Papers in Marine Biology and Oceanography, Suppl. to vol. 3 of Deep-Sea Research, pp. 261-280. 



Fish origins fresh or salt water? 



By Alfred Sherwood Romer 

 Museum of Comparative Zoology, Harvard University 



Summary — There are discussed various general considerations entering into palacontological study 

 of the problem as to whether the early evolution of fishes took place in fresh or salt water. Opposite 

 conclusions as to the typical habitat of Silurian fishes have been reached by the author and Grovl 

 on American evidence, on the one hand, and Gross, mainly concerned with European material, on 

 the other. An attempt is made to reconcile this difference. Part may be due to the very different 

 histories of the two continents in Silurian times. Close examination of the stratigraphy of European 

 fossil fish localities suggests that many deposits which are often regarded as marine are of continental 

 or near-continental nature. It is concluded that the evidence, taken as a whole, points strongly 

 towards fresh waters as the common Silurian fish habitat. 



INTRODUCTION 



It is generally agreed that the early chordates originated, as in the case of all major 

 animal groups, in the sea. But many of the more ancient geological records of fishes 

 are from sediments of continental type, and the possibility that fish evolved in inland 

 waters was long ago seen as a distinct possibility. T. C. Chamberlin in 1900 noted 

 theoretical considerations favouring a fresh water origin, but failed to give any detailed 

 consideration of the actual fossil evidence. In 1923 MacFarlane (a botanist!) wrote 

 an entire volume attempting to prove the fresh water origin of fishes; but this cannot 

 be taken too seriously as proof of the case, for the argument is a forced one in many 

 regards, and MacFarlane's general attitude is that of the King of Hearts in A/Ice In 

 Wonderland — " Verdict first, evidence afterwards "'. 



Two decades ago I determined to make a study of at least a considerable body of 

 the available fossil evidence. In this I associated myself with Dr. Brandon H. Grove. 

 My own inclinations at the time were mildly in favour of the hypothesis of a fresh 

 water origin; Dr. Grove (a student of invertebrate palaeontology) favoured, on the 

 contrary, an early marine habitat. Under the circumstances any conclusion, one way 

 or the other, which we could jointly reach, might be reasonably considered as attained 

 in objective fashion. We decided to restrict our study to an attempt to determine the 

 environmental conditions surrounding all fish finds recorded from North American 

 pre-Carboniferous rocks. Our upper time boundary was set for the reason that by 

 the end of the Devonian there definitely existed both fresh and salt water fish faunas 

 and there was no need to go further. We limited ourselves to North .America in a 

 belief that the American record would give us a representative sample of the total 

 world assemblage without our being forced to a consideration of European localities- 

 localities with which we were personally unfamiliar and concerning many of which 

 we would have had difficulty in obtaining exact data. 



The results of our study (Romer and Grove, 1935) led both of us strongly to the 

 belief that the early vertebrates were continental forms and that Devonian and later 

 marine types had invaded the seas from fresh waters. As summarized in the left-hand 

 column of Table I, all American Silurian finds were indicative of a fresh water environ- 

 ment; during the Devonian, however, there was an increasing trend toward the sea, 



261 



