Most naturalists recognize varieties of species; and many, like De Candolle, have come to conclude that varieties of the highest grade, or races, so far partake of the characteristics of species, and are so far governed by the same laws, that it is often very difficult to draw a clear and certain distinction between the two. Agassiz will not allow that varieties or races exist in Nature, apart from man's agency. Most naturalists believe that the origin of species is supernatural, their dispersion or particular geographical area, natural, and their extinction, when they disappear, also the result of physical causes. In the view of Agassiz, if rightly understood, all three are equally independent of physical cause and effect, are equally supernatural. In comparing preceding periods with the present and with each other, most naturalists and palaeontologists now appear to recognize a certain number of species as having survived from one epoch to the next, or even through more than one formation, especially from the Tertiary into the post-Tertiary period, and from that to the present age. Agassiz is understood to believe in total extinctions and total new creations at each successive epoch, and even to recognize no existing species as ever contemporary with extinct ones, except in the case of recent exterminations. These peculiar views if sustained will effectually dispose of every form of derivative hypothesis. Returning for a moment to De Candolle's article, we are disposed to notice his criticism of Linnaeus's "definition" of the term species (Philosophia Botanica, No. 157): "Species tot numeramus quot diversae formae in principio sunt creatae"-- which he declares illogical, inapplicable, and the worst that has been propounded. "So, to determine if a form is specific, it is necessary to go back to its origin which is impossible A definition by a character which can never be verified is no definition at all."