NORMS AND TRENDS IN EARNING ASSETS 29 districts, in the matter of the proportion of their earning assets represented by investments, is illustrated. A graphic summary of the ratios of investments to earning as- sets showing (1) the rates of change from year to year, (2) the percentage deviations of the yearly averages from the seven-year average for each district, and (3) the percentage differences of the yearly district averages from the yearly averages of the twelve districts combined is given in Chart 4. This chart is con- structed in the same way, and should be interpreted in the same manner,® as Chart 2. The direction of change in ratios of investments to earning assets for the twelve districts combined was generally downward from 1919 to 1921, and generally upward from 1921 to 1923, al- though there was a decline between 1923 and 1924. The greatest percentage fall was between 1919 and 1920, and the greatest per- centage rise between 1921 and 1922. The level in 1925 was al- most exactly the same as in 1919. The districts with rates of change closely corresponding to the average for the twelve districts combined are as follows: Boston, New York, Philadelphia, Cleveland, Chicago, St. Louis, Kansas City, and San Francisco; those with markedly different rates are Richmond and Atlanta. Dallas and Minneapolis follow the general direction of the country as a whole, but differ noticeably in certain years. All of the twelve districts had a downward trend between 1919 and 1921, and all of them (except Richmond) an upward trend between 1921 and 1925. The district having the smallest year-to-year percentage changes and the most uniform ratios over the whole period is Philadelphia; the one having the most violent year-to-year changes and the most varied history is Minneapolis. Between 1919 and 1925, the direction was upward for Boston, New York, Minneapolis, Kansas City; for the others it was down- ward, the fall being relatively large in Richmond, Atlanta, and Dallas—southern agricultural districts. Districts with ratios closely corresponding in size to those for the country as a whole are Boston, New York, Chicago, St. Louis, and San Francisco; those with ratios markedly different for the seven years from those for the country are Philadelphia, Cleve- land, Kansas City, and Dallas. In Richmond and Atlanta the ratios in 1919 and 1920 closely agree with those for the coun- % See page 21 for discussion of both points.