twilight and phosphorescence in many seas, and then sent a second paper to the Royal Astronomical Society. And that one they did print in the same year, with a very good engraving, too, of my several pairs of spectra just as they were simultaneously seen in my spectroscope-one member of each such pair being, for reference sake, invariably the bands and lines of the carbon- or hydrocarbon-spectrum, precisely as Dr. Watts is now proposing it in 1874, and as, too, I have since found that it was used before my time in telespectroscopy by the eminent Lewis M. Rutherford, of New York, and differing only in the manner of producing the spectrum from what Dr. Huggins actually accomplished on the comet of 1866 and on several others since then.

Truly, therefore, I have nothing to say against the principle of the method proposed now by Dr. Watts as his own; but I am sorry to differ from him in the instrumental description which he gives of the visible spectrum concerned, and which will affect the observations of all persons who use it as a standard of reference.

Mensuration of the Spectrum referred to.

The Doctor remarks, sagely enough, that for cometary work the reference-spectrum should be of "feeble intensity;" why then does he not examine it in that shape, viz. as given by the blue base of the flame of a small alcohol lamp, or the all but vanishing globule of flame when a common gas-light is on the point of going out from inanition?

But in place of so doing, he takes brilliant olefiant gas and burns it with oxygen "at the platinum nozzle of an oxyhydrogen blowpipe."

That in itself is quite enough to introduce some variations in the spectrum of a volatile chemical compound; and he gets still more when he examines it, not under the small dispersion usually employed on comets, but with the tremendous dispersion of six dense flint prisms. For to such an overpowerful instrument the appearance of the several "bands" (which is the characteristic feature under small dispersion) is almost annihilated, being replaced by more or fewer of the successive and almost isolated lines of which those bands are ultimately composed in nature, but which fine lines no one has ever seen, or need expect to see, in the spectrum of any ordinary comet.

The first step, therefore, in the Doctor's instructions to astronomers might well have been on the "bands," and should have been very complete with them; what will not, therefore, be any one's surprise to find that though the spectrum in question consists notably and notoriously (even when prepared from

sources of most feeble intensity) of five bands, viz. the orange, citron, green, blue, and violet, yet Dr. Watts, though from his intense (and even oxygen-excited) spectrum, indicates the existence of only three out of those five.

These three, which are really the citron, the green, and the blue, he names y, 8, and e rather confusingly; for in such case any one will ask, "where are the a and B of the system?" They are nowhere in the Doctor's paper of 1874; and even in his book of 1872, a very useful little book (The Index of Spectra'), and where a does appear as the orange group, there is no ẞ, and the violet band is most annoyingly called, among so many Greek letters, by the italic h, and there are many other lines only seen when oxygen and various other elements are used extravagantly. But letters, whether Greek or Roman, are sadly misplaced in their application to coloured "bands." They are not needed, because the names of the colours describe such bands, and their places, better; and Greek letters have been already appropriated to the bright lines of luminous and discontinuous spectra, in whose range they appear then consistently with the application of the same Greek letters to the brighter stars in any constellation, but never as a designation of a whole constellation of stars.

To make this point (of what are the bands and what are the lines) clear, I beg to introduce my present working diagram of the spectrum as a reference in astronomy (Plate I.); and while it will be seen that the "bands " are distinguished by the names of their colours, the lines in each band are distinguished by the numbers of their wave-lengths, precisely as Dr. Watts distinguishes them in his new paper of 1874; while they are also distinguished further in nature, and my drawings, by the brightest of each band being on the less-refrangible side, and going off gradually into invisibility on the more-refrangible side.

(And here I crave permission to write a rather long parenthesis, on a tangle needlessly introduced into the printed accounts of this spectrum, thus. As there is no allusion in the paper of 1874 to the Doctor's more important work of 1872 being cancelled, it remains as an authority, a competing authority, too, with the paper; whence, on putting the two side by side this singular discrepancy of terms comes out--viz. that the book calls "groups" what the paper calls "bands," the book again calling "bands" what the paper calls "lines," while finally the book calls "lines" some finer and intercalated modifications still, not mentioned in the paper at all. But the book does not venture to give the wave-length places of these extra-fine and interstitial lines belonging to it alone; and as they are certainly not

to be seen with all ordinary and some extraordinary forms of spectroscopes in any usual and convenient formula for preparing the carbohydrogen-spectrum of a moderate intensity, I propose not to take any further account of them now, except to say that, if I ever do meet them in future spectroscopy, I shall probably call them "linelets," or some such diminutive of lines, leaving therefore our present arrangement, and also Dr. Watts's present-paper arrangement, of bands and lines intact.)

Of these last most proper lines, then (in any ordinary spectroscope with a fine slit), let us speak now touching their appearance in Dr. Watts's new-paper account of them for 1874. They are given there as though they had never been observed or measured by any one else previously; and no one is entitled to object much to that, if they are now set forth in a better and completer manner (especially for the practical use of astronomers) than in all former accounts of them. Is that, however, the case?

No! I am sorry to say; for while thirteen distinct lines are recorded, there is not the slightest indication as to their being of any but equal visibility; yet one of them is an actual Sirius for brilliance, the very brightest of all the 1st-magnitude lines or stars, and others are as faint as 15th- and 20th-magnitude stars, seldom seen by any one. Yet of all that he has actually seen, I doubt not that Dr. Watts's measured wave-length places are always respectable for accuracy, though not perfect; and I can quite enter into his statement that the best-determined of them all are the lines 5165.5 and 5585.5.

Why or how that result came about is not explained by him; I will therefore proffer two reasons of my own, which will not decrease the interest or importance of the measures.

(1) The line 5165.5 is the first of the green band, the brightest line of the whole carbohydrogen-spectrum, the one which should have been decorated with an a; and it was much enlarged on, no less than nineteen years ago, by Professor Swan for its extreme beauty, its brilliance, its definition, and its capacity for being accurately measured. And

о

(2) The line 5585 5, not the first, but, strange to say, the second of the citron band, was found in 1870 by Piazzi Smyth to be so almost exactly coincident with the chief, the almost only, the Angström-discovered aurora-line, that he has recommended it repeatedly since then to all aurora-observers as affording a ready and instantaneous eye-proof whether there is any variety of spectroscopic character in the chief part of the light of successive auroras-he, too, having found none in nearly thirty auroras spread through two or three years, and embracing the widest external variations from arcs of bland light to needleshaped shooting rays, and from pale yellow-green or citron

colour up to almost any variety of red*. During the present year, indeed, Piazzi Smyth has been informed by Mr. Rand Capron, of Guildford, Surrey, that the coincidence is never quite perfect, the aurora-line being slightly more refrangible than the second line of the citron band; and Piazzi Smyth is quite ready to bow before this correction from the superior spectroscope with much larger dispersion, and the evident practical ability of Mr. Capron. But the correction still leaves the reference-use of the line 5585-5, though the second of the band (which gradually goes off into invisibility), more important than the first-and just as accurate as before, if each observer will remember that a normal aurora-line reads on the comb-like natural scale of the citron band of the carbohydrogen-spectrum not 20, but 2.2-the wave-lengths of all the lines in that interesting band reading thus, according to various authorities:

Lines.

Wave-length according to

Dr. Watts,

BAND

Compara

ness as

side.

Swan, Swan, in of carbo- Numbered tive bright-Angström, 1871, by hydrogen- from lessLielegg,&c. projection spectrum. refrangible with star collected of his anmagni- by Piazzi gular mea

in his

с

in his Index of

Spectra,' Magazine

for Novem

paper in the Philo

sophical

How, therefore, Dr. Watts came in 1874 to single out the second line of the citron band for super-accurate measurement of place by mere chance, and without knowing any thing more than appears in his paper of my having used it for years as a standard reference for an almost exactly coincident cosmical phenomenon in the night sky, is a strange problem. But its due investigation I must leave untouched on now, as there are other and more world-wide important matters to be discussed in the Doctor's last paragraph, where he states with no little positiveness both what the visible spectrum concerned is chemically the spectrum of, and of what it is not.

When the red colour is decided to the eye, a red line appears in the spectrum with wave-length 6290 nearly; but the 5585 citron-line is still always in its due place, and always brighter than the red line, even in the very reddest parts of the aurora.

to be seen with all ordinary and some extraordinary forms of spe troscopes in any usual and convenient formula for preparing th carbohydrogen-spectrum of a moderate intensity, I propose no to take any further account of them now, except to say that, if ever do meet them in future spectroscopy, I shall probably ca them "linelets," or some such diminutive of lines, leaving ther fore our present arrangement, and also Dr. Watts's present-pap arrangement, of bands and lines intact.)

Of these last most proper lines, then (in any ordinary spectro scope with a fine slit), let us speak now touching their appear ance in Dr. Watts's new-paper account of them for 1874. The are given there as though they had never been observed or mea sured by any one else previously; and no one is entitled to objec much to that, if they are now set forth in a better and complete manner (especially for the practical use of astronomers) tha in all former accounts of them. Is that, however, the case?

No! I am sorry to say; for while thirteen distinct lines ar recorded, there is not the slightest indication as to their bein of any but equal visibility; yet one of them is an actual Siriu for brilliance, the very brightest of all the 1st-magnitude lines o stars, and others are as faint as 15th- and 20th-magnitude stars seldom seen by any one. Yet of all that he has actually seen I doubt not that Dr. Watts's measured wave-length places ar always respectable for accuracy, though not perfect; and I can quite enter into his statement that the best-determined of then all are the lines 5165.5 and 5585.5.

Why or how that result came about is not explained by him I will therefore proffer two reasons of my own, which will no decrease the interest or importance of the measures.

(1) The line 5165.5 is the first of the green band, the bright est line of the whole carbohydrogen-spectrum, the one which should have been decorated with an a; and it was much enlarge on, no less than nineteen years ago, by Professor Swan for it extreme beauty, its brilliance, its definition, and its capacity fo being accurately measured. And

(2) The line 5585.5, not the first, but, strange to say, th second of the citron band, was found in 1870 by Piazzi Smytl to be so almost exactly coincident with the chief, the almos only, the Angström-discovered aurora-line, that he has recom mended it repeatedly since then to all aurora-observers as afford ing a ready and instantaneous eye-proof whether there is any variety of spectroscopic character in the chief part of the light of successive auroras-he, too, having found none in nearly thirty auroras spread through two or three years, and embracing the widest external variations from ares of bland light to needleshaped shooting rays, and from pale yellow-green or citron