l68 



NATURE 



[August i6, 1900 



little water, full of ordinary air, and provided with means of ex- 

 panding the air in the flask, and either returning the air to the 

 flask, or admitting filtered air. Go on repeating the process of 

 expanding and cloud-making in the flask. After this has been 

 done a number of times, the nuclei become fewer and fewer, 

 and at last only a very few are left in the air. Every one 

 must have noticed when making this experiment that the cloud 

 particles are very small on the first expansion, and that they 

 fall very slowly, almost imperceptibly, but that at the end of 

 the experiment, when the last dust particles become nuclei, 

 the water particles are large and fall rapidly like rain drops. 

 At the beginning of the experiment, with plenty of dust in the 

 air, there is almost no supersaturation, the nuclei being so close 

 the tension is relieved as soon as it is formed. When, how- 

 ever, only a few particles are present, there are large spaces 

 between the nuclei where supersaturation can take place, and it is 

 by falling through this supersaturated air that the drops, when 

 few in number, are able to grow so quickly and become so large. 

 It therefore seems probable that something of the same kind 

 will happen if ions were to become nuclei in supersaturated air. 

 Whenever an ion becomes active it will rapidly grow to the 

 dimensions of a rain-drop in the same manner and for the same 

 reason that the dust-nucleused drops do in supersaturated air. 

 These little drops evidently have a way of parting with the heat 

 of condensation at a very much quicker rate than Mr. Wilson is 

 disposed to admit. 



It is this capacity for rapid growth in supersaturated air that 

 makes it so improbable that ions can ever give rise to a cloudy 

 form of condensation. To form a cloud a large number of them 

 would require to become active at the same moment. But this 

 is evidently not possible in a rising column of air. The ions 

 which rise on the top of the ascending column will become 

 active first, and by falling through the lower supersaturated air 

 will grow with great rapidity and give rise to a rainy, but 

 cloudless form of condensation. 



There are some points connected with ions about which I 

 think the readers of Nature would be glad to have some 

 information, and which I think Mr. Wilson, with the aid of the 

 apparatus at his disposal, could give us. For instance, one 

 would like to know (i) how long ions remain in air in an 

 inclosed vessel, when both 4- and — ions are present ; (2) when 

 only -f or - ions are in the air ; (3) whether the presence of 

 dust has any effect on the duration of their life. For practical 

 purposes one would also like to know further (i) how many 

 ions are generally in the air near the ground ; (2) what amount 

 of electricity they carry with them. 



Finally, one would like to know how many ions will pass up 

 through a cloud and escape at the top ; as one would almost 

 expect, these ions, with their electric charges, will be more 

 likely to be cleared out of the air by rain than the dust particles, 

 and whether both kinds are equally liable to be washed out by 

 rain. If not, the inequality may help to explain some important 

 electrical phenomena. John Aitken. 



Ardenlea, Falkirk, June 27. 



The Melting Points of Rock-forming Minerals. 



In connection with the abstracts of papers read before the 

 Royal Dublin Society by Dr. J. Joly, F. R.S. , and myself, 

 given in Nature for July 12 (p. 262), I might perhaps be per- 

 mitted to draw attention to a few points. The same subject 

 has been recently dealt with by Mr. C. E. Stromeyer {Mem. 

 Manchester Lit. and Phil. Soc. , vol. xliv. Part iii. No. 7, 

 1900) and by Prof. Sollas, F.R.S. {Geol. Mag., July 1900). 



In the first place it may be noted that the " melting point " 

 of a substance under a definite pressure has a perfectly definite 

 meaning. The "softening point," on the other hand, obviously 

 depends on the magnitude of the distorting force with which the 

 softness is tested, as well as on the other conditions of experi- 

 ment. 



It is an established fact that the melting points of a very 

 large number of substances vary with the pressure. Bunsen, as 

 far back as 1850, perceived the geological application of this 

 phenomenon. In discussing the crystallisation of plutonic 

 rocks, it is the melting points of the minerals under enormous 

 pressures which really concern us. These pressures are 

 probably sufficient to alter the melting points through several 

 hundred degrees. There are then two ways open for us to 

 ascertain these melting points. Firstly, we might determine 

 them by direct experiment at the necessary large pressures ; or, 



NO. itC;, VOL. 62] 



secondly, we might measure the melting points at ordinary 

 atmospheric pressure and determine the rate of increase (or 

 decrease) of melting point with increase of pressure {dBjdp). 

 Considering the gigantic pressures with which we have to deal, 

 it seems decidedly easier to adopt the second method. The 

 agreement between the results obtained from the application of 

 the thermodynamic formula 



(ie_ e(v,- v,\ 

 dp L 



(where = absolute melting temperature; {vi-v,) = the change 

 of volume at the instant of melting; L = the latent heat in 

 mechanical units) with the results of experiments {e.g. M. A. 

 B&ltelW, /ourna/ de Phys., t. viii. p. 90, 1887), seems to justify 

 the application of that formula to the case of the minerals in 

 question, in the absence of direct experiment. It is true that 

 the formula was deduced for a reversible system, and that no 

 natural process is reversible. But a similar objection would 

 hold against the application of any theoretical formula to the 

 conditions obtainable in experimental work. In the present 

 case it is only claimed for the formula that it will afford an 

 approximate estimate of the melting points of minerals under 

 large pressures ; and after all, even direct measurement of such 

 high temperatures as are involved is always attended with un- 

 certainty. In order to apply this formula we require 0, {vi-v,), 

 and L. The melting points of the most important minerals at 

 atmospheric pressure have been determined by Dr. Joly and Mr. 

 R. Cusack {Proc. Roy. Irish Acad., Ser. 3. vol. ii. p. 38 ; vol. iv. 

 p. 399). A large part of the volume change on melting is, 

 I submit, afforded us by the difference in density between the 

 crystalline mineral and its fused glass. Now it is characteristic 

 of amorphous substances to pass gradually and continuously 

 from solid to liquid ((/ Preston, "Theory of Heat," pp. 270 

 and 286) ; and so it is highly probable that such a mineral glass 

 will pass without sudden volume change into the liquid state, 

 and it has, in fact, passed gradually in the inverse direction. It 

 is not contended that any given mineral ever existed as a glass in 

 the molten magma of an igneous rock, but only that it existed 

 as a liquid. 



In my paper, above referred to, I have shown how the 

 " fusibility " of a mineral must be connected with its latent 

 heat, and hence by a comparison of relative fusibility and 

 melting temperature we may often deduce the relative latent 

 heats of two minerals. Thus, for example, the "fusibility" of 

 labradorite is 3 on von Kobell's scale, and its melting point is 

 1229° C., whereas orthoclase has a melting point of only II75°C., 

 but is much less "fusible," viz. 5 on von Kobell's scale. 

 Hence I infer that the latent heat of orthoclase is decidedly 

 greater than that of labradorite. Similarly, the latent heat of 

 augite is less than that of orthoclase. But the volume-change 

 on melting of augite is greater than that of orthoclase. There- 

 fore dOjdp is greater for augite than for orthoclase. It is thus 

 possible to arrive at the order of melting points of minerals 

 under the pre.ssure of rock formation. If, after ascertaining 

 this order, it is still found to be inconsistent with the order of 

 crystallisation, as shown by microsopical examination, it may be 

 necessary to examine the more complicated influences of solu- 

 tion, &c., on the crystallising points of the minerals. 



In conclusion, I may point out that it must be a matter of 

 extreme importance in measuring the melting temperature of 

 quartz to make sure that the specimen used is pure, and in p:\r- 

 ticular free from the alkalis. Messrs. Shenstone and Lacell 

 (Nature, May 3, igcx), p. 20) have found that rock crystal 

 very often contains sodium and lithium, traces of which might 

 be expected to lower the melting point. Further, it has long 

 been known that quartz, with a density of 2 '66, passes into the 

 variety of silica with density 2*3 at a temperature below its 

 melting point {cf. Fremy, Eiic. Chiin. 6, p. 142). And similar 

 transformations are common among metals. Is it not possible 

 then that the phenomena observed by Dr. Joly may have nothing 

 to do with the fusion point of quartz, but are simply cases of 

 molecular transformation at a temperature below the melting 

 point? J. A. Cunningham. 



Royal College of Science, Dublin. 



Observation of the Circular Components in the 



" Faraday Effect." 



After repeated attempts to determine the nature of the 



" Faraday effect," I have succeeded in observing that ordinary 



light, when passing from a surface into a medium in such a way 



