Sept. IS. I9I9 Variation in Milk of Ayrshire Cows 313 



inherited genes for fecundity which determine the frequency and regu- 

 larity of ovulation. Corresponding to the physiological parallelism 

 in egg size and milk production is found corresponding asymmetry of 

 the variation curves, as well as a closer relationship between other of 

 the variation constants in the two cases than is found when milk produc- 

 tion is compared with egg yield. 



7. Considering the types of the curves, we find that 7 out of 10 curves 

 for mean weekly yield give upon analysis unlimited range curves — 

 in 6 cases the skew Type IV and in one case the symmetrical normal 

 curve. Something approaching the reverse condition is found with 

 respect to variation in fat percentage. Five out of the 1 2 distributions 

 for this character lead upon analysis to curves with the range limited 

 at both ends (Type I) and one to a cur\^e of Type III, which is limited 

 at the lower range end. The remainder of the curves are of Type IV, 

 but near the border line of passage over to the limited range types. It 

 would then appear that the physiological fact that variation in percentage 

 of fat content will necessarily tend to be confined within relatively nar- 

 rower limits than variation in total flow of milk is reflected in the dis- 

 tribution of the several curves in respect to type. 



8. The estimation of the range ends in the case of the limited range 

 curves is on the whole fairly good, leading in no case to absolutely impos- 

 sible values regarding the probable errors involved. The determina- 

 tion of the range ends in the Type I curv-^e is subject to rather considerable 

 probable errors. The most extreme range end estimation in mean 

 weekly yield is that given by the curves for 9-year-old cows. This 

 gives for the upper range end 67.2 gallons a week. This of course would 

 be an extraordinarily high average weekly yield, yet it probably can not 

 be regarded as physiologically impossible. It certainly would not be 

 for a single week. Indeed such a record is rather frequently exceeded 

 by Holstein-Friesian cows which on official tests may occasionally go 

 to a production over 100 gallons per week. In fat percentage the most 

 extreme range estimation is for the 1909 curve for 4- year-old cows, 

 which gives for the upper end of the range 16.6 per cent fat. Again 

 this figure, while of course extraordinary for an average test, probably 

 indicates no physiological impossibility for brief periods of time. That 

 such is the fact is indicated by some of the short period tests of Jersey 

 cows. 



9. In all of the curves for mean weekly yield the kurtosis is positive. 

 In other words, these curves show a tendency of greater or less degree 

 toward the leptokurtic condition. They are more peaked than would 

 be noi-mal curves of corresponding standard deviations. The value of 

 the kurtosis is probably significant in all the mean weekly yield curves, 

 with the exception of those for 3-year-old, 9-year-old, 10- year-old, and 

 12-year-old animals. In curves of variation in fat percentage there is 

 no such unifonn tendency in regard to the value of the kurtosis. The 



