THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 82 I 



degraded forest soil on clay. On the ridges still covered with virgin forest the 

 rock under the dark brown humus layer has been converted into a clay through 

 the weathering of the basalt and presumably also through the corroding action 

 of the roots. This profile was, for instance, observed on the steep slopes of Valle 

 Frances in 500 m altitude, and on the ridge separating Quebrada de la Piedra 

 Aguiereada from O. Laura immediately under the narrow crest. Even where the 

 gradient attains 40 — 45°, the forest may form a closed stand, as on both sides of 

 the Portezuelo pass; on the precipitous slope of the Villagra side, about 510 m 

 above sea-level, where the plant-cover was analyzed, the rock was a brick-red 

 clay-stone, easily cut with a knife, with the surface Cull of cracks and hollows 

 where a thin humus layer had accumulated, kept together by roots and fern 

 rhizomes. The trees must be well anchored in the fissures, for few signs of land- 

 slides were seen anywhere. More or less the same conditions were observed on 

 the upper slopes of Cordon Salsipuedes near the main ridge, whereas the hanging 

 gullies near the sea and in a lesser altitude have a drier, thinner and more sandy 

 soil supporting a poorer type of forest. 



In the fog-belt Dicksonia forms pure stands, quite locally on Masatierra, but 

 of great extension on the slopes of Los Inocentes on Masafuera. The dead trunks 

 and fronds give rise to a thick layer of peat, saturated with water, and the foot 

 sinks deep between the decaying stems. 



On the almost perpendicular escarpment of the summit ridges vascular plants 

 are entirely dependent on the narrow ledges, fissures and pockets for their ex- 

 istence. From a distance the vegetation frequently gives the impression of being 

 closed, but in reality it consists of stripes and patches separated by bedrock. 

 The same is true of the canyon walls of Masafuera. 



The table-land of Masafuera above the timber-line is inhabited by Anthoxan- 

 tJiian, the fern Lopliosoria and Gunnera as predominating species. As we approach 

 the border of the escarpment which forms the spectacular west wall of the island, 

 patches of heath make their appearance and put their stamp on the narrow fringe 

 above 1300 m. The country is strewn with boulders of all sizes and the ground 

 covered with stones, carpets of moss and lichen and mat-forming herbs and grass- 

 tufts. The humus layer is thin and mixed with stone fragments, the subsoil a clay 

 enclosing numerous stones. Here and there are shallow depressions with bare soil, 

 a light reddish brown, sandy clay, hard, dry and cracked after a period of dry 

 weather; I suppose that water stands here during the wet season in mid-winter 

 and that this is the reason why these places lack vegetation. 



Our slight knowledge of the soils of Juan Fernandez does not permit a safe 

 classification according to the soil system of modern textbooks. They would, 

 I think, be classified with Blanck's "Boden der feuchtwarmen gemassigten Zone 

 a) Gelberden und Gelblehme", but little information is given there on volcanic 

 soils derived from basaltic rocks. The term "yellow soil" and "yellow clay" does 

 not mean that the colour is always yellow, it may be brown or orange or approach 

 that of the lateritic red clays. The colour depends on the humus content and 

 on the mineralogical composition of the mother rock. 



Blanck's Handbook says: "Die Gelberden schliessen sich in den Mittelmeer- 



