320 FRANCIS B. SUMNER 



Faichney Company, of Boston. They were, in appearance, much 

 like ordinary clinical thermometers, but differed from these in 

 several important details. The scale was adapted to cover a 

 wider range of temperatures; the bulb was more slender and con- 

 tained a much smaller amount of mercury, and was connected with 

 the graduated stem by a long narrow neck. Furthermore, the 

 mercury in the stem did not detach itself, so as to register the 

 maximum point reached, as in the case of an ordinary clinical 

 thermometer, but was free to fluctuate up and down. It was 

 thus possible for the observer to follow the negative as well a? 

 the positive temperature changes. 



Two thermometers, referred to as 'number V and 'number 2,' 

 were used in taking the temperatures of the full-grown (or nearly 

 grown) mice. 'Number 1' was graduated in the Fahrenheit scale, 

 'number 2' in the Centigrade. The latter instrument was used in 

 the majority of these determinations. Its total length was 95 

 mm.; length of bulb, 10 mm.; length of neck, 12 mm.; length of 

 scale, 50 mm., ranging from 32° to 42°C. The diameter of the 

 bulb was 2 mm., the diameter of the neck being from 1.5 to 2 mm. 



In practically all of the experiments a rubber ring or collar was 

 placed on the narrow neck connecting the bulb and the graduated 

 portion of the stem. This served, on the one hand, to insure a 

 uniform depth of insertion (a point of considerable importance), 

 on the other to reinforce the anal sphincter in closing the entrance 

 to the rectum. This collar, in 'number 1,' was placed 13 mm. 

 from the tip of the bulb (about 5 mm. from its proximal end) ; in 

 'number 2' it was placed 16 to 18 mm. from the tip of the bulb 

 (about 7 mm. from its proximal end). 



A third thermometer ('number 3'), having a scale ranging from 

 20° to 40°C., was used almost exclusively for the very young 

 mice (Section 3). In the present series, it was used only in certain 

 cases where sub-normal temperatures were obtained. 



The errors of these various thermometers were determined by 

 careful comparison, in water of different temperatures, with a 

 standard thermometer which was borrowed from the Bureau of 

 Standards. The necessary corrections have, of course, been 

 applied to the original readings. 



