398 J. BARRELL INFLUENCE OF CLIMATES ON VERTEBRATES 



beginning until the Devonian. But it has been stated that the fauna of 

 brackish waters is more largely made by emigration from the sea. If, 

 then, the Ordovician and Silurian fishes of these waters came from the 

 protected land waters and were able to adapt themselves to the higher 

 salinity of brackish waters, coming to live with brachiopods and phyllo - 

 pods, what prevented them from an earlier spread into the truly oceanic 

 realm? The answer is found in the enemies of the open sea. Not until 

 the fishes were large, powerful, and formidable could they compete with 

 the cephalopods, which were then the rulers of those regions. Their 

 earlier invasion of brackish waters occurred while they were still repre- 

 sented by small and sluggish ostracoderms or small but active elasmo- 

 branchs. 



Another and independent line of evidence suggests that the primitive 

 fish arose in the land waters rather than in those of the sea. Chamberlin 

 has pointed out that the body plan of the fish, primarily adapted for 

 active swimming, is such as would be necessary for an organism living 

 in. the fresh waters, in order to stem the currents and maintain itself 

 Avithin its habitat. The primitive animals of the sea, on the contrary, 

 as seen in the invertebrates, are typically sluggish and possess more or 

 less rotund forms. 11 The comparison is best made between the floating 

 and not the bottom living forms. It does not apply to bottom dwellers, 

 and therefore is not an argument which can be used as to the original 

 environment of the pre-vertebrates, such as balanoglossus and amphioxus, 

 nor, in a measure, to the bottom-dwelling ostracoderms. It does apply to 

 the habitat of the primitive sharks. These must have existed in Silurian 

 if not even in Ordovician times. The sharklike Ccelolepida? seem to show 

 affinities between sharks and ostracoderms and are present in the later 

 Silurian. They have raised the question in the minds of some paleon- 

 tologists if ostracoderms may not be derivatives from primitive sharks. 

 If this be the case, the origin of elasmobranchs is pushed back at least to 

 the Lower Ordovician, although, as scattered spines, the actual testimony 

 of their existence has not been revealed until far later in the Silurian. 



In conclusion and review, it is seen that in the ages preceding the 

 Ordovician the early chordates may have played no higher part in life 

 than do now the annelids. In lowly wormlike forms they may have 

 spread over the earth and have left within the more permanent and stable 

 environments of the sea such lingering archaic representatives as balano- 

 glossus, ascidians, and the lancelets ; but against the ruling host of in- 

 vertebrates thev could not rise. Within the fresh waters wormlike forms 



11 T. C. Chamberlin : On the habitat of the early vertebrates. Jour. Geol., vol. vlii, 

 1900, pp. 400-412. 



