BY H. S. HALCRO WAKDLAW. 827 



thoroughly mixing the milk before taking a sample, if the sample 

 is required to give information about the total amount of milk 

 yielded by the cow at the particular milking under examination. 



The quantity of milk obtained from each cow was deter- 

 mined by weighmg with a spring balance. These weights are 

 correct to about 01 kg. 



The examination of the milk was commenced on the afternoon 

 on which it was collected. The milk was kept in an ice-chest, at a 

 temperature of about 5"C., during the following night. When- 

 ever possible, all the operations, which it was necessary to carry 

 out upon t?he fresh milk, were completed before the following 

 afternoon, i.e., before the milk was 24 hours old. During this period 

 the milk was kept for about 14 hours in the ice-chest, so that it was 

 exposed for less than 10 hours to ordinary temperatures. The 

 temperature of the milk never rose above 20''C., and Chapman 

 (1908) has shown that, even at temperatures as high as over 

 30°C., souring does not commence until about 1 1 hours have 

 elapsed since milking. Whenever it became impossible, for any 

 reason, to complete the examination of the milk in the unaltered 

 state within the specified time, 5 drops of formalin were added to 

 each 100 c.c. as a preservative. The addition of this small 

 amount does not measurably alter tlie properties of the milk under 

 investigation (Reiss and Sommerfeld, 1909). 



D e n si t y. — The density of the samples of milk was deter- 

 mined by weighing in a pycnometer of 25 cc. capacity, provided 

 with a thermometer. To save time, no attempt was made to till 

 the pycnometer at a constant temperature. The weights \vhen 

 filled at different temperatures were determined, and a gra23h was 

 drawn, from which the corresponding weight at 25X\ could 

 be read off. Observations were made at temperatures ranging 

 from 10'' to 30°C. Above 17''C., the points plotted lay upon, 

 or very close to, a smooth curve, the maximum deviation 

 being less than 2 mg. Below 17°C., the points obtained were 

 rather scattered. This behaviour is due to the fact, which was 

 noted as early as 1841 by Quevenne, that, at a temperature in the 



