140 REPORTS ON THE STATE OP SCIENCE. 



On page 124, under b, the reference to ' mica schists,' &c. , might 

 well be omitted, as the South African rocks are almost exclusively 

 horn-blendic and usually massive. 



Re Swaziland System, there is no evidence whatever of its relation- 

 sliips to the rocks of other parts of South Africa, and, moreover, where 

 are we to find a reliable and detailed field and petrographical account of 

 the typical area to provide a basis for comparison? The Malmesbury 

 beds are far more likely to be equivalent to the Band group than to the 

 Swaziland series, provided the latter do not prove to be metamorphosed 

 representatives of the former. 



Re term Archaean, it should, of course, be used in the comprehensive 

 sense in which it was originally proposed, and include therefore such 

 rocks as those of the Eand. Any divisions of the older rocks should be 

 considered merely as subdivisions (systems) of the ArchaBan group. 

 The term Algonkian is not very euphonious — Huronian would be far 

 better if such terms are needed, which I doubt. It has also to be 

 remembered that, barring a few doubtful ' Basement schists,' in 

 Bhodesia, at least, all the crystalline ' basal complex ' are altered igneous 

 rocks intrusive in the schists of sedimentary origin. The same fact was 

 accepted by the International Conference of the United States and 

 Canadian geologists, I notice, lately, for the relations of the Laurentian 

 and Huronian rocks. 



Re other matters, why not shorten ' Witwatersrand ' to the far 

 more euphonious ' Band, ' which everybody really uses ? I agree with 

 Brofessor Schwarz re the Vaal Biver or Ventersdorp series ; as igneous 

 rocks they have no claim to the rank of a separate system. I must 

 confess, indeed, that I do not like to use names for ' systems ' which 

 are unknown to other parts of the world, though we must, of course, 

 have local names for purely local purposes. 



I consider it premature to subdivide the Bhodesian coal-beds, as the 

 divisions made by Mr. Molyneux at Sengwe cannot be recognised even 

 at Wankies, which is a very short distance off, as things go in South 

 Africa. There also appear to me the strongest objections to including 

 the Forest sandstones even provisionally among the Karroo beds, though 

 the latter do appear to include nearly the whole Mesozoic period. The 

 Forest sandstones are apparently separated by a great unconformity 

 from the underlying upper beds of the Coal series (presumably Upper 

 Beaufort, or even later), and it is not very clear either that they do not 

 include beds of very different ages. It may be noted that the red beds 

 of the Forest sandstones occur above the basalts, or, in rare cases, 

 intercolated between them. 



The following is the Bhodesian sequence as at present known : — 



Somabula gravels and sands . 



Forest sandstones with interbedded basalts 



Coal series ( = Beaufort and Ecca beds) . 



Sandstones (? Waterberg) 



Dolomite ...... 



Conglomerate series, with interbedded lavas 



Banded ironstone series . . 



? Basement schists .... 



? Tertiary. 

 1 Cretaceous. 



Permian. 

 ? Lower Palaeozoic. 



Arehse&n (with intru* 

 sive granites, &c). 



