334 FRANCIS B. SUMNER 



part of the temperature tests here considered were made either 

 late in the morning or early in the afternoon, some extending 

 through the noon hour. A classification of such tests into 'morn- 

 ing' and 'afternoon' ones would be arbitrary. Finally, the mean 

 time of observation did not differ widely for the two contrasted 

 lots, being 11.30 a.m. for the warm-room experiments and 1.40 

 P.M. for the cold-room ones. That this difference of two hours 

 could materially affect the trend of my results would hardly be 

 contended. 



In making the comparisons now to be considered, I have in- 

 cluded only the first temperature reading for a given day. I 

 have excluded such of the tests with warm-room mice (table 7, 

 January 16; table 8, March 6) as were made at times of abnor- 

 mally low atmospheric temperature^" and such of the tests with 

 cold-room mice (table 10, January 26; table 11, March 1) as 

 were made at times of abnormally high temperature. I have also 

 excluded certain other tests (table 9, March 31) which were made 

 under conditions of abnormal excitement on the part of the mice. 



It so happens that nearly all of the cold-room determinations 

 were made on days when the temperature was some degrees 

 below zero, since it was my object at that time to determine the 

 greatest possible changes in body temperature which might be 

 attributable to atmospheric conditions. Thus, the mean air tem- 

 perature, at the time of the cold-room tests, was about — 3.°C., 

 while the mean temperature to which the animals had been sub- 

 jected for the preceding ten days was +7°, in the case of the 

 table 10 mice, and +1°, in the case of those of tables 11 and 

 12.31 



'" In reality, the March 5 tests of table 8 were made at a time when the room 

 temperature was considerably below the average. This, for the preceding ten days 

 had been 22.6°, i.e. 8° or 9° above the temperature when the March 5 readings 

 were made. To exclude these results, however, would be to seriously reduce the 

 number of available figures, and their inclusion is further justified by the fact that 

 the mean of those for the five adult males is very nearly the same as (in reality a 

 little higher than) the mean figure for the 9 other adult males of the warm-room 

 set. And it must be borne in mind that all of the cold-room tests were made at 

 times of unusually low temperature. 



'' The mean temperature of the cold room, during the entire period of my 

 experiments, was about +6°C. 



