I90I] TOWARDS THE CAPE 71 



On the rocky shore we scattered in various directions, some 

 of us climbing to a Hne of tree-ferns 1,200 feet above the sea ; 

 and the day passed pleasantly as we rambled about in search 

 of specimens of life peculiar to the island. Little of novelty 

 could be expected from a stay of six hours, but we had the 

 satisfaction of finding a few species new to science, of which 

 perhaps the most important was a new petrel, afterwards 

 named '■ /^strelata Wihoni^ after our zoologist. Dr. Wilson. 

 We left South Trinidad the same night and steered to the 

 south to get into the region of westerly winds. On the i8th 

 our coal supply was getting so short that I decided to proceed 

 under sail and husband what remained of our limited stock. 

 The wind proved very fitful, but by keeping well to the south 

 we received it from the westward, and made slow but sure 

 progress towards our destination. 



On October 2 we arrived within 150 miles of the Cape, and, 

 getting up steam, rounded Green Point and entered Table Bay 

 at four o'clock on the 3rd. On the 4th we refilled our bunkers 

 with coal, and that night put to sea once more, to take up our 

 quarters off the naval station at Simon's Bay. It was during 

 this short passage that we first appreciated the ' Discovery's ' 

 ability to roll : on meeting a heavy swell off the Cape Penin- 

 sula during the night, our small ship, without any sail to steady 

 her, was swung from side to side through an angle of 90°, and 

 as some of our furniture was not well secured, chaos reigned 

 below and discomfort everywhere. 



The main object of our stay at the Cape was to obtain 

 comparisons with our magnetic instruments. The instruments 

 which are used in a ship for taking observations of the various 

 magnetic elements are unfortunately subject to change, and 

 consequently the observations at sea are of little value unless 

 such changes are known. Whenever it is possible, therefore, 

 the sea instruments are compared with absolute values on land, 

 and by this means the sea observations are corrected. The 

 observations to be taken on our voyage to New Zealand were 

 an important part of our magnetic survey, and it was highly 

 desirable that the errors of the instruments to be used should 



