D.—ZOOLOGY. Sl 
sheet of water connected with the sea by a long, narrow neck, there is a 
Zoarces population with an average of 108 vertebrae, but in the neck 
the number gradually increases towards the sea. 
There can be no doubt that the fiords were originally populated from 
the outside, and it seems likely that the decreased number of vertebrae 
in the fiords is related to the lesser activity of the fiord fish. Evolution 
has proceeded to such an extent that the Zoarces of the Roskilde Fiord 
differs from that of the Kattegat more than does the European eel from 
the American, and these are generally regarded as good species. But the 
repetition of the same phenomenon in different fiords and the continuous 
gradation from one form to another make it impossible to recognise species 
here. 
Zoarces are very stationary, but possibly the young are more migratory 
than the adults. But if we suppose that these fishes move on an average 
a mile a year, or even less, and mate with the nearest fish of the opposite 
sex, we can understand how the tendency to form a pure fiord race is 
hampered by continuous interchange, and how the influence of the outside 
form gradually diminishes until in the innermost waters it is not felt at all 
and isolation is accomplished. In each fiord a series of intermediates, 
hybrids if we like so to term them, connect two well-differentiated com- 
munities, one in the sea, the other in the inner waters of the fiord. 
These detailed examples are sufficient to illustrate my view that some 
form of isolation, either physical or produced by localisation or by habi- 
tudinal segregation, is a condition of the evolution of a new species. 
The effects of physical isolation, due to the formation of a barrier, are 
well seen in comparing the fishes of the Atlantic and Pacific coasts of 
Central America, most of which can be paired, one species being found on 
the Atlantic side and its nearest ally on the Pacific side. The effects of 
habitudinal segregation are, as it seems to me, seen in the Cichlid fishes 
of Lake Tanganyika, where there are ninety species that appear to have 
evolved in the lake from two ancestral forms; the differences between 
these species in the form and size of the mouth and in the dentition are 
an indication that their diversity is related to specialisation for different 
kinds of food. 
_ The whole of my work leads to the conclusion that the first step in 
the origin of a new species is not a change of structure, but the formation 
of a community either with new habits or in a new or a restricted environ- 
ment. For some species we know fairly certainly what has happened, 
and where, when, and why; shall we ever know how? Experimental 
attempts to repeat the operations of nature might perhaps give us a clue, 
but I do not expect one from experiments of the kind that is so fashionable 
owadays. 
_ For example, if Salmo salar and Salmo obtusirostris could be bred 
together, it would not matter much whether the hybrids were sterile or 
ertile ; and if they were fertile, it would not interest me to know that the 
variation in their offspring could be squared with the factorial hypo- 
thesis by the ingenious assumption that there were several factors for 
both larger mouth and smaller mouth and for fewer gill-rakers and more 
ill-rakers. Even if the number of gill-rakers in either species could be 
increased or decreased by thyroid extract, I should still be unconvinced 
that we had got much nearer to the root of the matter. 
y 61925 G 
