MAINTENANCE AND WORKING REQUIREMENTS 51 



of course, be added to that required for maintenance 

 alone. The method of calculating the work-value of a 

 food by means of its starch equivalent is justified by the 

 results of experiment. Kellner made eighteen tests with 

 horses attached to a braked capstan, and taking 100 to 

 represent the observed amount of work done, he found 

 that the potential work (calculated from the starch 

 equivalents) of the food supplied varied in the various 

 experiments between 96*5 and 103*1, the average being 

 99 '8, a sufficiently near approximation for all practical 

 purposes. 



But now the question arises as to the constitution of 

 the working diet. It might at first sight appear as 

 though the extra food needed for work production could 

 all be supplied in the form of carbo-hydrate, or a mixture 

 of carbo-hydrate and fat. This, however, is not the case, 

 and for two reasons. Firstly, when an animal is put to 

 work after a long period of rest, the muscles are soft and 

 out of condition. With moderate work they become 

 gradually firmer and larger, and just as the athlete's 

 muscles develop as he gets fit, so do the muscles of a 

 horse, whether a thoroughbred in training, a hunter, or 

 a draught-horse. There is thus an actual growth of 

 muscle, which can only be supplied by protein in the 

 food, and, further, these largely developed muscles, kept 

 in use and at work every day, will require a greater 

 supply of protein to make good wear-and-tear than the 

 muscles of a resting horse. In the second place, we 

 have already seen that to get the fullest digestion of the 

 nutrients supplied in the food there must be a certain 

 proportion of protein, and this proportion (the nitro- 

 genous ratio) must not be less than about 1 : 10. Thus 

 the whole ration, made up of the maintenance and the 



