28 REPORT OF THE AGRICULTURAL RESEARCH 



error less than a certain magnitude that we desire to 

 estimate. 



11. Sugar Beet Roots. — Some samples of sugar beets 

 which had been grown near Peshawar by the Superintendent 

 of Farms were examined and the result is of interest because 

 one lot included beets of a high grade. The percentage of 

 sucrose varied in this lot between 17 per cent, and 20 per 

 cent, of sucrose in the juice, demonstrating that very good 

 sugar beet could be grown in Northern India if desired. 



12. Milk. — In co-operation with Mr. A. C. Dobbs a series 

 of tests were made during the year on the quantity and 

 composition of the milk of the Montgomery herd at Pusa 

 which have led to some interesting results- The work will 

 not be completed until next year, but the following will 

 indicate its nature. 



When testing cows' milk in India two sources of error 

 must be guarded against : {i) if the periods between milking 

 are not equal the composition of the milk will be influenced, 

 and {ii) the calf takes an indefinite quantity of the milk 

 unless this is specially controlled. In order to eliminate 

 these sources of error the cows were milked at 12 hour 

 intervals, and the calf allowed to take the whole of the 

 milk from only one side of the udder, the other half being 

 milked by hand. The cows were milked by hand for 24 

 hours on (say) the right hand side of the udder whilst the 

 calf took the milk from the left hand side, and then during 

 the succeeding 24 hours the left side would be hand-milked, 

 whilst the calf took milk from the right side. Moreover 

 the diurnal change from the right hand side to the left 

 hand side was made in the morning for one month, and in 

 the evening during the next month. During one period of 

 two months the milk of the fore and hind quarters of the 

 udder of three cows was separately examined. 



By these means great regularity was obtained in the 

 composition of the milk and several characteristic differ- 

 ences were observed. The most prominent was the differ- 

 ence between the percentage of fat in the morning and even- 

 ing milk; the former always contained from -5 to 15 per 



