Jan. 31, 1889] 



NA TURE 



319 



SnCl.,. Again, silica (as has been shown by Mendelejeff long 

 before Henry) certainly has a more complicated molecular 

 formula than SiO„, say wSiOj, otherwi^e it would be a ga?, 

 analogous to COg, for SiClj boils at 57° and CCI4 at 76°. 

 Hut would it be reasonable to say that the SiOa molecules are 

 held t( gether by the fifth valency of Si ? 



The only satisfactory answer to the above question is that in 

 molecules RoCIr the two KCI3 groups are held together by their 

 own— let us say residual— affinity. 



I think that the study of the question touched in the above 

 lines will greatly add to the development of our theoretical ideas 

 in chemistry, and deserves the careful attention of chemists. 



B. Brauner. 

 Bohemian University, Prague, January 8. 



In my letter, referred to by Dr. Brauner, I considered 

 only the molecular formulae of the chlorides, as indicated by the 

 experimental results. 



Dr. Brauner has now, in his interesting letter, enlarged the 

 field of discussion by introducing the question of the constitu- 

 tional formulae of these compounds, involving the vexed question 

 of the valency of the elements. Dr. Brauner points out the 

 importance of the proof of the trivalency of these metals in their 

 chlorides when in the gaseous state at high temperatures, and 

 he then discusses the valency of the metals when their chlorides 

 have the formula Ro^-Ig- 



Although I am inclined to believe that the views expressed 

 by Dr. Brauner may eventually prove to be correct, yet I cannot 

 help thinking that we are not yet in a position to speak with 

 certainty on one or two points. Dr. Brauner states that the 

 number of "bonds" by which the atoms of the metal in the 

 compounds RoClg are united is certainly not one, and again, 

 that Sn,Cl4 cannot be expressed by the formula Cl.,Sn"'— Sn"'Cl2. 

 r.ut is this quite certain? Indium, the metal next to tin in the 

 seventh horizontal series, is mono-, di-, and tri-valent in the 

 compounds InCl, InClg, InCls ; is it, then, quite unwarrantable 

 to assume that tin is di-, tri-, and tetra-valent in the compounds 

 SnCI,, Sn.,Cl4, SnCl4? 



As regards silica, «SiO.,, the properties of the compound 

 seem to indicate that « is a large number, and here, unless we 

 arrange the .*^i atoms in a ring, like the carbon atoms in benzene, 

 we must probably look on the compound as " molecular." 



It is difficult to find experimental evidence which bears 

 directly on the question ; perhaps such results as those ob- 

 tained by Dr. Hamsay and myself on the dynamical and statical 

 methods of measuring. vapour- pressures may throw some light on 

 the matter. We find that the two methods give identical results 

 with ammonium chloride, nitrogen peroxide, and acetic acid, 

 just as they do with all stable solids and liquids, but very 

 different results wiih all the other dissociating substances 

 examined. This appears to indicate some difference in the 

 molecular arrangement of the two groups of compounds, but it 

 is quite uncertain whether the difference is one between " mole- 

 cular" and "atomic" compounds, or between compounds 

 formed in a very simple manner, such as 20.2N = 02N — NO.^, 

 and those in which there is a breaking down of a stable mole- 



/II 

 cule, as m the case of chloral hydrate, CI3C— C^ -1- H.OH = 



,0H 



ci,c— ch: 



I am at present engaged in a study of the vnpotir pressures of 

 halogen compounds by both methods, and it is just possible 

 that some further light may be thrown on this question by the 

 results of the investigation. Sydney You.ng. 



University College, Bristol, January 15. 



Remarkable Rime and Mist. 

 You have a letter in Nature of January 17 (p. 270), signed 

 "Annie Ley," which induces me to add the following: — We 

 had here on January 6 an extraordinary rime formed at a tem- 

 perature varying between 2i°-5 and 25°7 (1° warmer on the 

 gr.iss than at 4 feet), the air being almost calm. This rime 

 increased in thickness and in length with the height above the 

 ground. The length measured of the deposit on a birch-tree 

 at 5 feet was \ of an inch; at 10 feet, i inch; at 15 feet. 



\\ inch; and at 25 feet, \\ inch. Tlie hoar was nearly 

 horizontal, pointing downwards at an angle of I5^ That 

 deposited on the grass, however, was perpendicular, rising with 

 a thin stem and having a large funnel-shaped head. Suddenly, 

 at 10 a.m. of the 7th (next morning), the whole of the rime 

 (still frozen) fell to the ground, and under the birch-tree of 30 feet 

 in height and 18 feet across (sparse of branches, and none for 

 10 feet), the fallen rime covered the ground to the depth of rather 

 more than 2 inches, and thi'^, when melted, yielded 0550 of an 

 inch of water (or i\ inches of rime to i of water). The rime 

 on the grass when carefully collected and melted yielded 0*033 

 of an inch. There was a dense mist whilst the rime was being 

 deposited. 



From this elevation (530 feet) we frequently look over 

 dense mists that cover the water of the Bristol Channel and 

 see the hills of Somerset, Devon, Monmouth, and Glamorgan. 

 On the 19th instant, with a hoar frost, over the Bristol Channel 

 was a dense dark mist apparently extending 100 feet into the 

 air. This mist, at 8.30 a.m., rapidly changed on its upper 

 surface to cirri clouds, and then to transparent vapour, and in 

 an hour the whole mist by this process had disappeared. Klse- 

 where the sky was cloudless. These mists of the Bristol 

 Channel change on their upper surface sometimes to cirro-strati, 

 sometimes to cumuli, and twice they have been known to change 

 to thunder-clouds during the last two years. The change to 

 cirri has only been seen once. E. J. Lowe. 



Shirenewton Hall, near Chepstow, January 22. 



Cercyonis alope and nephele. 



I.N his review (Naturp:, December 27, 1888, p. 193) of my 

 work on butterflies now in course of publication (in which I re- 

 ceive a far more generous treatment than I am accustomed to), 

 Captin Elwes thinks me illogical in holding to the probable 

 specific distinction of Cercyonis alope and nephele, and at the same 

 time the specific unity of the Eastern American forms of Cyaniris 

 pseudargiolus, hicia, violacea, and ncglecta ; and suggests as to 

 the former that climatic differences in the regions they re- 

 spectively occupy may have brought about the distinctions 

 noted. 



On general grounds, it seems in the highest degree probable 

 that climatic differences have had much to do with the origin of 

 the different forms in both cases, be they species formed or form- 

 ing. But surely Captain Elwes is confusing the judgment when 

 he fails to make a distinction between the successive seasonal 

 forms of a digoneutic butterfly, as in the case of Cy aniris and 

 de Niceville's Indian species to which he alludes, and the syn- 

 chronous variation of a monogoneutic species, like those (or that) 

 of Ceiryonis. SAMUEL H ScUDDER. 



Cambridge, U.S.A., January 10. 



MODERN VIEWS OF ELECTRICITY: 

 Part IV.— Radiation. 



Possible Accounts of the Faraday and Hall Effects. 



nPHE account I have given of the magnetic rotation of 

 ^ the plane of polarization has made it depend on 

 the phenomenon of hysteresis, in a way which may be 

 thus summarized. The value of yi for increasing mag- 

 netization is difiterent from that for decreasing magnetiza- 

 tion ; an electric displacement such as occurs in every 

 half-swing of a light-vibration is resolvable into two oppo- 

 site circular components, one of which increases, while 

 the other decreases, any magnetization already existing 

 in the direction of the ray : the value of p. affects the 

 speed of transmission of light ; hence the two circular 

 components will not proceed at the same pace, and the 

 direction of vibration will infinitesimally rotate. The 

 same thing is repeated at every half-swing, the elemental 

 rotations being all in the same sense, and so the ultimate 

 rotation of the plane of polarization in transparent bodies 

 is accounted for. 



' Continued from p, 13. 



