S6 Prof. B. Hopkinson on the 



in terms o£ the time, it is easy to calculate the rate at which it 

 is receiving heat from the gas. Let 0' be the temperature 

 and t the time, then heat must be supplied to the wire at a rate 



Jc-- — !-/(#'), where h is the capacity for heat of the wire, and 



f(&) the rate at which it loses heat by radiation in vacuo when 

 at temperature 6'. The radiation term depends on the nature 

 and extent of the surface of the wire and on the temperature; 

 it is not important except at very high temperatures — for most 

 purposes a sufficiently good approximation to its value can be 

 obtained from the numerous published results for polished 

 platinum, which are in fair agreement with one another, and 

 give the radiation per sq. cm. of surface up to temperatures 

 near the melting-point. 



The difference of temperature between the wire and the 

 gas at a distance from it will, under any given conditions, be 

 proportional to the rate of exchange of heat between wire 

 and gas. If 6' be the gas temperature, we have 



i/'J t +/(8')=Mo-e'), 



whence 6 can be obtained if X is known. The constant X is 

 the rate at which heat passes between the wire and the gas, 

 per degree difference of temperature between them, under the 

 conditions of temperature, motion of the gas, &c, which 

 actually obtain. These conditions vary from point to point 

 of the cycle in the gas-engine, and X also varies. It is 

 necessary, in fact, in measuring gas-engine temperatures in 

 this way to determine the relation between heat flow and 

 temperature difference at each point of the stroke. 



The experiments to be described in this paper consisted in 

 a determination of the temperatures in the cylinder of a large 

 gas-engine by means of a platinum thermometer, the gas- 

 engine being motored round with the gas-supply cut off so 

 that it simply compressed and expanded a charge of air. The 

 object of the experiments was to test the platinum thermo- 

 meter as a means of measuring rapidly varying gas-tempe- 

 ratures, rather than an enquiry into gas-engine phenomena. 

 The absence of explosions secures a fairly uniform temperature 

 throughout the cylinder contents. The mean temperature of 

 the charge given by the indicator diagram therefore does not 

 differ much from the local temperature in the neighbourhood 

 of the wire : and the wire temperatures can be checked by 

 reference to the indicator diagram. I have used a simple 

 method of getting the correction constant X which seems to give 

 o-ood results, and a description of this may be useful to others 



