The rule has its exceptions.  But, curiously enough, the most striking class of exceptions, if such they be, seems to us even more favorable to the doctrine of derivation than is the general rule of a pure and simple ascending gradation.  We refer to what Agassiz calls prophetic and synthetic types; for which the former name may suffice, as the difference between the two is evanescent.

“It has been noticed,” writes our great zoologist, “that certain types, which are frequently prominent among the representatives of past ages, combine in their structure peculiarities which at later periods are only observed separately in different, distinct types.  Sauroid fishes before reptiles, Pterodactyles before birds, Ichthyosauri before dolphins, etc.  There are entire families, of nearly every class of animals, which in the state of their perfect development exemplify such prophetic relations.

The sauroid fishes of the past geological ages are an example of this kind These fishes which preceded the appearance of reptiles present a combination of ichthyic and reptilian characters not to be found in the true members of this class, which form its bulk at present.  The Pterodactyles, which preceded the class of birds, and the Ichthyosauri, which preceded the Cetacea, are other examples of such prophetic types.”—­(Agassiz, “Contributions, Essay on Classification,” p. 117.)

Now, these reptile-like fishes, of which gar-pikes are the living representatives, though of earlier appearance, are admittedly of higher rank than common fishes.  They dominated until reptiles appeared, when they mostly gave place to (or, as the derivationists will insist, were resolved by divergent variation and natural selection into) common fishes, destitute of reptilian characters, and saurian reptiles—­the intermediate grades, which, according to a familiar piscine saying, are “neither fish, flesh, nor good red-herring,” being eliminated and extinguished by natural consequence of the struggle for existence which Darwin so aptly portrays.  And so, perhaps, of the other prophetic types.  Here type and antitype correspond.  If these are true prophecies, we need not wonder that some who read them in Agassiz’s book will read their fulfillment in Darwin’s.

Note also, in this connection, that along with a wonderful persistence of type, with change of species, genera, orders, etc., from formation to formation, no species and no higher group which has once unequivocally died out ever afterward reappears.  Why is this, but that the link of generation has been sundered?  Why, on the hypothesis of independent originations, were not failing species recreated, either identically or with a difference, in regions eminently adapted to their well-being?  To take a striking case.  That no part of the world now offers more suitable conditions for wild horses and cattle than the pampas and other plains of South America, is shown by the facility with which they have there run wild and enormously