266 Alfred Sherwood Romer 



Although believing strongly in a fresh water origin for vertebrates, I must confess 

 I am surprised, not at the fact that there are a certain number of reports of reputed 

 marine vertebrates from the Silurian, but rather at the fact that, when such reports 

 are carefully scrutinized, there are so few instances of unquestionable marine occur- 

 rences. During the Devonian and later periods there were various incursions from 

 fresh water into the sea ; similar incursions of ostracoderms may well have occurred 

 earlier. The record suggests that, on the whole, the Heterostraci were a relatively 

 euryhaline group and may have had marine representatives in the Silurian, as appears 

 rather definitely to be true in the Devonian. 



Biocenosis vs. necrocenosis 



Our object is to determine the conditions under which the early fishes lived. Study 

 of the beds in which fossil remains are found does not necessarily tell us this. Habitat 

 in life and place of burial may be far removed from one another. If a cadaver is 

 destroyed by predators, its remains may be buried on the spot. But if the trunk of a 

 dead fish remains intact, gradual decomposition results in buoyancy, and the cadaver 

 may be transported by currents far from its proper habitat before settling to the bottom. 

 The fact that water runs down hill works strongly to the detriment of one advocating 

 a fresh water origin. An inland stream-dweller will tend to be carried down to lower 

 reaches of a river system ; forms living in a coastal region may be carried out to sea, 

 with resulting ecological confusion. If fish specimens in a given formation are 

 abundant and well-preserved, a life habitat in or close to the spot where they are found 

 is strongly indicated. If rare and fragmentary, remains in a marine deposit strongly 

 suggest transportation by currents; but whether transportation from another marine 

 locality or from fresh water cannot, of course, be proven. 



Identification of fresh water deposits 



Marine sediments are frequently identifiable in ready and positive fashion; there 

 is generally an abundance of invertebrates of dominantly marine groups. Identifica- 

 tion of Palaeozoic fish-bearing beds as continental in nature is a much more difficult 

 task, for the evidence is essentially negative in nature. Sediments are no sure guide. 

 While limestones are highly suggestive of marine conditions, calcareous deposition 

 may take place in inland waters, but on the other hand shales and sandstones, although 

 presumably derived ultimately from continental areas, have been in large measure 

 deposited under marine conditions. It is only in the case of coarse elastics and 

 conglomerates that continental conditions, or an approach to them, are strongly 

 suggested by sediments. As to the associated fauna, the absence of typical marine 

 invertebrates suggests continental conditions. This does not, however, prove the case, 

 for there are numerous marine areas today in which invertebrates capable of fossil 

 preservation are absent, and there are numerous geologic formations which were 

 obviously laid down in a marine situation and yet contain few or no invertebrate 

 remains. Further, absence of marine fossils in fish beds may be attributed by those 

 advocating salt water environment to diagenetic removal, by solution, of carbonate 

 shells once present. But this argument is one of which one may be reasonably 

 sceptical if used too frequently. 



In many Devonian beds which are generally agreed to be of a continental nature 

 we find fishes associated with a sparse fauna including eurypterids, ceratiocarid 



