26 BULLETIN 1372, IT. S. DEPARTMENT OF AGRICULTURE 
better daughters on the whole are from the lower-producing dams 
and his lower-producing daughters from the higher-producing dams. 
In the list of a small number of daughters of a sire, the production 
of the daughters may not follow very closely that of the dams, owing 
perhaps to the sire's being more homozygous for the factors that 
will govern high-producing ability than the dams with which he is 
mated. But when a large number of daughters and dams are con- 
sidered, the higher-producing daughters will as a rule be found to 
have good dams. This is to be expected, for the high-producing 
dam is certain to have at least a part, if not all, of her germinal 
factors governing production, those that will determine high pro- 
duction, and she will therefore transmit high production to a part 
or all of her offspring. 
The evidence seems to point to both parents' contributing equally 
to the inheritance governing the milk and butterfat producing 
capacity of their daughters. But if one parent is homozygous or 
pure for the hereditary factors determining high production and the 
other parent is heterozygous in its inheritance, then the homozygous 
parent will have the greater influence on the producing capacity of 
the daughter; yet this daughter will transmit to a part of her progeny 
the inheritance for low production that she may receive from her 
heterozygous parent. From two heterozygous parents, it is to be 
expected that the daughters will show a great range in producing 
capacity, from very poor to very good. 
THE PERCENTAGE OF BUTTERFAT 
Roberts (3) found a significant negative correlation between the 
percentage of butterfat and the yield of milk for Jerseys, Guernseys, 
and Holsteins, but did not find so significant a correlation for Ayr- 
shires. That is, as the yield of milk increased, the percentage of 
butterfat in the milk decreased. Wilson (4) studying Ayrshire 
records, concluded that the yield of milk and the percentage of 
butterfat were independent of each other. Pearson (5) also found 
a small but significant negative correlation between percentage of 
butterfat and yield of milk. 
The material for the study by Roberts was made up largely of 
advanced-register and register-of-merit records of the various breeds. 
The animals in each breed were classed according to age, and the 
correlation was between the milk yield and the percentage of butter- 
fat for the group of animals in each class. The negative correlation 
was significant for the Jerseys and Guernseys; but the coefficient 
for the Holsteins, when judged by the probable error, was not sig- 
nificant in any of the classes. 
Evidence on the question whether a bull which has the ability 
to increase the milk yield of his daughters can also increase the per- 
centage of butterfat in the milk, or whether if he increases the milk 
yield he will decrease the percentage of butterfat (as is indicated in 
the results obtained by Roberts with Jersey and Guernsey records), 
is offered in Tables 14 and 15. These tables show that most of the 
sires making the greatest increase in milk yield also increased the 
average percentage of fat very materially, whereas a few sires that 
increased the average milk yield of their daughters decreased the 
percentage of fat. Several sires whose daughters showed an average 
decrease in milk yield also had a decrease in percentage of fat. In 
those cases where the daughters showed the greatest decline in milk 
