272 Cook: EmsryoLocy OF RHIZOPHORA 
be seen by a comparison of FIGURES 4 and 5, in which the smaller _ 
is the more advanced, or by comparison of FIGURES 21 and 22, in 
which the two structures are of about the same age. It was im- 
possible to determine positively the origin of the archesporium, 
but it was apparently sub-epidermal (FIGURE 18). One or two 
tapetal cells are cut off, but whether they ever exceed this number 
or not the author is unable to say. Karsten figures two each in 
R. mucronata and Bruguiera ertopetala and one each in B. caryophyl- 
loides and Ceriops Candolleana, At this time the two integuments - 
were just appearing (FIGURE 2). The archesporial cell enlarges 
with the enlargement of the nucellus, but the integument under- 
goes the most rapid growth (FIGURES 3, 19). The ovule con- 
tinues to increase in size and the archesporial cell divides into 
megaspores. My only satisfactory preparation of this stage 
showed three megaspore-cells (FIGURE 20), but it is impossible to : : 
say whether this is or is not the regular number. Karsten figures” 
four for both R. mucronata and C._Candolleana. The nucellus is 
always small as compared with the size of the ovule. My 
material was unsatisfactory for following the development of the 
embryo-sac; however, its formation results in the complete 
destruction of the nucellus (FIGURES 5, 21). Karsten figures the 
same condition in this species and in Carallia integerrima. In 
the eight-nucleate stage it was impossible to do more than work 
the details of the egg-apparatus (FIGURES 21, 22), which were in 
every way like the normal egg-apparatus. The author did not 
succeed in finding the polar nuclei or antipodal cells, but found 
what was apparently the primary endosperm-nucleus (FIGURE 23) 
lying in the micropylar end of the sac. It was impossible t° 
follow the early divisions of the endosperm, but from this point the 
preparations were quite satisfactory. 
EMBRYO AND ENDOSPERM 
The growth of the ovule and ovary is now quite rapid. Ti 
embryo-sac increases in size both as a result of growth 
ovule and by destruction of the inner integument, so that we now 
have a sac enclosed only by the outer integument, fille 
an enormous amount of endosperm, in the micropylat 
which is a small embryo (ricurE 6), At this stage the othef 
of the 
ee ESS tae ae 
‘ 
