May we not, then, safely generalize, and make the law universal? Is not this a sufficient ground for confident induction? Even though some facts are still inexplicable, is that a sufficient reason for withholding assent to a theory which explains so much? In all induction we first establish a law provisionally from the observation of a comparatively few facts, and then extend it over a multitude of facts not included in the original induction. If it explains these also, the law is verified. The law of gravitation was first based on the observation of a few facts, and then verified by its explanation of nearly all the facts of celestial motion. There are some outstanding facts of celestial motion still unexplained, but we do not, therefore, doubt the law of gravitation. The same principle applied in biology ought to establish the law of evolution, for it also explains all the facts of biology as no other law can. But inductive evidence differs from other kinds of evidence in one respect, which, in fact, constitutes its strength to the scientific, but its weakness to the popular mind. It is a kind of circumstantial evidence, but its force does not consist in a few strong circumstances easily appreciated, such as strike the popular mind, and force conviction, but rather in a multitude of small circumstances, each by itself insignificant, but all together pointing to one conclusion and demanding one explanation. Such evidence is, indeed, overwhelming, but only to the mind that masters it. The evidence for the law of gravitation is literally the whole science of astronomy. So also the evidence for the law of evolution is the whole science of biology. Neither of these laws can be proved in a debating society, but only by a course of study. In the one case the law has been universally accepted--not, however, on evidence, for there are few indeed who appreciate the evidence, but on the authority of scientific unanimity. In the other case there has not yet been time enough for the already established unanimity to have its full effect. Thus much, we believe, will be generally admitted as a very moderate claim. Evolution is certainly a legitimate induction from the facts of biology. But we are prepared to go much further. We are confident that evolution is _absolutely certain_. Not, indeed, evolution as a special theory--Lamarckian, Darwinian, Spencerian--for these are all more or less successful modes of explaining evolution; nor evolution as a school of thought, with its following of disciples--for in this sense it is still in the field of discussion--but evolution as a law of derivation of forms from previous forms; evolution as a law of continuity, as a universal law of becoming. In this sense it is not only certain, it is axiomatic. It is only necessary to conceive it clearly, to see that it is a necessary truth. This may seem paradoxical to some. I stop to justify it.