A GEOGRAPHICAL SKETCH OF THE JUAN FERNANDEZ ISLANDS I 41 



a very irregular rock close to the beach in Loberia. The promontory in the 

 middle is Buque Varado. Groups of trees are seen only above this place and 

 above Loberia Nueva. 



Ulloa found the island very inhospitable: "La Isla de afucra de Jiam Fer- 

 ftajtdea es toda muy alta, y tan escarpada, y escabrosa que no tiene parage 

 conmodo para desembarcar" (p. 287). On a former occasion (3 p. 796) I quoted 

 the narratives of Byron and Carteret. The island was well wooded in 1765. 



The hydrographic expeditions despatched by the Oficina in Valparaiso paid 

 little attention to Masafuera. When Lopez reports that the island was "cubierta 

 de arbolado" in 1875 (p. 6^) he must have included the Dkksonia jungle on 

 the slopes of Mt. Inocentes. Johow (p. 96) asserts that the luma was common 

 all over the island except near the. coast and in the highland, but this does not 

 necessarily mean that it formed extensive woods. We find the same statement 

 in Instrucc. naut. of 1896. It is not known when the destruction of the forest 

 started in earnest, but it is probable that the exploitation of the sandal-wood, 

 which led to its complete extermination, had serious consequences for the forest 

 as such. The names of two valleys and a place on the south coast testify that 

 a species of Sanialuni grew on the island. The possibility that forest fires have 

 ravaged the woods must not be forgotten. Foreign grasses spread into the cleared 

 spaces and prevented the germination of the seeds of the native trees. The 

 direct influence of the goats, introduced, I believe, during the 17th century, 

 remains to be found out. They greedily devour the arboreous Compositae and 

 the endemic herbs, but I cannot tell if they eat the leaves of the luma, naranjillo 

 and canelo. I can testify from my own experience that there was better forest 

 in 1908 than in 191 7, an undisputable consequence of the activity of the con- 

 victs during the intermediate period. Nevertheless Giinther (1920) repeats the 

 old statement that Masafuera was covered with trees. 



In 1895 the Oficina Hidrografica published the first map showing the prin- 

 cipal topographical features. New dates had been provided by Johow. During 

 our visit in 191 7 we soon discovered that this map was too defective to be used 

 as a basis for a map of the vegetation, and we did what we could to correct 

 it. The result was a sketch put at the disposal of the Oficina, which used it 

 for a new chart. When the flight over the islands was undertaken in 1952, Masa- 

 fuera was covered by clouds. A new map is urgently needed. 



Main geographical features. 



In shape and general appearance Masafuera differs profoundly from her 

 sister island. The island (fig. 54) is a solid rectangular block with slightly rounded 

 corners, tilted NE so that the escarpment along the west side is very much 

 higher than on the east side, where it is quite respectable (figs. 55 — 57). I he 

 table-land is traversed by numerous parallel deep valleys running NK — L (figs. 

 57, 58), the high west wall by a number of precipitous gorges; this side has 

 much the same appearance as the south side of Masatierra E of Mt. Yunque 

 (fig. 52). 



