| THE PALEONTOLOGIC RECORD |
By Dr. W. D. MATTHEW
AMERICAN MUSEUM OF NATURAL. HISTORY
CONTINUITY of development in a broad sense hardly calls for discussion here. The paleontologic evidence in its favor is so extensive and so universal that the perfection of the proof is merely a question of the completeness of the evidence. The question for discussion is rather as to the method of race development and specific change—whether continuous, by the slow accumulation of minute individual variations, definite or indefinite, through the influence of natural selection or of other causes—or discontinuous by the sudden appearance of distinct mutations or sports, usually of subspecific or specific value, sometimes of generic value. This question is much debated nowadays, and" it would seem that the evidence from paleontology ought to be of the first importance in deciding it.
It is very commonly asserted that this evidence is strongly in favor of discontinuous development. This would mean that new species and even genera appear, as a rule, suddenly at certain levels, and that the record of a phylum is not usually a slow continuous change from one species into another as we pass upward from stratum to stratum; but that one species has a certain vertical range and is then supplanted by another species, this in turn by a third, and so on, each successive stage being an advance over the preceding, but the species overlapping instead of grading.
I think that there is no question but that in vertebrate paleontology the evidence taken at its face value does appear to be very distinctly in favor of discontinuous development. Where we are able to follow a phylum of Tertiary mammalia through a series of strata in one locality, we find that the successive stages appear, as a rule, full formed at certain levels, supplant and replace the more primitive stages, and are in turn supplanted and replaced by more advanced stages. In former years, when the records of locality and level were less exact, it was possible to arrange a series of gradations from one stage to another among the specimens pertaining to a particular phylum, and to assume that this gradation corresponded to the levels in the formation at which the specimens had been collected, and that the specific change was through continuous gradation. The more exact records of locality and level and the more extensive and complete collections in recent years have in general failed to confirm this arrangement. In the great majority of cases, so far as the record shows, new species appear already
