20 BOTANICAL GAZETTE [JANUARY 
few root hairs; and others that are less white, somewhat thinner 
but decidedly longer, with many lateral rootlets, and amply pro- 
vided with root hairs. No special arrangement of these two 
forms was noticed, but the thick ones were perhaps the more 
abundant. Both forms appeared to have developed in the same 
year, and I was unable to detect any pronounced difference in 
their internal structure that might warrant any further distinc- 
tion than the one observed in their external appearance. 
The thick white roots show a very weak structure in general, 
which seems to imply that they are of short duration, and prob- 
ably do not last beyond one season. The epidermis is very thin 
walled, and root hairs are present, but rather scarce. The root 
hairs observed were all single, not in pairs, as Van Tieghem 
observed in species of Paepalanthus (Lachnocaulon?). Poulsen 
did not find any root hairs at all in E. helichrysoides, the speci- 
mens of which had perhaps not been lifted with sufficient care. 
Within the epidermis is a hypodermis of a single stratum, the 
cells of which are of the same size as those of the epidermis, and 
are also thin walled. The cortical parenchyma is quite broad, 
but almost totally collapsed radially, excepting the innermost 
stratum which borders on the endodermis, and which is often 
somewhat thick walled in contrast with the other portion of the 
cortex. There are thus several very wide lacunes in the cortex, 
and these contain horizontally placed diaphragms composed of 
star-shaped, thick walled cells, which contain chlorophyll. The 
lateral roots do not immediately break through the epidermis, 
but traverse the cortex for some distance downwards before they 
penetrate the hypodermis and epidermis. A similar course of 
the lateral roots was observed by Poulsen in Carpocephalus cau- 
lescens Kth. 
The endodermis (jig. z, Z) is usually thin walled, and the 
spots of Caspary are plainly visible. However, in some roots 
the cell walls of the endodermis are slightly thickened, especially 
in the basal portion of the root. The pericambium (figs. 7, 2, P) 
is very thin walled, and consists of a single layer, which in no case 
was observed to be continuous, although a very large number 

