SEASONAL FLUCTUATIONS 7 
It is evident that the peak in building employment does not 
coincide with the peak of contracts awarded or of permits granted, 
but occurs some three or four months later. Consequently, though 
they have a wider geographical scope and extend over a longer 
period of years than the available direct statistics of employment 
or unemployment, it does not appear desirable to utilize the con- 
tract or permit figures as indices of seasonal variation in employ- 
ment. 
In Fig. B, the curve for New York building and street labor is 
repeated, and curves are added to indicate the seasonal variation 
in Wisconsin building construction during the three years from 
1922 to 1924, inclusive, and also employment as reported by con- 
tractors in twenty-six cities of the United States in 1922. (See 
footnotes to Table 57 for sources.) To obtain a composite estimate 
of seasonal fluctuations in building employment, weights of SIX, 
three, and one were assigned to the New York, Wisconsin, and 1922 
series, respectively, and the result is plotted as Curve “h” in Fig. C 
of Chart 53. 
Our index of seasonal variation in employment on highway cons- 
truction (Curve “g”, Fig. C, Chart 53) is also admittedly only an 
estimate based upon fragmentary data. The statistics used cover 
the years 1922, 1923, and 1924, and include (1) the number of 
common laborers, by months, on highway projects receiving Federal 
aid in the fourteen states included in the New England, Middle 
Atlantic, and East North Central sections, and (2) employment on 
highways in Wisconsin. In computing the index of highway cons- 
truction these two series were weighted by the relative population 
of the states represented. 
Finally, a composite index of employment in construction (Curve 
“1” in Fig. C, Chart 53) was computed by combining the building 
and highway indices, assigning a weight of six to the building and 
one to the highway series. Obviously, this index of construction 
employment must be taken as a rough approximation. It indicates 
small activity in the first quarter of the year ; increasing employ- 
ment in the second quarter; maximum activity in the third quarter; 
and then a decline in October, November, and December. 
In some of the subsequent comparisons, the index for railway 
maintenance and the index for construction have been weighted 
by the estimated numbers employed in these occupations in 1909 
and combined into an index of “Selected outdoor industries” (Curve 
“fof Fig. C, Chart 54). 
2C