out the millions of cubic miles of water which compose that mighty ocean? Of course, the mere statement of such a case is sufficient to show its absurdity. But the improbability that the ardent naturalist would secure the prize in the way I have described is not one whit greater than the improbability that, even if there were a central sun, it should lie within the domain of our scrutiny.
There is another line of reasoning by which the theory of probability will often give us invaluable information, which is not this time merely of a negative kind. There are many instances which might be taken of the principles now to be employed. I shall, however, adopt that particular one which presents, perhaps, the greatest interest to astronomers. The question often arises as to whether two objects which appear to us to lie near each other on the surface of the heavens are really neighbours in space, or whether their contiguity is only apparent. It often happens, for instance, that two stars appear very close together through the telescope, and we desire to know whether the two bodies are indeed allied by any bond of physical association, or whether the appearance may not be a mere accidental coincidence. The latter would be the case if the line joining the two stars happened to be so nearly directed towards the earth that, though in reality one of the stars is so much more remote than the other, yet that from our point of view the two happen to be projected on the same part of the sky. We are generally at fault in determining this question by direct observation, because it is usually impossible to find the actual distances by which the earth is separated from the objects, and, therefore, we are deprived of any direct assurance that those distances are so far equal as to enable
