June 2, 1892] 



NA TURE 



lOI 



motion Maxwell's law of distribution of energy would fail, 

 assuming that law to consist in the ultimate equality of the 

 energy of different parts of the system. He has thus shown the 

 necessity for more accurate language than is commonly employed 

 in the enunciation of that law, and a consideration of his prob- 

 lem may help to determine the limits to whiih it is subject. 



The following statement, whether co-extensive with Maxwell's 

 law or not, will probably be accepted as true as far as it goes — 



If there exist a very great number of material systems, the state 

 of each being defined by certain coordinates and momenta, and 

 if at a given instant all combinations of the co-ordinates and 

 momenta are represented among them with frequency propor- 

 tional to e-'''^^'^^\ then that distribution will be permanent — 

 that is, will not be disturVed by the mutual action of the 

 systems, or by any forces in the field of which they are placed, 

 provided all the forces concerned be conservative. 



The further question as to how far the solution thus found 

 for the permanent state is unique, has been treated by Boltz- 

 mann. He shows that a certain function, which in stationary 

 motion must be positive and constant, neces>arily diminishes 

 with the time, so long as any small deviations exist from the 

 above described state. It is obvious that this proposition of 

 Bolizmann's cannot be applicable to all cases of stationary motion. 

 Periodic motions are exceptions, and so is the system described 

 by Lord Kelvin. The question is what assumptions underlie 

 Boltzmann's demonstration. It will be of great advantage if 

 one speaking with Lord Kelvin's authority will assist in defining 

 the limits to which the proposition is subject. 



Maxwell, although he may at times have expressed himself 

 incautiously, was aware that the theory was subject to limita- 

 tion?. The statistical, as distinguished from the historical, 

 method was from his point of view of the essence of the theory. 

 A distinction may be drawn between systems, such as Lord 

 Kelvin's, to which the statistical method is inapplicable, and 

 those in which the stationary motion, when attained, is what is 

 called thermal motion — that i?, the relative motions are in all 

 direciions indifferently, and of that irregular character in which 

 heat is supposed to consist. 



It may be that we shall be driven to the conclusion that 

 Maxwell's law has no application except to this class of sy>tems ; 

 that it is, in fact, only the limiting state to which 'a material 

 system approaches as we increase indefinitely the number of 

 its degrees of freedom. 



It does, at all events, appear that in cases where the law 

 fails, its failure is due to the introduction of some restrictions on 

 freedom of motion, especially as regards direction. Maxwell 

 pointed out that daemons— or, shall we say, beings endowed with 

 free will— miyht by directing the courses of individual mole- 

 cules cause a sy-^tem to violate, not only the law of distribution 

 of energy, but even the second law of theimodynamics. What 

 these bemgs might be suppo.-ed to do, that Lord Kelvin in fact 

 does once for all for his system, by prescribing a pHori \.\i^ 

 directions of motion and other conditions of the problem to suit 

 his purpose. H. W. Watson. 



S. H. BURBUKY. 



The Former Connection of Southern Continents. 



With reference to the very interesting question treated in Mr. 

 Mellard Reade's letter of your issue of May 26 (p. 77), as to the 

 former connection of southern continents, it may be worth while 

 calling attention to the fact that a great circle, which I may call 

 the Kaffraria Great Circle, connects that coast line with the 

 Falkland Island and the South Georgia Island. It may be pre- 

 sumed that these two islands are the remaining summits of what 

 was once a chain of mountains in connection with the continent 

 of South America. Some of the points through which or near 

 which this great circle passes are as follow — the above- 

 mentioned islands, Port de Sta. Cruz, Patagonia ; it traverses 

 the Pacific, runs parallel to the southern branch of the Aleutian 

 Islands, and cuts Kamtchatka somewhat south of Klienchewskaia 

 Volcano, and traversing A^ia emerges by the Island of Cutch, 

 so interesting on account of the earthquakes which occurred 

 there. It is of interest to note that South Georgia Island is 

 antipodal to the northern extremity of Saghalian Island. 



J. P. O'Reilly. 



Royal College of Science for Ireland, 

 Stephen's Green, Dublin, May 30. 



NO. I 179, VOL. 46] 



ON THE RELATIVE DENSITIES OF 

 HYDROGEN AND OXYGEN} 

 I N a preliminary notice upon this subject (Roy. See. 

 ■'• Proc, vol. xliii. p. 356, February 1888), I explained 

 the procedure by which I found as the ratio of den- 

 sities 15*884. The hydrogen was prepared from zinc and 

 sulphuric, or from zinc and hydrochloric, acid, and was 

 liberated upon a platinum plate, the generator being in 

 fact a Smee cell, inclosed in a vessel capable of sustain- 

 ing a vacuum, and set in action by closing the electric 

 circuit at an external contact. The hydrogen thus pre- 

 pared was purified by corrosive sublimate and potash, 

 and desiccated by passage through a long tube packed 

 with phosphoric anhydride. The oxygen was from 

 chlorate of potash, or from mixed chlorates of potash 

 and soda. 



In a subsequent paper " On the Composition of Water " 

 (Roy. Soc. Proc, vol. xlv. p. 425, February 1889), I 

 attacked the problem by a direct synthesis of water from 

 weighed quantities of the two component gases. The 

 ratio of atomic weights thus obtained was 1589. 



At the time when these researches were commenced, the 

 latest work bearing upon the subject dated from 1845, 

 and the number then accepted was I5"96. There was, 

 however, nothing to show that the true ratio really 

 deviated from the 16 ; i of Prout's law, and the main 

 object of my work was to ascertain whether or not such 

 deviation existed. About the year 1888, however, a 

 revival of interest in this question manifested itself, 

 especially in the United States, and several results of 

 importance have been published. Thus, Prof. Cooke and 

 Mr. T. W. Richards found a number which, when cor- 

 rected for an error of weighing that had at first been 

 overlooked, became 1 5*869. 



The substantial agreement of this number with those 

 obtained by myself, seemed at first to settle the question, 

 but almost immediately afterwards there appeared an , 

 account of a research by Mr. Keiser, who used a method 

 presenting some excellent features, and whose result was 

 as high as I5'949. The discrepancy has not been fully 

 explained, but subsequent numbers agree more nearly 

 with the lower value. Thus, Noyes obtains 15 "896, and 

 Dittmar and Henderson give 15 866. 



I had intended further to elaborate and extend my 

 observations on the synthesis of water from weighed 

 quantities of oxygen and hydrogen, but the publication of 

 Prof. E. W. Morley's masterly researches upon the 

 "Volumetric Composition of Water" {Amer. Journ. 

 Set., March 1891) led me to the conclusion that the best 

 contribution that I could now make to the subject would 

 be by the further determination of the relative densities 

 of the two gases. The combination of this with the 

 number 2 0002,- obtained by Morley as the mean of 

 astonishingly concordant individual experiments, would 

 give a better result for the atomic weights than any I 

 could hope to obtain directly. 



In the present work two objects have been especially 

 kept in view. The first is simplicity upon the chemical 

 side, and the second the use of materials in such a form 

 that the elimination of impurities goes forward in the 

 normal working of the process. When, as in the foriner 

 determinations, the hydrogen is made from zinc, any im- 

 purity which that material may contain and communicate 

 to the gas cannot be eliminated from the generator ; for 

 each experiment brings into play a fresh quantity of zinc, 



' "On the ReU live Densities of Hydrogen and Oxygen. II." Abstract 

 of a paper by Lord Rayleigh, Sec.R.S., read at the Royal Society on 

 February 18, 1892. 



^ It should not be overlo ked that this numbir is difficult to reconcile with 

 views generally held as t.> the applicability of Avogadro's law to very rare 

 gases. From what we know of the behaviour of oxygen and hydrogen gases 

 under compression, it seems imp.-obable that volumes which are as a 0003 : i 

 under atmospheric conditions would remain as 2 : i upon indefinite expaiuion. 

 According to the formula of Van der Waals, a greater change than this in 

 the ratio of volumes is to be expected. 



