330 Shaw . — The Seedling Structure of Araucaria Bidwillii. 
As mentioned above, it was not possible, as a general rule, to trace 
the relationship of the cotyledonary bundles to those appearing in the 
hypocotyl. This was probably due to the large amount of secondary 
thickening in all the specimens ; had younger seedlings been available there 
is no doubt that the connexion, of which there were strong indications 
in two specimens (Diagram III), would have been substantiated. It will be 
remembered that in these two seedlings each pair of bundles in the hypocotyl 
was traced back, in the one case to a pair of bundles in the cotyledonary 
tube, in the other to a single arc-shaped strand which divided. The proto- 
xylems of each pair then fuse, the secondary xylems and phloems divari- 
cating to fuse with those from adjacent pairs. Ultimately diarch structure 
is obtained in the root. 
Now this does not differ essentially from the type described by Hill 
and de Fraine in the Cupressineae. Here a single bundle enters the 
hypocotyl from each cotyledon and splits into a pair of bundles. The 
xylems of each pair then twist towards one another and fuse, with their 
protoxylems external ; the phloems divaricate to fuse with those of 
adjacent pairs. Thus, if the seedling had three cotyledons, a triarch root 
should arise ; if two cotyledons, a diarch root is formed. 
Now Araucaria Bidwillii has two cotyledons and ultimately forms 
a diarch root, each pair of bundles in the hypocotyl behaving very much as 
a pair of bundles in the hypocotyl of a seedling Cupressas. If the remain- 
ing Araucarieae agree essentially with this type, then the transition pheno- 
mena in Coniferae will show a strong general uniformity. Araucaria 
Bidwillii , however, while agreeing with the general plan, shows interesting 
numerical variations — numerous pairs of bundles derived from the cotyle- 
dons enter the hypocotyl, and diarch structure is attained in the young tap- 
root by reduction from an earlier (embryonic) stage, when numerous proto- 
xylems are present. Thus, in a species remarkable for its swollen hypocotyl, 
there is in the root-structure at the base of the hypocotyl a large increase in 
the number of protoxylems, the primary root becoming diarch by reduction. 
This at once suggests a comparison on anatomical grounds with the 
seedling of Eranthis hyemalis . In this plant there is a tetrarch structure 
in the tuberous hypocotyl at the base of the cotyledonary tube, this structure 
becoming reduced to diarch as the hypocotyl merges into the main root. As 
is well known, this tetrarch structure has been compared with that occurring 
in Anemarrhena , and it has been made the basis of wide phylogenetic 
speculation. It was, however, early suggested (Tansley, 5) that, since 
diarch structure prevailed almost exclusively in the hypocotyl of non- 
tuberous species of Ranunculaceae, the increase in number of xylem groups 
in the tuberous Eranthis should probably be correlated with the increased 
need for vascular tissue in such a structure rather than be regarded as 
an ancestral trait. 
