Revieics — A. Smith Woodward' s Vertebrate Paleontology. 371 



change ; this advanced form then seems to be driven from all the 

 areas where the dominant ancestral race reigns supreme, and evolu- 

 tion in the latter becomes comparatively insignificant. Meanwhile 

 the banished type has acquired great developmental energy, and 

 finally it spreads over every habitable region, replacing the eifete 

 race ' which originally produced it. Another ' expression point ' 

 (to use Cope's apt term) is thus reached, and the phenomenon is 

 repeated. The Actinopterygian fishes furnish an interesting illus- 

 tration. The earliest known member of this order [CheiroJepis) 

 appears as an insignificant item in the Lower Devonian fauna, 

 where Crossopterygian and Dipnoan fish.es are dominant. When 

 the latter begin to decline in the Lower Carboniferous, the suborder 

 to which Gheirohpis belongs (Chondrostei) suddenly appears in 

 overwhelming variety. By the period of the Upper Permian another 

 fundamental advance has taken place — the Protospondyli have 

 arisen ; but only a solitary genus is observed among the hosts of 

 the dominant race. In the Trias the new type becomes supreme, 

 and at the same time the next higher suborder, that of the Iso- 

 spondyli, begins to appear. This lingers on in the midst of the 

 dominant Protospondyli during the Jurassic period, and then in 

 the Cretaceous this and still higher suborders suddenly replace 

 the earlier types and inaugurate a race which has subsequently 

 changed only to an insignificant extent. The Mammalia afford, 

 another illustration of the same phenomenon. The reptilian class 

 shows the closest approximation to that of the Mammalia at the 

 dawn of the Mesozoic epoch, when it is just beginning to replace 

 the older Batrachia. In rocks of this age, on all four continents — 

 Europe, Asia, Africa, and America — there are numerous remains 

 of the mammal-like Anomodontia (as they are termed, p. 144). 

 Afterwards not a trace of these 'missing links' is known; and 

 with the exception of the insignificant small jaws of possible Proto- 

 theria and Metatheria in the Jurassic and Cretaceous of England 

 and North America, mammals do not appear either in Europe, 

 Africa, or North or South America until the base of the Eocene, 

 when they suddenly became dominant and are already differentiated. 

 " Parallelism in Evolution. — As nothing is yet known of the 

 supposed refuges to which the incipient new races have betaken 

 themselves to differentiate and acquire vital energy, we can merely 

 assume them as a tentative hypothesis. But even when the facts 

 are abundantly manifest, it is often difficult to settle the most 

 elementary questions by direct reference to them. The problem of 

 parallelism in evolution is one of these. It is necessary to 

 determine whether the same animal can be evolved simultaneously 

 in more than one region from distinct series of ancestors. Are the 

 pumas and jaguars of America, for example, wandering cousins of 

 the lions and leopards of the Old World, or have they been evolved 

 on the other side of the globe through a distinct set of carnivorous 

 ancestors? The case of the horses is often cited as suggesting that 

 such a parallelism in evolution may have occurred ; because the series 

 of ancestral horses traced through the Tertiary strata of Europe is 



