404 



NATURE 



IFed. 28, I i 



air as nuclei is e-iseotial to the condensation of aqueous vapour, 

 it is by no means i nprobahle that ice may l)e associated with 

 these phenomena. For, as these lofty regions mus', even within 

 the tropics, be far above the plane consiitutin^ the lower 

 boundary of the term of perpetual congelation, the condensed 

 vapour must necessarily assume ihe form of aggregations of ice 

 around these nuclei. Hence the diffractive corona; may be 

 associated with imperfectly developed ice-crystal halos. 



It seems to me scarcely necessary to invoke — as Mr. Rowell 

 has done (Nature, vol. xxix. p. 251) — the repulsive agency of 

 electricity to account for the per^istent suspension of the volcanic 

 dust, even in these region; of rarefied air. If the attenuation be 

 sufticienily great, there will bi no sensible .subsidence of the 

 dust-particles. Faradsiy found that even metallic gold, when 

 minutely divided, required months to suliside when suspended in 

 water; an 1 some form; of insjlu le mineral matter remain 

 suspended in water for an almost indefinite period. Now, the 

 du-t-particles constituting the nuclei of condensation for fogs and 

 clouds are a.hs,o\Me\y iilira-miiroscopic in smallness ; hence their 

 suspension, even in rarefied air, may be prolon ed almost inde- 

 finitely. Moreover, it is po.ssible that air may posse-s some 

 degree of viscosity; in which case the indefinitely attenuated 

 dust-particles mi^ht have no tendency to subside, and could only 

 be removed from the atmosphere by those meteorological 

 agencies, — such as the condensation of vapour, — which tend to 

 augment their size. 



Mr. D. Wetterhan (Nature, vol. xxix. p. 250) refers to Mr. 

 Kes-elmeyer's hypothesis of the atmospheric origin of meteorites 

 put forth some twenty years ago, which ascribes them to the 

 condensation of metallic and other vapours issued from vol- 

 canoes. If I am not mistaken this hypothesis was advanced by 

 Biot near the beginning of the present century. The hii^/i 

 velocities of meteorites is overvvhelminglyyj;^a/ to their teyrestria/ 

 oriein. John Le Co.nite 



Berkeley, California, February i 



The recent sunsets were nearly or quite as remarkable in the 

 Rocky Mountain region "S they were in Europe, and the 

 phenomena were very similar. There M'as the same peculiar 

 fire-red after-glow continuing for two hours after sunset, &c. 

 ThciC unusual appearances be^an to attract attention soon after 

 the middle of November. They were most brilliant during the 

 last week of November, but continued at intervals until early in 

 January. The carefully kept meteorological record of Prof. 

 F. H. Loud, of Color.ado College, shows that the atmospheric 

 pressure varied considerably during the latter part of November, 

 but there was no apparent accompanying change in the after-glow. 

 The sunrises were aUo quite briUiant, but leis so than the sunsets. 

 Late in November I began to observe the wide chromatic belt 

 w Inch surrounded the sun, and at midday usually reached from 

 near the sun to the horizon. Somewhat similar appearances 

 and chromatic halos are not uncommon here, and it was not until 

 after several weeks of comparison of colours that I became con- 

 vinced that the tints seen around the sun during the time of the 

 remarkable sunsets were somewhat different from those ordin.arily 

 seen. By degrees the brick, or fire-red, and other abnormal 

 tints of the Hvilight hours have given place to the ordinary 

 prismatic colours, and a similar but les marked change could be 

 seen in the colours observed near the sun during the daytime. 

 These day colours were brightest when the sky was overcast 

 with thin clouds or filmy cirri, though plainly visible when there 

 was no cloud to be seen. The prevailing day tint is usually a 

 peculiar dull purple, but during the time of the red after-glow 

 the common colour w as duller, more like a yellowish brick-dust. 



Colorado College, February S G. H. Stone 



"Probable Nature of the Internal Symmetry of Crystals" 



In reply to the important criticisms offered by Herr L. Sohncke 

 on my new theory puljlished in NATURE of December 20 and 27, 

 18S3 (pp. 1S6 and 205) — 



Taking first those relating to the geometry of the subject ; 

 the following explains why only the five symmetrical arrange- 

 ments of points in space described in my paper are taken as the 

 basis of the theory. 



If it is the case that, prior to the act of crystallisation, the 

 cheinioal atoms of a body fall into some symmetrical arrange- 

 ment, it is n.atural to suppis? that they do so through some 



influence they exert on one another — such, for example, as mutual 

 repul-ion — and that a similar influence is exerted by each atom 

 of the same kind on atoms around it. And if this be so, there 

 wi'l be no stiMe equilibrium of the forces thus exerted until the 

 atoms are vtry eiienly diitribuled t/iroug/iout the space allotted to 

 tliein. 



Now although, as Herr Sohncke has shown, there is a large 

 variety of symmetrical arrangements of points in space in which 

 the points are disposed around every one point of the system in 

 jjrecisely the same manner as around every other, it would 

 appeir that only four of the-e regular systems, the first four 

 described in my paper, signally fultil the re')uirenient of even 

 distribution, these four systems bein' distinguished from all the 

 rest by the proptrty that, if the nearest points i^roitpeU around any 

 point of either of these four systtjiis are joined ^ the solid thus out- 

 lined has its edoes all equal. 



And further, although the fifth system described in my paper 

 is n t one of Herr Sohncke's regular systems, its points are 

 more evenly distributed through space than those of any of these 

 systeurs except the four just referred to. In this system the 

 property is found that either lines joining the nearest points 

 around any point of the system, "r lines joining the next nearest, 

 in all cases outline a solid whose edges are all equal. 



As the five systems I have in my paper too vaguely distin- 

 guished as " very symmetrical " thus stand alone, and moreover, 

 if my views are adopted, they appear to be adecjuate to all cases 

 of crystallisation, I still incline to think that the chemical atoms 

 of bodies about to crystallise always have one or other of these 

 five kinds of symmetrical arrangement. If I am wr^mg in this, 

 and some other symmetrical arrangements are admissible, the 

 general lines of the new theory will not however be affected. 



Next, as to the bearing of the theory on chemical valency and 

 the usual conception of a chemical molecule, it may be remarked 

 that, while there is no clear knowledge of the nature of the union 

 betueen the different sorts of atoms in a compound by which to 

 test the new theory, this theory appears to receive support from 

 the phenomenon of electrolysis. For the fact that one ion is 

 liberated at one pole, the other at the other, while no apparent 

 altcialion takes place in the fluid between the poles, goes to show 

 that any particular atom can change its partners without dissolving 

 the chemical ties subsi-ting between the several atoms of the 

 comi>ound, and thus favours the view that similar atoms equally 

 near to a particular atom are similarly related to it. 



As to my supposition that the expansion, or contraction, 

 occurring in the act of crystallisation, is due to the increa.sed or 

 diminished repulsion exerted by some only of the atoms of a body 

 on surrounding atoms, it i^, perhaps, interesting to notice that if 

 this conception could be extended to the gaseous state, and 

 the expansion to the state of gas of any compound attributed 

 to the agency of certain atoms in each molecule, or ideal unit, to 

 the exclusion of the rest, the simple relations found subsisting 

 betueen the volumes of compounds and the volumes of their 

 uncoinbined constituents might in this way be accounted for : — 

 Thus the fact that aqueous vapour has a volume two-thirds that 

 of the added volumes of the hydrogen and oxygen of which it is 

 composed would be explained if all the gaseous expansion of this 

 compound is due to the hydrogen atoms only. 



Muswell Hill Wm. Barlow 



"Mental Evolution in Animals" 



Mr. Faraday does not seem to have quite understood one 

 point in my comment on his letter. I said that whether the 

 action of the skate was accidental or designed, " in either case, 

 under the conditions, and more especially the ' attitude ' de- 

 scribed, seizure of the food at the proper moment can only be 

 ascribed to the sense of smell." When we remember the form 

 of a skate, it is certain that, under the conditions described, the 

 animal could not see the approaching food, and therefore Mr. 

 Faraday's illustration from the cricketer would only hold if the 

 cricketer continued to hit the ball after he had been blindfolded. 



I do not care to continue this ditcussion ; but I may say that 

 as the glass wall of a tank is not an object upon the solidity of 

 \\hich a skate would be likely to calculate, and as the sense of 

 smell in this animal is so highly developed that it might easily 

 give rise to "the appearance of co-ordination" described, I still 

 "think that the incident was probably accidental. Any other 

 piece of food happening to approach the mouth would no doubt 

 have been seized in just the same way. 



George J. Romanes 



