Seed in the Scitctmineae, 
15 
growing part of the embryo, and some of the accessory roots 
quickly outstrip the primary one. A discussion of the 
various interesting features of germination in Canna is reserved 
for a future occasion. 
The endosperm. The nucleus, formed by the fusion of the 
polar nuclei, has been said by Hofmeister (’49, J 61) to give rise 
to no endosperm ; the statement is repeated by Tschirch (’90), 
and by various text-books. But a short time after fertili- 
zation a definite and readily separable protoplasmic layer 
with numerous nuclei may be found in the micropylar part 
of the embryo-sac, precisely as it occurs in plants with 
abundant endosperm (csp., Fig. 9). As development proceeds, 
this layer may be followed downward until, when the seed 
reaches its full size, the entire cavity may be found to be 
lined by it. There now appear cell-walls within the layer, 
but not so regularly as in many plants, since several nuclei 
are commonly included in a single block of cytoplasm 
(Fig. 22 ). Ultimately each cell becomes uninucleate, ap- 
parently by a process of nuclear fusion similar to that 
observed by Strasburger (’80) in Corydalis (Fig. 23). In the 
ripe seed this layer assumes all the characters of the so-called 
£ aleurone-layer,’ now known in the seeds of so many plants, 
and lies between embryo and perisperm, in contact with 
both. Traces of this layer seem to have been recognized by 
Hegelmaier (’7 4) in early stages of the seed ; but he failed to 
discover its presence in the ripe seed. Overhage (’87), on the 
other hand, correctly describes its occurrence in the ripe seed 
and suggests that it may be a remnant of endosperm, but 
without presenting any evidence of the correctness of his 
suggestion. This appears to be the first recorded case of the 
normal reduction of the endosperm to a single persistent 
aleurone-layer, and points to this tissue as one of considerable 
importance. Guignard has lately shown (’93) the very wide 
distribution of such a layer in so-called exalbuminous seeds ; 
but in these it appears to be always a remnant of a more 
voluminous endosperm. In other cases in which the develop- 
ment of endosperm does not exceed the formation of a single 
