EVOLUTION OF THE INFLORESCENCE. 545 
heads, found in the genera Echinops, Lagascea, and Cesulia, in which the 
individual capitulum is reduced to a single floret. 
We have passed in review several families and individual genera possessing 
inflorescences midway, as it were, between the true cymose and racemose 
types. Points strongly suggesting the origin of the latter from the former 
have been brought out and emphasized. 
The reverso way of regarding the relationship between these two main 
types of flower-arrangement seems to the author unsupported by the facts 
gleaned from comparative morphology. 
It appears unnecessary to proceed to any great length with the study 
of the inflorescence, once it has assumed the racemose form. By the 
suppression of the internodes, either of the lateral or main axes or of both, 
the different varieties of racemose inflorescences (spike, umbel, head, ete.) 
can be derived from the raceme; and I see no reason to depart from this 
generally accepted view, except in so far as the internodal suppression 
may occur at times before the inflorescence has completely lost its cymose 
character. This, perhaps, may be found to oecur chiefly in spicate types of 
clustering. A spicate cyme or pleiochasium may evolve into a true racemose 
spike, thus eliminating the raceme condition. Consequently, all spikes may 
not necessarily have been preceded first by true racemes. 
One further remark. As racemose inflorescences arise from cymose ones 
of the pleiochasial type through the arresting of the upper part of the 
inflorescence, it would seem, as has already been pointed out, that a further 
development is shown in the reduction of this part to a mere filament or blunt 
protuberance ; and, finally, to a loss of all trace of the former occurrence, 
resulting in the production of a fixed number of flowers to the inflorescence, 
or at any rate a marked tendency that way. The inflorescence, in fact, 
becomes very definite, and shows the unsuitableness of the old terms, definite 
and indefinite, as applied to cymose and racemose inflorescences respectively. 
VI. Continuous DicHasia, Monocnasia, AND SYMPODIAL 
CYMOSE INFLORESCENCES. 
In the preceding two sections of this paper one line of evolution of flower- 
clusters from the simple dichasium, which arises from the solitary terminal 
flower, has been traced: namely, that advance brought about through the 
increase in the number of lateral (secondary) floral shoots, whereby the 
dichasium becomes a pleiochasium. From the latter, by a further increase in 
the number of laterals, racemose inflorescences have been shown to evolve. 
There is another mode of advancement from the simple dichasium which 
now claims our attention. It has already been pointed out that tertiary 
2P2 
