If we sail straight across from Northern Scandinavia to Greenland, we traverse an exceedingly deep marine basin; but if we examine the sub-marine bank which runs all along the coast of the former country from south to north, we find that it does not end when the extreme north of the land is reached. The bank extends much farther north, and is continued as far as Spitsbergen. As I have said before, the latter, as well as Bear Island, must be looked upon as the remains of a large mass of sunken land--the ancient Scandinavia stretching far into the Arctic Circle. Professor Nathorst speaks of Spitsbergen as a northern continuation of Europe, not only geographically, but also botanically and geologically. However, this northern land must have stretched even farther--not perhaps farther north, but farther west. Here lay the old land-connection between Scandinavia, Greenland, and North America (Fig. 13). One of the highest authorities on the geographical distribution of plants, Professor Engler, maintains that the arguments in favour of this Arctic connection of America with Europe are more weighty than those for a land-bridge between Greenland, Iceland, the Faroes, and Great Britain. Moreover, he is of opinion that a certain number of species of plants belonging to the Alpine flora of Arctic Siberia have travelled from Scandinavia _viâ_ Greenland and North America to Eastern Asia, and not direct from Scandinavia to Siberia (p. 143). [Illustration: Fig. 13.--Map of Europe, indicating approximately the distribution of land and water during the earlier stages of the Glacial period--shortly after the period represented in Fig. 12, p. 156. The darkly shaded parts indicate the areas covered by water, and the white portions what was land at the time.] That this ancient Arctic land-connection existed almost throughout the Glacial period appears to me probable. It has often been suggested that such a land-barrier was one of the principal causes of the production of the glacial phenomena in Europe, and as such it must have existed intact certainly during the earlier stages of the Glacial period. The barrier must then have gradually subsided in one or two places; and once a breach was formed, the complete union between the Atlantic and the Arctic Oceans could not have been long delayed.