Take first the class of Fishes.  We have seen that in the Jurassic periods there were none of our common Fishes, none corresponding to our Herring, Pickerel, Mackerel, and the like,—­no Fishes, in short, with thin membranous scales, but that the class was represented exclusively by those with hard, flint-like scales.  In the Cretaceous epoch, however, we come suddenly upon a horde of Fishes corresponding to our smaller common Fishes of the Pickerel and Herring tribes, but principally of the kinds found now in tropical waters; there are none like our Cods, Haddocks, etc., such as are found at present in the colder seas.  The Fishes of the Jurassic epoch corresponding to our Sharks and Skates and Gar-Pikes still exist, but in much smaller proportion, while these more modern kinds are very numerous.  Indeed, a classification of the Cretaceous Fishes would correspond very nearly to one founded on those now living.  Shall we, then, suppose that the large reptilian Fishes of the Jurassic time began suddenly to lay numerous broods of these smaller, more modern, scaly Fishes?  And shall we account for the diminution of the previous forms by supposing that in order to give a fair chance to the new kinds they brought them forth in large numbers, while they reproduced their own kind less abundantly?  According to very careful estimates, if we accept this view, the progeny of the Jurassic Fishes must have borne a proportion of about ninety per cent, of entirely new types to some ten per cent, of those resembling the parents.  One would like a fact or two on which to rest so very extraordinary a reversal of all known physiological laws of reproduction, but, unhappily, there is not one.

Still more unaccountable, upon any theory of development according to ordinary laws of reproduction, are those unique, isolated types limited to a single epoch, or sometimes even to a single period.  There are some very remarkable instances of this in the Cretaceous deposits.  To make my statement clearer, I will say a word of the sequence of these deposits and their division into periods.

These Cretaceous beds were at first divided only into three sets, called the Neocomian, or lower deposits, the Green-Sands, or middle deposits, and the Chalk, or upper deposits.  The Neocomian, the lower division, was afterwards subdivided into three sets of beds, called the Lower, Middle, and Upper Neocomian by some geologists, the Valengian, Neocomian, and Urgonian by others.  These three periods are not only traced in immediate succession, one above another, in the transverse cut before described, across the mountain of Chaumont, near Neufchatel, but they are also traced almost on one level along the plain at the foot of the Jura.  It is evident that by some disturbance of the surface the eastern end of the range was raised slightly, lifting the lower or Valengian deposits out of the water, so that they remain uncovered, and the next set of deposits, the Neocomian, is accumulated