(3) Adaptational convergence is the occurrence of parallel modifications due to analogous specialization in unrelated forms, whose phylogeny has been obviously diverse. “Also, animals belonging to quite distinct groups,” says Woods, “may, when living under similar conditions, come to resemble one another owing to the development of adaptive modifications, though they do not really approach one another in essential characters; thus analogous or parallel modifications may occur in independent groups—such are the resemblances between flying reptiles (_Ornithosaurs_) and birds, and between sharks, icthyosaurs and dolphins.” (_Op. cit._, p. 16.) As this type of convergence has been discussed in a previous article, with reference to the mole and mole-cricket, it need not detain us further. All these types of convergence, but especially the second type, are factual evidence of the compatibility of resemblance with independent origin, and the fact of their occurrence tends to undermine the certainty of the phylogenetic inferences based on fossil evidence; all the more so, that, thanks to its bad state of preservation, and the impossibility of dissection, even superficial resemblances may give rise to false interpretations. And, as for the cases of radical convergence, there is no denying that they strike at the very heart of the theory of descent. The third difficulty for Transformism arises from the discontinuity of the geological record. It was one of the very first discrepancies to be discovered between evolutionary expectation and the actual results of research. The earliest explorations revealed a state of affairs, that subsequent investigations have failed to remedy: on the one hand, namely, a notable absence of intermediate species to bridge the gaps between the fossil genera, and on the other hand, the sudden and simultaneous appearance of numerous new and allied types unheralded by transitional forms. Since Darwin had stressed the gradualness of transmutation, the investigators expected to find the transitional means more numerous than the terminal extremes, and were surprised to find, in the real record of the past, the exact reverse of their anticipation. They found that the classes and families of animals and plants had always been as widely separated and as sharply differentiated as they are today, and that they had always formed distinct systems, unconnected by transitional links. The hypothetical “generalized types,” supposed to combine the features of two or three families, have never been found, and most probably never will be; for it is all but certain that they never existed. Occasionally, it is true, palæontologists have discovered isolated types, which they interpreted as annectant forms, but a single pier does not make a bridge, and only too often it chanced that the so-called annectant type, though satisfactory from the morphological standpoint, was more recent than the two groups, to which it was supposed to be ancestral.