MR. O, A. DEIUJY OX MCrin^LIXK-UOCKS IN BRAZIL. 2511 



entirely subacrial, cut tliroiigli tlie upper and iniperfectly-con- 

 sctlidiited deposits of ii volcanic cone. It- is interest in f>- to note 

 that this Caldas rock shows the same tendency to phonolitic struc- 

 ture and the same polyhedral segrcijations (pseudo-crystals), to 

 be described farther on, as does the Tingiia rock. It may also be 

 remai"ked that, but for the railway cutting the mountain-side from 

 top to bottom, the effusive types of eruptive rocks (tuffs and pliono- 

 litic lava-flows) would be no more prominent at the Caldas locality 

 than they are at Tingua ; while the extremely significant type of 

 basic eruptives (leiicitite) would have escaped observation altogether. 



The general fragmentary condition of the Tingua foyaites, 

 suggestive of a breaking-up through undermining of the eruptive 

 nuisscs, finds a ready explanation in the hypothesis above presented, 

 which presupposes an underlying mass of easily-decomposed and 

 presumably more or less incoherent material. In the marginal 

 zone where the broken-up foyaite rests upon gneiss, this non- 

 existent cushion may have been, in part at least, the soft material 

 resulting from the decomposition of the latter rock previous to the 

 eruption. It may be noted in this connexion that, so far as seen, 

 the present drainage-level, whether in the surface-streams or subter- 

 ranean channels, is always below the soil-cap with foyaite boulders. 



Several peculiarities of structure of the Tingua foyaites also 

 point in the same direction, indicating a comparatively superficial 

 origin. The type, as regards both mineralogical composition and 

 structure, is extremely variable. Herr Graeff, in his excellent 

 paper, has described a considerable number of varieties, all of which 

 were collected in a small area, and might be duplicated, with many 

 additions, at any point where the rock occurs. The four types of 

 composition established by him on the bisilicate element (horn- 

 blende, biotite, augite, and aegerine) are difficult to distinguish iii 

 the field ; and, from the mode of occurrence in loose blocks, the 

 deteimination of their mutual relations is out of the (]uestion. So 

 far as can be judged from a superficial examination, the more 

 purely hornblendic type is most characteristic of the rocks of nearly 

 uniform grain, the jegerine type of those that are most distinctly 

 porphyritic. As regards structure, all possible gradations, from a 

 normal granitic type of uniform or nearly uniform grain (always, 

 however, with a porphyritic tendency) to a very pronounced porphy- 

 ritic type, which only differs from phonolite in the coarser granula- 

 tion of the elements of the groundmass, occur mingled in inextricable 

 confusion. 



An irregularity of structure roughly suggestive of stratification 

 (SchHeren (?) structure) is beautifully brought out by atmospheric 

 weathering. AVhen exposed to water-action or embedded in the 

 soil, the blocks of foyaite are rounded like a normal intrusive rock. 

 "Where exposed to the atmosphere on the hillsides, however, there 

 is a very general tendency to become fluted, as represented in the cut 

 on the next page ( tig. 2). This peculiar style of weathering does not 

 appear to be confined to any ])articular type of structure or compo- 

 sition ; and on the worn or freshly-broken surface no appreciable 



