GASEOUS SPECTRA UNDER HIGH DISPERSION. 455 



lesiler-liiie and its precursor shadow-pair from any good specimen conjunction of them on your 

 rr sp, and to apply it successively to all the leader lines from near the "green giant " backwards 

 across the crossing-point and on into the thick of the melee that precedes it, right up to the first 

 edge of the band. The result was, to my joyful surprise that it accounts instantly, and in totofor 

 every single line of the band laid down on your map ! The band is simply two exactly similar 

 single-rank line progressions laid over each other displacing one of them slightly on the other ; 

 and while one consists of single strong lines, the other is formed of fainter, closely double ones. 

 You will see this by the enclosed card strips* along the top edge of one of which your map of the 

 band is exactly copied, while under it the members that compose the close double, or fainter series 

 are prolonged so as to produce a linelet progression by themselves. On another card the rest of 

 the band's linelets are figured, also in a single-ray progression of (in red ink) strong single lines ; by 

 applying this card with its left hand leader at No. 5 line of the natural delineation, you will 

 see that it includes all the lines not prolonged downwards or abstracted from that stripe to 

 form the partial stripe of duplex lines ; and by then shifting it leftwards till its beginning 

 coincides with ISTo. 1 line of the natural band, you will see too that it then exactly covers all 

 the duplex line series of the natural band. 



Besides the two constituent bands' precise resemblance to each other, there is also this 

 link of connection between them, that the ruler of the following band is not placed anywhere, 

 but ON No. 5 line of the foremost one. And again there is this simplicity about the single- 

 rank or partial bands themselves, that their intervals are quite distinctly an arithmetical 

 progression of spaces denoted by the series of natural numbers 1, 2, 3, 4, &c. I have plotted 

 in, under each (singular and duplex) portion of the band a true simple progression of this kind, 

 so that the eye can judge how nearly each of the two tributary bands satisfies it, and there is 

 no question, I think that it is, with some very slight disturbances here and there, the simple 

 rule of formation of them both ? Instead of being, therefore, a linelet band of the most 

 curiously involved complexity, as it at first sight looks by its " crossing " lines and close pack of 

 crowded lines near the front edge, the ruled CO green band is really the very simplest in its 

 mode of construction that I think has yet been met with in Spectroscopy ! The way in which 

 your sharp resolution of the two " crossings " lines themselves into a minute triplet and a 

 minute doublet respectively agrees with the conjunction is by itself a wonderful corroboration 

 of the structure. But without the clear and precise resolution of all its lines throughout with 

 the most accurate autographic measurement that you have effected, it would evidently have 

 been quite impossible to recognise and establish it in its microscopic mixture! A good 

 example of the powerful discrimination that you have used upon the band occurs in its very 

 first line, which I had noted " broad," only, in my little sketch of measurement, but which you 

 have mapped as a pair Nos. 1 and 2 of the band, just as accurately placed as the other equi- 

 vocal looking linelets of the band are all clearly and exactly broken up and divided into their 

 proper places in the dual band. 



The displacement between the band's two parts is 10 (1 + 2 + 3 + 4), unit spaces of the 

 structure, which is neither an indifferent interval nor an indifferent number of unit-intervals of 

 its structure; so that the two parts can't be described as two independent overlying bands 

 belonging possibly to two different gases. But yet the duality is singular, as if either sever- 



* These very ingenious card strips of Prof. Hekschel's, being unsuited to book-illustration,— it occurred to me, to 

 prepare Plate LXXVIIL, including them both and the manner of working them, but in one statical view. This Plate 

 afterwards had his approval, though with a proviso touching ideal accuracy, which he has touched on in Appendix II. 

 page 43.— C. P. S. 



VOL. XXXII. PART III. 4 D 



