304 



SCIENCE. 



[Vol. XVI. No. 408 



pass south of the sun, it will not be so easy to use buildings as 

 screens in the northern hemisphere, and special means must be 

 devised. The luminous ring will be as bright and conspicuous as 

 in 1866, and the first appearance of the prolongation of the cusps 

 may be looked for about the 24th of November. 



It is now evident that similar opportunities will happen on the 

 1st of December, 1898, when least distance of centres will be about 

 1°, and about the 38th of November, 1906, when least distance of 

 centres will be about li", the planet in both cases south of the 

 sun. In each case the least distance of centres will be less than 

 the limit within which the formation of the luminous ring is pos- 

 sible, but the duration of the ring will be successively less as the 

 least distance between centres becomes greater. No other oppor- 

 tunities will present themselves until near the end of the next 

 century, when they will occur in June. 



Similar opportunities must have occun-ed in years preceding 

 1866; that is, on the 14th of December, 1858, and also on the 16th 

 of December, 1850; but it does not appear that either was used. 

 This last date is only nineteen months after Madlers observations 

 in May, 1849; and, if any one properly situated as to time had 

 endeavored to repeat Madler's observation on the day of conjunc- 

 tion, he would almost ctrtainly have seen the luminous ring. 



Lewis R. Gibbes. 

 Charleston, S.C., Nov. 13. 



A Problem in Physics. 



An experiment was tried by Joule nearly fifty years ago which 

 has attained a world-wide reputation, and which has crept into 

 nearly every text-book of physics. The commonly accepted in- 

 terpretation of it, however, would seem not entirely satisfactory. 

 I will quote from Tail's description of the experiment. 



" Joule took a sti'ong vessel containing compressed air, and con- 

 nected it with another equal vessel which was exhausted of air. 

 These two vessels were immersed each in a tank of water. After 

 the water in the tanks had been stirred carefully, ... a stop- 

 cock in the pipe connecting the two vessels was suddenly opened. 

 The compressed air immediately began to rush violently into the 

 empty vessel, and continued to do so till the pressure became the 

 same in both; and the result was, as every one might have ex- 

 pected, that the vessel from which the air had been forcibly extruded 

 fell in temperature in consequence of that operation. It had ex- 

 pended some of its energy in forcing the air into the other vessel; 

 but that air, being violently forced into the other vessel, impinged 

 against the sides of that vessel, and thus the energy with which it 

 was forced in through the tap was again converted into heat. 

 On stirring the water round these vessels, after the transmission 

 of air had been completed and the stopcock closed. Joule found 

 that the number of units of heat lost by the vessel and the water 

 on the one side was almost precisely equal to the quantity of heat 

 which had been gained on the other side." Tyndall gives the fol- 

 lowing (let B represent the vessel in which the air was compressed 

 to 23 atmospheres, and A the vessel which was exhausted): — 



"Now, the air, in driving its own particles out of B, per- 

 forms work, . . . and the air which remains in B must be 

 chilled. The particles of air enter A with a certain velocity, to 

 generate which the heat of the air in B has been sacrificed; 

 but they immediately strike against the interior surface of 

 A, their motion of translation is annihilated, and the exact quan- 

 tity of heat lost by B appears in A. The contents of A and B 

 mixed together give air of the original temperature. There is 

 no work performed, and there is no loss of heat." Tyndall gives 

 an illustration of a cylinder having a piston in the centre, and the 

 space above the piston a vacuum. Suppose the air below the 

 piston is heated up from 0° to 373° C. "If the pressure were 

 removed, the air would expand, and fill the cylinder. The lower 

 portion of the coIuqju would thereby be chilled, but the upper 

 portion would be heated ; and, mixing both portions together, we 

 should have the whole column at a temperature of 373°. In this 

 case we raise the temperature of the gas from 0° to 373°, and 

 afterward allow it to double its volume. The temperatures of the 

 gas at the beginning and at the end are the same as when the 

 gas expands against a constant pressure, or lifts a constant weight ; 



but the absolute quantity of heat in the latter case is 1,431 times 

 that eoiployed in the former, because, in the one case, the gas 

 performs mechanical work, and in the other not." 



The following quotation is from Balfour Stewart, and bears 

 upon this question : — 



" The prevalent idea is, that when air expands it becomes 

 colder, and that when condensed it becomes hotter; but Joule, by 

 experiment, has shown that no appreciable change of temperature 

 occurs when air is allowed to expand in such a manner as not to 

 develop mechanical power. It follows as an inference, that, 

 when air is coQipressed, the rise of temperature is scarcely at aU 

 due to the mere diminution of the distance between the particles, 

 but almost entirely to the mechanical effect which must be spent 

 on the air before this condensation can be produced." 



A final quotation is taken from Ganot's " Physics: " — 



"A strong metal box is taken, provided with a stop-cock, on 

 which can be screwed a small condensing-pump. Having com- 

 pressed the air by its means as it becomes heated by this process, 

 the box is allowed to stand for some time, until it has acquired 

 the temperature of the surrounding medium. On opening the 

 stop cock, the air rushes out; it is expelled by the expansive force 

 of the internal air : in short, the air drives itself out Work is 

 therefore performed by the air, and there should be a disappear- 

 ance of heat; and, if the jet of air be allowed to strike against a 

 thermopile, the galvanometer is deflected, and the direction of its 

 deflection indicates a cooling. . . . Joule placed in a calorimeter 

 two equal copper reservoirs, which could be connected by a tube. 

 One of these contained air at 33 atmospheres; the other was ex- 

 hausted. When they were connected, they came into equilibrium 

 under a pressure of 11 atmospheres; but, as the gas in expanding 

 had done no work, there was no alteration in temperature." 



I have given these quotations rather freely from standard 

 authors, in order to present the problem as clearly as possible. 

 In order to arrive at just the action taking place in this experiment, 

 it seems to me a phenomenon first described by Faraday in 1837 

 should be mentioned. Gas compressed to 80 atmospheres was 

 allowed to suddenly enter a cylinder 30 feet long, in which the 

 gas was at atmospheric pressure presumably. It was found, that, 

 where the gas rushed in, the cylinder was much cooled, while at 

 the other end it was heated. It would seem that in this case the 

 heating was not produced by the particles of gas impinging upon 

 the end of the cylinder. If a piston were placed in front of the 

 expanding gas, the whole of the gas on the other side of the piston 

 would be compressed and heated. If, now, instead of a piston, 

 we open a stop cock at the end of the cylinder, the gas would 

 stream in and compress that already there, and heat it; but the 

 gas, expanding violently as it enters, would be much cooled, and 

 this would more than counteract the heating where it enters. 

 Thus the farther end would show a heating, while the end at the 

 orifice would show a cooling as observed. Have we not precisely 

 analogous phenomena in Joule's experiment ? For a very small 

 fraction of a second (perhaps .0001) after the stop-cock was 

 opened, there would be a partial vacuum in A, into which the air 

 streams ; but after that the particles would not impinge upon the 

 sides of A, but would have their velocity diminished and finally 

 overcome by striking other particles. In imparting this velocity, 

 the particles in B would be slightly chiUed. The air, in streaming 

 out of B, would be cooled by expansion after an instant, and 

 would serve to cool the end of A near the orifice, as we have just 

 seen; also the chilled particles in B would stream into A, and thus 

 cool it still more. Whatever may be the action in these vessels, 

 it is certaiii that the final heating in A, and cooling in B, would 

 be exceedingly slight as shown by Joule's experiment, though it 

 does not seem that the popular explanation is entirely correct. 



It seems to me this question of the action of air in Joule's two 

 vessels is an intensely interesting one. The conclusion that the 

 chilling of the air in the vessel due to the work of imparting a 

 velocity to its particles is very slight, corroborates in a marked 

 manner the experiments tried by the present writer, in which he 

 found a cooling of four degrees, while the dynamical cooling 

 should have been ten times greater. The quotation from Ganot 

 shows precisely an analogous case. H. A. Hazen. 



Washington, Nov. 17. 



