50 



supply of energy was corresponding!}' large. The value for energ}^ 

 is, however, so low that it seems not unlikely that owing to an over- 

 sight or misunderstanding some articles were omitted in recording the 

 daily diet. 



In experiments with the respiration calorimeter made by Atwater 

 and his associates, it has been found that subjects who were performing 

 ahnost no external muscular work expended as much energ}' per day as 

 Subject K received from his diet on an average, and when their mus- 

 cular activity was increased their output of energy was correspond- 

 ingly larger. In other words, they required as much energy for the 

 simple processes of living as was supplied by the recorded diet of this 

 subject, who had considerable muscular exercise in connection with his 

 usual daily occupations. It is difficult to believe, in the light of the law 

 of the conservation of energy, that he could have subsisted on such a 

 diet without some drain on his body, and certainly it can not be sup- 

 posed that he could continue to draw upon his body throughout the 

 college year without some eflfect upon it. 



DISCUSSION OF RESULTS. 

 METHOD OF COMPUTING THE DIETARIES. 



The practical impossibility of weighing and sampling all the cooked 

 food on its way to the subjects, under conditions existing at Randall 

 Hall, led to a method of making dietary studies difi'erent from that 

 generally followed. As stated earlier (p. 5), records were kept of the 

 number and kind of the portions of food served to each subject. The 

 amount of food in each portion was estimated by weighing sample 

 portions and finding an average value from these weighings, which 

 was used to calculate the weight of the portions actually served to the 

 men. This ma}^ be called the method of average weights, inasmuch as 

 it rests upon the assumption that the weights of successive portions 

 of any given cooked food, on account of accidental variations, swing 

 around an ideal value, which can be determined by weighing a sufficient 

 number of the portions and averaging the figures so obtained. That 

 this is the case within certain limits of error was shown by the agree- 

 ment between weighings of consecutive portions of any given article 

 of food. This is nothing more than is to be expected where the same 

 sized dishes and serving spoons are used da}" after day for serving the 

 same foods For example, to obtain the average weight of an order 

 of white bread, nine portions were weighed as they came from the 

 serving room on different days, with the following results: 66, 56, 63, 

 69, 64, 70, 62, 63, and 6(^ grams, the average being 64.3 grams. While 

 the difi'erence between the extremes is appreciable, the average varia- 

 tion from this average is only 3 grams, or 4.7 per cent. 



