part 1] GEOLOGICAL SECTIONS THROUGH THE ANDES. 35 



and is remarkable for its pronounced zonary structure : a characteristic 

 feature of the Comanche rock (see op. cit. pi. vii, fig-. 5). It tends to be 

 porphyritic in habit, having crystallized in general before the hornblende ; 

 numerous small crystals are also frequently enclosed in poecilitic fashion 

 by plates of the latter. 



A yellowish-brown biotite is present, in addition to the green horn- 

 blende ; they both, show ragged outlines, and are but seldom idiomorphic. 

 This, however, is a feature apparently due to subsequent partial solution 

 in the magma, after they had assumed the crystalline form. 



Magnetite is fairly abundant, and there is also a considerable amount 

 of calcite. 



Apatite occurs sparingly in small prisms. 



Although the rock has undergone so profound a modification, all 

 the essential underlying features of the Comanche diorite can at 

 -once be recognized. 



The absorption of so large an amount of extraneous material 

 would appear to suggest a batholitic mode of intrusion, rather than, 

 as appears to be the case where it penetrates the Cretaceous lime- 

 stones, one involving actual displacement of the strata. The latter 

 rocks, however, are intensely folded, while the underlying Devonian 

 quartzites, although dipping steeply, approach the diorite without 

 any sign of such disturbance. 



It is conceivable, then, that the magma gradually stoped its waj r 

 .by batholitic replacement through these ancient and resisting rocks, 

 becoming much acidified in the process, and on reaching the over- 

 Iving Cretaceous deposits, which, if we judge from the neighbouring 

 districts, were at that time little folded, it encountered a less resis- 

 tant mass capable of being removed from its path by the more 

 drastic and rapid method of bodily displacement. 



Much of the silica, derived according to this theory from the 

 Devonian quartzites, must have become dissolved in the igneous 

 magma, thereby rendering it increasingly acid, and causing a partial 

 resorption of those minerals which had already begun to crystallize : 

 of these the hornblende and biotite were most readily attacked, and 

 their idiomorphic contours have been almost entirely destroyed. 



Finally, the residue of the acidified magma consolidated as an 

 intergrowth of quartz and plagioclase, the latter of a rnore acid 

 variety than the phenocrysts. The rounded outlines of many of 

 the quartz-grains, however, make it probable that they represent 

 original sedimentary material that has been derived and corroded, 

 but not fully dissolved. 



The alluvial pampa of Mara villas ends in a somewhat narrow 

 river-gorge, flanked on each side by cliffs of limestone and crimson 

 marl ; a short distance farther on, however, the valley once more 

 expands to form the pampa of Taya Taya, and from this point to 

 Lake Titicaca the fall of the river is very slight (about 1 in 1250). 



On entering the pampa we meet for the first time with Palaeozoic 

 rocks. These are of Devonian age, and form an unbroken series 

 on the north of the valley as far as Cabanillas, with an almost con- 

 stant dip of 45° east-north-eastwards (see PI. V). 



d2 



