Frost found in his survey of 130,033 individuals that the percentage of the population attacked varied from 15 per cent. in Louisville, Ky., to 53.3 per cent. in San Antonio, Texas, the aggregate for the whole group being about 28 per cent. He remarks that this agrees with scattered observations in the first phase of the 1889–1890 epidemic, when the attack rate seems to have varied within these limits. In five of the localities studied, geographically widely separated, the incidence rate varied only within a narrow limit, from 200 to 250 per thousand. Variations in attack rate showed no apparent consistent relation to geographic location or size of community, or to the rapidity of development of the epidemic. In a house-to-house survey of 10,000 individuals in Boston the author found that in the winter of 1918–19, 19.71 per cent., or one-fifth of the entire population had developed the disease. It should be pointed out that while the standards used in this survey are entirely comparable to those used by Frost, the author has, contrary to Frost’s method, not included in his group of positive cases those classified as “doubtful.” This would raise the total incidence to a certain extent, but we feel convinced that by omitting the doubtful cases we have approached nearer to a correct picture of the epidemic as it actually occurred. As will be seen from Chart XVI there was no great variation in the different districts studied, with the exception of Districts IV and V. Districts I, II and III were in the tenement section of the city, while District VI was in one of the finest residential parts of Brookline. Districts IV and V were midway between these two extremes as regards economic and sanitary status, as well as extent of crowding. The lowest incidence was in the Irish tenement district. The highest in a middle class Jewish population. CHART XV. [Illustration: 1 and 2 Influenza incidence in 10,000 individuals [both epidemics]] CHART XVI. [Illustration: 1918 Incidence of Influenza by sex in Districts surveyed 1920 Incidence of Influenza by sex in Districts surveyed Individuals having Influenza in both Epidemics by sex] In the 1920 recurrence we found that 9.55 per cent., or one-tenth of the entire population, suffered from the disease, and the arrangement of districts in order of incidence was very little changed. The Irish community suffered least; the two middle class communities most. The well-to-do district in Brookline had the next lowest incidence in 1920. That the high recorded incidence in middle class districts was not due to more accurate or more thorough work on the part of the inspectors is indicated by the fact that a great part of the work on Districts IV and V was done by the same individuals who inspected Districts II and III. One-fifth of the population studied developed the influenza in 1918–19, and one-tenth of the same population suffered in 1920.