Syncotyly and Schizocolyly . 809 
logical tricotyls, and in the present species at least is connected with their 
origin, the two smaller cotyledons representing the products of fission, 
as indicated by the anatomy. The mode of folding of the three cotyledons 
in the seed helps to explain their persistent differences in size and shape ; 
this is shown in transverse section in Fig. 30, the ventral surfaces of the 
cotyledons being indicated by thicker lines. The two smaller cotyledons 
are also slightly falcate when unfolded, as shown in Fig. 31 (taken, however, 
from Tricotyl E): and the venation differs somewhat from that of the whole 
and symmetrical cotyledon ; the two smaller seed-leaves being practically 
related to one another as object and mirror image. 
The root was diarch, but began to become triarch at the collet : the 
mode in which this was effected was somewhat different from that seen 
in the above hemitri- and hemitetracotyls. The protoxylem, instead of 
dividing equally, did not divide at all : a protrusion of metaxylem appeared 
just below the collet (Fig. 32), and just above the collet this metaxylem 
acquired a new protoxylem having no connexion with any group lower in 
the axis (Fig. 33). Corresponding with the protrusion of metaxylem, the 
superjacent phloem group divided radially. A normal triarch stele was 
thus organized just above the collet, and this behaved in a precisely ana- 
logous fashion to the diarch stele in a dicotyl : three triads being produced, 
one serving each of the three seed-leaves. 
It is noteworthy that this exceptional mode of origin of the third proto- 
xylem is that considered as characteristic of what Hill and de Fraine term a 
‘ subsidiary * cotyledon. 1 I shall return to this later. 
Tricotyl E. The root was triarch (Fig. 34), and the whole transition 
was perfectly symmetrical, each cotyledon taking a triad derived from a single 
root-pole. The first whorl of plumular leaves was also trimerous ; and there 
were three main plumular traces in the upper part of the hypocotyl alter- 
nating with the three triads (Fig. 35). 
This seedling shows the anomaly in its most complete form : there is 
nothing to indicate its relations with dicotyly, save possibly the slight 
inequality of the cotyledons (Fig. 31). 
Other Tricotyls. Nine other completely tricotylous seedlings were 
examined : of these, four had diarch, five triarch roots ; no corresponding 
differences being noticed in the cotyledons. One triarch seedling had only 
two primordial leaves, but the others, diarch and triarch alike, had a 
trimerous plumule. 
Hemitetracotyl F. This seedling bore two whole cotyledons and a third 
divided below half-way. The root was diarch, becoming triarch by division 
of one protoxylem % cm. below the collet ; the three protoxylems thus 
formed supply the three seed-leaves. The other original root-pole forked 
about one-eighth up the hypocotyl and behaved just like the dividing strand 
1 T. G. Hill and de Fraine (’08), p. 708. 
