188 Racial Studies in Fishes 



corresponds with the gradation in the average number of vertebrae. 

 To conclude from that fact that this character is directly determined 

 by the salinity or by a factor acting parallel to it — which several in- 

 vestigators have been inclined to do — is however not justifiable as I 

 have already explained in my first paper in ihxs Journal (Vol. vii, 1918, 

 pp. 115-117). 



VII. Concluding Remarks. 



The investigations and experiments with Zoarces described above 

 have provided new evidence of the correctness of the view that the race- 

 characters in fishes are of a hereditary nature, i.e. that they are first and 

 foremost determined by internal factors (Sections III and IV). That 

 they are however on the other hand to a certain extent influenced by 

 external factors, is here for the first time experimentally proved for one 

 of our northern species of fish. We have thus seen that the average 

 number of vertebrae which under natural conditions remained very 

 constant in six consecutive annual classes of the same population 

 (Section II) was altered, if not strongly, at any rate perceptibly, by a 

 transplantation of the population (Section V), 



In view of the investigations hitherto undertaken the gradation of 

 the average number of vertebrae has to be interpreted not as a direct 

 effect of the external conditions (the salinity or another factor acting 

 parallel to it), but as the result of a selection. 



TABLE I. 



Station 31, Ise Fjord, Denmark, 1914 — 1919. Decade Investigations. 

 Average number of vertebrae (a) in the six annual classes 1914 — 1919. 



n number of decades, a average number of vertebrae, <r standard deviation for average 

 value of a decade, P.E.A. probable error of average, P.F.A. probable fluctuations of 

 average. 



^ Average of six averages (1914 — 1919). 

 ^ Calculated from exact decade averages. 



