Fish origins— fresh or salt water? 263 



to extend our palaeontological study to the European deposits, much richer in 

 Silurian and Devonian fish locaHties than North America, but had hesitated to do so 

 because of the amount of time it would have taken to familiari/e myself with the 

 complex stratigraphic situation involved. To my delight, this topic is now covered in 

 an excellent comprehensive work by Gross (1950). In this he summarizes the pre- 

 Carboniferous fish faunas of the entire world; since but few finds have been made in 

 continents other than Europe and North America, the major new contribution is his 

 careful review of the numerous and varied European assemblages. As one topic. 

 Gross considers the fresh vs. the salt problem. Although he had previously voiced his 

 belief in a marine origin for fishes (Gross, 1933, 134, etc.), he nevertheless treats the 

 matter in a properly objective fashion. All finds from each locality are listed by genus 

 or species, his conclusion as to the marine or continental nature of the desposit is 

 stated, and totals for the two environments are computed. 



For comparison, I have listed his summarized figures for the Silurian and the three 

 subdivisions of the Devonian parallel to my own in Table I. For the Devonian stages 

 we are in complete agreement. We both find that in the early Devonian, fishes were 

 dominantly continental in habitat, but that there was a strong trend toward establish- 

 ment of an important marine fauna before the end of the period. 



As regards the Silurian, however, there is radical disagreement between our two 

 studies. The American " sample " suggested that Silurian fishes were nearly or 

 completely fresh water in habit, with the first marine invasion taking place only about 

 the beginning of the Devonian. Gross, on the other hand, believes Silurian fishes to 

 have been predominantly marine; following this, his figures indicate that there was a 

 sharp shift to a fresh water life in the early Devonian and then a strong reverse trend 

 toward the sea.* 



It is to a more detailed consideration of the European Silurian record that (following 

 consideration of some general questions) the present paper is devoted. Not so much 



Table I 

 Percentage of forms in seemingly fresh water deposits at successive 

 horizons. Left, Romer and Grove i\935) for finds from all North 

 American localities, right, for world totals by species as given by 



Gross (1950, table IV) 



* The sharp marine-continental shift indicated by Gross's figures would be turthcr accentuated 

 if the American finds were excluded and if, as is done by many workers today, the Dountonuin and 

 its equivalents were included in the Devonian. In this case, the apparent mcrcasc in Ircsh Nvaicr 

 fishes from Upper Silurian to Lower Devonian instead of bemg from 36 per cent to M per cent 

 would have been from close to per cent to about 90 per cent ! 



