South African Agamas allied to Agama hisj^ida and A. atra. 
287 
The tropical African form nearest to this prototype is clearly A. mossam- 
bica (N. Kliodesia, Njassaland, Portuguese East Africa), which is represented 
to the West (Angola, Damaraland) by the more depressed A. planiceps, and 
the most extremely remote from it the typical A. hispida from the Cape, the 
other forms of the latter species filling up the gap, more or less, from north 
to south. 
Now this is exactly what one should have expected on the assumption, 
founded on the absence of the Agamidae from Madagascar, which seems to 
show that, like the Lacertidae, Agama did not extend its range south of the 
Equator until after a connection between the continent and the great island 
had ceased to exist.* If, as we believe, the forms of Agama now spread over 
South Africa are, geologically speaking, of recent origin, it is quite 
legitimate to suppose that they represent the actual steps through which the 
genus has passed in its evolution and gradual dispersal. 
The directions of the lines of evolution are not only traceable in 
A. hispida in the broad sense, but also in A. atra, the most extreme form of 
which, in accordance with these principles, is the var. radis of the extreme 
south, but having probably thence migrated again northwards along the 
coastal region. It seems, if we consider A. knoheli as the initial form of the 
A. atra group, that this group has spread from the north-west to the south- 
east, while A. hispida has followed two parallel lines — an eastern (var. 
armata) and a western (var. aculeata). 
* Cf. Boiilenger, C.R. Ac. Sci., Paris, 191S, p. 594. 
[Note. — A Map showing the distribution of the genus Aguma heis been 
omitted and will appear in a subsequent Part of the Transactions.] 
