219 
Seedling Structure of Gymnosperms. II. 
form the diarch root-structure in the manner already indicated. The 
account given by Seward and Ford for this same plant does not agree with 
that given by Dangeard in respect to the union of the cotyledonary bundles 
at the base of the seed-leaves. It appears, however, that there may be 
some variation as regards this particular feature, for we have described 
a tangential union of the seed-leaf-traces in A. Cunningkamii (Series C), and 
on one side of the seedling in A . brasiliensis (Series A). With this exception 
there is no material difference, and a comparison of the descriptions and 
corresponding illustrations warrants the conclusion that A. imbricata does 
not differ in any material feature from A. brasiliensis. 
If the rearrangements of the vascular strands in these seedlings of 
Araucaria be traced in the opposite direction, from the root upwards, they 
may be summarized thus : — each pole of the diarch root gives off two 
bundles, which, during their passage to the cotyledons, first rotate to bring 
the protoxylems into the endarch position and then dichotomize, once in the 
case of A. Cunningkamii , and ideally twice in the case of A. brasiliensis , 
and some other hypogeal forms, so that the former has four bundles in each 
seed-leaf and the latter eight, subject to reduction in some instances. 
If reference be made to Podocarpus } it will be seen that there is 
a marked resemblance in the actual transition between that plant and 
Araucaria ; in each case the upper regions of the hypocotyl are occupied by 
four cotyledonary bundles, two on each side, which rotate and give rise to 
a diarch root-structure, a fact which is clearly brought out in the relative 
figures. The only difference is that each seed-leaf in Podocarpus has two 
bundles, while in Araucaria there are typically either four or eight. 
Conclusions. 
The chief immediate conclusion arrived at by the study of the 
seedling-structure of the Taxaceae and Araucariaceae is that the poly- 
cotyledonous condition, in the majority of cases, has been attained by the 
longitudinal division of the pre-existing single members which were in 
all probability two in number. 
Before reviewing our evidence upon which this conclusion is based, 
some mention must be made of earlier work, which directly bears upon the 
subject. 
The botanists of the earlier part of the nineteenth century were divided 
in their opinions regarding the nature of polycotyledony. Adanson and 
Jussieu both believed that the multiplicity of seed-leaves was more apparent 
than real, and that it had arisen by the division of two cotyledons, which 
opinion was not shared by Salisbury and Richard. 
Duchartre 2 , on the evidence summarized above (p. 189), shared the 
belief of the two former investigators. 
i Part I, loc. cit., p. 694. 
loc. cit. 
