132 Sargant . — The Reconstruction of a Race of 
Thus the comparison of Angiospermous embryo-sacs with each other 
has given no clue to their primitive structure. What suggestions are offered 
by Gymnosperms ? 
Among Gymnosperms, with the exception of Gnetum and perhaps 
Welwitschia, the mature unfertilized embryo-sac contains archegonia em- 
bedded in prothallial tissue. Archegonia are never found in Angiosperms, 
and we have seen that the exact equivalent of the prothallus cannot be 
determined. The absence of archegonia from the embryo-sac of Gnetum 
naturally suggests an approach to Angiospermous structure. 
The three genera included in the Gnetaceae differ widely in the struc- 
ture of their embryo-sac. In Ephedra , undoubtedly the most primitive 
form, it is completely Gymnospermous ; but certain features in its develop- 
ment both before and after fertilization have suggested to Miss Berridge ( 11 ) 
a new interpretation of the Angiospermous embryo-sac. 
The facts on which this hypothesis is based have been lately published 
by the same author in conjunction with Miss Sanday ( 12 ). They are, 
shortly, as follows. The upper part of the mature embryo-sac in Ephedra 
distachya is partly divided from the lower region by a constriction in the 
wall. The tissue which fills it is on the whole looser in texture, more 
spongy, than that of the lower cavity. But near the apex of this upper 
region is a ‘compact conical mass of archegonia and jacket-cells’ (12, 
p. 129). The initials of archegonia and jacket-cells are alike: they are 
found as a pyramidal group of tubular alveoli in the young embryo-sac 
( 12 , Fig. 3). Each archegonium initial gives rise to the primary neck-cell 
and the central cell. Within the central cell a nuclear division, apparently 
amitotic, separates the nucleus of the ovum from the ventral canal nucleus. 
A ventral canal cell is not formed. 
‘ At the time when the primary neck-cells are being cut off from the 
initials, the tubular cells between the latter are undergoing a series of divi- 
sions, and forming rows of jacket-cells arranged in a regular manner. The 
likeness of these to the central cells, except in the matter of size, is apparent 
in Fig. 4/ . . . ‘ This tissue keeps pace with the growth of the central cell, 
first by cell division, and later by individual growth of the cells. They soon 
become crowded with food material, and the nucleus divides into two by 
direct division long before this occurs in the central cells’ (12, p. 1 67). 
The jacket-cell nuclei are capable in E. distachya of giving rise to apo- 
gamous proembryos after fertilization has taken place in the archegonium 
they enclose. Before fertilization, nuclear fusions are common in the jacket- 
cells. A nucleus from one jacket-cell migrates to another and fuses with a 
nucleus within it. A nucleus may even migrate from a prothallial cell into an 
adjacent jacket-cell ( 11 , p. 282). Mitotic figures with twenty-four chromo- 
somes — the sporophytic number — are found within the jacket-cells at this 
