92 DR. A.. HOLMES OX THE PRE-CA.MBRIA.N [vol. lxxiv, 



is probably immediately demonstrable, for Falconer ascribes the 

 elliptical outline of the hills to the modification of original phaco- 

 lithic intrusions (op. cit. p. 246). The generalization made by 

 Passarge implies that in Adamawa and in South-West Africa there 

 are similar differences between the rocks of the two land forms- 

 To the Rovuma type, however, this generalization only applies in 

 special circumstances. Throughout German and Portuguese East 

 Africa it is rather the exception to find the inselberg rocks differing 

 markedly from their neighbours of the surrounding plains. 



The passage of banded gneisses into granite in certain peaks 

 (as in the Etipoli Hills) and in the gneissic anticlines or elongated 

 domes of others (as, for example, the Mwima and Chica Ranges ;. 

 see PI. XI) are clearly cases where the discriminating hand of 

 denudation has picked out slight differences of structure and 

 mineral composition, and has revealed them in the exaggerated 

 form of an inselberg landscape. The whole country abounds with 

 evidence that strongly -foliated or banded rocks lend themselves to 

 attack more readily than granulitic rocks, and that mafic rocks and 

 minerals are removed more rapidly than their felsic associates. 



A few isolated peaks and domes, and some of the individual 

 peaks of the mountain-blocks, differ from their surroundings in 

 being built mainly of a grey granulitic or porphyritic granite. 

 Figs. 12 & 13 (pp. 90-91), for which I am indebted to my friend. 

 Mr. E. J. Wayland, illustrate the remarkable appearance of some of 

 these granitic inselberge. This second type is matched in Nigeria, 

 for Falconer {op. cit. p. 70), in describing the rocks of Borgu, says : 



' The occasional inselberge are composed invariably of grey porphyritic 

 granite.' 



A third type of inselberg is also considered to owe its survival 

 to granite intrusion ; in this case not to the exposure of ancient 

 stocks or granitic foci, but to the riddling of the gneisses with 

 tongues and apophyses that are probably connected with a more 

 continuous form of intrusion below. Many inselberge and groups 

 of peaks seem to be identical with the surrounding country as 

 regards the composition, structure, and dip of their gneisses, the 

 foliation being independent of the contours of the hills. They 

 differ, however, in being scaffolded and fortified by quartz-felspar 

 veins of extraordinary abundance. The Mhala Hills of Nakavala 

 may be taken as a t} T pe of such inselbftrge. 



There are, then, in Mozambique at least three kinds of insel- 

 berge that owe their survival to peculiarities of structure or 

 composition : — 



(a) The gneissic dome or anticline type. 



(b) The granulitic-granite type. 



(c) The injection type. 



In each case the greater resistance offered to denudation is due 

 to the presence of less foliated (or unfoliated) and more felsic 

 rocks than are found in the adjacent areas. It may be that the 



