Pearson''s criterion of the probability that 2 independent distributions of frequency are really samples of the same population was employed in the comparison of the egg production of Rhode Island Red, White Wyandotte and White Leghorn birds. This method has the advantage, as compared with those employed earlier, of basing the test of differentiation on the frequency distributions of the 2 classes as a whole instead of on the means, standard deviations and coefficients of variation of these distributions. The results, taken in connection with earlier investigations of the same problem, demonstrate that in general the frequency distributions of monthly egg production are significantly different in the 2 classes of birds. The distribution of the values of x" (which measures the deviation of the frequency distribution of monthly egg productions of birds which survive from that of those which die) throughout the year roughly parallels the mean monthly egg production and the monthly percentage death rates. Thus, as measured by this criterion, the differentiation of the birds of the 2 classes with respect to egg producing capacity is, generally speaking, greatest during the months of the year in which the average production and the % death rates are highest. The biological significance of this parallelism is discussed. To what extent the measured differences between the egg production rec-ords of the 2 classes of birds are due to the influence of pathological factors on egg production, and to what extent they represent innate constitutional differences of a kind which may be significant in their relation to the problem of natural selection, must be determined by future research.