746 TRANSACTIONS OF SECTION M. 
In connection with the variation and sampling of oat straw and using data 
from a hundred individual straw analyses the probable error was very great, and 
varied according to whether it was calculated on the percentage of nitrogen, the 
total weight of nitrogen, or the dry matter of individual straws, respectively. 
Except in the case of the total weight of nitrogen the frequency curves were 
abnormal. Similar variations were found in the probable error and frequency 
curves calculated on the different constituents of the mangel. 
4. Feeding Values, Practical and Scientific. By Davin Witson, D.Sc. 
5. The Investigation of Feeding for Milk: Its Problems and its Method. 
By Cuarues Dovucuas. 
The practical problem is the economical and profitable production of milk. 
This, however, cannot be fully investigated in such a way as to yield a result 
generally applicable, since the conditions are very various. Cost of production 
varies with the cost of feeding-stuffs, labour, and rent, and with the milking 
capacity of cows. The value of the product alters with the price of milk at 
different times and places, and with the value of the manurial residues. These 
factors must be combined in any attempt to apply practically any knowledge of 
the milk-producing effect of system of feeding. 
The scientific problem of the actual influence of foods on milk-yield is 
separate from and subordinate to this practical issue. 
The essential method of investigation is the actual recording of milk-yields. 
Purely chemical calculations cannot be regarded as final, but are valuable 
chiefly as yielding hypotheses. The ‘scientific ration’ is no more than an hypo- 
thesis; and it is one which leaves out of account essential factors, physiological 
and psychological, which may be summarised in the phrase the individuality of 
the crganism. This is illustrated in the contrasting records of cows and in 
recent experiments on the influence of flavour, &c. 
The hypotheses must be tested by actual recorded results. The necessity for 
this is illustrated by results recently obtained in relation to the influence of 
ventilation and temperature. 
Very large numbers of experiments and observations must be recorded in order 
to obtain reliable results. 
The extent of the necessary survey must be kept in view. It is of little 
value to ascertain the temporary effect of a ration on milk-yield. The subject of 
investigation must be the whole lactation, including in that term the capacity of 
the cow to yield milk in the succeeding lactation. 
In view of the extended character of this problem, sufficient investigation must 
be made to eliminate the influence of certain disturbing factors, such as the 
variations of different cows, and of cows in different lactations, and the influence 
of accident and season. 
It is necessary, in fact, to break up the problem and to investigate separately 
the influence of different rations on the milk-yield of cows (a) of various 
capacities and (6) under various conditions. 
6. Note on the Use of Cotton-seed Oil as a Substitute for Butter-fat 
in Calf Feeding. By James Henpricx, B.Sc., F.I.C. 
In most experiments on feeding calves with separated or skim milk and an 
oil in substitution for the butter-fat, cod-liver oil has been the oil used. Con- 
sequently the opinion is prevalent that this is the only oil which can properly 
be used. The author recently carried out some experiments which were intended 
mainly as practical demonstrations on the economy of using separated milk and 
oil in substitution for whole milk in the feeding of ordinary commercial calves. 
In these cotton-seed oil was used as well as cod-liver oil. Cotton-seed oil was 
chosen as a comparatively cheap and easily obtained vegetable oil which is exten- 
sively used in human food, and which is known to be wholesome. Another 
