150 A. H. Graves, 
In fig. 102 appears evidently the beginning of the formation of 
the dermatogen, at least in the terminal segment at the left. Fig. 
103 shows the segment lines more irregular, the dermatogen has 
became more pronounced, and in fig. 104 shows a distinct differ- 
entiation. A point of interest at this stage is the considerable in- 
crease in size of all the cells, which are, however, only shghtly 
greater in number than in fig. 103. The embryo here measures 
about 0.075 mm. in diameter, as against the 0.05 mm. of the pre- 
ceding one. 
Up to this point the embryo has exhibited a globular form, but 
in fig. 105, where it measures about 0.085 mm. it has commenced 
to elongate. It is at this period that the divisions are initiated in 
the terminal segment, appearing here at the left, which signalize 
the approach of cotyledonary development. 
It has been shown by Schaffner (1897, II, pp. 263-265) in Sagit- 
taria that of the proembryo of three cells, the uppermost develops 
the cotyledon, the middle divides transversely, and of the two result- 
ing segments the upper develops the stem apex and the lower the 
hypocotyl, root, and secondary suspensor. The lowest cell of the 
three-celled proembryo remains undivided, forming the basal sus- 
pensor-cell, 
Obviously the primary segments in Ruppia do not all have des- 
tinies similar to those of Sagittaria. For, to begin with, no second- 
ary suspensor is here found, and only a minute rudimentary root, 
as will be shown later. Again, it seems quite probable, although 
for the reasons stated above no conclusions can be certainly drawn, 
that the terminal segment produces the stem-apex as well as the 
cotyledon. Campbell (1897, p. 49) finds such a condition in Zanni- 
chellia, and the appearances shown in fig. 106 point to such a 
situation here. Solms-Laubach (1878), as noted by Campbell, has 
also reported a terminal origin of the stem-apex in some of the 
Commelinaceae and Dioscoreaceae. This leaves the second segment 
(which, as has already been indicated, probably divides transversely 
to form the second and third segments, as in Sagittaria) for the 
development of the hypocotyl, extraordinarily large in Ruppia, and 
the small rudimentary root adjacent to the suspensor-cell. 
Pl. XIII, fig. 105, however, represents the oldest embryo in which 
the primary segments can be determined with accuracy. In the 
next figure (fig. 106), it is quite impossible to make out definitely 
the line separating the terminal segment from the rest of the 
embryo, to say nothing of any other segment-lines. Nevertheless, 
in this embryo, cell divisions are active in the terminal region 
