THE PETROLOGY OF SOUTH GEORGIA. 833 



mountain-range of Graham Land, known as the " Antarctic Andes." According to 

 Suess,* the Andes, after turning to the east in Tierra del Fuego, form a great 

 eastwardly directed loop, of which the West Antarctic islands are the unsunken 

 remnants. This loop passes through South Georgia, and returns through the South 

 Sandwich Islands and the South Orkneys to Graham Land, where the trend becomes 

 north-east and south-west. This loop of islands Suess considered to be homologous 

 to the Antilles, connecting the tectonic lines of North and South America, and he 

 therefore proposed to call them the " Southern Antilles." 



In an interesting discussion of this question 0. Nordenskjold t points out that, 

 in contradistinction to the Central American Antilles, there is here no continuous 

 island-chain, and that no submarine connection exists between the islands, although 

 this fact does not prove that they are geologically unconnected. Nordenskjold is 

 of the opinion that the South Georgian sedimentary series is not represented in the 

 continental mountain-ranges. J. G. Andersson told him that he had found tuff-rocks 

 externally similar to those of South Georgia at Harberton in Tierra del Fuego, but 

 Nordenskjold does not press this identification.^ According to the very few obser- 

 vations made, the South Sandwich Islands appear to be composed, of basaltic lavas of 

 recent extrusion. § 



The rocks of the South Orkneys, however, possess a decided affinity to some of 

 those of South Georgia. They consist of folded, massive, greenish or greyish grey- 

 wackes, with coarse conglomerates, slates, and shales, in which a graptolite, Pleuro- 

 graptus, has been found, thus establishing their Silurian age.|| These rocks are 

 very similar to the arenaceo-argillaceous series of South Georgia, especially the 

 lower division of the Cumberland Bay Series, as I have had the opportunity of 

 proving by the examination of thin sections of some rocks collected by the Scotia 

 Expedition. Amongst these rocks was a small rounded pebble of a typical spilite, 

 a lava-rock which is characteristically associated with Palaeozoic sediments. 



The eruptive rocks of Graham Land are strikingly analogous with, those of the 

 Andes.1I The sedimentary rocks of the mountain-ranges of this region, according to 

 Nordenskjold, form a connected series of greywackes or sparagmite-like rocks with 

 clastic texture, consisting of mineral and rock-fragments, hardly any of which are 

 eruptive, embedded in a carbonated groundmass. These are followed by fossiliferous 

 dark shales and light-coloured tuffs. As shown by plant-remains at Hoffnungs Bay, 

 these rocks are of Jurassic age. As far as the brief petrographical descriptions go, 

 these rocks are similar to some of those of South Georgia, which, according to the 



* E. Suess, The Face of the Earth, English trans., vol. iv, 1909, p. 489. 



t Die Schicedisch Sud-Polar Exped. und Hire Geographische Tatigkeit, 1911, pp. 211-13. 



| Und., p. 212. 



§ Ibid. 



|| J. H. Pirie, "On the Graptolite-bearing Rocks of the South Orkneys," Proc. Roy. Soc. Edin., xxv, 1905, 

 pp. 463-70. 



1 0. Nokdenskjold, " Petr. Unters. a. d. Westantarctischen Gebicte,''' Bull. Geol. hid. Upsala, vi, 2 (1900), 

 pp. 234-40 ; also see " Antarctis,'' Handbuch der Begionalen Geologie, Band viii, 6, pp. 5-9. 



