A GEOGRAPHICAL SKETCH OF THE JUAN FERNANDEZ ISLANDS I I 9 



sondern urn eine teilweise erhaltene Kraterbildung handelt." Finally, "Ob das west- 

 lich von dem Portezuelo sich offnende cirkusformige Tal audi eine Kraterbildung 

 darstellt, lasst sich nicht entscheiden. In einer vorlaufigen Mitteilung habe ich diese 

 Moglichkeit angedeutet, und vieles scheint mir darauf hinzudeuten, dass urspriinglich 

 ein Doppelkrater hier gestanden hat, wobei gerade der enge Rucken des Portezuelo 

 die Scheidemauer bildet." This was written after our visit in August 1908, and at 

 that time I shared my companion's opinion (i p. 136). During my later expedition, 

 when I came to know most of the island better, I arrived at a different conclusion 

 (see e.g. 2 p. 57). The geological structure is the same all the way from Pto Frances 

 to Bahi'a Juanango, we find the same agglomerate beds more or less well exposed 

 in the valley bottoms, but nobody would be inclined to deny that they are typical 

 erosion valleys where denudation, thanks to the scant supply of running water, is 

 very slow; in many cases abrasion has worked faster than erosion. In Pto Frances 

 the stream has barely managed to lower its bed down to sea level near the coast. 

 In Pto Ingles, where the surrounding ridges are very high, the distal part of the 

 valley has been widened and levelled, and quite some alluvial soil has been formed. 

 The same procedure is, I think, responsible for the formation of the Cumberland 

 valley system; there is nothing that speaks in favour of a crater theory or against 

 its origin as the result of erosion. The streams drain the wettest part of the island, 

 where erosion is greater and faster and accumulation a factor of some consequence. 

 As I said, my first impression was that the Colonial valley was the old crater and 

 its wall formed by Cordon Central, Mt. Piramide and the Salsipuedes ridge, and 

 I think this also was what glimmered in Quensel's mind. However, the thick, strati- 

 fied deposits of brick-red, deeply weathered tuff with "Bruchstiicke von Olivin- 

 krystallen, Lapilli-artige Lavabruchstiicke, Erzkorner und Glas" also occur in 

 Anson's valley, separated from Valle Colonial, the supposed crater, by the over 

 500 m high Cordon Central which runs down to the shore of the bay. I mentioned 

 above that my photograph of the section through an agglomerate bed published 

 by Quensel was taken in Anson's valley. Either this is another broken-down crater, 

 or the deposits are ejections from the "Colonial" volcano. This would make us 

 postulate that the Anson valley had been excavated even to a lower level than the 

 present one when the volcano was active, and this seems questionable. In his 

 second paper Quensel quotes his earlier discussion (p. 53) and adds: "It is over 

 40 years since I visited the locality and naturally I cannot now rely on any personal 

 recollection", and "But the composition of the formation, as well as my notes from 

 the field, offer indications that the tuffaceous material of Bahia Cumberland also 

 represents pyroclastic sediments of recent volcanic origin". He refers to the sub- 

 marine eruption 1835 off Pta Bacalao. I willingly admit that I lack the necessary 

 training to get to the bottom with a geological problem, but my observation near the 

 foot of Mt. Yunque (see above p. 99) suggests that the agglomerate bed underlying 

 the basalt is of the same nature as the formation in Cumberland Bay. With regard 

 to Villagra, there is no semicircular valley corresponding to the opposite one, for 

 only on the north and east sides are ridges resembling a crater wall. Until a geo- 

 logist-vulcanologist has had an occasion to study the island it is better to leave the 

 question of the nature of Cumberland Bay open. 



