Rev. B. Baron — Bock Cavities in Granite. 17 



will be less favourable and have less chance of surviving. This 

 will account for the greater length of time the Palseozoic species 

 existed compared with the Mesozoic species, as the seas would have 

 greater extent and be less broken up by land. 



III. — EocK Cavities in Granite in Madagasoak. 



By the Rev. Richard Bakox, F.L.S., F.G.S., Tananarive, Madagascar. 



I WAS much interested in the account of the ' erosion ' of rocks in 

 Corsica given by Mr. Tuckett and Professor Bonney respectively 

 in the January and August numbers of the Geological Magazine 

 for 1904, more especially as I have been long acquainted with similar 

 phenomena in Madagascar, and have, equally with the authors 

 mentioned, been long puzzled to find an explanation of their origin 

 at all satisfactory. 



Perhaps a brief statement of the occurrences of such rock ' erosion * 

 in the form of cavities as I have met with in Madagascar may be 

 worthy of a place in the Magazine, as being another small con- 

 tribution to the subject. 



I shall first of all give, in very few words, a description of the 

 rock in which the cavities occur, and then proceed to a short 

 account of a few of the most striking of the cavities themselves. 



The rock (apparently a boss) in which these cavities occur is 

 porphyritic granite, occupying, roughly speaking, a circular area 

 of, say, fifteen or sixteen miles in diameter, and weathering, as granite 

 usually does, into small and large, more or less rounded or oval- 

 shaped hills, which are scattered all over the surface. The western 

 edge of the mass is some ten or twelve miles to the east of 

 Antananarivo, the capital of the island. 



The rock is, strictly speaking, a porphyritic hornblende-granitite 

 (i.e. hornblende-biotite-granite). In colour it is grey, blotched, and 

 streaked with black. Fully three-fourths or four-fifths of its bulk 

 is felspar. This occurs in the form of porphyritic crystals, the 

 longest of which attain sometimes two inches in length. The felspar 

 includes orthoclase-microperthite (pure orthoclase is apparently 

 quite absent, or at any rate rare), microcline, and another plagioclase 

 (? oligoclase) ; these exist in not very unequal proportions. The 

 felspars often show the central accumulation of extremely minute 

 dusty particles, more frequently seen perhaps in more basic rocks 

 (gabbros and norites especially). Micropegmatite, as well as 

 ordinary quartz-felspar mosaic, occurs as a coating round some of 

 the grains. Both the felspar and the quartz (the latter with abundant 

 planes of gas and liquid inclusions) show signs of some amount 

 of molecular strain. The mica is ordinary brown biotite, and the 

 amphibole ordinary green hornblende. These occur in irregular 

 flakes and grains, but are often so intimately associated as to lead 

 to the supposition that the one has been derived from the other. 



As accessory minerals, occur sphene (pretty abundant), magnetite 

 (often intergrown with iron pyrites), apatite, zircon, and calcite, 

 possibly also a few minute garnets. 



DECADE V. — VOL. II. — NO. I. 2 



