378 
DUNCAN S. JOHNSON 
G. Abnormal Embryo Sacs 
In addition to the exceptional occurrence of peripheral cells and 
of a second synergid, which were referred to in discussing the normal 
development of the embryo sac, because they show a resemblance, of 
P. hispidula to other species, there is a less common,probably decidedly 
rare, aberration still to be mentioned. This has not been described 
'Or figured in discussing the normal development for the sake of avoiding 
confusion. This peculiarity is the occurrence of archesporial cells, 
megaspore tetrads and embryo sacs in pairs or trios in the same nucel- 
lus. The occurrence of more than one archesporial cell in a nucellus 
lias been noted in a large number of species of angiosperms since they 
were discovered by Strasburger (1879) in Rosa, and by Fischer (1880) 
in a half dozen other genera of the Rosaceae. Not only is this true, 
but in the cases of Fagus, Corylus and Carpinus (Benson, 1894), of 
Delphinium and Senecio (Mottier, 1893, 1895), and some others, two 
or more embryo sacs may even mature until ready for fertilization. 
The sort of doubling of the megaspore and embryo sac just referred 
to occurs somewhat rarely in P. hispidula. In the youngest stages 
seen parietal cells were already formed. The most frequent type of 
doubling found is that where the twin megaspore mother cells or em- 
bryo sacs are side by side and each has essentially the structure of the 
ordinary single cell or sac (fig. 113). In other cases, however, two 
archesporial cells may lie one above the other in the axis of the ovule 
(fig. 114), and, still more rarely three of these cells may occur in one 
longitudinal row (fig. 115). The appearance of these groups of cells, 
especially the pairs or trios in longitudinal rows, suggests of course the 
possibility of there being rows of megaspores formed by the division of 
a mother-cell. But this possibility is practically excluded by the fact 
that the stage shown in figure 113 is too early to show such a division, 
while in the cases given in figures 114 and 117 the nucleus of one cell in 
each case shows clearly the typical synapsis characteristic of the ma- 
turing megaspore mother-cell. There can be no doubt that the cases 
mentioned really have twin megaspore mother cells. It is probable 
also that the cases shown in figures 115 and 116 are really groups of 
megaspore mother cells rather than of megaspores themselves. 
The further fate of these multiple archesporial cells is also proof 
of their real nature. A considerable number of examples were seen of 
the further development of the paired archesporial cells of both types. 
Thus figure 118 shows two superposed embryo sacs which have evi- 
