5o6 



NATURE 



[June 24, 1909 



pressure, bul, owing to variation of specific heat with 

 tempirature, a correction must be made, which in the 

 present slate of knowledge is very uncertain, as the dis- 

 tribution of temperature and the variation of the specific 

 heat are not accurately known. 



The condition of the gas as regards chemical equilibrium 

 has been the subject of much speculation and research. 

 Calculations by tlie fvii law of the rise of pressure due to 

 the amount of thermal energy liberated give very much 

 greater pressures than are found by experiment, and 

 various theories have been put forward to account for this. 



The view that dissociation of the constituents of, the 

 ignited charge will account for the discrepancy appears 

 improbable, as there is no conclusive evidence to show that 

 either steam, COj, or nitrogen arc split up to any extent 

 at the temperatures and pressures obtained, while the fact 

 that in weak mixtures the discrepancy between calculation 



vibratory motion in the gas, but it appears unlikely that 

 this has any elfict on the mean pressure shown by: tht 

 gauge. .Mihouyh the difiicuUies which arise in ihe.detci;- 

 minution of 'he internal energy are so great, the. results 

 of independent observers are in very fair agreement, as 

 is shown by ihe accompanying diagram, Kig. i, in which 

 various deterniinulions of the internal energy of a gas- 

 engine mixture are marked and compared with the values 

 for an ideal gas having an internal energy of 49 calories 

 per gram molecule per degree. 



The experiments of Clerk are particularly interesting, as 

 he used a method of a novel character, which pcrmittcil 

 the study of the working fluid in the gas-engine cylinder 

 itself. 



An indicator diagram of a gas-engine cycle gives in- 

 formation as to the time of ignition, the work done, and 

 the compression and expansion of the charge. It is 



TemhersCbuj'e, Centi^racLe. 



and experiment is about the same lends no support to the 

 dissociation theory. The cooling effect of the walls plays 

 some part, but the experimental evidence of explosions in 

 vessels of various forms and capacities shows that this 

 cause alone is quite insufficient to account for the differ- 

 ence. 



Another view, due to Clerk, is that the combustion of 

 the gas is not complete at maximum pressure, so that in 

 mixtures of all strengths, but especially in weak ones, 

 there is a suppression of heat which materially affects the 

 maximum explosion pressure, and the cooling effect of the 

 walls will have a considerable time effect on the combus- 

 tion process. It cannot be doubled that combustion is 

 greatly retarded in the neighbourhood of cold metal walls, 

 and some direct evidence is available that such phenomena 

 are mainly of a surface character. Profs. Bone and Dixon 

 are of the opinion that if the effect of cold bodies may 

 be disregarded, the combustion of the charge in the 

 presence of air is practically complete before the attain- 

 ment of maximum pressure. 



The effect of the explosion is also to set up intense 

 NO. 2069, VOL. 80] 



possible by rearranging the valve gear to shut in an ex- 

 ploded charge, and the indicator card obtained while the 

 engine is coming to rest affords further information con- 

 cerning the specific heat and the rate of heat loss to the 

 walls. 



A portion of such a diagram is shown in Fig. 2, in 

 which the curve BC represents the first compression of the 

 charge after llie valves are shut down, and CD represents 

 the following expansion curve. Usually, the first five or 

 six cycles are distinct, but they ultimately merge into 

 one another ns the engine comes to rest. If the gas be 

 comprc!jScd ov expanded without gain or loss of heat, the 

 specific heat .it constant volume can be readily obtained 

 by a consider. II ion of the work areas of the diagram and 

 the end temperatures, but on account of the heat ffow 

 to the walls .1 correction must be made, which can be 

 obtained by a successive approximation process. This heat 

 loss is divided between the expansion and compression, 

 and Mr. Clerk divides it on the assumption that if the 

 mean temperatures in compression and expansion arc the 

 same, the heat loss will be the same. The mean tempera- 



