70 UNIVERSITY OF WISCONSIN STUDIES 
of the experiments show a decline in the saponification value 
and Reichert-Meissl number of the butter fat, and an increase 
in the iodin-absorption value. The melting point, however, 
in some cases was increased, in some cases was not affected, 
and in other eases declined. Another striking feature of 
the results is the extreme yariations in the chemical fat con- 
stants obtained in a great many of the experiments. In 
fifteen of the twenty-three experiments given in the tables 
the maximum iodin-absorption value rose to a figure of 40.0 
or above and in twenty-one of the twenty-three experiments 
to a figure of 36.0 or above. The extent of these variations is 
emphasized when they are compared with the average normal 
iodin value of 32.7 for all the experiments. Similarly, 
Reichert-Meissl numbers below 24.0, the legal standard for 
butter in the United States, were obtained in ten of the twen- 
ty-three experiments. Reichert-Meissl numbers below 28.0, the 
Belgian legal standard for butter, were obtained in twenty 
of the twenty-three experiments. Again, the extent of these 
variations is emphasized by comparison with the average 
normal value of 29.1 for the twenty-three experiments. 
Application of the results of underfeeding studies. The 
riarked variations in the chemical and physical constants 
of the milk fat which are found to accompany a subnormal 
plane of nutrition, even in cases of mild underfeeding, or 
when the plane of nutrition is merely reduced from super- 
normal to normal, at once indicate the importance of taking 
into account this important factor in both human nutrition 
and in experimental work involving the influence of specific 
factors on the composition of milk fat. 
It is certain that at times milk comes to the market from 
herds in such a condition of underfeeding that the fat will 
have the somewhat abnormal composition noted in the experi- 
mental work reported. It is of interest in this connection 
to raise the question as to the possible relation between the 
character of the fat and human nutrition, especially of infants. 
Unfortunately up to the present no information is at hand 
concerning this possible relation althongh it seems entirely 
possible it may be a matter of some importance. The im- 
