Seed in the Scitamineae . 
19 
the canals of Thalia represent sterile embryo-sacs. This idea 
was shown to be mistaken by Gris (’59, ’60, ’60), who con- 
sidered them to be ingrowths of the chalazal tissue into the 
nucellus. Eichler (’84) added only a few details concerning 
the adult structure of certain seeds, and regarded the canals 
as chalazal ingrowths. Tschirch (’90) refers to them merely 
as ingrowths of the seed-coat, which they certainly are not. 
In the absence of full details of their development, we are not 
in a position to decide finally the question of their mor- 
phology. Yet the fact that they extend inward for a con- 
siderable distance from the chalaza, on their first appearance, 
and that they seem to arise as gradual differentiations in the 
nucellar tissue, directly opposes the view of Gris and Eichler, 
except in so far as the common basal portion of the canals in 
Thalia , which is probably present in other forms also, is due 
to the enclosure of chalazal tissue by the growth of the 
nucellus, as in the Centrospermae. The conception of the 
chalaza as an actively penetrating organ , like the embryo-sac, 
rather than as a mere region of the seed will, I think, hardly 
bear criticism. It seems, then, most probable that the peri- 
sperm-canals are in part specialized portions of the nucellus, 
and in part chalazal tissue included by the growth of the 
latter. Of what use they are to the seed future studies must 
attempt to explain. 
In a single Marantaceous seed, whose development I have 
not been able to study, the embryo is not hooked, but nearly 
straight, and lying parallel with it is the perisperm-canal 
(Fig. 32 A). Whether we have here a seed developed with- 
out the campylotropic growth usual in this family, it is 
impossible to say. If not, then, for some reason, the exten- 
sion of the embryo-sac into the later-formed portion of 
the nucellus has been here suppressed. If so, we have evidence 
that this growth is not necessary to the formation of a peri- 
sperm-canal. The seed here mentioned, whose development 
promises to prove of considerable interest, is that of Phrynium 
capitatum , Willd. 
Returning to our study of Thalia , we may note some details 
