514 



NATURE 



yApril 3, 1690 



of paper, is apt to lose its purity and acquire confusing lines 

 nnd bands of ordinary copper-chloride, by oxidation, which the 

 preparation then undergoes spontaneously, before igniting it. 

 For pyrotechnists, therefore, it seems scarcely probable that the 

 subchloride of copper, with its pure cerulean flame, will ever 

 be of any very useful valae. Bat as a parallel example of a 

 coloured-fire composition, it may be mentioned here, that 

 powdered V.al Traversite (a bituminous linestone found near 

 Neuchatel, in Switzerland), on account of its prodigious natural 

 richness ia bitumen, when mixed with sufficient chlorate of 

 ])otash, also burns self-supportingly, with a fine orange-red 

 flame in which the familiar spectrum of calcic oxide is, of 

 course, most vivid. Were hot asphalt, pitch, or bitumen, instead 

 of hot glycerine, used to dissolve or to "masticate" the dry sub- 

 chloride of copper when it is freshly made, a copper-chlorinated 

 mass would be produced which would probably be capable of 

 resisting atmospheric action, and whose mixture with chlorate of 

 potash would, like the similar Val-Traversite mixture, probably 

 also not suffer by keejiing and exposure, and would furnish a 

 source of blue flame and of the significantly simple spec rum of 

 subchloride of copper, not less vividly true and fixed in their 

 distinctness, than the orange.-red light and strongly pronounced 

 calcic-oxide spectrum of the other combination of chlorate of 

 potash with a bituuien-contnining substance. 



As regards the blue salt-flaine, whose spectrum in its purity 

 shows no conspicuous lines, or bands of greatest brightness, 

 it can hardly be doubted that the element chlorine, from the 

 positions of its own principal line groups, contributes mainly to 

 produce the blue color.uion, at a temperature, in the fire, which 

 is not high enough to dissoci'te the sodic chloride and liberate 

 sodium vapour, with its tell-tale yellow line, from its chemical 

 union. In the green flame of chloride of copper the colour- 

 ing groups of lines show a more detailed resemblance than this 

 to the chief colorific lines in the elementary chlorine spectrum,' 

 while in copper subchloiide's "bluest of blue" flames, the 

 wide green light-bands of copper chloride fade out, leaving the 

 colorific light concentrated almost entirely in three close pairs, 

 or in six bright lines, which, if they do not coincide in place 

 with, are at least not far distant in position from, three chief 



' A very suggestive example of a substance's detect! m by rec gnition of 

 its spectrum was described, with a drawm^; of 'he rec Tied spectra, by Mr. 

 A. Percy Smith, in a short notice uf a series i>f observations on the sp-ictra 

 of chlorides, and on the h\\xt fla ne . f c imm n sa't. in the Chemical News, 

 V )1. 39, p. 14' (1S79). An examination of the flime-spectra ^f several 

 different chlorides, e .abled the author of that noti e to recognize a comnon 

 similarity am >ng them a'l to the spark- or flame-spectnim > f hydroch'oric 

 acid gas. This gas showed a belt of green line-ban Is which ajrerd in ih-ir 

 man positions with the green portio.i of a long array of ban l-pair shown 

 with much constancy by several different alka in- and earthy chlorides, and 

 especially by ammonium chio ide, and by merciirous chloride (or ca'omel. 

 where tlie agreement was also verified by a direct c imparison), in a Bun.sen 

 flame ; but no lin--c lunterpar s to the equally bright, blue-lined portion of 

 the same constant spectral striaiion were obs-^rvable in the hydrochi ric acid 

 spectrum. 



From the easy conveision of ch'orides into the correspinding oxides in a 1 

 air-gas flame, when the flam: is not kept artific ally saturated with hydro- 

 chloric acid gas, we might prety certainly assume that in the flame's 

 ordinary condition, the heated chl irides would always di^eng ge suffic.ent 

 chlorine to produce by comb na ion wiih hydr gen in tlie coal-gas of the 

 flame, traces of the stable product, hydrochloric acid gas. among the gaes 

 of the flame's combustion ; and the different chlorides would thus, by sup- 

 positions whic'n may not perhaps be unlike y an i ina imis.sib e, all supply the 

 flame alike with the ^ub.stantlal fac.or needed, for tlie appearance of the 

 green line portion of the constant .spe trum. 



At the same time new carbon-compounds w. u'd be formed by dehydration 

 of the flame's gaseous hydrocarbons, to furnish hydrogen 10 the liberated 

 chlorine, and ■■ome constant carbon-gases then, of 11 t yet known descrip- 

 tions, might be conjecturel just as comprehens'blv and fi ly, 10 be c n- 

 currently productive in the constant chloride-r.uik's illumination, of the blue- 

 line porti Ml of its bands, • f which no spectral counterparts could be detected 

 in the hydrochloric aCid sjectruni. 



But whether ihe interestii'u figure and decrlption given by Mr. A. Percy 

 Sm'th in the ab ve paper, of his long series of experiment-, may or may not 

 admit of such a simple spectro ch-m cal int rpre.ation the conflicts <f con- 

 tending chemiral affinities of wh.ch the spectro.scopic recognition ot hydro- 

 chi .ric acid in flames fed with d ff rcn. chlorides lurnLshes such a wonderful 

 example, give weight and va'ue to the notes of the discovery recorded by 

 Mr. A. Percy S uith, in a ne*- wide field T the .spectroscope's utility, wh ch 

 areof much deep=rinteres' tban any single theory to account only f jr th.s 

 particular reiogn't'on and discovery itself. 



Mr. A. Percy Smth'sowii c ipitally based, and clearly prjved deductions 

 from his numerous expert. ne. its, were acco-dingly. in prospect of their 

 further pro.ecuti jn, expressed thus, quite ge.ierally :— that the blue flime of 

 common salt in a hot fire owe; its col ritio 1 to reactions eiiher exactly or 

 very neai-ly sim lar to those which produce resemblance uf a nearly constant 

 spectral type in different chloride flames, to that of iiydrochljric acid ; and 

 that, aga n, among the partly undetermined, an i perhaps t) some ex;ent 

 variable re<ctims which pr iduce tne similar.ty, there appear to be .so ne 

 which disturb and mo tify th; odiaa'y app-.aranc; of ih; h .'drochloric 

 acid spectrum, and wh.ch would appear t> superadd to it a series of 

 blue line-bands which, as it is presented in a fl ime, or elec rically in vacuum 

 tii'^ps. the spectrum of pure hydrochloric acid gas a^one does not usuilly 

 exhibit. 



line-pairs in the ordinary spectrum of chloride of copper. There 

 is much in these resemblances which betokens some kind of 

 continuity of connection with the primary features of the chlorine 

 spectrum itself; the evidences of "hich, although thus displayed 

 by cop er and chlorine in the spectroscope, may perhaps be 

 sensibly regarded as having some near relation of analogy to the 

 appearance of variable chemical combining power under the in- 

 fluence of light, between silver and chlorine, presented in 

 photography. But there is also, undoubtedly, a very marked 

 distinctioT between the "spectroscopic reactions" of these two 

 different copper chlorides ; and, similarly, there are in the 

 apparently mutable photochemical affinity between silver and 

 chlorine in photography, two fairly stable delimitations of its 

 range, in the "subchloride" (or as it has been termed by Mr. 

 Clement Lea, the " photochloride ") of silver, and inordinary 

 silver-chloride. Further discriminations of the copper-chloride 

 spectra in intermediate forms which they seem to comprise 

 transitionally between the two definite ones of the chloride anrl 

 subchloride, would perhaps extend and strengthen this analogy, 

 and may not impossibly help, at some future time, to explain and 

 illustrate it, if there is any real soundness in it, more fully and 

 completely. 



The example of fluoride of calcium is a curious one in spectrum 

 analysis, where s])rinkling fluor-spar dust in a Bunsen-flame 

 produces, in addition to the normal calcic oxide spectrum of one 

 orange red and one green band, a second bright and narrow 

 green one at a distance from the first about equal to that of the 

 red band from it. There are no other distinguishable bands. 

 But if the pair of normal ones is reallv due to calcium-oxide 

 vapour produced by decomposition in the flame, it is not very 

 easy to conjecture to what other product of decompositi )n the 

 adclitional, sharj^ly defined and brilliant, solitary green band can 

 be ascribed. The spectrum of hydroflunsilicic acid gas presents 

 a very gorgeous band-array of violet-blue lines, whose lustrous 

 group is pr. ibably indicaive of near neighbourhood in place to 

 some bright line concentration in the spec rum of fluorine itself; 

 but if so, the collection of its colorific strength in the single 

 additional green line of the fluor-spar spectrum, seems to imply 

 a freedom from uniformity in fluorine's power of imparting 

 spectral coloration to its C'unpounds, jut opposite to the sensible 

 continuity and kinship o{ spectral clusterings, above described, 

 which the presence of chlorine appears to impose upon its com- 

 pounds by common resemblances discernible in the blue light- 

 ascendencies of the fire-flames of common salt, chloride and 

 subchloride of copper, when they are spe^trosc<)]:)ically analyzed. 



A. S. Hei-ischel. 



Observatory House, Slough, March 3. 



Brush-Turkeys on the Smaller Islands North of 

 Celebes. 

 The reviewer of Dr. Hickson's book, "A Naturalist in North 

 Celebes" (March 20, p. 458), believes that the brush-turkey or 

 moleo, Megacephalon inalto, has never been recorded as occurring 

 in the smaller islands north of Celebes. I beg to remark that in 

 the year 1879 I recorded this species from Siao, and in the year 

 18S4 from Great Sangi, on botli of which islands, besides, occuis 

 a Alegapodhis peculiar to them, viz. M. sanghirens's, Schlegel, 

 representing there iVl. gUberti, Gray, from Celebes (see the Ibis, 

 1879, p. 139 ; his, 1884, pp. 6 and 53, &c.). Perhaps Mr. 

 GuiHemard did not c mprise Siao and Great Sangi under the 

 head of "smaller islands," but Dr. Hickson himsell ([>. 95) re- 

 cords two brush-turkeys from the smaller island of Tagulanclang, 

 a larger and a smaller one, and these must be Aiegaccphalon 

 maleo and a Megapodiiis, Tagulandang is situated between 

 Celebes and Siao, and much nearer t > the latter island. F'rom 

 the volcano islet of Ruang, o|jposi;e and within about a mile 

 from Tagulandang, he only records (p. 41) one brush-turkey, and 

 this, of course, may be either the Megacephalon or a Megapodiiis, 

 if both do not occur, as appears rather probable. When I 

 visited Ruang in 1871 after the heavy eruption in March of that 

 year (see Nature, vol. iv. p. 286), ne.irly the whole of its 

 forest was destroyed and burnt down, and I do not believe 

 that a living brush-turkey then remained on the islet ; but it has 

 since been repeopled froin its near neighbour, Tagulandang, where 

 both species occur, and therefore, if the one could reach Ruang,, 

 the other may have reached it loo. This is of no consequence at 

 all. Dr. Hickson's follov\ing remark as to brush-turkeys on 

 Tagulandang (p. 95), "The larger bird is perhaps the Megapo- 

 dius sanghirctnis of Schlegel, a brush turkey, which is l>igger 

 than the AJegarepIia'on, and extends over the Sangir Islands," 

 contains a mistake, as M. sanghirensis is much smaller than 



