46 Mr. P. J. Hartog and Dr. J. A. Harker on 



cause the radiation to the calorimeter to vary. We were 

 able to regulate the amount of this radiation at will ; but, 

 of course, too small a flame made the determination too 

 slow, too large a flame the initial and final corrections too 

 high. The calorimeter and its jacket, and the thermometer 

 were protected from excessive radiation by means of screens 

 of asbestos board. 



The calorimeter itself consisted of a copper vessel 

 weighing 282 grammes. The general arrangement of the 

 calorimeter and jacket was practically identical with that 

 employed by Berthelot. 



The stirrer consisted of a ring of copper pierced with 

 several holes, and moved up and down on glass guides fixed 

 into a light wire frame, which served to protect the glass 

 worm from any accidental blow. It was moved by means 

 of an electro-motor, and a wheel and crank mechanism. 



The thermometer used was one by Baudin, of Paris (No. 

 12,771, metastatic), divided into xo-ths of a degree centigrade. 

 It was compared with an instrument calibrated by Dr. 

 Schuster, and compared by him with a thermometer 

 standardized at the International Bureau of Weights and 

 Measures at Sevres. It was read by means of a telescope, 

 and the 20th part of a division, i.e., the x^ooth part of a 

 degree was estimated. 



The actual modus operandi was as follows : — The water 

 was first boiled in the flask with the valve at C closed, and 

 the steam escaping (not condensed). The thermometer 

 was read from this time forward at the end of each half 

 minute. When the march of the thermometer had become 

 regular for some minutes, the valve was opened and the 

 exit F closed. The steam then condensed in the worm. 

 When it was thought that sufficient steam had been con- 

 densed, the total rise being from 3 to 4 degrees, the exit 

 was opened, and the valve closed, and thermometric obser- 

 vations were only discontinued when the march of the 



