Proceedings 53 
inconsistent results obtained by examining the structure at 
this early age have discredited embryology among botanists. 
When the seedling struggles out of the seed-coats, its 
woody skeleton begins to develop. The veins of the leaf for 
example are complex structures, called vascular bundles by 
anatomists. A few of them are present in the embryo—those 
of the cotyledon are almost always there—but at that age 
they consist of soft tissue only. Immediately after germina- 
tion the vascular bundles becomestiffened by the hardening 
of some of their elongated cell-walls, and this of course 
gives firmness to the cotyledon and leaves as a whole. 
Similar bundles are found in the stem and roots. 
When examining the young seedling of various Monoco- 
tyledons I paid particular attention to the skeleton formed 
by these bundles, tracing their course with the utmost 
care in cotyledon, stem, leaves and roots. The first leaf of 
all these seedlings always showed a well-marked midrib 
with one or more smaller bundles on either side of it. The 
cotyledon on the other hand very seldom possessed a 
typical midrib. Often it had two bundles only, parallel to 
each other throughout the length of the cotyledon, and 
sometimes so near together as to touch along a narrow 
line. If other bundles were present they were arranged on 
either side of the two main bundles, just as if the two to- 
gether represented a midrib, In the very few instances in 
which the cotyledon possessed a single median bundle, 
the corresponding bundle in allied species would show the 
dual structure, or there would be two distinct bundles close 
to each other. 
If there were no other indication of comparative an- 
tiquity, we could as well suppose the cotyledon with two 
bundles to be derived from an ancestor with one as the 
uninerved from the binerved. But the vascular skeleton of 
the species I examined seemed characteristic not only of 
species but even of genera and tribes within the Liliacez 
—the family most thoroughly studied. Series of types were 
linked together by intermediate forms, and it was often 
clear that a single type had been modified in two or three 
different directions, thus giving rise to several distinct 
