558 VITALITY AND EFFICIENCY WITH RESTRICTED DIET. 



Squad A. Sjpe, owing to illness (see p. 360), was not present at the 

 Laboratory after December 8, and his place was not filled. In Squad 

 B, Fis was in the hospital for an operation for hernia on December 

 15. McM was not taken into Squad B until after their first session on 

 October 6, when he replaced Kon. In the early part of January he had 

 some digestive difficulties which made it seem unwise for him to serve 

 as a subject during the severe food reduction for this squad. His 

 place was taken by Kim, who had served on January 5 in place of Lon. 

 Lon on that date was ill and it was possible that he might not be able to 

 continue as a subject in Squad B. Mac was called into the Govern- 

 ment service during the Christmas vacation, receiving his notifica- 

 tion only the day before it was necessary for him to appear at camp. 

 We were not informed of his call until after he had been at camp several 

 days. An arrangement was made with the commandant to release him 

 until the first of February. The subject thought that this short 

 interval in college would be of little value, and, as he was well started 



* This average does not include Kon, Spe and Fre. 

 ^ The one normal experiment, Sept. 29, is not included in the average. 



3 It is worthy of notice that Vea, who made this best score and one of the best average scores, 

 suffered most from cold hands during the low-diet period. 



in his work, he did not see fit to leave it to carry on the experiment. 

 A substitute was found in Sch, who served in the last four sessions of 

 Squad B. Tho, in returning from his Christmas vacation, was de- 

 layed, and because of transportation conditions could not reach the 

 Laboratory on the evening of January 5 until after the group work 

 had been completed. These breaks or irregularities in all our psycho- 

 logical data occur in the succeeding tables for other measurements 

 and will not need special mention there. 



The values given in tables 147 and 148 show the average number 

 of errors per line. Each average represents five lines, as the test 



