A GEOGRAPHICAL SKETCH OF THE JUAN FERNANDEZ ISLANDS I 39 



On Anson's map of 1741 Santa Clara is shown as covered with some kind 

 of arboreous vegetation. Walter does not tell if a landing was made, but the 

 name Goat I. shows that there were goats on the island, and Ulloa's map of 

 1742 leaves it treeless. Just as Johow we found the place very barren and the 

 vegetation mostly formed by weeds with Ai/ejia in dominance, all very dry 

 during the summer. When Giinther says that there were "algunos arboles en la 

 pendiente del este" he either did see some specimens o^ Dendroscrjs on iha c\\f(s 

 above the water or refers to Morro Spartan (also called M. de los Alelfes; 

 fishermen are said to have collected seeds of Matthiola in B. del Padre and to 

 have scattered them on Santa Clara). Relics of the original flora found a refuge 

 on the morro, which is separated from the island by a channel less than 10 m 

 wide and blocked by huge boulders which are exposed at low tide. The current 

 rushes through the narrow channel, and to judge from the vegetation the goats 

 are unable to cross. I believe that Johow (p. 261) was right in assuming that, 

 before the introduction of this pest, Santa Clara must have looked much like 

 M. Spartan. The altitude of the island is sufficient to catch the trade clouds, as 

 shown on fig. 4, and we are told that rains are not unfrequent during the winter 

 months (Johow p. 261), giving origin to an abundant vernal flora of annuals. 

 During our brief visit in January we did not see any trace of water; but we 

 read in Instrucc. naut. p. 230 that a stream leaps over the cliff at the NW point, 

 a statement repeated by Guzman who calls it Chorro de Doiia Maria (p. 50). 



To judge from our passing observations, the geology is the same as of 

 western Masatierra, uncounted, light yellowish or brownish volcanic beds tra- 

 versed by numerous vertical dikes of a hard gray basalt. I regret having neglected 

 to bring samples of the lava beds for comparison with the rocks from Puente, 

 The dike rock was described by Hagerman p. 28 and found to be identical with 

 the basalt from the top of Co Negro. The sample came from a dike on M. Spartan. 

 We landed at the foot of this dike which forms a flight of steps leading to the 

 table-land of the morro; see Skottsb. 3 p. 924 and PI. 103. The dikes project 

 as flat slabs above the softer beds. 



Masafuera, former surveys and maps. 



Until our visit in 19 17 Masafuera was much less known than her sister island. 

 It had been inhabited more permanently only during the period of the penal 

 settlement 1909 — 19 13, a misfortune that befell this ocean castle a second time 

 in 1927 — 1930. Only few scientists have visited Masafuera. 



Of the early navigators few paid a visit to this island. One of Commodore 

 Anson's captains, on his way to Cumberland Bay, happened to come up under 

 Masafuera and reported that the island was not, as former navigators had imagined, 

 a barren rock, but "almost every where covered with trees and verdure, and 

 was near four miles in length". He had not been able to land, but added that 

 "it appeared to him far from impossible, but some small bay might be found 

 on it, which might ofl"er sufficient shelter for any ship desirous of refreshing 

 ashore" (Walter p. 134). As four ships of the squadron were missing the Com- 



