224 



NATURE 



\ynly 8, 1880 



length scale, between one part of the spectrum and 

 another, as compared with an average refraction or prism 

 representation, viz., some sixteen times, was found to defy 

 all accuracy by any ordinary pen or pencil, and to mislead 

 or confound the eye, as to the mere physiognomy of 

 groupings of the lines. Then, worse still, nature herself, 

 and spectrum-forming nature too, was being fought 

 n^ainst, in having scales increasing their numbers for 

 dispersion one way, when the prismatic deviations whicli 

 produced these dispersions were going the other way. 

 So at last it was determined that whatever the scale a 

 pure theorist may eventually prefer to put a few spectrum 

 places in at last, for his own purposes, the spectrum 

 observer, in order to observe well, quiclcly, and safely 

 throughout the whole spectrum must have : — 



1. A scale according to nature, as to the direction of 

 increase of its numbers. 



2. Increasing therefore these numbers from red to 

 violet, both because the prismatic deviations do the same, 

 and because, when the temperature of bodies is gradually 

 raised, from that of the air in which we live up to such 

 point that they begin to be luminous, the first light given 

 off is red ; and they only attain to violet light in the latest 

 and most e.vtreme degrees of heat eventually obtained. 



3. Red therefore being the natural beginning of the 

 spectrum, and all spectral numbers arranged as above, 

 increasing towards the rest of the spectrum, the said red 

 end requires to be placed on the left hand, so that every 

 spectrum map may be told off as all writing and printing 

 is made to read in all European countries, viz. from left 

 to right, never from right to left. 



4. Seeing that prisms will always be employed by some 

 observers of the solar spectrum, and gratings by others, 

 the scale to be used should be one whose general form, 

 in equal parts, should divide the immense difference of 

 physiognomy u'hich exists between the spectra offered by 

 these two instrumental methods; that is, not compressing 

 the red end so much as the prism does, nor compressing 

 the violet so much as the grating does ; and this end is 

 obtained most neatly, on an equally absolute foundation 

 with wave-lengths, and in a handy set of whole figures by 

 adopting the number of such waves to the inch, British. 



The above points having been all fairly arrived at, after 

 great sacrifices of both time and labour in the other direc- 

 tion, the Edinburgh experimenter proceeded without any 

 further compunction to alter his spectroscope once more, 

 and make it conform in all respects thereto, i.e., to show the 

 red end of spectra towards the left, and to increase 

 spectral readings from left to right ; while he further 

 applied new scales to his collection of spectrum maps in 

 terms of wave-numbers. And then came the reward ; 

 for not only did the same eye and pencil succeed in 

 applying a wave-number scale more accurately than a 

 w'ave-length one to prism-observed spectra, and make 

 the correspondences between prism and grating spectra 

 more numerous, perfect, and easily apprehensible, but 

 the wave-number scale was found more suited naturally 

 to the absolute requirements of the solar spectrum in 

 itself. Or thus, while the wave-length scale, as repre- 

 sented in Angstrom's grand normal solar, but diffraction 

 spectrum stretches out the red end to such a degree that 

 the lines there are so few and far between as to waste the 

 very paper on which they are drawn, the wave-number 

 method gently compresses them, or brings them twice 

 as close together ; while again, if at the violet end the 

 lines are so numerous, and closed packed in Angstrom's 

 map that they have hardly standing room, and can 

 scarcely be separated one from the other — the wave- 

 number method gives them twice as much space there, in 

 a map measuring, on the whole, from red to violet, only 

 the same length as Angstrom's. 



But there was a still higher reward to the experimenter, 

 who, adopting the scale of wave-number, and finding he 

 had more room for the violet end of the spectrum, began 



to pay more attention thereto ; for^ he then found that, 

 crowded as were the violet lines in Angstrom's diffraction 

 map, they were not half crowded enough ; or rather that 

 there were really in that part of the solar spectrum three 

 or four times as many more lines still ; far more indeed • 

 than could have been inserted on the engraved plates of 

 the Swedish philosopher, and many more than his diftrac- 

 tion grating was probably able to show. While therefore 

 all strong lines throughout Angstrom's map are believed 

 to have been most admirably measured, and the far more 

 numerous thin lines are also most truthfully rendered in 

 the earlier and middle parts of the spectrum — the violet 

 termination, what with the imperfect showings of his 

 grating, and the contracted space of the wave-length 

 scale map, has not been done justice to. 



Yet this is a very material point in the physics of the 

 sun ; for according to the preponderance of violet, over 

 red, light, so may be assumed the intensity of the tem- 

 perature of that light's origin. Whereabouts then did 

 the increased number of lines in the violet observed by 

 the Edinburgh experimenter with his prisms, over Ang- 

 strom with his grating, place the photosphere of the sun 

 as to temperature ? 



This point, described by the experimenter in the 

 Transactions of the Royal Society, Edinburgh, vol. xxix., 

 for 1S79, '^^s approximated to by him in this manner : — 

 Having collected from various sources several thousands 

 of spectrum place observations, he reduced them all to- 

 wave-number scale, and then arranged them according to 

 the temperatures of their sources of origin, or, as Mr. 

 Norman Lockyerhas since then termed it, their respective 

 "heat-levels," and the following series was obtained : — 



Source of origin of spectral light, when at freezing 

 point as in telluric absorption fpectra, has its 



maximum of lines at W.-N. place =39,000 



Chamber absorption spectra at temp. 68° F. at ... 41,000 



Flame lines at lamp-flame temperatures at 47,000 



Gas-vacuum tubes illuminated by i inch induction 



spark 49,oco 



Chemical lines in 2 inch sparks 49,000 



Chemical lines in 6 to 10 inch sparks intensified ... 51,000- 



And Angstrom's diffraction solar spectrum S5'°°° 



But the solar spectrum, as observed on this occasion in 

 Portugal, showed its maximum of lines at 61,000 of the 

 same scale ; or indicated that the temperature of the solar 

 photosphere may be as much above the highest temperature 

 yet attained by man, even with assistance of electricity 

 in its condensed form, as that is above the freezing tem- 

 perature of the upper strata of the earth's atmosphere. 



Lastly, Gaseous .Spectra. — Under this term are included 

 both flames, especially blow-pipe flames, in the open air ; 

 and electric illuminations inside so-called gas-vacuum 

 tubes, such as those of Geissler and Plucker combined. 

 But in all these cases the experimenter, finding that 

 faintness of the light was the crying evil, changed the 

 usual transverse inethod of looking at lines, or cones, of 

 light, for an end-on view of the same. 



Trying this first for the blow-pipe, whose flame of coal- 

 gas urged by a stream of air could then, by a collimating 

 objective applied to the anterior telescope, be safely 

 looked into, though directed right towards the slit— the 

 increased number of lines, their steadiness and definiteness 

 in all the several hydro-carbon bands — and then the 

 resolving of the mere haze in the field of view into closely 

 ranked little lines or linelets, proved an inimitable reward, 

 as well as a priceless source of the best kind of reference- 

 data in all his subsequent inquiries ; especially too because 

 these advanced results were procured without increasing 

 either the temperature, or size, or combustion material of 

 the flame at all. 



Next applying the same principle to the Geissler- 

 Plucker tubes, by having their form modified by M. 



