Origin and Development of the Composite. 229 
Ainslicea type reached South Africa at a much later date (lower 
Pliocene) by means of a greater development of the pappus which 
made migration easier. During the journey it underwent a few 
other slight changes and became Dicoina. 
To the north, in the Mexican region, the end of the Oligocene 
saw a descent from the mountains by the progeny of Guaphalium. 
Such a migration to the moist wooded plains rendered the closely 
aggregated capitula of that genus unnecessary. Larger and less 
densely involucrate heads were developed, with some rounding off 
of the tips of the style branches, and a weedy habit was acquired. 
The result was Plucliea, another wanderer, which spread and 
developed largely on the other side of the Pacific. 
Miocene Period. 
Beyond a larger development of the Cynareae, with the initia¬ 
tion of the Cardnus type, in response to the continued evolution of 
the semi-desert area of the eastern Mediterranean region, the lower 
Miocene was comparatively uneventful. The only other events 
were confined to the Mexican region. 
Guaphalium again descended from the mountains, this time 
into the Californian valleys and, undergoing changes similar to 
those which originated Plucliea in a similar environment, it became 
Filago. This type, which like Plucliea spread to Asia, finally 
became well-developed in the Mediterranean and other regions 
lying on the path of migration. 
In the Mexican region also, about the same time, the descend¬ 
ants of Spilantlies having reverted to their distant Lobelioid 
ancestors in producing receptacular bracts, again came into line 
with the rest of the Compositae by omitting these accessory struct¬ 
ures. Continued existence in the favourable mesophytic environ¬ 
ment of the Mexican woods rendered the numerous involucral 
leaves of those forms unnecessary and the result of these two 
changes was the initiation of the Layia-Madia group, 
The middle Miocene, a geologically active period, was more 
eventful for the Compositae. The orthogenetic elongation of the 
appendages of the style branches continued in the Aster plexus 
and the production in the open Brazilian woods of forms which, 
with the abundant food supply, produced dense masses of capitula 
had an interesting result. The crowding of the heads gave little 
room for ray florets and the food supply of the dense inflorescence 
being limited by the conductivity of the peduncle, each capitulum 
would get a distinctly limited supply. These factors led to the 
