[Footnote a:  P. 484, Engl. ed.  In the new American edition, (Vide Supplement, pp. 431, 432,) the principal analogies which suggest the extreme view are referred to, and the remark is appended,—­“But this inference is chiefly grounded on analogy, and it is immaterial whether or not it be accepted.  The case is different with the members of each great class, as the Vertebrata or Articulata; for here we have in the laws of homology, embryology, etc., some distinct evidence that all have descended from a single primordial parent.”]

[Footnote b:  In Bibliotheque Universelle de Geneve, Mars, 1860.]

[Footnote c:  This we learn from his very interesting article, De la Question de l’Homme Fossile, in the same (March) number of the Bibliotheque Universelle.]

This raises the question, Why does Darwin press his theory to these extreme conclusions?  Why do all hypotheses of derivation converge so inevitably to one ultimate point?  Having already considered some of the reasons which suggest or support the theory at its outset,—­which may carry it as far as such sound and experienced naturalists as Pictet allow that it may be true,—­perhaps as far as Darwin himself unfolds it in the introductory proposition cited at the beginning of this article,—­we may now inquire after the motives which impel the theorist so much farther.  Here proofs, in the proper sense of the word, are not to be had.  We are beyond the region of demonstration, and have duly probabilities to consider.  What are these probabilities?  What work will this hypothesis do to establish a claim to be adopted in its completeness?  Why should a theory which may plausibly enough account for the diversification of the species of each special type or genus, be expanded into a general system for the origination or successive diversification of all species, and all special types or forms, from four or five remote primordial forms, or perhaps from one?  We accept the theory of gravitation because it explains all the facts we know, and bears all the tests that we can put it to.  We incline to accept the nebular hypothesis, for similar reasons; not because it is proved,—­thus far it is wholly incapable of proof,—­but because it is a natural theoretical deduction from accepted physical laws, is thoroughly congruous with the facts, and because its assumption serves to connect and harmonize these into one probable and consistent whole.  Can the derivative hypothesis be maintained and carried out into a system on similar grounds?  If so, however unproved, it would appear to be a tenable hypothesis, which is all that its author ought now to claim.  Such hypotheses as from the conditions of the case can neither be proved nor disproved by direct evidence or experiment are to be tested only indirectly, and therefore imperfectly, by trying their power to harmonize the known facts, and to account for what is otherwise unaccountable.  So the question comes to this: