The Embryo-Sac of Angiosperms. 381 
The above is a mere outline of Pearson’s main argument; 
further evidence must be sought in his paper. It has been too 
recently published for adequate comment to have been made on it, 
but whatever criticisms may he in store, the facts he describes will 
always be of far-reaching importance in any discussion of the 
morphology of the embryo-sac, and in his argument we find both 
an important clue to the origin of the endosperm and the most 
probable phylogeny yet suggested for the embryo-sac of the 
Angiosperms. 
In the above brief sketch of current ideas on the morphology 
of the embryo-sac, we have purposely avoided much debatable 
ground and many disputed facts by writing as if the authors of 
these theories had only to take into account the normal eight- 
nucleate and bipolar sac. Now we must consider what bearing 
certain exceptions to this normal rule have on these questions. In 
the first place, there have been described several sacs which are not 
bipolar, and in which free nuclear division stops at the fourth 
instead of at the third generation, so that they contain sixteen nuclei 
instead of eight. At first sight these would seem to be instances of 
the preservation of a relatively primitive type of embryo-sac, and 
as such, these cases have been used to uphold most of the theories 
noticed above. But on the other hand, there has been some 
discussion as to whether these forms are not to be regarded as 
specialised from the normal type rather than primitive, and since 
this question is still unsettled, we have reserved them for discussion 
at this point. 
The following are the cases referred to 
1. In Peperornia 1 the first two divisions in the embryo-sac 
give rise to four nuclei, arranged “like the spores of a tetrad,” 
whose further divisions result in sixteen nuclei, scattered through 
the sac. One of these functions as an egg-nucleus; of the remainder, 
eight (usually) in P. pellucid a, P. arifolia and P. sitensii and fourteen 
in P. hispidula, fuse to form the primary endosperm-nucleus and 
the remainder ultimately degenerate. 
2. In three genera of the order Penaeaceae 3 (the remaining 
genera have not been investigated), the first two divisions form a 
tetrad of nuclei arranged just as in Peperornia ; these separate, and 
each again divides twice, so that four groups of four nuclei each are 
formed at the periphery of the sac. Cells are organized round 
three of the nuclei in each group, in the form of an egg-apparatus, 
and the four remaining nuclei move to the centre and fuse to form 
the primary endosperm-nucleus. Euphorbia procera, at the earliest 
stage known, shows four nuclei at the periphery of the sac, whose 
further behaviour is exactly that of the first four nuclei formed in 
the Penaeaceae; but the earlier history of these nuclei has not been 
followed out. 
3. In Gunnera the first four nuclei are again tetrahedrally 
arranged. They undergo four successive divisions, and of the 
resulting sixteen nuclei, three form a micropylar egg-apparatus, six 
form an antipodal cell-group, often in two sets of three cells, and 
the remainder fuse to form the primary endosperm nucleus. 
On the hypothesis that these types of sac are primitive, they 
1 Campbell, 1899, 1901 ; Johnson. 1900, 1907 ; Brown, 1908, 
2 Stephens, 1909. 
