FEEDING OF ANIMALS. 329 



and in that found in the excrements. There was, too, a much 

 greater loss indicated by the results of the direct nitrogen 

 determinations in the urine dried with an excess of oxalic 

 acid, than when the nitrogen was calculated from the amount 

 of urea found daily in the fresh urine. As, however, nitrogen 

 determinations (by soda-lime and platinum salt) were made 

 by two analysts, whose results agreed very fairly, it may be 

 concluded that the loss was connected with the methods of 

 collection, sampling, and preparation for analysis, rather than 

 with those of the analysis ; and it is probable that the same 

 remark applies to the results obtained with the faeces. In 

 illustration of the range of loss of nitrogen indicated, it may 

 be stated that when the nitrogen in the urine was reckoned 

 from the amount of urea, the loss ranged in the four experi- 

 ments between 20 and 30 per cent of that in the food, and 

 when by direct nitrogen determinations in urine as well as in 

 fseces, from under to over 40 per cent. However, in the case 

 of each food, whether the nitrogen in the urine was deter- 

 mined, or calculated from the urea, there was considerably 

 less loss indicated over the ten-day than over the shorter three- 

 day period ; again connecting the error with the collection, 

 sampling, and preparation, rather than with the analysis. 



In view of these unsatisfactory results, and of the evidence Further ex- 

 that much at any rate of the loss was probably due to experi- ^™J£ ents 

 mental difficulties and errors, the subject was taken up again 

 in 1862. The pigs were kept in frames as before, and the 

 voidings were collected in the same way; but they were 

 sampled morning and evening, instead of only once in the 

 twenty-four hours, as in 1854. Advantage was also taken of 

 the previous experience in regard to various other points of 

 manipulation. Lastly, the direct nitrogen determinations were 

 made by soda-lime as before, but with titration instead of 

 platinum salt. 



Two animals were experimented upon, each for a period Food used. 

 of ten days, and after an interval of a few weeks for five days 

 more. The food of one consisted of three parts bean-meal 

 and one part bran, and of the other of three parts barley-meal 

 and one part bran. 



In the case of the pig having the highly nitrogenous bean- Nitrogen 

 meal and bran, the nitrogen balance for the ten days showed ^°<3no* 

 a gain of 4.04 per cent when direct nitrogen determinations accounted 

 were made in the urine, and of only 2.32 per cent when the f w ' 

 nitrogen in the urine was calculated from the amount of urea. 

 On the other hand, over the five-day period there was a loss 

 indicated of 3.35 per cent with the direct nitrogen determina- 

 tions in the urine, and of only 1.61 per cent when the nitrogen 

 was calculated from urea. In the latter case, therefore, the 



