Bie) 
are always sheathing kataphylls ; though when two are present 
the second is much smaller than the first and does not protrude 
beyond the sheath of the cotyledon (Plate LIII, Fig. 14). -In the 
axil of each kataphyll there is early formed a large bud which 
grows into the secondary stem, and this in turn produces below 
ground two contracted internodes bearing two sheathing kata- 
phylls, one or both of which subtend a single axillary bud, the 
process being repeated indefinitely. Even in those species 
retaining only one kataphyll below ground at the base of their 
primary axis—that derived from the plumular meristem directly 
—there are two in the case of the secondary one and all that 
follow, though as a rule only the second of these has in its axil 
a bud. In this way the sympodial rhizome so characteristic of 
the genus is built up from the bases of the aerial shoots. 
Let us now examine the changes which have taken place in 
the other parts of a somewhat advanced seedling. The suctorial 
end of the cotyledon (Plate LIII, Figs. 1 and 5) gradually enlarges 
within the seed as the reserve food-material of the endosperm 
becomes dissolved and absorbed. This is a slow process however, 
and the seed remains attached to the plant for a considerable 
period. The primary root in seedlings three or four weeks old 
has become more or less thickened—especially so in A. medeoloides 
(C, Fig. 3), many of the roots of which function solely as storage 
organs—and seems to act largely asa pull-root. It never branches, 
becomes wrinkled, and in fact presents all the peculiarities 
typical of such roots. Soon there arises, usually from the point 
where the hypocotyl and stem meet, a lateral root which branches 
and remains unthickened (C, Fig. 3). This is followed by other 
lateral roots proceeding from the rhizome as it is formed—for it 
must be remembered that two or more joints of this are produced 
each year—some of which develop in the same way as the primary 
root. The hypocotyl (Plate LIII, Fig. 15) remains short during 
the whole of its life, and, as I have already said, I have never, 
under any circumstances, seen one which had elongated. Thestem 
apex in older seedlings is always protected by two or more scale- 
leaves, which later on are left behind, but not until others have 
been produced to take their place. At first they appear opposite 
in pairs; then, as the internodes lengthen, they assume their 
true position, alternating ina } spiral. The first few scale-leaves 
of the stem—not including the basal kataphylls—sometimes 
have apparently no axillary buds, for example A. comorensis 
(Plate LIV, Fig. 16) ; but in other cases, such as A. medeolordes, 
may subtend lateral aerial branches or even cladodes (C, Fig. 3). 
In a longitudinal section of the stem apex in the latter plant 
(Plate LIII, Fig. 6) it is interesting to note how the development 
of the large leaf-like cladodes is almost as rapid as that of the 
