202 REPORTS ON THE STATE OF SCIENCE. 



might play a part in the heat transmission, he suggested that a more 

 important agent was radiation from the roof of the furnace which 

 received heat by direct contact with the hot gas and so reached a very 

 high temperature. He showed by calculation that a comparatively 

 small excess of temperature in the roof over that of the floor would 

 cause a sufficient flow of heat. 



But though the discussions on the Siemens furnace and the work 

 of Helmholtz show that the idea that a flame, even if non-luminous, 

 might radiate large amounts of heat was a familiar one to many people 

 twenty years ago, its possible importance in causing loss of heat during 

 and after a gaseous explosion and in determining the heat flow in a 

 gas-engine does not appear to have been appreciated until quite recently. 

 Professor Callendar was probably the first to draw attention to its sig- 

 nificance in this connection. In the discussion on a paper about 

 explosions, read before the Eoyal Society in 1906, he said that he 

 had found a non-luminous Bunsen flame to radiate 15 to 20 per cent, 

 of its heat of combustion, and expressed the opinion that the loss 

 from this cause in a closed-vessel explosion would be of the same order. 1 

 Professor Callendar 's note dealing with this matter is published in full 

 in Appendix A, and it is only necessary to state here that he was 

 led to study the subject by his work on the efficiency of the petrol 

 motor. 



There are, in fact, several points about the behaviour of gas-engines 

 which suggest the importance of radiation as a cooling agent. The 

 particular matter which attracted Callendar's attention was the effect 

 of speed on thermal efficiency. His experiments showed that a part of 

 the loss of efficiency in an internal-combustion motor, as compared 

 with the corresponding air-cycle, was independent of the speed at which 

 the engine was run. The loss of heat per cycle could, to a first approxi- 



mation, be represented by an expression of the type Ah — where n 



n 



is the number of revolutions per minute and A and B are constants. 

 The term A represents a constant loss of heat per explosion, and 

 among the many causes contributing to this constant loss of heat, radia- 

 tion from the flame is probably important. 2 



Another phenomenon which is difficult to explain, except as the 

 result of radiation, is the effect of strength of mixture on heat-loss. 

 The following table shows some results which were obtained by 

 Hopkinson upon a 40-h.p. gas-engine 3 : — 



Percentage of gas in cylinder contents ... 8-5 11-0 per cent. 



Total heat-loss per minute 1,510 2,300 B.Th.U. 



Total heat-loss as percentage of total heat-supply . 29 34 per cent. 



Temperature of piston 300° C. 430° C. 



It will be observed that the proportion of heat-loss to the walls 

 increases very materially as the strength of mixture is increased. If 

 the transfer of heat were wholly due to conduction it might be expected, 



1 Hopkinson, Proc. Roy. Soc. A., vol. lxxvii., p. 400. 



2 Proc. Inst. Automobile Enq., June 1907. 



3 Proc. Inst. C.E., vol. clxxvii. (1909). 



