20 Humphrey . — The Development of the 
of the development of the tissues of the seed. The single- 
seeded fruit of this genus is indehiscent, and the seed is 
apparently set free by the decay of the rather thin pericarp. 
The outer integument of the ovule is about six cell-layers in 
thickness (o. i., Fig. 28), of which the outer and inner layers 
begin early to take a distinctly cuboidal form. The cells of 
the intermediate layers increase irregularly in number, and 
constitute a nutrient layer which shows compression and 
obliteration of its cell-cavities as the seed ripens (nut., Fig. 31). 
In the ripe seed the cells of both outer and inner layers are 
large and thin-walled, while the latter show traces of compres- 
sion in the wavy folding of their radial walls (ext., int Fig. 31). 
The inner integument is here pressed into a film in which no 
cellular structure is recognizable (i. i., Fig. 31). The infolding 
of the inner integument before mentioned (m. c., Fig. 29) 
produces the same result as in Canna — the formation of an 
enclosing testa-collar about the apex of the embryo-sac 
(m. c., Fig. 30). In many Marantaceae a distinct germinal lid 
is said to be developed over the end of the cavity. In Thalia 
I have been able to recognize no such differentiation, possibly 
from lack of sufficiently old seeds. Thalia dealbata possesses 
no trace of an aril ; but in most members of this family, which 
Eichler (’ 84 ) has remarked to be those with dehiscent fruit, 
the thickened funiculus, with the tips of the integuments, 
forms a fleshy arillar appendage to the seed, which has been 
shown by Fritz Muller (’ 83 ) to play an important part in the 
dehiscence of the fruit and the expulsion of the seeds. Eichler 
remarks (’ 84 ) that the origin of the aril in this family is 
peculiar; but just those parts are here concerned which we 
shall find contributing to the aril in other Scitamineous 
families. The most unusual feature here is the extreme 
thickening of the funiculus. 
The large amount of nucellar tissue that is neither destroyed 
by the embryo-sac nor used for the perisperm-canal becomes 
filled with starch to form a perisperm. The embryo ulti- 
mately attains the form and size of the embryo-sac, usually 
tapering towards its cotyledonary end. In young seeds of 
