The fishes of the Upper Silurian are such a prophecy. They can claim no parentage in the older rocks, and they appear at once as kings of their class. With reference to the Silurian itself, they are of little consequence; and in the midst of its gigantic forms of invertebrate life they seem almost misplaced. But they predict the coming Devonian, and that long and varied reign of vertebrate life which culminates in man himself. No such prophetic ideas are represented by the giant crustaceans and cuttle-fishes and swarming graptolites. They had already attained their maximum, and were destined to a speedy and final grave in the Silurian, or to be perpetuated only in decaying families whose poverty is rendered more conspicuous by the contrast with the better days gone by. The law of creation provided for new types, and at once for the elevation and degradation of them when introduced; and all this with reference to the physical conditions not of the present only but of the future. Such facts, which cannot be ignored save by the wilfully blind, are beyond the reach of any merely material philosophy. The little that we know of Silurian plants is as eloquent of plan and creation as that which we can learn of animals. I saw not long ago a series of genealogies in geological time reduced to tabular form by that ingenious but imaginative physiologist, Haeckel. In one of these appeared the imaginary derivation of the higher plants from Algæ or sea-weeds. Nothing could more curiously contradict actual facts. Algæ were apparently in the Silurian neither more nor less elevated then in the modern seas, and those forms of vegetable life which may seem to bridge over the space between them and the land plants in the modern period, are wanting in the older geological periods, while land plants seem to start at once into being in the guise of club-mosses, a group by no means of low standing. Our oldest land plants thus represent one of the highest types of that cryptogamous series to which they belong, and moreover are better developed examples of that type then those now existing. We may say, if we please, that all the connecting links have been lost; but this is begging the whole question, since no thing 'but the existence of such links could render the hypothesis of derivation possible. Further, the occurrence of any number of successive yet distinct species would not be the kind of chain required, or rather would not be a chain at all.