J. Alan Murray 159 



Tlie figiu'os from wliicli the graph was constrnctcd were taken 

 from Kellner's Erniihrung. They are not the results of direct determina- 

 tions of food requirements, but, if the author is not mistaken, are 

 derived from measurements of the relation between the body surface 

 and live weight of certain dogs ; and they involve the assumption that 

 the requirements for maintenance vary directly as the extent of body 

 surface. It is well that this should be recognised, but there is good 

 reason to anticipate that when it is put to the test of experiment the 

 assumption will prove to be well founded. Over a year ago the author 

 published a formula "^ for maintenance rations of oxen in terms of "total 

 digestible nutrients with an albuminoid ratio of 10 — 1." Assuming 

 that 9-3 lb. of such nutrients is the maintenairce ration for oxen of 

 1000 1b. live weight, this formula may be expressed as follows: 



log iV = I log M - 1-03152. 



This view is confirmed by independent data of a purely mathe- 

 matical kind. It also derives a certain amount of support from the 

 fact that the requirements of animals of 40 to 170 lb. live weight, 

 calculated by this formula, are consistent with experimental data 

 relating to the maintenance rations of pigs of that size-. The argument 

 cannot be applied to the formula in terms of starch equivalent because 

 the foods that are suitable for maintenance of pigs are of nearly "full 

 value," i.e. they contain no thermic energy over and above the dynamic 

 portion represented by the starch equivalent. It might be supposed 

 that it could be applied in the ease of other ruminants such as sheep, 

 which can subsist on much the same kind of diet as oxen. Such, however, 

 is not the case, because, owing to their thick coats of wool, these animals 

 do not lose heat by radiation so rapidly as pigs and oxen in proportion 

 to their body surface. 



Wood and Yule have assessed the maintenance ration for sheep 

 of 100 lb. live weight at about 6-| lb. of starch equivalent per week; 

 but, on the ground that this shows a higher percentage utilisation of 

 the food for fattening, they suspect that the allowance is too small. 

 If the starch equivalent referred to was calculated by the formula 

 given by them, i.e. if it includes the whole of the available energy 

 of the food, there can be little doubt that it is too small. On the other 

 hand, if it means Kellner's starch equivalent, it is probably too large. 



' The Chemistry of Catlle Feeding and Dairying (Longmans), pp. 128 and 142. 

 - Sanborn's Expers. Bui. 28, Mo. Agr. Col. C'f. Chemistry of Catlle Feeding and 

 Dairying, p. 206. 



