THE GLACIAL EPOCH. 31 



down in a wedge along the mountains of Greenland. The area of distribution was Emigration 

 originally a disk ; the accumulation of ice round the pole made it a ring, and the glaciers of the second 

 Greenland cut the ring in two, so that the birds of Scandinavia could no longer interbreed ^^^^^^^ 

 with those of Labrador. This partial discontinuity of the area of distribution allowed 

 Evolution to modify the species at each end of the horseshoe in different directions ; but the 

 differentiations could only be subspecific, as interbreeding must have taken place along the 

 whole line, the birds of each bay intermarrying with those in the next, so that the result 

 was probably a Scandinavian species, which appeared to be quite distinct from its ally in 

 Ijabrador, until it was found that the two species were connected together by a complete 

 series _of intermediate forms, the Kamtschatkan form being perhaps midway between the 

 two extremes. 



As the severity of the Glacial Period increased, the cracked ring must have been broken 

 into two pieces by glaciers, one coming down from the Rocky Mountains, and the other from 

 Nova Zembla or the mountains of Eastern Siberia. The shore birds were thus isolated in 

 three colonies — one on the Atlantic coast of America, one on the Atlantic coast of Europe, 

 and the third on the Pacific coast of Alaska and East Siberia. 



Later on, when the Arctic ice filled the Polar basin, the birds retreating before it on 

 the shores of the Pacific must also have been isolated in two parties, one being compelled 

 to follow the coast of Asia, and the other that of America. 



During the Prae-Pliocene glacial period the Charadriidse were differentiated into genera, 

 duriiig the Post-Pliocene into species. To try and find out the locality where each genus 

 was differentiated is too hazardous a task, but to attempt the discovery of the locality where 1 

 each species was isolated and differentiated is seldom entirely hopeless. It seems probable 1 

 that the habits of migration, formed originally when the birds extended their range beyond ■ 

 the Arctic Circle, and were compelled to wander in search of light during the three months' 

 night, and developed to a much greater degree when they were compelled, by the increasing 

 severity of the winters, to absent themselves from their breeding-grounds for six months in 

 search of food, became so deeply impressed upon the species, that they were never wholly 

 abandoned, even during the warm climate which followed the last glacial epoch. It seems i j^gga^^u 

 to be a reasonable hypothesis that the locality where a species is now most abundant in i where each 

 winter was its winter home during the glacial period, where it was isolated and difi'erentiated, 4soiated^nd 

 and whence it has migrated to its breeding-grounds every spring for the last hundred 'differentia- 

 thousand years. The pertinacity with which ancient routes of migration are retained is 

 exemphfied in the history of many species of Charadriidae. The Common Dotterel 

 (Charadrius morinellus) appears to have gradually extended its range during the breeding- 

 seasons from the Atlantic to the Pacific, but its ancient winter-quarters of North Africa are 

 still so far as we know, the only winter-quarters of the species. On the other hand, the 

 Curlew Sandpiper [Tringa subarquata) winters in South Africa, India, and Australia. As 

 no difference can be found in the Curlew Sandpipers wintering in these three locaUties, we 

 must infer one of two hypotheses, either that India and South Africa have only recently 



