176 MEMOIRS FROM THE DEPARTMENT OF BOTANY OF COLUMBIA COLLEGE. 
anum. After passing a somewhat different epidermis and hypoderma, we meet with 
quite similar starch-rings and stereome cylinders, the latter being very compact and 
rigid, the elements small, angular, and of much the same size throughout; again passing 
a difference in the arrangement of the phloem, the xylem appears strikingly similar in 
both, with its system of tracheides and vessels disposed in the same manner; the xylem 
merges into the pith through a cylinder of the same kind of cells, and in both species the 
latter tissue appears about the same in cross section, its difference in long section being 
mentioned under the head of the subgenus Tovara. 
Third: This division will include the subgenera AvicuLArtA and Duravia. These 
two are not thrown together on account of the similarity of their respective tissues, but 
on account of the stereome development in the primary cortex. The phloem is peculiar 
in each case, while in AvicuLArtA the xylem is normal, but in DurAvta it is not what 
we should expect to find. 
Fourth: P. scandens falls in here with its large-celled epidermis, its often scanty 
hypoderma and starch-ring, a characteristic stereome cylinder, its non-radial fascicles of 
phloem and weak form of xylem with few tracheides and vessels. ‘The inner sides of the 
bundles also are strongly wedge-shaped. 
Fifth: This, the last division, embraces subgenus Ecurnocauton. P. arifolium as 
representive, shows a double epidermis, the only one found in this study, a very irregular 
primary cortex, a strong and characteristic stereome cylinder, a scanty supply of phloem 
and a remarkably strong and diversified system of woody fibres, together with numerous 
and conspicuously marked tracheides and vessels. The pith elements in this case are 
unusually large. 
After arranging and grouping the subgenera in a natural sequence based on their 
anatomy, reference to pages 18 and 19 will show that this classification corresponds ex- 
actly to the sequence and grouping there based on their morphology. 
Although the eight subgenera group themselves into five divisions as arranged 
above, there is more or less dissimilarity among some of the tissues of those falling into 
the same divisions. These variations suggest a number of interesting points. 
First: There is a difference in the development of the tissues. I have discussed this 
with special reference to the classification and will attempt to give reasons for these differ- 
ences. The general similarity of the tissues of each subgenus has been given above, and 
this serves to show the distinction between the divisions, but a glance at the plates, 
especially the cross sections, will show that more or less difference exists betwéen the 
members of the same division. 
Second: Are there any tissues wanting? I have stated that the same tissues are 
