120 EvaANs: TAXILEJEUNEA PTEROGONIA 
third inflorescence with a single sterile innovation after bearing 
two underleaves with their corresponding leaves. A female branch 
system on another specimen (Evans 3) bore eight underleaves 
before the inflorescence, and the latter showed a single long sterile 
innovation with sixteen distinct underleaves. Another branch- 
system on the same specimen bore the first inflorescence after 
only two underleaves; the single innovation gave rise to eleven 
underleaves and a male branch before bearing a second inflores- 
cence; and this in turn gave rise to a single long sterile innovation 
with ten distinct underleaves. These examples, selected more 
or less at random, will give some idea of the variations encountered 
and indicate that long branches and long subfloral innovations 
are of frequent but by no means of constant occurrence. 
The other distinctions between T. pterogonia and T. jamaicensis 
are likewise associated with the greater variability of the latter 
species. The characters derived from the leaves, underleaves 
and floral parts are similar to those of T. pterogonia but less defi- 
nitely realized. In the leaves, for example, the auricles at the 
base, although often as distinct as in T. pterogonia, are sometimes 
less marked; the thickenings of the leaf-cells tend to be less 
developed, and the surface-verruculae are often more difficult 
to demonstrate. The auricles of the underleaves may likewise 
be very distinct (Fic. 4), but the same specimen (Fic. 1) will 
often show underleaves which are rounded at the base. In the 
bracts (FIGs. 5,6, 8 and 9) the teeth are usually much less distinct, 
and bracts with entire margins are of more frequent occurrence, 
while the bracteoles (Fics. 7, 10 and 11) show no differences of 
importance. It is in the perianths, however, that the greatest 
range of variability is met with. Of the perianths figured (Fic. 1) 
the one at the left is almost beakless and the wings of the angles 
are reduced to scattered projecting cells; the perianth at the 
right shows a distinct beak and interrupted wings two cells wide 
on the two ventral keels, these wings bearing rounded teeth; 
the lateral and dorsal keels of this same perianth (not clearly 
shown in the figure) bear very narrow, indistinct and entire 
wings. In one perianth on another specimen (Evans 3) the short 
wings on the ventral keels run out into sharp points, but the dorsal 
_and lateral keels are scarcely winged at all; in another perianth 
