420 SCIENCE PROGRESS 



415. and 525., are deducted, the net food costs are 144s. and 83s. 

 respectively. But according to the results of the experiments, 

 the soya bean cake is worth 25s. more and therefore its relative 

 food value becomes 1085. when the food value of linseed cake is 

 144s. Thus the food value of soya bean cake was three-fourths 

 that of the linseed cake and its purchase value, taking linseed 

 cake at £9 55., would be f (185s. — 415.) + 525. = £8 per ton. 



If the results of the coconut cake, bran and linseed cake 

 experiments already described are dealt with in the same way, 

 it will be found that the consuming value of both coconut cake 

 and bran is 62*6 per cent, that of linseed cake. 



These experiments, so far as they go, indicate that compo- 

 sition and energy value are not the only things to be taken into 

 account in feeding. It appears that certain foods either have a 

 peculiar feeding value apart from that indicated by their compo- 

 sition or that certain substances combine to make a good ration 

 and other substances do not. 



THE DISCREPANCY BETWEEN THE RESULTS 



ACTUALLY OBTAINED AND THOSE EXPECTED 



FROM CHEMICAL ANALYSIS 



(Dr. F. Gowland Hopkins) 



It seems characteristic of the present moment in science that 

 fundamental conceptions, which we had looked upon as estab- 

 lished, concerning which our teaching had become dogmatic, 

 should prove to need revision. 



The science of animal nutrition, though no one has pre- 

 tended that, in any of its departments, the data are exact, has 

 certainly developed its own quota of dogma. 



We have long taught, for instance, that satisfactory criteria 

 of the efficiency of a dietary (assuming the presence of the 

 necessary inorganic constituents) are furnished by its content of 

 protein and energy considered solely from the quantitative 

 standpoint. A dietary, to be efficient for this or that animal, we 

 have taught, must contain a certain, rather vaguely known, 

 minimum of protein and a more exactly determined minimum of 

 total energy. We have commonly been content to evaluate the 

 protein by multiplying estimated nitrogen values by a numerical 

 factor; the energy from calculations based upon calorimetric 



