THE ANATOMY OF THE NODE IN ANGIOSPERMS 3II 
POLYGONALES 
We have already spoken of the Polygonaceae as a typically multi- 
lacunar family. In Miihlenbeckia, Coccoloba, Fagopyrum and most 
species of Rumex and Polygonum examined there are a considerable 
number of strands passing off into the sheathing base of the leaf from 
the entire periphery of the stem {jig. 6). In young plants of these 
forms, however, there are only three bundles to each leaf, and this 
trilacunar condition, which also prevails in the dioecious species of 
Rumex {fig. 5) and the climbing species of Polygonum, is probably 
primitive for the order. 
Centrospermae 
In this great order a large number of species were investigated in 
all families save the Cynocrambaceae and Bassellaceae, and in every 
instance the departure of the foliar strand was found to cause but a 
single gap in the cylinder. In the Chenopodiaceae, Amarantaceae, 
Nyctaginaceae and Phytolaccaceae the stem structure is often anoma- 
lous, but even here the complex leaf- trace comes off from one definite 
region and is not inserted in various places around the stem {figs. 8 
and 10). The simpler species in these groups {fig. 9), however, and 
the very young plants {fig. 7) display a single vascular ring with no 
anomalous bundles. Although the leaf-trace in such forms and in all 
the rest of the order (the Aizoaceae, Portulacaceae and Caryophylla- 
ceae) never leaves but a single gap, it is almost always three-parted 
or three-lobed, thus indicating its probable origin from an approxi- 
mation of the three strands of a trilacunar type. The fact that such a 
very large series of plants are characterized without exception by a 
single type of nodal structure is strong evidence for placing them 
together, especially since the unilacunar condition which they display 
is possessed by very few of the other lower Archichlamydeae. The 
Polygonales, with which some or all of the Centrospermae have fre- 
quently been included, even in the recent system of Hallier, possess 
a nodal anatomy so unlike that of the Centrospermae as to preclude 
any very close relationship, if the structure of the node is to be regarded 
as of much taxonomic value. Evidence from this region certainly 
supports the view that the Centrospermae as defined by Engler con- 
stitute a very natural order which is somewhat removed from the other 
lower Archichlamydeae. 
