsions, 239 ; sex and age mortality from, (Tables), 239, 254-264 Brazil, census of (Table), 128, 129; development of, 128; migration to (Table), 48,51, 130 ; race-constitution of (Table), 129, 130 Brisbane, growth of population, 157 British Empire, abstract of census of, desirable, 222-225 British Islands, inter-migration between, 38-41, 43 British in United States (Tables), 95-101 Bronchitis, and associated diseases, 312; causes of death studied in connection with, 871; geographical distribution of (Table), 380-382; increased mortality from (Tables), 374,875, 441,442 ; weekly Anctuations of mortality from (Plates), 376-378; yearly fluctaations of mortality from (Plates), 371, 372, 376 Brown Institution, rabid Dogs at (Lable), 419-423 Buchan and Mitchell, Messrs., analysis of London weekly mortality, 277, 278, 283, 286, 319-321, 369, 373, 375; graphic method, 285, 311; influence of weather on mortality, 277, 278, 280, 281, 283, 286; seasonal prevalence of continued fevers (Table), 898, 401 Buchanan, Dr., causation of diarrheea, 283 ; method of comparing certain curves, 282, 322 Buenos Aires, city, b'rth-places (Tables), 184 ; education in (Table), 135; migration to (Table), 136; nationalities of foreigners (Tables), 134, 185; population of (Table), 133, 134; rapid increase of, 159; sex-proportions in (Table), 134 Buenos Aires, province, census of (Table), 183; nationalities in (Table), 133 Burgess, Mr. Arthur, cartoon showing orowth of Greater London, 175,177 of, 127; migration from United Kingdom to, 114, 115; national and religious animosities in, 122, 123; nationalities of immigrants to, 121; negroes in (Plate), 150; origins of the people (Table), 119-121 ; population, rapid increase of, 114; provisional districts, increase of population, 116; provisional districts, race - constitution of people, 116; railway communication in, 127; Scandinavians in (Table), 80; Scotch in, 119, 120, 122; and United States, mixture of races compared, 117, 118, 125 sanadians, character and prospects of, 125-127 ; in United States (Table), 74-78, 101, 102 Jancer, age and sex-mortality from (Tables), 239, 254-264 ; hereditary tendency to, 240; increased mortality from (Tables), 230, 239-241, 248, 249, 254-264, 267, 268, 441 Jape Colony, census of (Table), 137; dependencies of (Table), 138; migration to (Tables), 188-140; neglect of statistics in, 137; raceconstitution (Tables), 137, 138 Case-mortality in Metropolitan Asylum Hospitals (Table), 410, 411 Causes of death, according to sex and age, 1861-1870 and 1876-1880 (Tables), 254-264 ; changes in age ind sex-mortality summarized, 241, 244, 252, 253 ; classification of, 228, 229; difficulties from changes in zlassification, 232 ; improved certification of, 228, 229, 231, 235, 252; rise or fall in (Tables), 229, 230, 254-264, 267 Jensus, Bill of 1753, opposition to in House of Comnions, 194, 195; Bill of 1753 thrown out by House of Lords, 195; Bill of 1800 passed without opposition, 195, 196; of British Empire, abstract of, desirable, 222-225 ; cost of, in United Kingdom (Table), 2056-207; cost of, in United States, 207, 208; importance of, for all statistical measurements, 199-205 ; influence of Statistical Society on arrangements for, 197 ; intervalsin various wountries, 198, 199; operations in weient times, 191-193; opposition to, obsolete, 197; quinquennial, 1ecessity for, 199-205, 223 ; recom~ mendations of Statistical Society for, 208-213, 222; results, tabulation of. 214-223: scope of. 208-CANADA, birth-places of population (Table), 117,118; censuses of, 113, 114. 193, 198; climate of, 126; Dutch in, 78 ; French in, 74, 75, 119-125; future government of, 124-126 ; geographical structure nf 1928. 127: internal navieation INDEX 214, 222-924: of United States, reason for elaboration of, 208 Censuses, of Canada, 113,114, 193,198; taken in 17th and 18th centuries, 198, 194; of United Kingdom, machinery for taking, 196, 197 Chamberlin, Mr. J. E., on foreign elements in population, 112 Chapels, additional, required before 1901, 28, 29 Charity Organization Society, 179 China, small migration of women from, 69, 73, 108, 109 Chinese, migration of, to Queensland, 144; in United States, 68, 69 “holera, epidemic of, a cause of sanitary legislation, 17,18; epidemics, affect of on diarrhoea mortality (Plate), 285 ; epidemics of, exceptional, 16; simple or English, a forin of sinner diarrhea, 281 Churches, additional, required before 1901, 28,29 Circulatory organs, diseases of. See Heart Diseases Cities, tendency to aggregation in, 156. See also Towns Civil registration, commencement of, 12 Classification of diseases, according to effects of weather, 369, 370 ; danger of carrying too far, 323; in Registrar-General’s Reports. changes in, 324 Colam, Mr. J. C., statistics of rabies (Table), 419-421 ; Mr. M., statistics of rabies (Table), 419-421 Coleman, Prof., sex-incidence of rabies, 430 Colonies, British and Foreign (Table), 53 ; British, increase of population in proportion to mother country (Table), 150, 151; countries possessing, 52, b3 Jonsanguineous marriages, and idiocy, deaf-mutism, &c., 213, 214; as a subject for census inquiry, 213, 214 Continued fevers, See Fevers, Continued Convulsions in relation with brain diseases, 239 Cornwall, diminution of population in, 36 Counties, changes of population in (Table and Plate), 35-37 Jountry districts, migration from, to towns, 24, 25, 164 Sourtney, Mr. Leonard, on increase of foreigners in Ireland, 41, 42 Jox, Rev. W. E., epidemic of pneumonia in 1746/7. 443 447 Crouch, Mr., migration of Bohemians to United States, 106 Croup in relation with diphtheria, 316 Oroup. See Diphtheria and allied diseases Jurve, technical use of term, 7 Jynanche Maligna. See Diphtheria and allied diseases DaxnEs in Utah (Table), 104 Davidson, Dr. J. T. R., on a source of diphtheria, 355 De Beaucaire, M. Le Vicomte, on migration into towns, 164 Death rate, curve of (Plate), 12; definition of, 5; effect of birth-rate on, 11; effect of changes in, on increase of population, 251, 252; sffeet of epidemics on, 15-17, 32; effects of season, fanine, war, and pestilence on, 81; influence of age on, 6; influence of sex on, 6; low, in recent years, 226, 264,165 ; as a measure of health, 5: variations of, 15 Jeath-rates, according to sex and age, [876-1880 and 1881-1885 compared (Table), 265; according to sex and age, 1881-1885 and 1886-[588 compared (Table), 265, 266 ; according to sex and age, changes in 50 years (Table), 268, 269 5 cor-‘ection of, for sex and age, 10, 246 ; affects of weather on, 206, 322; recorded and corrected, 9; in town and country compared, 9-11, 245, 246 ; at various ages (Table and Plate), 6, 7 Deaths, in England and Wales, 12; from various causes, recent rise or fall in (Tables), 229, 230. 254-264, 267 Nebt, local, recent increase of, 30 denmark, tabulation of occupations in, 221. See also Scandinavia Depopulation, alleged, of rural districts in England, 36 ; of rural districts in France, 164 Developnental diseases, decrease in mortality from (Table), 230, 234, 235 diagnosis, importance of accuracy in, 393 ; increasing accuracy of, 235 ; statistical errors caused by mistakes in, 272 diarrhoea, age and sex mortality from (Table and Plate), 283-287, 308, 309; and associated diseases (Plate), 280, 281, 283, 312; available statistics of. 271, 273: and density of INDEX population (Table), 353-355; epilemics of, 16, 17; geographical distribution of (Table), 273-277, 279, 280, 290-308, 442; inquiry into causation of, 277, 288, 289; in London in relation with temperature of Thames (Piate), 282 ; wortality, compared with birth-rates (Table), 298, 299; mortality, method of comparing different towns, 293; and sanitary reform, 300; predisposing causes of, 271, 276, 277 ; probably a specific cause of, 271 ; seasonal prevalence, 286, 287 ; and weather conditions (Plate), 278, 279, 281-283, 289, 290, 205-293, 300; weekly flnctuations of, in London (Plate), 277-279, 295, 296; whether a ““zymotic” or a “climatic” disease, 20; winter, distribution of, 279, 280; winter, sonietimes caused by arsenical poisoning, 281; yearly fluctuations of (Plate), 281 Diarrhoea, summer, causation of, opinions of several medical offi cers, 304-308; a communicable zymotic disease, 283; connected with putrefaction, 288, 290, 295, 297, 309; death-rates from, in several towns (Diagram), 273-276 ; a distinct disease, 280, 283, 287 289 ; increasing mortality from, ir certain towns (Table), 306, 307 mortality, 1870-1879 and 1880 1887, in seventy towns, compared (Table), 300-303 ; a town disease 274, 276, 277, 287 ; untenable theo ries as to causation of, 288 Diarrheeal diseases, variations in mortality from, 237 Dilke, Sir Charles W., Problems of Greater Britain, 138 Diphtheria, admissions to Metropolitan Asylum Hospitals, 410; and allied diseases, fluctuations in mortality from, 315, 316; and density of population (Tables), 339-355, 358, 360-867; districts with highest and lowest mortality, 350, 351; first distinguished In Registrar Generals returns in 1855, 314; first distinguished in Registrar-General's returns for Londen in 1859, 320; fluctuations in mortality from, 236, 287, 267; geographical distribution of (Tables), 338-367, 442; increasing prevalence in towns difficult to explain, 346: male and female mortality from, compared (Tables), 335-337, 356, 357 ; mortality from, compared with general death-rates in certain districts, 851; mortality from, in dense, medimm, and sparse districts Tables), 343-355, 361-367; morsulity from, and general death-rate, in Registration counties and divisions (Tables and Plate), 838-342, 358, 360; mortality from, greatest in villages and small towns, 349 ; as a “new disease,” 315; possible sources of, 352, 353, 365; recent increase of mortality from (Table), 526, 334; a rural disease, 274; weekly variations of, in London (Plate), 819, 320 Diseases, classification of according to effects of weather, 369, 870; conditions for forecasting prevalence of, 243, 244 ; possibly modified by the individual attacked, 823, 324; sex and ave incidence of (Tables), 248, 249, 254-264 Divorced persons, probably imperfect returns of in census, 210 Dixey, Mr. F. A., diagram showing varying temperature of Thames, 282 Dogs’ Home, Battersea (Tables), 419-423 Dropsy, decreased mortality from, probably only apparent, 236, 267 ; in relation with heart diseases, 238 Ducie, Sir R., estimate of population of London in 1631, 170 duncan, Dr. Matthews, on puerperal fever and erysipelas, 327-330 Dunn, Dr. H. P., on increase of cancer, 240 Duration of life, data as yet insufficient to forecast, 247, 248; Mr. Noel A. Humphreys on, 226, 242; probable future changes in, 245, 253 Durham, growth of population, 28 Dutch, in Canada, 78; in United States, 7 EpucATION, as a subject for census inquiry, 213 Elizabeth, Queen, proclamation to restrict growth of London, 170, 176, 180 migrants, sex-proportions among (Tables), 108, 109, 148, 149 Emigration, centres of, 44; dangers of aiding, 180, 181; effect of on birth-rate, 252; effect of on sexconstitution of population, 249; recent increase of, 25 England. daily increase of nopulation. 23-25; growth of large towns in (Table), 160 Enteric fever,mortality from,in London 1869-84 (Table), 396, 397, 410; seasonal prevalence of (Tables and Plate), 349-405, 407, 411, 412 ¢Epidemic,” a vague term, 327, 328 Ipidemies, death-rates from (Plate), 18; effect of, on general death-rate, 15-17; permanent injury from, to survivors, 17: since 1838 (Table), 16 Trysipelas, geographical distribution of (Table and Plate), 330, 331 ; recent fluctuations in mortality from Table), 325, 832; in relation with puerperal fever (Table and Plate), 312, 316-319, 321, 827-331; in re-‘ation with rheumatic fever (Table and Plate), 317, 329-331, 442 Turope, summary of migration from, 152-154; Western, calculated populations at various periods (Table), 183, 184 ; Western, recent growth of population (Tables), 182, 183; wheat exporting and importing countries of (Table), 184, 185 Furopean countries, migration from, compared (Table), 49 Farr, Dr., classification of diarrheea, 20; formula for calculating casemortality, 411; on summer diarrheea, 282, 283 Females, excess of (Tables), 8, 9, 249. 250, 253 Fever, decrease in mortality from (Tables), 230, 235, 236, 241, 254-264, 267, 268. See also Enteric fever, Simple continued fever, and Typhus Fevers, continued, mortality from, in London 1837-84 (Tables), 395, 396; when first separately distinguished and classified, 394 Food supply, effect of, on growth of population, 22 Forster, Mr., conscience clause in Irish Census Act, 212 france, depopulation of rural districts in, 164; emigration from, to South America, 75; foreign-born in Tables), 38, 42 ; migration from, to Argentine Republic, 136; migration to, 88; nationalities of forsigners in, 165; restriction of native population, 838; small emigration from, 50, 51 French in Canada (Table), 74, 75, 123-125: in United States. 74-78 INDEX 449 FERHARD, on classification of continued fevers, 394 Jerman colonization, 53 Germans in New York City (Tables), 88, 92; in United States (Tables), 82-89, 91 Fermany, emigration from (Tables), 48, 60, 106-109 Giffen, Dr., on einigration to Australasia, 47; on net emigration, 46; on prospects of food supply, 188, 189 Graphic method, explained, 7; value of, in statistical inquiries, 310-312 Freater London, 169; growth of (Table), 174 aresswell, Dr. Astley, rainfall and scarlet fever, 826, 827 sreville, Mr. George, in support of Census Bill of 1753, 195 Griffin, Mr, Martin J., on religious census in Canada, 212 Srimshaw, Dr., on census of religious profession, 212 Frigualand. Bast, estimated population of (Table), 138; West, estimated population of (Table), 138 Hauirax, low diarrhoea mortality in (Diagramn and Tables), 275, 276, 290, 292, 299, 301-303 Jarvests in England (Plate), 15 Haviland, Dr., on relation of dropsy to heart disease, 238 Hawaii, migration from Portugal to, 51 Hayden, Professor, estimate of wuncultivated arable land in United States, 188 Health, Acts dealing with, 19, 347, 409 ; measured by death-rate, 5 Jeart diseases, increased mortality from (Tables), 230, 238, 289, 241, 254-264, 267, 441; in relation with dropsy, 238; sex- and age-mortality from, 238 Hill, Miss Octavia, and industrial dwellings, 179 Hime, Dr., liability of children to Liydrophobia, 428 Hole, Rev. W. H., epidemic of pneumonia in 1746/7, 443 Holland, migration from, 48 Horsley, Mr, Victor, statistics of rabies (Table), 418, 419 House accommodation, census inquiry into, 209, 210, 223 Judson’s Bay Company, 157 full, high diarrhea mortality in (Table), 290, 300-303 Huniphireys, Mr. Noel A., on alleged increase of lunacy, 247; method a G 450 INDEX of estimating inter-migration, 39-41; on recent decline in deathrate, 226 . Hungary. See Austro-Hungarian Uydrophobia, in England and Wales (Tables), 414-416, 435; in London (Table), 414-416; communicated by other animals than dogs, 430, 431; comparative Immunity of London from, 434, 437; and density of population (Plates), 434; effect of police regulations in diminish. ing, 420-423; geographical distribution of (Tables and Plate), 432. 437; increase of, in recent years, 436; influence of weather on, 418, 423, 424 ; mortality in London in 1885 unprecedented, . 413; occu pations 6f persons attacked by (Table), 431, 432, 437 ; predisposition to develop the disease, 429 ; and rabies (Table and Plate), 418, 422, 423 ; seasonal prevalence of, in London (Table), 416,417,437 : sex: and age-mortality from (Lables) 426-430. 4837 [LLEGITIMATE births, 14; rate practically constant 1845-1865,18: recent fall in, 13 [mmigration, centres of, 54; as a factor in growth of population, 55, 56 Imperial Federation (Table), 151, 152 Indians in United States (Table), 66-68 Industrial dwellings, 179 Infectious Diseases (Notification) Act, 347, 409 Institutions, additional, required before 1901, 29 Intemperance, increase of kidney diseases due to, 241 Inter-marriage in United States, of British (Table), 98-101; of Germans (Tables), 86-89; of Irish, (Tables), 92-95 Inter-migration, 34, 37-43 Intra-migration, 34-37 [reland, first census of, 136 ; reasons for decrease of population, 45, 46 ; and Scotland, natives of, in different countries (Table), 148 Irish in Australasia, 146 ; emigration of, 45-47 ; “land hunger,” 45, 91; in New York, 92; in United States (Tables), 89-95 Italy, migration from, 50, 51. 106-109, 1364 JENNER, Dr., on classification of continued fevers. 394 Johnston, Dr. W., on summer diarrhoea, 283, 288 Kimpxey diseases, increased mortality from (Tables), 230, 239, 241, 243, 248, 249, 254-264, 267; sex- and age-mortality from (Tables), 239, . 254-264 Koch, Dr., discovery of tubercle bacillus, 368 Korosi, M., 166; on census-taking, 190; on tabulation of census results, 214-218 LA Prats. See Argentine Confederation Lancashire, bad sanitary reputation of, 293 ; hydrophobia mortality in (Tables and Plate), 432-434, 437 ; rapid inerease of population, 27 Latin countries, migration from, 50, 51 Leeds, high diarrhea mortality in (Table and Diagram), 290, 300-302 Leicester, high diarrhea mortality in (Table and Diagram), 274, 275, 290-292, 299-303; summer diarrhea in, 283 Little, Mr. W., on timber supply, 189 “ocal debt and rates, recent increase of, 30 “ondon, birthplaces of people (Tcble), 174, 175; comparative immunity of, from hydrophobia, 434, 437; contains most of foreigners in United Kingdom, 150; dangers of aiding migration from, 180, 181; day census of, 172 ; depopulation of central area (Table), 171, 174; sstimated population in 1631, 170; astimated population in 1682, 170; axtent of, 179 ; Fever Hospital, admissions to, and mortality in (Tables and Plate), 399, 402, 403 ; in fourth century, 169; immigration ‘rom country (Table), 24,25; Inner Ring, increase in (Table), 171,174 ; Local Government of, 179, 180; mortality of, compared with other large towns, 179; mortality of, n 17th century, 180; natives of, in provinces, 178; occupations of foreigners in, 178 ; occupations of people, 175, 176, 178; old bills of inortality, 173; Outer Ring, increase in (Table), 171, 174 ; past growth of, 169-171 ; poor, sick, “ afflicted,” and criminals in, 179; population compared with those of several countries (Table). 168. 169 : Queen INDEX Elizabetl’s proclamation to restrict growth of, 170, 176, 180; rapid mnerease of population, 27; with Outer Circle of towns, 169 Loua, M. Toussaint, on increase of rural and urban populations, 163, 164 Lubbock, Sir John, on census of consanguineous marriages, 213, 214 Lunacy, alleged increase of, 247 Lunatics and imbeciles, increase in, 29 Lung diseases, increased mortality from (Tables), 230, 237, 241, 248, 249, 254-264 ; sex- and age-mortality from (Tables), 237, 238. 254-204 MacDoxAaLD, Sir John A,, 126 MacKellar, Dr. A. O., Police returns of rabies (Plate), 424 Macoun, Professor, estimate of uncultivated arable land in Canada, 188 Male infants, excessive mortality of, in first year, 9 Malthus, lssay on population, 182, 187, 196 Manchester, fluctnations of diarrhoea mortality in, 298, 294; growth of, 163 ; unhealthy condition of, 9 Manitoba, growth of a city in, 116,117 ; increase of popu'ation (Table), 115, 157; race-constitution of people, 115 Marringe-rate, curve of (Plate), 12; definition of, 5; effect of, on birthrate, 13-15, 33, 252; influence of depression of trade on, 33; influence of Irish famine on, 18; influence of Russian war on, 15; influence of price of wheat on, 15; variations of, 14 Marriage-rates in European countries, variations of, 14 Marriages, consanguineous, and idiocy, deaf-mutism, &c., 213, 214; as a subject for census inquiry, 213,214 Marriages in England and Wales, 12 Marshall, Prof, on census of employs ment, 207, 220, 221 Measles, diseases resulting from, 237 ; apidemics of, frequent, 16; fluctuations in mortality from (Tables), 230, 237, 254-264, 207 ; in relation with lung diseases (Plate), 371, 872 Meat, countries exporting, 184 Melbourne, population of (Table), 158 Mental afllictions, Imperfect returns of, in census, 209, 223 Merthyr Tydfil, low diarrhea mortality in (Diagram and Tables), 275, 276. 200. 292. 268. 269 303 Metropolitan Asylum Hospitals, admissions compared with fever mortality in London (Table), 409 ; admissions to, a measure of prevalence of: fever, 403; admissions and mortality (Tables), 899, 400; 402-412; case-mortality in (Table), 410,411; deaths from fevers, compared with totals in London (Table), 408-412 ; increasing proportion of fever cases in London dealt with, 410, 412; recent admission of diphtheria cases, 410 Mexicans in United States (Table), 78,79 ° Middlesborongh, rapid growth of (Table), 160, 162 Migrants from Kuropean countries, percentages. of females among (Table), 108 Migration, balance of, 249; from Europe, summary of, 152-154 ; important influence of, on political world, 439, 440; influence of, on growth of population, 34 ; net loss by, in England and Wales (Table), 42; returns of] liable to error, 43; statistics, lessons of, 154, 155 Minor, Dr, Thos. C., on erysipelus and child-bed fever, 317-319 ditchell, Mr. See Buchan and Mitchell Moltke, Von, on prospects of next great war, 31 Monte Video, net migration at (Table), 131 Montreal, French difficulty in, 122; population of, 158 Mortality, at different ages, 8; effect of weather on, 20; of males and females, 89; recent changes in, 441, 442, See also Death-rate Murchison, Dr., Treatise on continued fevers. 394. 897-399 NATAL, migration to (Tables), 139, 140; population of (Table), 138 National Debt, recent reduction of, 29, 30 Nationality, as a subject for census inquiry, 213 Negroes, and self-government, 70, 71: in United States, 69-71, 112 New South Wales, growth of pojulation (Tables), 141-145; marriage-, birth-, and death-rates in (Table), 58; town population in (Table), 157, 158 New York, Germans in (Tables), 88, 92; Irish in (Table), 92 New Zealand, growth of population (TNDEX in (Tables), 141-145; Maoris in, 146 ; marriage-, birth-, and deathrates in (Table), 58; town population in, 157 Norway. See Scandinavia DocUPATIONS, classification of, for census, 210, 211, 224; difficulty of tabulating, 219-222; tabulation of, in Denmark, 221 Ogle, Dr. W., on depopulation of rural districts, 36 ; on mortality of buys and girls from violence, 429; on marriage-rates and marriage ages, 15 Orange Free State, census of, 138 Owen, Sir Hugh, method of estimating populations for Local Governinent Act. 1888, 201 ParkEs, Dr. E. A., manual of practical hygiene, 185 Pennock, classification of continued fevers, 394 Petty, Sir William, estimate of population of London in 1682, 170 Phthisis, causes of death studied in connection with, 871; decreased mortality from (Tables), 230, 233, 234, 254-264, 267, 268, 373-375, 141, 442; evidence insufficient to prove it communicable, 385, 393; mortality of females at ages over fifteen, 889 ; mortality of males at ages over fifteen, 887; not an epidemic disease, 372, 374; prooable deaths of husbands and wives from, in combination (Table), 392; probably a distinct disease, 234, 241 ; in relation with tabes mesenterica and tubercular meningitis (Plate), 374, 383; weekly variations of mortality from (Plate), 378, 876-378; yearly fluctuations of mortality from (Plates), 371, 372, 376 Pidgeon, Mr. D., Old World ques tions and New World answers, 65, 74 Pleurisy in relation with rheumatism of lreart, 375, 376 Pneumonia, causes of death studied in connection with, 371 ; decreased mortality from (Table), 375; excessive mortality of males from (Table and Plate), 378-380, 383; geographical distribution of (Table), 380-382; infectious, an epidemic of, in 1746/7, 443; not in general an epidemic digease. 383: weekly variations of mortality from (Plates), 376-378; yearly fluctuations of mortality from (Plates) 371,372,376 Police regulations as to dogs in streets, 420-422, 424 Polynesians, migration of, to Queensland, 144 2ondoland, 138 Population, causes tending to diminish, 30; density of, in counties (Table arid Plates), 26-29; of England and Wales, daily increase of, 23-25; estimates of, after long interval from census, 203-205; estimates of, need verification, 191; growth of, as affected by changes in deathrate, 251, 252; growth of, effect of epidemics on, 82; growth of in geometrical progression, 32; importance of studying growth of, 438, 439; influence of migration on, 34; precedence of Great Powers by, 57; probable increase of 1881-1901 (Table), 26; recent growth of, 22, 23, 25, 30; sex- and age-constitution of, effect on death-rates, 945, 246 ; of Western Kurope, growth of (Tables), 182-184 Populations of certain counties, diminution in (Table), 35, 36 Portugal, migration from, 51 Portuguese in United States, 79 Preston, high diarrhea mortality in (Diagram), 274, 275, 290, 291, 298-303 Preventive medicine, province of, 253 Price of wheat, influence of, on marriage-rate, 15 Orice- Williams, Mr. R,, on growth of London, 161 ) Priestley, Dr., definition of word epidemic,” 328 Public Health Acts, operation of. 235, 236 Puerperal fever, geographical distribution of (Table and Plate), 330, 831; more complete returns of, since 1880, 325; recent increased mortality from (Table), 325, 332; in relation with erysipelas (Table and Plate), 312, 316-319, 321, 327-331; in relation with rheumatic fever (Table and Plate), 517, 320-831; and scarlet fever, essential differences between, 329 Pymmia, apparent rise in mortality from. 316 Quenrc. decadence of, 122 INDEX Queensland, growth of population (Tubles), 141-145, 157 Quinquennial census, necessity for, 169-205. 2923 RABIES, cases admitted to Dogs’ Homes “Pables), 419-423, 437; effect of police regnlations in diminishing, 420-423, 437; geographical distribution of, in 1887-1889 (Table), 425; and Lydrophobia, 418; influence of weather on, 418, 423, 424; seasonal prevalence of (Tables), 118-424, 437; sex-proportions of animals attacked by, 430, 437; value of isolation as a means of checking, 436 Races, amalgamation of (Tables), 86-89, 92-95, 98-101 Radcliffe, Dr. J. Netten, on sanitary fundamentals, 18 Rainfsll in connection with disease (Table and Plate), 281, 282, 296, 822, 323, 326, 327, 334 Rates, birth- and death-, value of depends on shortness of census intervals, 203 ; meaning of, 4 Ravenstein, Mr. E. G., on the laws of migration, 37, 149, 156 Registrar-General, value of returns published by, 438 Registration, new Act, 12 Religious profession as a subject for census inquiry, 211, 212, 222, 224 Respiratory organs, diseases of. See Lung diseases Rheumatic fever as a cause of heart disease, 238, 239 ; in relation with arysipelas (Table and Plate), 317, 399-331, 442; in relation with puerperal fever (Table and Plate), 317, 329-331 Rheumatism, geographical distribution of (Table and Plate), 330, 331; of heart in relation with pleurisy. 875, 376 Ridley, Mr. Matthew, opposition to Census Bill of 1753, 194, 195 Risley, Mr. H. H., Census of Thibet in 18th century, 193; on unwise grouping of religions, &c., in Indian census, 219 Russia, essentially an Asiatic power, 140 ; increasing migration from, to United States (Tables), 106-109 ; migration from, 50; neglect of statisties in, 440 Russiun war, influence of, on English matriage-rate, 15 453 SALFORD, fluctuations of diarrhea mortality in, 293,294, 300, 301 Sanitary legislation, 19, 347, 409 Sanitary reform, and diarrhcea mortality, 800; effect of, on fever mortality, 235, 236 Jaskatchewan, See Canada, provi sional Districts Scandinavia, migration from (Tables), 49, 50, 106-109 jcandinavians in Canada (Table), 80 ; in United States (Table), 79-82 jcarlatina, and as~ociated diseases (Table, Plate, and Diagram), 312-334; decreased mortality from, 441 ; effect of isolation, 826 ; epidewics of, 16, 17,814 ; geographi--al distribution of, 355 ; Increasing proportion of cases dealt with in Metropolitan Asylum Hospitals, 110,412; possibly spread through medium of some lower animal, 327 ; and puerperal fever, essential {ifferences between, 329; recent dacrease of (Tables), 230, 236, 254-264, 267, 268, 326, 333 ; in relation to density of population (Table), 353-855; seasonal prevalence of (Tables and Plate), 400, 405, 411; yearly fluctuations of (Plate), 314 chools, additional, required before 1901, 28, 29 Scotland and Ireland, natives of, in different countries (Table), 148 Seeley, Professor, on settlement of America, 52 Segregation of races in United States, 66 Jewell, Mr,, statistics of rabies, 422 Jex, as a factor in death-rates, 245 ; influence on death-rate, 6 Jex-proportions of children born, 249 Sickness, as a subject for census inquiry, 213 $'mon, Sir John, on sanitary legislation, 19 imple continued fever, diagnos’s, probable errors in, 397, 398, 407, 408, 412; in London 1869-1884 (Table), 396, 897, 410, 411; mortality from, disputed, 397, 398, 407; seasonal prevalence of (Tables and Plate), 401, 402, 407, 412 3mall-pox, epidemics of, now rare, 16 ; recent low mortality from, 236, 287 Smith, Adam, 196 Snow, Dr. H., on recent increase of cancer, 240 Socialism, tendency to, in modern leeislation. 29, 80 South Austialia, growth of population (Table), 141-145 South Wales, growth of population, 28 Spain, migration from, 51, 136 : Spaniards in United States (Table), 78, 79 Statistical Society, recommendations of, for census, 197, 208-213, 222 Statistician, qualifications of, 2 Statistics, inexact data of, 440, 441; necessity of, 1; neglect of, in Cape Colony, 137; neglect of, in Russia, 440; tend to increased knowledge of important facts, 442; tendency of errors to balance each other, 4:41 Stewart, classification of coutinued fevers, 394 Stockport, high diarrhea mortality in, 292 Summer diarrhea. See Diarrhoea Sumner diseases, 312 Swansea, low diarrhea mortality in, 292, 298, 301 Sweden, See Scandinavia Switzerland, emigration from, 49 ~ Sydney, population of (Table), 158 Tanks MESENTERICA, in relation with phthisis (Plate), 374 Tabulation of census results, 214-223 Tasmania, aborigines extinct, 146; growth of population (Table), 141-145 Tate, Mr., improvements in arithmometer by, 5 Taxation, local, recent increase of, 30 Tembuland, estimated population of (Table), 138 Tennant, Major, and the rabies Order of 1889, 422 Thibet, census of, in 13th century, 193 Thomas, M.. arithmometer of, 5, 227, 311 Thornton, Mr., opposition to Census Bill of 1753, 194 Thresh, Dr. John C., estimated population of Bssex, 243 Toronto, population of, 158 l'own and country populations compared, 10; death-rates depressed by immigration from country, 35; population in new countries. 156-159 Towns, aggregation of population in (Table), 24, 25, 164, 246, 247, 439, 440; conditions of life in, 35; deathrates in, higher than in country, 245. 246 : degeneration of race in, To INDEX 216, 247,253 ; excess of marriages in, 10; excess of women in, 10; growth of, in Awmerica (Table), 159; growth of, in Australasia Tables), 157, 158; growth of, in Austria (Table), 166; growth of, in Belginm (Table), 165, 166; growth of, in Canada, 157-159; growth of, in England (Tables), 160-163 ; growth of, in Europe, 163-167; growth of, in- France Table), 163-165; growth of, in Germany (Table), 166; migration to, from country districts, 24, 25, 164 Trade, condition of, influence on marriage-rate, 33 Transkei, estimated population of (Table), 138 Transvaal, state of government in, 140 ; surrender of, 139, 140 Tubercle bacillus, discovery of, 234, 368 Tubercular meningitis in relation with phthisis (Plate), 374, 383 “yphus, difficulties of nomenclature, 16; effect of sanitary measures on, 235, 236; mortality from, in London, 1869-1884 (Table), 896, 397, 410, 411 ; seasonal prevalence of (Tables and Plate), 899-404, 406, 407, 412; term formerly used to describe all continued fevers, 394. 395 JLTRA-MIGRATION, 34, 43-51 Jnemployed, imperfect returns of, in census, 211 Tnited Kingdom, birthplaces of people (Table), 39 ; distribution of natives of (Table and Plate), 149; migration from (Tables and Plates), 44-47, 160; increase of population (Table), 46; inter-migration (Table), 41-45 ; natives of, in United States at successive censuses (lable), 101 Inited States, Aboriginal Indians in, 66-68; becoming less British in race, 107, 110-112; British in (Tables), 95-101; and Canada, mixture of races compared, 117, 118, 125; Canadians in (lable), 74-78, 101, 102; Chinese in, 68, 89; distribution of foreign-born (Tables), 102-106; Dutch in, 78; estimated natural increase in, 57-82; French in, 74.78; foreigners in, as foreign-born and native-born (Table), 73, 74; foreign-born in larze cities. 105: Germans in Tables), 82-89, 91; growth of population and wealth in, 54-62; unmigration from European countries (Tables), 106-110 ; increase of coloured population, 61; Irish in {Tables),R89-95; Mexicansin (Table), 78,79 ; migration to, 47, 48, 54-56, 60,62 ; migration from Germany, 60; from United Kingdom, 60; nationalities of immigrants (able), 62; natives of United Kingdom in, at successive censuses (Table), 101; neglect of vital statistics in, 57; Negroes in, 69-71, 112 ; parentage of people (Tables), 64, 65, 71-74, 77, 82, 85-88, 92-95, 97-100, 103 ; population by race and birthplace (Table), 62-64; Portuguese in, 79; probable future increase of population, 188; race constitution of people, 110, 111; rapid growth of cities in (Table), 159; reason for elaboration of census, 208; Scandinavians in (Tables), 79-82; segregation of races in, 66; Spaniards in (Table), 78, 79; statistics of parentage (Tables), 64, 65, T1-73,76, 77, 82, 84-88, 92-95. 99-101, 103-105 Urinary system, diseases of See Kidney diseases Traguay, census of (Table), 130, 131; future prospects of, 131, 132; migration to (Table), 131, 132; race-constitution of (Table), 130. 131 Mah, birthplaces of population (Table), 104; Danes in (Table), 104; English, Welsh, Scotch, and Irish in (Tables). 96-98. 104 Victoria, growth of population (Tables), 141-145; marriage-, birth-, and death-rates in (Table), 58; town population in, 158 Violence, changes in age distribution of mortality from, 240 7ital statistics, national system of, defended, 227, 228; neglect of, in United States. 57 INDEX 455 Wags as a subject for census inquiry, 21 Walker, General, on migration into towns, 164 Wars in past and present times compared, 30, 31 Weather, influence of, on disease, 266, 278, 279, 281-283, 289, 290, 295-297, 300, 312, 313, 322, 326, 327, 369, 370, 373, 376, 382, 383 Weir, Dr. A. McCook, on cansation of summer diarrhoea, 297 Western Australia, growth of population (Table), 142-145 Wheat, consumption «f, per head, 185, 186; countries hmporting or exporting (Table), 184, 185; price of, influence of, on marriage-rate, 15 Wheat supply, area required for cultivation, 186-188 Whitelegge, Dr, B.A, on weekly variations of mortality from croup, 321 Whooping-cough, diseases resulting from, 237; epidemics of, frequent, 16; fluctuations in mortality from, 231, 254-264, 267, 268; in relation with lung diseases (Plate), 371, 372; sex-mortality from (Tables), 2537, 254-264 Widows, excess of (Tables), 9, 250 Wigan, fluctuations of diarrhea mortality in (Diagram and Table), 274, 275, 290, 291, 299, 301-303 Winnipeg, rapid increase of, 157 Winter, diseases of, 312; method of measuring severity of, 376 Women, excess of (Tables), 8, 9, 249, 250, 253 ; longevity of, 8; married, excess of (Table), 9 YORK, increased diarrhcea mortality in, 292 Young, Sir Frederick, on state of government in Transvaal, 140 Ustrad-y-fodwg, rapid growth of (Table), 160, 162, 163 ZymoTIC diseases, decreased mortality from (Tables), 229, 230, 254-264 ; possible sources of infection, 327, 352. 353. 855 THE END. LONDON : PRINTED BY EDWARD STANFORD, 26 & 27 COCKSPUR STREET, CHARING CROSS, 8.W. —— mA WATNKRA I Ay