DESCRIPTION OF THE ISLANDS 147 



Wind and set. The current in the Southern Ocean from lat. 50° to 60° S and long. 

 24° to 42° W has proved to be north-easterly and approximately 12 miles per day: it 

 is constant and can be depended upon. The prevailing winds are north-west and south- 

 west; both blow with violence, the former coming with a falling barometer and the 

 latter when it is rising. As a rule neither blows for more than four days and is im- 

 mediately followed by the other. Swell and sea rise very quickly. At the South Sandwich 

 Islands the wind was mainly north-west or south-west, but sometimes it blew from the 

 north-east, bringing fog and mist. 



Contrary to that experienced at South Georgia the set round the islands is clockwise. 

 This we attribute to the trend of the coast-line, which bears more north and south than 

 at South Georgia, thereby deflecting the usual north-easterly set to the northward and 

 giving a clockwise motion. At the Candlemas and Southern Thule Groups, where two 

 or more islands lie close together, a strong southerly set of about i knot was experienced 

 in the channels between them (Nelson and Douglas Straits). 



Magnetic variation. The opportunity of swinging ship for magnetic variation oc- 

 curred but once during our survey, and then, though taken advantage of, was not as 

 thorough as we might have wished. This was at Montagu, when, during a short clear 

 period between snowfalls, we swung (to port only, unfortunately) and took observations 

 for compass error on 16 points. We obtained a variation of 10° 31' W, and think this 

 may be taken as fairly accurate, though the continual rolling of the vessel made observa- 

 tion rather difficult. 



Local attraction. Compasses were adjusted on February 21 when leaving South 

 Georgia, and with observations made when the ship was steady, the standard compass 

 showed little or no deviation. Later on, under similar conditions at the South Sandwich 

 Islands, and allowing the variation given on chart no. 3775, it was likewise found that 

 little or no deviation still existed. This, we consider, may be taken as evidence that 

 magnetic conditions are quite normal. At no time was any local attraction observed. 



Wireless time-signals. Reception of wireless time-signals was found to be very good 

 at all points in the group with the exception of Bristol and Southern Thule. Signals 

 received were mainly from Washington, but Nauen was heard on one or two occasions. 



Anchorages: Landings. The islands are without harbours, but at most of them there 

 are anchorages where a vessel may obtain some shelter. The best is Ferguson Bay in 

 Thule Island, but since this is in the extreme south it will very often be inaccessible 

 on account of ice. The next best are in Nelson Strait in the Candlemas Group ; in 

 these protection from any wind can be obtained and they are much more likely to 

 be free of ice during the summer months. Even in the finest weather there is a heavy 

 swell and the coast is generally precipitous and rocky: where beaches exist they 

 shelve steeply and there is a strong undertow. As a consequence landings can only be 

 made with difficulty, and on some islands they are probably impossible. Water could 

 no doubt be obtained on any of the group by melting snow or ice. Owing to volcanic 

 heat the ground on some of the islands is warm, and on one — -Vindication Island— small 

 streams from the melting glacier run down to the shore. 



