208 T. Holm — North American Terrestrial Orchidece. 



rays of hadrome with a corresponding number of leptome- 

 groups (fig. 4). The endodermis and pericambium (End. and 

 P. in fig. 4) are constantly thin-walled, and the latter was 

 found to be continuous in some, but interrupted in others of 

 these small mestome-cylinders within the same tuberous root. 

 A very small, central pith was observed in P. ciliaris and 

 eristata, but not in the others. 



If we compare this structure of the tuberous with the 

 slender roots of these same species, there seems to be no other 

 difference than there being a much smaller number of mestome- 

 cylinders in the latter. 



Of the three types of roots observable in our terrestrial 

 Orchidem, the third category emphasizes those in which several 

 mestome-cylinders are present instead of but one, and this 

 peculiarity may be briefly described in connection with the 

 anatomical data, mentioned above. The fact that these tuber- 

 ous roots contain several, isolated cylinders provided with a 

 special endodermis and pericambium, has given rise to various 

 views regarding their morphological identity : whether the 

 " tubers," as they are frequently called, might represent 

 1) the basal, swollen part of the bud-axis, 2) a single root, 

 3) a concrescence of several roots or 4) a concrescence of a 

 stem-portion with leaves and roots. Of these the most gen- 

 erally accepted theory is the one which explains the origin of 

 the tuber as being a concrescence of several roots, very ably 

 discussed by Yan Tieghem and others. 



But the definition tending to explain the tuber as being the 

 result of a concrescence of a stem-portion with leaves and 

 roots, as proposed by Germain de St. Pierre,* has not been 

 approved by others. Nevertheless, as will be shown in the 

 following pages, this definition does not only seem to be well 

 founded, but is, indeed, the only conceivable one, as far as 

 concerns the tuberous body beneath the hibernating bud in 

 North American Ophrydece ; we may illustrate this by the 

 rhizome of Platanthera dilatata (fig. 2). The rhizome of 

 this species is relatively slender and the hibernating bud is 

 prominently removed from the mother shoot by the descend- 

 ing stolon {St. in fig. 2) ; the bud itself (b) appears as if it 

 were lateral, since the stolon gradually passes over into the 

 long, tuberous body (r) underneath the bud, the so-called 

 " tuber " of most authors. The bud, however, is terminal and 

 its apparently lateral position is due to the growth of the 

 stolon, the direction of which is neither horizontal nor vertical, 

 but simply descending. As may be seen from the figure the 

 basal region of the bud with its rudimentary leaves and young 

 roots is located on the upper, the dorsal, face of the stolon, 



* Bull. Soc. Bot. de France, vol. 2, p. 659, 1855. 



