AFFINITIES OF DAVIDIA INVOLUCRATA. 317 
epidermis and not penetrating inwards: the branches are claw-like and the fine 
capillaries end in the abraphal side and are not localized in the chalazal region of 
the seed. The cells of the outer zone change in structure during development by 
various wall modifications: those of the external epidermis possess a thin outer 
cuticular membrane and their radial walls become gradually covered with fine, roughly 
parallel, radial strie. The walls of the subjacent cells become marked with numerous 
elliptic pits or are striated. The cortical system is associated with an elaborate peripheral 
vascular system, and consists externally of cells unoccupied with solid food material, 
which possess pitted and striated walls, and internally of smaller cells with denser 
protoplasmic contents. It thus forms an efficient water-jacket with collecting-cells 
ending against the internal epidermis during the development of a definite endosperm 
tissue. 
The terms “tapetum,” used by Billings * and Balieka-Iwanovska f, and * epithelial 
layer " used by Ducamp 1 and others imply certain physiological functions, and Billings 
uses the term tapetum in a purely physiological sense. It seems better in this paper to 
use the expression “ internal epidermis” and to ascribe to it a particular function. The 
cells of the internal epidermis are distinguished from their fellows by their shape, small 
size, aud denser staining properties of their contents. These properties are shared some- 
what by the immediately subjacent cells. Soon after fertilization time the inner wall of 
the internal epidermis is limited by a fine cutinous membrane, which is easily distin- 
guished when the endosperm is already some cells in diameter. The cellulose portion 
of this inner wall is minutely perforate. The use of sulphuric acid and iodine shows 
that the protoplasmic threads of the endosperm and of the internal epidermis 
end up against the cutin. In Aralia, according to Ducamp, the “ assise épithéliale " , 
(internal epidermis) and the embryo-sac are shut off from one another by a cutinous 
membrane. Concerning this layer he states: ^ Elle s'isole d'ailleurs du sac embryon- 
naire en cutinesant sa surface et ses parois latérales"; and “le sac embryonnaire 
et protégé par une laure isolante de cutine contre laction digestive des cellules 
épithéliales." The mere presence of a cutinous membrane is insufficient to disprove 
translocation of liquid. In the case of gaseous and liquid diffusion between two adjacent 
cell-layers, the inhibitive factor is one of relative thickness of the intervening membrane, 
since we are dealing with a problem of increased resistance to “ flow" dependent upon 
length and diameter of the membrane passages. The formation of cutin does not 
necessarily involve the loss of function of the “connecting threads,” but the relative 
thickness of such a membrane may act as a kind of regulator during development, 
à certain maximum thickness in given cases inhibiting translocation. | 
The occurrence of antipodal haustoria in certain ovules might lead to the belief that 
nutriment is passed into the sac at the ehalazal end, but it must be borne in mind that 
the haustoria have only been found in connection with sterile organs. The seeds in 
which pro-embryos were found show no indication of the presence of persistent antipodals. 
* Billings in * Flora,’ 1901, pp. 254 &c. T Balicka-Iwanoveka in * Flora,’ 1399, pp. 47 &c. 
i Ducamp in Ann. Sci. Nat., Bot. sér. 5, xv. (1902) pp. 311-402, tt, 6-19. 
