any part of the heavens this was not the case, the shape of the system towards that part could not be determined. It is necessary to notice, however, that the general uniformity of distribution by no means implied the non-existence of clustering aggregations of stars, or of streams, branches, and nodules of stars within the limits of the system. On the contrary, Herschel, so early as 1785, clearly indicated his recognition of such varieties; and all that he insisted upon at that time was that such peculiarities were themselves so distributed as to produce within the system, regarded as a whole, a general uniformity of distribution. It is absolutely essential, if we would understand Herschel's earlier views, to take his own preliminary description, which somehow appears to have escaped the notice of commentators,-unless we suppose the difficulty of grasping Herschel's real meaning to have caused them to misunderstand the passage. "It will be best," Herschel says, "to take the subject from a point of view at a considerable distance, both of space and of time. Let us suppose, then, numberless stars of various sizes scattered over an indefinite portion of space in such a manner as to be almost equally distributed throughout the whole. The laws of attraction, which no doubt extend to the remotest region of the fixed stars, will operate in such a manner as most probably to produce the following remarkable effects": "Form I. In the first place, since we have supposed the stars to be of various sizes, it will frequently happen that a star being considerably larger than its neighbouring ones, will attract them more than they will be attracted by others that are immediately around them; by which means they will be, in turn, as it were, condensed about the centre; or, in other words, form themselves into a cluster of stars of almost a globular figure, more or less regularly so, according to the size and original distance of the surrounding stars." "Form II. The next case, which will also happen almost as frequently as the former, is where a few stars though not superior in size to the rest, may chance to be rather nearer each other than the surrounding ones; for here also will be found a prevailing attraction in the combined centre of gravity of them all, which will occasion the neighbouring stars to draw together, not, indeed, so as to form a regular, or globular figure, but, however, in such a manner, as to be condensed towards the common centre of gravity of the whole irregular cluster. And this construction admits of the utmost variety of shapes, according to the number and situation of the stars which first give rise to the condensation of the rest." "Form III. From the composition and repeated conjunction of both the foregoing forms, a third may be derived,* when many large stars, or combined small ones are situated in long-extended, regular, or crooked rows, hooks, or branches; for they will also draw the surrounding *Here the words " may be derived" are not intended to imply doubt as to the fact that the groups of the third form exist. The context ones, so as to produce figures of condensed stars, coarsely similar to the former, which gave rise to these condensations." "Form IV. We may likewise admit of still more extensive combinations when, at the same time that a cluster of stars is forming in one part of space, there may be another collecting in a different, but perhaps not far distant quarter, which may occasion a mutual approach towards their common centre of gravity." "Form V. In the last place, as a natural consequence of the former cases, there will be formed great cavities, or vacancies, by the retreat of the stars towards the various centres which attract them; so that upon the whole, there is evidently a field of the greatest variety for the mutual and combined attractions of the heavenly bodies to exert themselves in." After considering the possibility of catastrophes during the evolution of the forms here described, Herschel proceeds to consider the position of the terrestrial observer in his "own retired station, in one of the planets attending a star." He shows that to such an observer, placed in a far extending stratum "or branching cluster of millions of stars, such as may fall under Form III.," considered above, the following appearances will be presented :-To the naked eye, "The heavens will not only be richly scattered third form from considering that both the other forms must be compounded and repeatedly conjoined. It is important to notice this, because "Form III." is the key of the whole passage, being the form which Herschel attributed to our Milky Way at this stage of his researches. over with brilliant constellations, but a shining zone or Milky Way will be perceived to surround the whole sphere of the heavens, owing to the combined light of those stars which are too small, that is too remote, to be seen." Let this passage be particularly noted before we proceed, as on its right comprehension depends our entire judgment as to Herschel's earlier views. He here presents the sidereal system as a far-extending stratum or branching cluster of millions of stars, of Form III., and therefore including within its limits many subordinate clusters and nebula of Forms I. and II.; while he regards the light of the Milky Way as resulting from the extension of the system towards that zone much farther than in other directions.* This must be borne carefully in mind in reading what immediately follows. "Our observer's sight," proceeds Herschel, "will be so confined that he will imagine this single collection of stars, of which he does not even perceive the thousandth part, to be the whole contents of the heavens. Allowing him now the use of a common telescope he begins to suspect that all the milkiness of the bright path which surrounds the sphere may be owing to stars. He perceives a few clusters of them in various parts of the heavens, and finds also that there are a kind of nebulous patches; but still his views are not extended so far as to * In fact, his views at this stage corresponded closely with those which had been advanced by Lambert nearly a quarter of a century earlier. In the papers of 1784, Herschel presents views more nearly resembling those which Wright of Durham had advanced half a century earlier, and which Kant adopted a year or two before Lambert reach to the end of the stratum in which he is situated, so that he looks upon these patches as belonging to that system which to him seems to comprehend every celestial object. He now increases his power of vision, and applying himself to a close observation, finds the Milky Way is indeed no other than a collection of very small stars. He perceives that those objects which had been called nebulæ are evidently nothing but clusters of stars. He finds their number increase upon him, and when he resolves one nebula into stars he discovers ten new ones which he cannot resolve. He then forms the idea of immense strata of fixed stars, of clusters of stars, and of nebulæ, till going on with such interesting observations he now perceives that all these appearances must naturally arise from the confined situation in which he is now placed. Confined, it may justly be called, though in no less a space than what before appeared the whole region of the fixed stars; but which now has assumed the shape of a crookedly branching nebula, not one of the least, but perhaps very far from being the most considerable of those numberless clusters that enter into the construction of the heavens." It cannot be denied that the passage just quoted is not very easy to understand. At one stage, or rather throughout the greater part of the passage, it seems abundantly clear that Herschel is describing our sidereal system as including multitudes of subordinate clusters and nebulæ. But then at the end, he describes it as itself a nebula, greater than some, but less than others, of numberless clusters, composing the sidereal heavens. And the per |