x TRANSACTIONS OF SECTION ¥ 731 
(a) Selection of the Figure. 
(i) When normal lactations only were considered, all the above figures showed 
coefficients of variability (ae almost identical with that 
of lactation totals. 
(ii) When all cows were included, however, the variability of totals was 
increased 25 per cent., while that of the other three figures remained almost un- 
changed. : 
(iii) In order to show that these figures really were closely connected with the 
inherent capabilities of a cow, and so with the yield the owner might expect 
to get under ordinary circumstances, correlation coefficients were determined 
between them and lactation totals, both for normal lactations and for various 
groups of similar abnormal lactations. These were found to vary between :— 
+ 800 + ‘007 
+ 904 + °005 
(iv) All were sufficiently high to justify the use of any of the figures, but a 
comparison of the coefficients gave an almost constant indication of their respective 
merits. 
(b) Lffect of Lauterior Influences. 
By the choice of any of above figures :— 
(i) Two important causes of variability have been eliminated, namely :— 
Number of weeks in milk. 
Interval between calving and subsequent service. 
(ii) The influence of food, weather, and general treatment has been minimised. 
(iii) But :— 
Age of cow, 
Length of rest before calving, 
Season of year of calving, 
must be statistically dealt with. 
If a standard figure could be agreed upon for the expression of a cow’s milking 
capabilities, it seems that the problems of the inheritance of milk yield, and of 
its possible correlation with other characteristics, would be materially advanced. 
They could, at any rate, be dealt with on the same basis, and so with strictly 
comparable results, in all parts of the country. 
3. On the Effect of Heavy Root-Feeding on the Yield and Composition 
of Milk. By Avexanper Lauper, D.Sc., and T. W. Paaan, M.A. 
The object of the experiments was to test the effect of a ration consisting 
largely of roots on the yield and composition of milk, and incidentally to deter- 
mine how far turnips could profitably replace the more concentrated and expen- 
sive foods commonly employed in feeding dairy cattle. 
Three separate experiments have been made—the first in 1909, when twenty 
cows were under experiment; the second in 1910-11, with twenty-two cows; and 
the third in 1911-12, with eighteen cows. For each experiment two lots of cows, 
as nearly equal as possible in respect of age, period of lactation, and yield and 
quality of milk, were selected from the herd. Before the experiment began there 
was a preliminary period of trial of about three weeks in length, in order to 
complete the adjustment of the two lots. During this preliminary period all the 
foods to be experimented with were fed. The cows were then gradually put on 
to the experimental rations. In experiment III., after the experimental rations 
had been given for eight to ten weeks respectively, the rations were crossed. The 
change was effected gradually, so as to introduce as little disturbance as possible, 
and then the experiment was continued for another four to six weeks. 
