Campbell .— The Embryo-sac of P and anus. 787 
the formation of division cells, while in Pandanus cell-formation does not 
take place until a later period. 
It is in the subsequent history of the embryo-sac that Pandanus shows 
the greatest departure from the usual angiospermous type. The extra¬ 
ordinary development of the antipodal cells at the time of fertilization 
exceeds that found in any other form. In Peperomia pellucida and in 
Gunnera , where there are sixteen nuclei in the adult embryo-sac, there are 
six antipodal cells, and in some of the Grasses 1 there may be as many 
as thirty-six, and the number is often increased in many Compositae. But 
in all of these except Peperomia and Gunnera there is first developed a 
typical embryo-sac with three antipodal nuclei. These later undergo 
secondary divisions resulting in the increased number of antipodal cells. 
The increased number of antipodal cells, subsequent to fertilization, has 
been observed in a considerable number of forms, both Monocotyledons 
and Dicotyledons. 2 
In Pandanus , unlike these other forms, the mass of antipodal tissue is 
preceded by the formation of a considerable number of free nuclei, and the 
solid mass of antipodal tissue is formed by the subsequent division by cell- 
walls between these free nuclei exactly as is the case in the formation of 
the endosperm. This emphasizes the homology which undoubtedly exists 
between the two types of gametophytic tissue developed before and after 
fertilization. 
The formation of the endosperm nucleus from the fusion of several 
nuclei is not peculiar to Pandanus. The same thing occurs regularly in 
Peperomia and Gunnera , and there are numerous other instances that have 
been observed in other plants. It is pretty clear from a study of these 
forms that the theory of the fusion of the polar nuclei being a sort of 
sexual process is quite untenable. Just what it means is hard to say. 
While there are marked differences between Pandanus and Sparganium , 
especially in the condition of the embryo-sac at the time of fertilization, 
and the much better developed embryo in Sparganium , still the general 
morphology, and particularly the great development of antipodal tissue 
which finally is formed in Sparganium , confirm the view that the Pan- 
danaceae and Sparganiaceae do really belong near together in the system. 
There seems no valid reason for supposing that the condition of the 
embryo-sac found in Pandanus is not really primitive. The genus on 
other grounds has been placed near the bottom of the series of Mono¬ 
cotyledons, and the structure of the embryo-sac certainly confirms this 
view. The arguments that have been brought against assuming that the 
1 Cannon, W. A. : A Morphological Study of the Flower and Embryo of the Wild Oat, Avena 
fatna , L. Proc. Calif. Acad., 3rd Ser., Bot., vol. i, 1900, pp. 329-64. 
2 For a full account of the occurrence of more than three antipodal cells, see Coulter and 
Chamberlain, Morphology of Angiosperms, 1903, pp. 97-102. 
