Evolution of Branching in the Filicales. 17 
begins its development the lateral plant of Nephrolepis cordifolia in 
its ontogeny omits some of the earlier stages in stelar evolution, 
which are passed through normally by the sporeling. It would be 
agreed that, inasmuch as it passes through a series of stages 
parallel to that passed through by the main axis during its 
ontogeny, the lateral branch of Alsophila excelsa more nearly 
corresponds to our idea of an individual, than does the branch of 
Polypodium. 
We have taken as one criterion of individuality the degree 
to which the ontogeny of the branch resembles that of the corres¬ 
ponding main axis. Hence in a case like Alsophila excelsa, where 
this resemblance is rather close, the structure of the branch should 
be susceptible of the same phylogenetic interpretation in the light of 
the recapitulation theory, as the structure of the developing main 
axis. And this view should hold in spite of the probability that the 
the simpler structure of the branch at the base is of relatively 
recent origin, because the ontogeny of the branch still resembles 
the ontogeny of the main axis. 
The specialization of the branch has, however, gone further 
than the stage of Alsophila excelsa. The branch in Nephrolepis 
(Fig. 1, G) has been subjected to physiological conditions very 
different to those affecting the main axis. The result is that 
a portion of the branch, the stolon, has assumed an appearance 
and structure so widely different from that of the leaf-bearing 
stock, that anything short of an organic connexion between the 
two would not have convinced one that they both belong to the 
same plant. The apex of the primary stolon is endowed with the 
power of long-continued growth, and of producing a large number 
of lateral dormant buds, each the beginning of a potential 
individual. 
c. The Relation Between Branch and Leaf. 
We may pass now to another aspect of the evolution of the 
filicinean branch, namely, the relation of the branch to the leaf, 
which is so conspicuous a feature in relatively wide circles of 
affinity. We know that this relation existed in some of the most 
ancient known branching Ferns, the Zygopterideae, and it is seen 
to-day in the Ophioglossaceae Hymenophyllaceae, Cyatheae, and 
sporadically among Ferns of a still higher order. 
It is to be noticed that the relative positions of branch and 
leaf vary in different groups of Ferns. The branch (or the rudiment 
of it which may fail to develop) is sometimes adaxial (Ophio¬ 
glossaceae, Hymenophyllaceae) sometimes abaxial (Cyatheae) to the 
