6 INFLUENCE OF BORAX AND 



The faeces whenever passed were collected in a weighed dish, 

 the mass thoroughly desiccated over a water-bath, and the dry 

 weight ascertained. The dried material was then pulverized 

 and the nitrogen-content as well as the ether-soluble matter 

 determined in sample portions. The nitrogen determinations 

 were always made in duplicate by the Kjeldahl method and 

 rarely varied more than 0.05 per cent. Whenever, as some- 

 times occurred, hair accumulated in the cage it was likewise 

 collected and the nitrogen determined. The ether-soluble 

 matter was determined by extraction of the dried faeces in a 

 Soxhlet-apparatus. 



The animals were fed during the experiments on a mixed 

 diet composed of fresh lean beef, cracker dust, lard, and water. 

 The meat was prepared as follows : fresh lean beef, freed as 

 far as possible from all adherent fat and connective tissue, 

 was run through a hashing machine, after which it was 

 enclosed in a bag of thin cloth, placed under a heavy press, 

 and kept there under increasing pressure for several hours, 

 the bloody fluid which drained off being thrown away. By 

 this method there results a mass of tissue free from surplus 

 moisture, and which, when enclosed in a bottle, will keep 

 perfectly fresh on ice for seven to ten days without separation 

 of fluid. Several advantages accrue from this method. Thus, 

 we have a perfectly homogeneous mixture which can be drawn 

 from for at least a week with surety that its nitrogen-content 

 is constant. There is therefore no necessity for a daily 

 determination of nitrogen in this portion of the diet, for each 

 sample can be analyzed when prepared and the data accepted 

 as long as the meat keeps fresh. Further, meat prepared in 

 this manner at different times, if subjected to essentially the 

 same pressure, varies only slightly in its content of nitrogen. 

 We have invariably analyzed each lot when prepared to avoid 

 any possibility of error, but, as the following results show, the 

 differences in composition are very slight and necessitate very 

 little alteration in the proportion of meat when changing from 

 one lot to another. The following results are a few of the 

 many obtained: 



