345 
Campbell . — Studies on the Araceae , III. 
exist in other genera. While this fact may not be of much importance as 
indicating an exceptionally primitive position for the Araceae, it must be 
taken into account in trying to assign these plants to their proper position 
in the system. 
The most striking fact brought out in this series of investigation is the 
extraordinary variability shown in the structures of the embryo-sac itself. 
So far as the writer knows no family of angiospermous plants (unless 
perhaps the Piperaceae) shows so great variation. 
While in some species (e.g. Dieffenbachia Seguine , Anthurium violaceum ) 
the embryo-sac conforms entirely to the usual angiospermous type, in 
other species (e. g. Lysichiton , Aglaonema commutatum , Spathicarpa , and 
Nephthyiis) there are more or less notable deviations from the type. In 
Spathicarpa and Lysichiton these differences are secondary, consisting in 
a remarkable development of the antipodal cells subsequent to fertilization. 
This marked development of the antipodal cells is, perhaps, most nearly 
paralleled by that of Sparganiurn 1 , where there is a similar growth and 
division of the antipodal cells after fertilization. 
The most puzzling type is that of Nephthyiis and Aglaonema commit - 
tatum ) where in the same species extraordinary variation is encountered. 
In neither of these forms was the ordinary angiospermous type found, 
although approximations to this type were sometimes met with. In some 
cases, as we have shown, the number of nuclei in the mature embryo-sac 
is reduced to four, or possibly even two ; definite synergidae may be 
wanting and antipodal cells may be entirely absent. 
On the other hand, the number of nuclei may exceed the normal, 
perhaps in some cases being double the typical number ; but it is by no 
means easy, in all cases, to decide whether the augmented number may not 
be due to a coalescence of two or more young embryo-sacs. The polarity 
shown by the typical embryo-sac seems to be often wanting in these forms, 
and although nuclear fusions may occur, these often seem to include more 
than two nuclei, and more than one of these fusions may take place in the 
same sac. In these cases it is not always easy to determine which is 
destined to become the egg-nucleus, and it not infrequently happens that 
the embryo is developed at the side, or even at the chalazal end of the 
embryo-sac. 
How far these variations are normal it would be hard to say; but they 
are hardly paralleled in any other angiospermous plants. Perhaps a study 
of material of these species collected in their natural habitat, or of other 
showing the same characters, may make it possible to decide what the 
normal condition is. As it is they resemble more nearly than any other 
1 Campbell, The Flower and Embryo of Sparganinm. Proc. Calif. Acad. Sciences, Botany I, 
No. 9, 1899. 
