K P. Mennell — ConstitHtion of Igneous Rocks. 213 



show liow largely exceptional types are represented by analytical data 

 among rocks of all classes. No one is likely to dispute the enormous 

 preponderance of felspathic rocks over those containing the so-called 

 felspathoids, yet if we were to measure their abundance by the number 

 of analyses that have been made we should arrive at a very different 

 conclusion. But this is precisely the method we are asked to accept 

 as correct. I do not know what would be thought of a mining 

 engineer who took all his samples of a gold reef from the points where 

 he thought it likely he would get exceptional results, and then 

 calculated the average value on such a basis, especially if he left out 

 of account the widths which his samples represented ! 



In my former paper I only gave the data from a very limited 

 area mapped in Rhodesia, and though this is representative of the 

 conditions throughout the whole country, it may be well to give 

 the results arrived at from the mapping of an area ten times as 

 great in central Matabeleland. Of 20,000 square miles, approximately 

 2,340 are sandstone, 5,700 Archaean schists, etc., and 11,960 granite 

 and other igneous rocks. These last comprise portions of a number 

 of plutonic masses, among which are two developments of syenite 

 with about 63 per cent, of silica, covering some 40 square miles. 

 There is also an important intrusion of a pyroxenic rock, which may 

 be regarded as largely picrite and as containing on an average 

 45 per cent, of silica. It is probably a gently inclined sheet injected 

 along a thrust plane, and its outcrop is much wider therefore than 

 its true thickness. I have, however, taken the appai'ent width as 

 the real one in order not to unduly favour my argument. Nearly 

 all the rest of the igneous area is granite, with an average silica 

 percentage of at least 70. It must not be thought that there are 

 few basic rocks ; on the contrary, thej' are well represented by 

 dolerite dykes, basalt flows, etc. We shall nevertheless be making 

 a generous allowance for them if we suppose there are 10,000 dykes 

 a mile long and 5 feet thick, and 100 square miles of basalt 20 feet 

 thick. We may further assume that they contain about 50 per cent. 

 of silica. The volume of the dolerites maj' be subtracted from that 

 of the granites, in which they are commonly intrusive. The basalts 

 are intercalated among the sandstones that have been referred to. 



Not to make too great an assumption regarding what may happen 

 at a depth, we will suppose that each dyke and plutonic mass has 

 vertical boundaries (though we have evidence that the latter class 

 tends to spread out and encroach on the schists below the surface). 

 We may then calculate the bulk of the various classes down to sea- 

 level (i.e., nearly 5,000 feet, or, say, roughly, 1 mile) as follows: — 



Total . 11,960 



