1906] CHRYSLER—NODES OF GRASSES 13 
has been retained in the regions where it is of use. On the other 
hand, it may be argued that we have here the first appearance of 
a feature which in the dicotyledons becomes prevalent. But why 
should the cambium appear only at the nodes, where it is of use ? 
It may be regarded as almost axiomatic that the need for a struc- 
ture is not a sufficient cause for its appearance. So it seems more 
reasonable to read the evidence in the way first suggested, viz., 
that we have here a relic of a structure which was present in the 
ancestors of the grasses, but has disappeared from most parts of 
the plant and from most families of monocotyledons, and_ is 
retained above the nodes of grasses in connection with their power 
of bending at these regions: Thus the evidence favors the derivation 
of the grasses from ancestors having a cambium. 
The stele of the grass stem has evidently departed widely from 
the primitive protostele or siphonostele. It has been repeatedly 
shown that reproductive axes are able to retain ancestral characters. 
An examination of this region in seventeen species of grasses belong- 
ing to seven of the tribes has failed to disclose any instances where 
the stele presents the primitive type described by PLOwMAN (11). 
This result seems to confirm the opinion derived from other con- 
siderations, that the Gramineae represent a more specialized group 
than the Cyperaceae. These considerations may now be stated 
categorically: 
(1) The grass family has adapted itself to every habitat, from 
salt marsh to pampas, and shows every gradation in habit from 
the bamboo downward. The sedges are prevailingly hydrophytes, 
and few of them attain a considerable size. 
(2) The hollow stem characteristic of most grasses has probably 
been derived from a solid stem such as is present in the sedges and 
most monocotyledonous families. 
(3) Amphivasal bundles are not found in so large a proportion 
of species nor are. they as numerous in an individual in the grasses 
as in the sedges. 
(4) In practically all of the grasses the leaf-trace bundles are of 
at least two ranks, while the sedges show no such distinction. 
(5) The floral axis of the grasses dees not present the simple 
type of stele shown by some sedges. 
