2i6 FOOD REQUIRED [chap. 



increase, and they are based upon experiments with 

 ruminant animals — sheep and oxen. They do not hold 

 exactly for fattening pigs, nor do they refer to the 

 production of energy in working horses. However, the 

 energy which becomes available for work purposes is 

 derived from the digestible constituents after deduction 

 of the work spent in digestion in just the same way as 

 the materials available for increase, so that no consider- 

 able error is introduced if we take the starch equivalent 

 of a food as representing the energy that would be 

 available for work purposes as well as the surplus that 

 will be available for making increase of live weight. 



One other point, however, we must bear in mind. The 

 starch equivalent takes into account only the fat-making 

 power of the food, and pays no attention to the nitrogen 

 it contains nor the requirements of the animal for 

 nitrogen. Suppose we have arranged a certain ration 

 that will supply the requirements of the animal as 

 regards energy or fat production, basing the ration upon 

 the starch equivalent of the foods ; we have then also to 

 make a second calculation of the albuminoid ratio of the 

 ration in order to make sure that the animal is getting 

 the proper amount of nitrogen. But with all the know- 

 ledge that hcis been derived from the experiments upon 

 feeding animals, knowledge which is summed up in the 

 starch equivalents, it would be extremely unwise to 

 begin to construct rations for farm animals on a priori 

 principles, considering that they require a certain number 

 of starch equivalents per diem and a particular 

 albuminoid ratio, and assuming that it is a matter of 

 indifference how these units are arrived at The better 

 plan is to take as our starting-point certain well- 

 recognised rations which have been justified in practice 

 for the particular purpose in view, and see how they can 

 be modified to secure equal eiHciency but greater 



