certain Penaeaceae. 
37i 
confined to the south-western region of South Africa, a region in which 
4 an extraordinary number of species, many of them belonging to a few 
genera and orders elsewhere rare, are massed . . . between the sea and 
a desert interior , . 1 Hooker has pointed out 2 the striking affinities that 
this flora shows with that of Australia, especially South-West Australia, 
and the less marked relationships of these two floras to that of the Antarctic 
regions. This connexion between these three southern floras, combined 
with their separation from the tropical floras which succeed them to the 
north, seems to him to indicate that they may have been 4 members of one 
great vegetation, which may once have covered as large a southern area as 
the European now does a northern \ 3 Schonland, recently re-examining 
the evidence derived from the distribution of the plants composing these 
floras , 4 comes to the conclusion that there was a direct land-connexion 
between Australia and South Africa at the time when the types common to 
the two developed, and a later one between Australia and South America. 
The isolated groups of this now dismembered flora are being encroached 
upon everywhere by northern forms, through whose usurping tendencies 
the small local genera are beginning to disappear . 5 Bolus says : — 4 Few 
botanists who, like the present writer, have spent many years in South 
Africa, and especially in the south-western districts, have not been pene- 
trated by a gloomy impression that the south-western flora is dying out 
and is doomed to extinction. ... In general, species of the Bruniaceae, 
Penaeaceae, and Proteaceae, so peculiar to this region, seem to have become 
much more rare .’ 6 He adds : — £ No weight can be attached to this, for 
it is wanting in adequate evidence ’ ; but as regards the Proteaceae, 
Phillips’s recent investigations 7 show that the power of setting seed is in 
many species on the wane, and in the case of the Penaeaceae, it may 
be significant in this connexion to note that although they set seed freely, 
all attempts to germinate these seeds have failed, and in the genera 
examined careful search on many occasions has failed to discover any 
seedling plants. Moreover, even young plants less than about a dozen 
years old seem extremely rare ; I have not found above five such in all the 
species examined. 
This order then may probably be regarded as belonging to an ancient 
flora, and its restriction to the one region within that flora would seem 
to indicate either that it is a 4 survival ’ which has lingered on in this 
south-western region, while dying out in other regions where the flora to 
which it belonged still flourishes ; or else that it is a new type developed 
since the isolation of this region. Two considerations would seem to point 
to its being an original member of the flora rather than a lately-evolved 
1 Hooker, 1859, P* xcii. 2 loc. cit. 3 Hooker, 1859, p. civ. 
4 Schonland, 1909. 5 Hooker, 1859, p. cv. 6 Bolus, 1905, p. 235. 
7 Ballantine, 1909, p. 16 1, 
