174 
AMERICAN JOURNAL OF BOTANY 
[Vol. 8 
Discussion 
Structures such as the thick-walled root hairs under discussion are 
commonly associated with xerophytic conditions, and at first it was thought 
that perhaps desiccation would explain their presence. When they were 
found in various types of habitats including bottomland forests, however, 
and when they were found deep in the soil as well as in the superficial layers, 
the idea that their presence could be explained as due directly to desiccation 
was quickly dispelled. However, Gleditsia triacanthos has a rather wide 
specific range of tolerance with respect to habitat. In Illinois it occurs 
most commonly in bottomland forests along streams, where it is a charac- 
teristic tree. But it is also a characteristic tree of the dry barrens of Ken- 
tucky and of dry hills elsewhere. For this reason, it is the opinion of the 
writer that these thick- walled hairs, together with the thorns which are also 
so characteristic of this species, are relics of a time when the tree grew only 
under more arid conditions. (It should be added here that Gleditsia triacan- 
thos inermis, which lacks the thorns, produces thick- walled root hairs as 
abundantly as the species). In the case of Gymnocladus and Cercis, both 
of which have a lesser specific range of tolerance with respect to habitat 
than has Gleditsia, there have probably been greater constitutional changes 
during the course of evolution, and they have retained the characteristics 
due to an arid climate to a much lesser extent. 
As previously stated, my observations indicate that in general, except 
when broken off by growth movements of the roots in the soil, the root hairs 
remain for as long a time as the rest of the epidermis. In some cases this 
may be only a few weeks while in other cases it may be several months. 
McDougalP has shown that the roots of trees grow at any time during the 
year when they can absorb a sufficient amount of water. Some growth 
would take place, therefore, and some root hairs would be formed late in 
the growing season. These would undoubtedly persist through the winter 
and into the next year until such time as there was sufficient growth to 
rupture the epidermis and finally cause it to be cast ofT. Indeed, it is not 
improbable that in some cases these hairs could rightly be spoken of as 
perennial. 
Harrison and Barlow^ state that Gleditsia triacanthos, Gymnocladus 
dioica, and Cercis canadensis never have root nodules (bacterial) but 
that mycorrhizas are always present on these species. I have never been 
able to find mycorrhizas on any of the leguminous trees, and it is my opinion 
that the roots bristling with thick-walled root hairs were mistaken by Harri- 
son and Barlow for mycorrhizas. The absence of both root nodules and 
1 McDougall, W. B. The growth of forest tree roots. Amer. Jour. Bot. 3: 384-392. 
1916. 
2 Harrison, F. C, and Barlow, B. The nodule organism of the Leguminosae — its 
isolation, cultivation, identification and commercial application. Centralbl. Bact. 19: 
264-272; 426-441. 1907. 
