26 REPORT OF THE AGRICULTURAL RESEARCH 



Pusa was made in collaboration with my colleague 

 Mr. A. C. Dobbs, then Imperial Agriculturist. The pre- 

 cautions which we employed in order to avoid experimen- 

 tal error as far as possible are dealt with in a succeeding 

 paragraph. 



Although the analytical part of the work was naturally 

 completed day by day at the time, it was necessary, in order 

 that the data obtained might have their full value, to 

 examine them carefully and ascertain the probable errors 

 mathematically. This portion of the work was done during 

 the hill recess 1913, and explains why the subject was only 

 briefly referred to in my last Annual Report. 



Altogether three series of tests were made; with three 

 cows in December and January ; with twelve cows in April 

 and May; and with ten cows in September to November. 

 Most of these cows had calved recently and were in " full 

 milk " or " fairly full milk." 



Regarding the yield of milk, one half was obtained by 

 hand milking (see succeeding paragraph) and from these 

 data the total production of each cow was ascertained. This 

 varied considerably, being as small as 8 lbs. in one case and 

 as high as 18 lbs. and 21 lbs. respectively in two others. 

 The majority produced from 10 to 14 lbs. per 24 hours. 

 The morning and evening yields were found to be approx- 

 imately equal in all cases; the periods between milking 

 were 12 hours each. In respect of fat the percentage 

 varied generally from 4 to 6, but was as low as 3-5 in one 

 or two instances. The percentage of solids-not-fat varied 

 from 8-2 to 9-0. 



The morning and evening milk was not equally rich, 

 the morning milk being systematically richer by about 1-0 

 per cent. ; in some cases the difference was as small as -3 per 

 cent., in others as high as 15 per cent. ; the mean difference 

 for the 25 cows was -91 per cent. 



Tests of the yield and composition from different parts 

 of the udder showed that : — 



(i) Usually the yield is approximately equal from the 

 right and left side respectively, though there 



