388 
H. F. Wernham. 
remarks and from the diagram. Although we may compare the 
Monocotyledons and other groups with the ranalian progeny in 
respect of the mere order of their appearance in descent as reflected 
by the fossil-record, we cannot associate this comparison with 
conclusions of relative primitiveness. In other words, we may not 
say that a grass or a poplar is more primitive than a Magnolia, 
because the former precede the latter in the fossil-record ; for they 
lie on different branch-systems, which, diverged from a common 
ancestor before the separation of Dicotyledons and Monocotyledons ; 
they are incapable of comparison from this standpoint. 
The crux of the matter is this, then, that if a form is referred 
to as “ relatively primitive,” this statement must be treated as 
a strictly comparative one, and both sides of the comparison must 
be made clear; that is to say, the ancestor must he considered in 
relation to its progeny. 
Many arguments have been adduced in favour of the primi¬ 
tiveness of certain amentiferous trees, but insufficient attention has 
been accorded to the nature and limits of the groups which are to 
be reckoned among their descendants. We have already, in our 
opening chapter, hinted at the probability that considerable 
reduction must have occurred in the course of descent to produce 
a solitary sporangium shared between two or more sporophylls 
—a common condition among Amentiferae; we have, moreover, 
traced the history of a similar process in more cases than one—in 
the flower of Composite, for example. But we conclude that, in 
the case of the Amentiferae in question, this reduction must have 
occurred at a remote period ; if, that is, they are to be regarded as 
“ relatively primitive.” They lie upon a branch-system different 
from the one which has been the subject of our examination ; they 
lie, therefore, outside our present province. 
An important result of all this is that Incomplete, called by 
that or some other name, reduced considerably, it may be, from 
its present proportions in the system of Bentham and Hooker, 
must, in contrast with Sympetale, be retained as a separate 
group in any natural scheme of classification; a group as 
isolated from the progeny of the ranalian ancestor as Mono¬ 
cotyledons from Dicotyledons; and, moreover, there may be 
several such groups; just so many, in fact, as there are lines of 
descent, Dd x , Dd 2 , .... persisting above Y and bearing existing 
or fossil forms upon their subsequent branches. Our ultimate 
conclusion, therefore, is that the Synipetalce should not exist as a 
