Vol. 70.] THEOUGH THE AXDES OF PERU AND BOLIVIA. 27 



(b) The Mesozoic and Palaeozoic Rocks of the 

 Altaplanicie and the Line of Dioritic Intrusions. 



The vast covering of ashes and tuff described above gradually 

 thins out, and disappears about 3 miles bej^ond the bridge over 

 which the railway crosses the Mauri River. 



Its limit is marked by an abrupt change in the nature of the 

 landscape, the fiat-topped sides of the river-valley, cut through the 

 horizontal volcanic deposits, giving place to scarped hill-faces 

 formed by steeply-dipping stratified rocks. Before discussing the 

 latter in any detail, it is necessary to describe briefly the section as 

 far east as Coniri (see map, PL X, fig. 2). 



At the point where the Mauri tuffs are last met with appears 

 a series of red gypsiferous sandstones, cropping out from beneath 

 the volcanic rocks under which they dip westwards. These sand- 

 stones can be traced in a series of folds, past the village of Calacoto 

 across the Desaguadero River to Comanche, where they are broken 

 through by a line of dioritic intrusions. About 5 miles beyond 

 this point they are seen to overlie, with apparent conformity, a 

 second series of darker chocolate-coloured sandstones and conglo- 

 merates, which continue as far as Coniri, where they end abruptly 

 along a fault-line against vertical shales and vellow sandstones of 

 Devonian age. 



This thick series is remarkable for the complete absence through- 

 out (or, at least, extreme rarity) of any organic remains, with the 

 exception in one or two localities of indeterminable fragments of 

 plants and a few derived fossils in the Coniri conglomerate. In 

 consequence, there has been considerable difference of opinion as to 

 the correct determination of its geological age. 



A. d'Orbigny colours the beds as Devonian. Carboniferous, and 

 Trias ; Forbes, comparing them with similar deposits in Europe, 

 assigns them to the Permian or Trias ; Prof. Steinmann considers 

 them as the equivalent of his fossiliferou* Cretaceous ' Puca ' 

 Sandstone of more southern localities, while many local mining 

 experts regard them as of still later date. With the exception of 

 D"Orbigny's classification, the general tendency has been to place 

 all these beds in one continuous series, but the view that I have 

 adopted since studying the country farther north is that they are 

 divisible into two groups, which may be defined briefly as follows: — 



(1) A younger, red gypsiferous sandstone and marl series of Cretaceous 



age, resting with pseudoconforaiity upon 



(2) an older, chocolate-coloured sandstone and conglomerate series of 



probably Permian or Permo- Carboniferous age. 



(1) The first group consists for the most part of brick-red and 

 crimson sandstones containing abundant gypsum and rock-salt, 

 variegated red and tea-green marls, and thin paper-shales, the whole 

 series being strongly reminiscent of the British Trias and evidently 

 formed under somewhat similar conditions. 



This fact at first led me to regard Forbes's determination as the 

 correct one ; but, after working northwards through Peru, and noting 



