112 Racial Studies in Fishes 



Our population analyses have thus shown that Zoarces is sjjlit up 

 into numerous " races," often highly localised, and diflering widely one 

 from another. In this respect, however, there is no difference of prin- 

 ciple between Zoarces and the other well investigated fish species such 

 as herring and plaice. The difference is only one of degree, depending 

 upon the fact that Zoarces is an even more " variable " and even less 

 migratory species than either nf the othei- two. 



One species of fish, however, to which I have also devoted particular 

 attention, viz. the common fi-eshwater eel {Angidlla vulgaris), does (\\^ev 

 in principle, and a comparison with this species, which has about the 

 same number of vertebrae, etc. as Zoarces, is therefore highly instructive. 



On going through a large amount of material, it was found that in 

 contrast U> Zoarces, which, as we have seen, is split up into numerous 

 local races, all the eel populations of Europe are identical (see graphs. 

 Fig. 7). The explanation of this most remarkable fact must be sought 

 in the great biological diftei'enee between the two species; Zoarces 

 spends the whole of its life in the same very restricted area, whereas 

 all the freshwater eels of Europe midertake migrations of thousands 

 of miles out into the Atlantic to spawn, the young thereafter journeying 

 the same way back. Long before the young, or rather larvae {Lepto- 

 cephalus brevirostris), reach the coasts of Europe, they have developed 

 their full complement of vertebrae, etc., and we have here a natural 

 explanation of the fact that all the eel populations of Europe are 

 identical, oi- belong to one and the .same "race." 



As mentioned above, Zoarces gives birth to a great number of 

 young, which have already, long before they are born, developed their 

 full number of vertebrae, fin rays, j)igment spots, etc. This renders it 

 possible for us to cany out individual offspring analyses with this 

 species, the offspring of each gravid female being separately investigated. 



Examples of such individual offspring analyses will be found in the 

 graphs (Fig. 8). The material is from the same pcjpulation (St. 31) as 

 the three population analyses for 1!)14, 1915 and 191(3 in Fig. 4. On 

 comparing Fig. 4 with Fig. 8 we find that the average number of 

 vertebrae in various offspring samples can exhibit considerable variation 

 among themselves, and differ greatly from the average for the popula- 

 tion. Thus in the case of the three offspring samples in Fig. 8 ; the 

 mean values for number of vertebrae were about 117-83, 112-36 and 

 109-48 respectively, whereas the number for the population was in all 

 the three years 1914-16 between about 113-2 and 113-4. 



