The telescope serves for purposes of localisation, the spectroscope for purposes of analysis. The novel principle thus introduced that of investigating the nature of self-luminous bodies through the quality of their radiations goes very deep, and leads very far; it has implications still dim, and consequences still remote; its horizons are indeed illimitable; and they drew onward from the first all the more adventurous spirits in the astronomical camp.

None responded with more alacrity to this calling of the unknown than Sir William Huggins (to give him his present title). To apply Kirchhoff's method to the stars,' was the ambition that laid hold of him. Under its impulsion he became the founder of sidereal chemistry. It was no easy task that he undertook. Stellar light-analysis strains the resources of human ingenuity, even now that its advance is, so to speak, through a settled country. But in those early days every step was tentative; no tracks had been laid down, no survey executed; it was exploratory work, pure and simple. The difficulties besetting it may be illustrated by the statement that from Vega, one of the three brightest stars in the northern hemisphere, we receive about one forty-thousand-millionth part of the light imparted by the sun. And what was aimed at and substantially attained was to derive from beams thus emaciated by distance and diffusion, information comparable to that afforded by the copious rays of our own splendid luminary. Not a mere superficial inspection of stellar spectra was undertaken, but their thoroughgoing interpretation. Hence special contrivances had to be devised for their comparison with light from terrestrial sources, by which to test the presence in the stars of materials occurring on the earth. At the outset of his investigations Sir William Huggins was aided by Dr. William Allen Miller, Professor of Chemistry in King's College, London, a man of sound intellect and considerable experimental skill, although lacking the vis vivida of his younger colleague. But enthusiasm must have been contagious when nearly every observation revealed a new fact, ' and almost every night's work was red-lettered by some 'discovery.' Their preliminary results, communicated to the Royal Society in January 1863, established the im'portant, and, at that time, new fact, that all the stars belong to the same order of bodies as our sun, and consist of matter identical, at least in great part, with the 'chemical substances which form the material of the solar 'system.'

A pioneer rarely keeps the lead in a second generation as Sir William Huggins has done. We do not, however, propose here to follow the long course of his inquiries, or to enumerate the attendant disclosures. An authentic account of them is contained in the historical section of the admirable volume of which the title is prefixed to this article. It is designed to be the first of a series of publications emanating from the Tulse Hill Observatory; and we can only say that unless by its successors its merits are unlikely to be surpassed. The combination which it presents of artistic beauty with scientific value is, within our experience, unrivalled. From the mere bibliophile's point of view the book is a treasure. Paper, type, binding, are all choice and delightful, while a feature of unique interest is added in the numerous incidental illustrations from Lady Huggins's drawings, forming the headings and initials of the various chapters. They show us the picturesque aspect of the astrophysical scene of action--the wide sky, the trellised porch, the fair garden with its lovelinesses and curiosities, including an armillary sundial more than two centuries old, and a hive for the abode of exemplary bees, austerely inscribed • Nil nisi labore.' But the roses of forty years ago are no more; they have perished, asphyxiated by the insidious. approach of mephitic London. And their decease is symptomatic of a deterioration in the transparency of the air. The 'seeing' at Tulse Hill is no longer what it was. From open country the district has passed to the suburban condition; it has consequently forfeited its rural claim to exemption from city fogs, and impracticable nights succeed each other there with exasperating monotony. Yet the temptation to transplant the observatory was most prudently resisted. The gain in such changes is problematical; the sacrifice certain. Habits of work are dislocated; inestimable time is lost; delicate human machines are perhaps deranged, to say nothing of accidents, incalculable in kind, inevitable of occurrence. Truncated careers have too often resulted from observational breaches of continuity, which should accordingly be sedulously avoided except under the 'force majeure' of necessity.

6

The results embodied in this volume cover thirty-eight years of uninterrupted progress-of progress in many directions, since the tree planted by Kirchhoff shows its abundant vitality by sending forth suckers and ramifications innumerable. And not a few owe their start to the eminent observer who may be said to have taken all the sidereal

heavens for his province.' Not, indeed, to the exclusion of the solar system. He originated the spectroscopic study of the planets; recognised the carbonaceous nature of comets; anticipated the prismatic method of observing prominences at the sun's edge, first realised by Janssen and Lockyer; and invented, in daylight coronal photography,' a plan of research which may yet prove one of the most effective weapons in the armoury of solar physicists. The discovery of gaseous nebulæ falls into line with the main trend of his activity, no less than his momentous application of the spectroscope in 1868 to determine the radial or line of 'sight' velocities of the stars. Our immediate concern, however, is with his work in photographing stellar spectra. He was the first to attempt the task; he was the first to succeed systematically with it; his skill in its execution has never been surpassed. The magnificent'Atlas' just issued by him the designation emphasises the fundamental importance of the graphical section of the volume-bears unimpeachable testimony to this skill. Nor to his alone. The name of Lady Huggins appears on the title-page, and it has a significance there which only a minority of readers can fully penetrate. Since 1875 she has been, on equal terms, her husband's coadjutor, and while content to merge her initiative in his, she has known how to make its effect and influence tell as essential factors in the joint product of their labours. Apportionment of credit would be equally invidious and impossible. Suffice it to say that all belongs, in the truest and deepest sense, to each.

'Spectrography' is the complement to spectroscopy. The examination of spectra through the medium of the camera does not supersede, but essentially completes, their direct observation, the chemical retina' being mainly affected by impressions to which the eye is blind. Invisible, like visible, light is subject to specific absorption by interposed vapours or gases, and thus comes to us from the stars variously inscribed with signs of determinate meaning. But the messages are shrouded and inaccessible to sight. They can only be read by the use of a sixth sense, the organ of which is the sensitive plate. Autographically printed they become legible, while, apart from some such intermediary process, they should remain for ever unread, and their content of knowledge undivulged. Hence the urgent necessity for the developement of the new method.

It is applicable besides to a considerable part of the visible spectrum, and can, by certain modifications, be rendered

applicable to the whole. And, even in direct competition with the eye, it asserts a superiority scarcely qualified by some minor drawbacks. Visual impressions are evanescent; the memory may play fast and loose with them. Photographic records are permanent; they can be measured at leisure, consulted again and again, and examined at long intervals, for the resolution of questions that had not yet arisen when they were obtained. They are, moreover, objectively true; they escape the influence of subjective bias and personal error, and largely eliminate the effects of air tremors that baffle the eye in its efforts to interpret delicate spectral details.

Sir William Huggins's adoption, in 1875, of dry gelatine plates secured the future of celestial photography. For, by their substitution for the wet collodion plates previously employed, indefinite lengths of exposure became feasible, and the accumulative faculty of silver-bromide gained free scope. More, however, was needed for the satisfactory registration of the long range of ultra-violet light waves. Glass strongly absorbs them; it had therefore to be abolished in the special apparatus constructed for their analysis. A reflecting telescope with metallic specula replaced the customary refractor as the agent for their concentration; Iceland-spar prisms and quartz lenses-substances transparent to excessively short undulations-served for their dispersion and transmission. Star spectra of unprecedented actinic extension were by these means delineated, and proved highly characteristic, especially those of white stars,' like Vega and Sirius; for in them, on a brilliant background, a set of strong dark lines stood out, disposed in a definite harmonic order, at intervals narrowing continuously upward. Unmistakeably sequential to the four visible hydrogen rays, they formed with them the first recognised spectral series. That it should have been discovered in the stars is one of the most curious facts in scientific history.

A spectral series is a group of ethereal vibrations connected by a numerical law so definite that from the wave lengths of the first pair those of all the others can be inferred. They occupy places in the spectrum strictly calculable and surely occupied. A gap in their rhythmical progression is wellnigh unthinkable. Crowds of such series, belonging to a variety of chemical substances, are now known. Their occurrence is evidently rooted deep among the essentials of molecular structure and movement. They hold the keys to baffling enigmas; their study engages the attention of

experts; their explanation fascinates the assiduous thoughts of profound speculators. And this fresh start in terrestrial physics, so full of promise, so strangely suggestive, was due to the spectrographic impression of a hydrogen-star! Who could have ventured the forecast?

In the paper containing this announcement, presented to the Royal Society in December 1879,* Sir William Huggins essayed an evolutional' arrangement of the stars on grounds supplied by his photographs; he recurred to the topic in his address as President of the British Association in 1891, and it is treated of with some fulness in his recent work. That sidereal bodies run through some course of developement is undeniable. They undergo life changes' due to the gradual diminution, terminating in total ex'tinction, of their "vital force," as expressed by their temperature, actual and potential.' † Stars are cooling bodies. They continually and lavishly expend energy in the form of radiation, and no agency apparently exists by which any appreciable part of it can be restored, They pursue accordingly a downward slope towards darkness, which is the death of suns. No arrest of decay is possible.

'Noctes atque dies patet atri janua Ditis.'

The difficulty is to determine the successive gradations by which they descend thither. Here authorities are at variance. Some attribute comparative youth to classes of stars held by others to be verging on decrepitude; the order of progression laid down by one investigator is inverted by the next; there is total disagreement as to which are the 'hottest' stars. In a subject so new and complex, unanimity of opinion, it is true, could scarcely be expected; yet needless divergences have arisen through the neglect of some obvious principles adverted to by our present authors. Under their guidance, a few broad conclusions are, we believe, safely within reach; although many frontier problems remain, which we should vainly attempt to solve by a speculative coup de main. Formal approaches must be made to them by the regular methods of experimental attack.

The four spectral types, defined some thirty-five years ago by Father Secchi, still prescribe the main outlines of stellar classification. They have been variously and