Vol. 70.] THEOUGH THE ANDES OF PERL* Al\D liOLIVTA. 



25 



and mode of fossilization with numerous specimens preserved 

 in the British Museum (Natural History) from the Santa Cruz 

 Beds of Patagonia, which are of Miocene age. A horizontal section 

 of the silicified tree has been kindly examined for me by Miss N. 

 Bancroft, who reports that the wood is that of a Dicotyledon and 

 appears to agree fairly closely with one of the Lauraceae, being 

 very similar to Cinuamomum glanduliferurn and also to Nectan&ra 

 from British Guiana. According to Engler & Prantl, Brazil was 

 one of the principal centres of distribution of the Lauraceae, and 

 their appearance in Bolivia would therefore not be improbable in 

 Miocene times. It would be unsafe to base any definite conclusion 

 as to the date of the deposit from a single specimen of this 

 nature ; but the occurrence of Nesodon is sufficient to prove that 

 most of the volcanic beds of the Mauri River were formed 

 during or subsequent to Miocene times, and, as it will be Shown 

 later that they are overlain by river-gravels containing numerous 

 remains of Pleistocene mammalia, a sufficiently approximate 

 estimate of their age can thus be arrived at. 



These Miocene or post-Miocene volcanic rocks are occasionally 

 found to have been penetrated by still later igneous intrusions, 

 which occur in the form of dykes or small necks cutting through 

 the overlying beds. In common with the sequence noted from 

 the coastal region, these later rocks have, as a rule, a more basic 

 composition, acid dykes being of somewhat rare occurrence. 



There are only two localities, where such intrusions are met 

 with, that are of sufficient interest to deserve mention. The first 

 is situated on the River Cano, about a mile above its junction with 

 the main stream. Here, pale tuffs of the Mauri Volcanic Series 

 overlie steeply-dipping beds of porphyritic conglomerate, the latter 

 being composed of well-rounded pebbles which consist chiefly of 

 basic igneous rocks. In the narrow railway-cutting at this point, 

 both are seen to be cut across and metamorphosed by a broad 

 dyke of rhyolite, with a chilled vitreous margin and well-marked 

 flow- structure. 



The rock}* outcrop of this dyke can be traced for some distance 

 south-eastwards. On the opposite side of the river rises a rounded 

 hill, having the appearance of a small volcanic neck, composed of a 

 •coarse porphyritic andesite, which has the following distinctive 

 characters : — 



(A 31 ) Andesite; River Cano (Bolivia). (PL VII, fig. 6.) 

 Macroscopic characters: A compact, dark reddish-brown 

 rock, containing big porphyritic crystals of glassy clear felspar, 

 hornblende, and augite in a microcrvstalline base. 



Microscopic characters: Plagioclase, near andesine or 

 acid labradorite, occurs sparingly as large clear porphyritic pheno- 

 •crysts, and abundantly as a second generation in the ground-mass. 

 Hornblende is abundant as large idiomorphic phenocrysts of 

 a dark brownish-green with strong pleochroism, often twinned 



