296 THE ROTHAMSTED EXPERIMENTS. 



must have had some other source than the fatty matter and 

 the proteids of the food. 

 Views as to The reconsideration of the results with pigs fully confirmed 

 ^confirmed ^ ne y i ew that, in many cases, much more fat had been pro- 

 duced than could possibly have been derived from trans- 

 formed albumin of the food. We concluded, therefore, that 

 we were not called upon to institute new experiments ; and 

 decided instead, again to direct attention to the results which 

 had already been published. 

 Paper read Accordingly, we gave a paper on the subject in the Section 

 burgle' f° r Agriculture and Agricultural Chemistry, at the meeting of 

 1876. the Naturforscher Versammlung, held at Hamburg, in 1876, 



at which there were present a number of the chief agricultural 

 chemists of Germany. The results given in Tables 69 and 

 70 were discussed, and it was pointed out that, even according 

 to the mode of calculation adopted, which would imply about 

 62 parts of fat to be producible from 100 parts of nitrogenous 

 substance, the experiments 4 and 5, in which the proportion 

 of the non-nitrogenous to the nitrogenous constituents in the 

 food was the most appropriate for fattening, showed that 

 about 40 per cent of the produced fat could not have had its 

 source in the nitrogenous substance consumed ; whilst if, 

 according to Henneberg and Voit, it were assumed that 100 

 parts of albumin can at most yield 51.4 of fat, the results 

 would be much more striking still. They would, of course, be 

 still more so if, as has more recently been estimated, only 42 

 instead of 51.4 parts of fat can be derived from 100 of albumin. 

 It was next considered what amount of error in the esti- 

 mates would have to be admitted to turn the scale, and to 

 show that the whole of the produced fat might have been 

 derived from the albuminoids of the food. After going into 

 considerable detail on the point, it was concluded that any 

 such range of error was simply impossible. 

 A test ex- Further, it was maintained that, in the case of pigs fatten- 

 penment. m g rapidly on their most appropriate fattening food, the 

 amount of fat stored up in proportion to the amount of fat 

 and nitrogenous substance consumed was so large that the 

 question of whether or not the carbohydrates contribute to 

 fat-formation might be conclusively settled by a properly 

 conducted feeding experiment with those animals, without 

 any analysis of the faeces or the urine, or any determination 

 of the products of respiration. It was stated that it was only 

 necessary to select two animals of a breed of good fattening 

 quality, and as nearly alike as possible in character and in 

 weight; a convenient size and weight being — say about 90 lb. 

 per head. Each should then be fed with ground barley of 

 good quality, giving it, by degrees, until both weighed about 



