147. In a very remarkable paper by Struve,[49] an attempt was made to settle the question, _Is the ether perfectly transparent?_ or, as we may now put it, Is any radiant energy absorbed by the ether, whether to produce other forms of energy, or to be dissipated by radiation in all directions? Long ago it had been pointed out by Olbers and others, that if the stars be infinite in number, and be distributed with anything roughly approximating to an average density through infinite space, the sky ought, night and day, to be all over of a brightness of the same order as that of the sun. Is the number of stars, then, finite; or does the ether absorb their light? Now, it need not in the least surprise us to find that the number of stars is _finite_, even though matter be infinite in quantity, and distributed with something like uniformity through infinite space. For only a finite portion of it may yet have fallen together so as to produce incandescent bodies; or, the other extreme, only a finite portion of it may be left incandescent. Either of these altogether different hypotheses is perfectly reasonable and scientifically justifiable; so that, from this point of view, we are not at present likely to obtain any information. Struve’s reasoning, which, by the way, is not accepted by Sir J. Herschel, introduces another consideration, viz., _the number of stars of each visible magnitude_. To apply this: suppose for a moment we make the assumption (actually measured values of annual parallax show it is certainly at best a very rough one) that the brighter stars are the nearer, and that a set of stars, on the average one-fourth as bright as another set, are on the average twice as far off, etc. A great deal of what we know to be certainly false is here assumed as true, but it is possible that the general accuracy of the results of the reasoning from it may not be thereby much affected. On the supposition of a sort of rough uniformity of distribution through space, we can easily calculate approximately what ought to be the relative numbers of the stars, classed by astronomers as of the various different magnitudes, once we have obtained (as it is not difficult to do) an estimate of the relative brightness of typical stars of these (arbitrary) magnitudes. From their brightness we calculate at once their relative distances, and thence (according to our hypothesis of approximately uniform distribution) what ought to be _the relative numbers of each magnitude_. When this is done, it appears that there is a great excess of the calculated over the observed numbers, at least for telescopic stars, and the greater the smaller the magnitude. This is the gist of Struve’s method, and he arrives at the result that the light of stars of the sixth magnitude (the smallest visible to an ordinary unaided eye, and whose average distance from us is supposed to be somewhere about ninefold that of stars of the first magnitude) loses about eight per cent.