should be made for a vessel with coal and fresh food to meet her there. Here the special object is to carry on the bathymetrical survey in the region where opinion is divided as to whether deep or relatively shallow water exists, namely, that portion cautiously marked in the Scottish chart (Scottish Geographical Magazine, vol. xxi, 1905, pp. 402–412) lying between the south end of the "Scotia Rise" and the Sandwich Group. This is of vital importance in the study of continental connections. A short time would be spent in the South Sandwich Group, especially with a view of obtaining a knowledge of the geology and natural history of the islands. Having filled up with coal, a cruise eastward to Bouvet Island should be made to determine more definitely whether or no there is a "rise connection" between the Sandwich Group and that island, and also with the south end of the "Scotia Rise." From Bouvet Island a southerly course should be steered towards the southern boundary of the Biscoe Sea and a thorough connection made between the Valdivia and Scotia bathymetrical surveys. In March it would be necessary to decide whether the expedition was to winter in the south, but in no circumstances, if it can possibly be avoided, should the ship winter. She is there for oceanographical research, and must not be turned into a harbour hulk. Acci-
