THE FEEDING OF MILCH CATTLE 337 



one unfavourable, and the other two had no effect 

 one way or the other. With cotton-seed meal one 

 trial was favourable, another unfavourable, and 

 the third showed no effect upon the milk yield or 

 the percentage of fat. These contradictions can 

 hardly be explained in any other way but by the 

 extraordinarily different behaviour of the animals 

 towards the various foods (p. 310). Without ex- 

 ception, in these investigations, too small a number 

 of cows were used, and so the influence of individu- 

 ality was not equalised. In contrast to the above- 

 mentioned observations, there have been other ex- 

 periments in which not less than 200 cows have 

 been used in each case, and still no specific action 

 of the food-stuffs that have so far been tested has 

 been detected. Here a mixture, half oats and half 

 barley, both coarsely ground, was fed, and the effect 

 of this was compared with (1) ground maize, (2) a 

 mixture of oil cakes (J rape, J palm-nut, J- sun- 

 flower-seed), (3) wheat bran, (4) crushed wheat, 

 (5) a mixture of wheat bran and palm-nut cake with 

 molasses, (6) mangels. In none of these extensive 

 series of experiments was the replacement of the 

 ground cereals by the other foods attended by a 

 change in the percentage amount of fat in the milk. 

 The yield of milk certainly did rise when the mix- 

 ture of oil cakes took the place of an equal weight 

 of the ground cereals, but this must be ascribed to 

 the greater quantity of protein in the former (p. 325). 



