688 



NATURE 



[February i8, 19 15 



obtained with mixtures of all inflammable gases or 

 vapours with air or oxygen, and very similar pheno- 

 mena occur with like mixtures of inflammable dust. 

 The rise and fall of pressure is studied by means of 

 sensitive indicator devices, which produce tracings on 

 a rotating drum. 



The pressure rise is caused by the formation of 

 flame at the ignition point — in this case an electric 

 spark — and the spread of this flame throughout the 

 vessel increases the temperature of the gaseous con- 

 tents, and thus increases the pressure, as in a closed 

 vessel the volume remains constant. The rate of 

 pressure rise is thus, broadly, the measure of the rate 

 of the travel of the flame through the mass; and, if 

 it be assumed that maximum pressure is attained 

 when the whole vessel is filled with flame, then it is 

 possible to determine the flame velocity by these 

 measurements. 



In these experiments the flame velocity for the weak 

 mixture is about 2 ft. per sec, and for the strong 

 mixture 13 ft. per sec. These velocities are attained 

 in closed-vessel experiments without a moving piston. 

 In actual engine cylinders, where the charge has been 

 introduced at a high speed through an inlet valve, 

 the flame velocity is much higher, and it varies with 

 different types of engine from about 20 to 100 ft. per 

 sec. The cause of this variation for similar mixtures 

 has been the subject of investigation by Clerk and 

 Hopkinson, and it has been definitely proved that the 

 higher flame velocity found in working engine 

 cylinders is due to the residual turbulence of the 

 mixture after compression caused by the high rate 

 of flow into the cylinder through the inlet valve. In 

 some engines the gases forming the charge flow 

 through the inlet valve at a rate as high as 160 ft. 

 per sec. Without this increase of flame velocity in 

 the engine cylinder it would have been impossible to 

 run high-speed engines, such as petrol engines, in 

 an economical manner. At the lower flame velocity 

 of the closed vessel the exhaust valve will be open 

 before the pressure had time to rise. The effect of 

 turbulence within the cylinder is very marked in 

 other respects. The rate of loss of heat from a 

 heated platinum wire to the mass of air in the com- 

 bustion space of an engine is much greater for any 

 given point of the stroke when the engine is running 

 fast than when it is running slow, and if the charge 

 be trapped, so as to allow several compressions and 

 expansions before measuring the heat loss, the turbu- 

 lence has so died down that the conductivity 

 diminishes. 



Another point of interest in gaseous explosions is 

 found in the fact, predicted bv Clerk many years 

 ago, but proved by Hopkinson, that the flame entirely 

 fills the vessel before maximum pressure is attained, 

 that is, the combustion, even in an explosive mixture 

 of gas and air, is relatively slow as the chemical com- 

 bination approaches completion. 



Another point proved by Hopkinson 's experiments 

 in a large closed vessel is that at whatever point in 

 the vessel the ignition be started, that point is the 

 point of maximum temperature during the subse- 

 quent pressure rise, and at that point the temperature 

 rises about 500° above the temperature of combus- 

 tion, due to adiabatic compression of the hot gas. 



The investigations of Profs. Callendar, Dalby, and 

 Coker by means of platinum resistance thermometers 

 and platinum alloy thermal couples has proved the 

 temperatures attained in ordinarv engines to vary 

 from about 1800° C. to 2500° C. ' The new method's 

 of direct thermometric measurement of the tempera- 

 ture flame have amply proved, however, the general 

 correctness of the older method of deducing mean 

 temperature by pressure change. 



NO. 2364, VOL. 94] 



The investigations of Callendar, Hopkinson, and 

 David (Prof. Hopkinson 's pupil) on radiation are of 

 great importance, and prove that this source of heat 

 loss is only second in magnitude to heat loss by con- 

 vection and conduction. All the radiation experi- 

 ments clearly show the existence of an unexpected 

 transparency of flame to its own radiations. The 

 radiation work throws much light upon heat distribu- 

 tion in the rapidly succeeding explosions used in 

 internal-combustion engines. 



It was long ago observed by Hirn, Bunsen, and 

 others that the rise of temperature in gaseous ex- 

 plosions could not be calculated from the then 

 assumed specific heat of the constituent gases and 

 the known calorific value of the inflammable gas. 

 The deficit of temperature was found to be about 

 50 per cent., and many attempts were made to explain 

 this deficit, Hirn advocating the theory of heat loss 

 on the rising line and Bunsen supporting the idea of 

 a limit to temperature due to dissociation. Later, the 

 French observers. Mallard and Le Chatelier, main- 

 tained that at least part of the deficit could be 

 accounted for on the assumption of increase of specific 

 heat of the gases. Investigations of the members of 

 the Committee have dealt, not only with the points 

 which have been here discussed, but with all these 

 questions — heat loss on the rising line, specific heat 

 of the constituent gases, heat loss on the falling line, 

 and dissociation of the combining gases. 



Specific heat work has been in progress by Clerk, 

 by Callendar and his pupil, Swann, and much of this 

 work has not yet been published. Dissociation has 

 been discussed by Dr. J. A. Harker, Prof. Smithells, 

 and Dr. Bone, and both internal energy and dissocia- 

 tion have been discussed by Hopkinson. Ignition 

 temperatures of gases have been dealt with by Prof. 

 Harold Dixon, and Dr. Watson has studied the 

 nature of the exhaust gases from the petrol engine. 

 Many experiments, too, have been made on the law 

 of cooling and heating of gases under compression in 

 cylinders by Hopkinson, Dalby, Callendar, and Clerk. 



As a result of this work, the conclusion has been 

 arrived at that, so far as explosions in internal-com- 

 bustion engines are concerned, dissociation has but 

 little to do with the limit reached. This limit is 

 partly due to increased specific heat at high tempera- 

 tures, to heat loss to the walls, and to radiation from 

 the explosion. Varying specific heat and increasing 

 radiation account for most of the deficit. Allowing 

 for all these things, however, it appears now to be 

 established that combustion is not quite complete 

 even at maximum temperature, and Watson's experi- 

 ments on the spectrum of an explosion flame appear 

 to support this view. 



All these matters are still under examination, and it 

 Is hoped that in the near future a much more com- 

 plete knowledge may be gained than at present exists. 

 Much is known in a qualitative way, and some 

 quantitative knowledge has been attained, but much 

 still remains to be done in the way of quantitative 

 determinations of matters at first apparently so simple 

 as specific heat. 



CLIMATE AND TREES. 

 T N an article on "Woods and Trees of Ireland," in 

 -*■ the Co. Louth Archaeological Journal for 1914. 

 Prof. Augustine Henry states that in Ireland, as in 

 Scandinavia, the climate prevailing in neolithic times 

 was drier and warmer than that of to-day. Many 

 facts are adduced in support of this improvement, 

 which in Scandinavia amounted to an increase of 4° F. 

 in the average annual temperature. The occurrence 

 of this optimum cHmatic period is confirmed by the 



