100 Mr. J. S. Macdonald. Studies in the 



driven to it from a constant-pressure supply through coils of tubing situated 

 in ice-filled tanks, and arrives at a temperature sometimes less, sometimes 

 greater, than 5° C. It passes out into a balance, such as that described by 

 Benedict. This balance is duplicate, and automatic arrangements provide 

 that the filling of the collecting pan suspended from one beam shall imme- 

 diately be followed by admission of the stream of water into the similar 

 collecting pan suspended from the second beam. The change over is made 

 known by the ringing of an electric bell, and the weight then ascertained by 

 the observer thus warned. Thus, this is the only set of readings which is not 

 arranged in five-minute periods, and average rates of flow have to be 

 accepted for each of the included five-minute periods. The intervals of 

 such readings have varied from 12 to 20 minutes. 



The remaining readings are taken by the observer every time warning is given 

 by a bell attached to a five-minutes clock, in order somewhat as follows : — 



(a) The wet and dry bulbs in the entering and leaving air. 



(b) The thermometers in the entering and leaving water. 



(c) The resistance of some 570 ohms of iron wire arranged on a series of 

 coils within the calorimeter, by means of which its temperature is assessed. 



(d) The number of revolutions of the cycle as read upon a cyclometer 

 driven by an electromagnet from contacts on the cycle. 



(e) The watts due to the lamps (wattmeter). 



(/) Voltmeters and ammeters connected with the circuits of the fan and 

 of the cycle-brake. 



{g) The surface temperature ; and 



(h) the rectal temperature of the subject, as evidenced by galvanometric 

 deflections due to suitable thermocouples. 



(i) The temperature of an incubator in which are placed the " constant 

 temperature junctions " of these thermocouples. 



The data collected in this way have been dealt with on a uniform plan 

 throughout the whole series of experiments here included. The only details 

 of this plan which perhaps should be dealt with briefly here are (1) the 

 means of applying the correction for the temperature of the calorimeter, 

 and (2) for estimating the storage value of the air space in the calorimeter. 



By suitable calibration experiments it has been found that an alteration of 

 0"01 ohm in the resistance of the calorimeter " thermometer " is equivalent 

 to an addition, or subtraction, of 1/6 kilo-calorie (or kalorie, as written 

 elsewhere in this paper). Now, in making up my accounts, I have expressed 

 all the results in rates of change per hour, so that a change of this value 

 observed in a five-minutes period (1/6 kalorie gained or lost in five 

 minutes) is expressed as a change of 2 kalories per hour in the rate of heat- 



