582 MR. W. C. WORSDELL ON THE DEVELOPMENT 
Further transverse divisions would then appear to occur in the 
upper portion, by which this part becomes much elongated, and, 
contrary to what is described for Orobanche, much greater in 
length than in breadth (Pl. XXI. figs. 25, 26). About this time 
also longitudinal divisions occur in two directions (figs. 26-29). 
There does not seem to be in these plants such a clear deli- 
mitation of the hypophysis from the rest of the embryo as is 
afforded by Orobanche; but in some embryos its position was, 
nevertheless, fairly easy to trace, as in figs. 31-35, h. It is 
doubtless separated off by the first transverse wall which is 
formed in this part of the embryo. It subsequently becomes 
divided into two parts by another transverse wall. A longi- 
tudinal wall is then formed in each of these cells. Eventually 
the lower portion of the hypophysis, consisting of two small 
narrow cells, becomes destroyed, while the upper portion becomes 
divided up into several cells by longitudinal walls. About the 
time that the longitudinal wall occurs in the hypophysis the 
dermatogen begins to be cut off in the upper part of the embryo, 
the first walls appearing in the lower portion of this part (fig. 30). 
The number of the cells of the dermatogen is increased by anti- 
clinal divisions. Numerous divisions then occur in the interior 
of the embryo, by which it becomes greatly increased in size, so 
that it at length comes to assume somewhat of an oval or even 
semi-orbicular shape (figs. 31-37). 
The developmental structure of the embryo of Christisonia is, 
therefore, in its main features similar to that of the embryo of 
Orobanche as described by Koch. It differs in its minor details 
from that of the latter in being apparently rather more rudi- 
mentary, and in the different parts being less sharply defined. 
This would agree with what we know of the vegetative structure 
of the adult plant, which certainly is more reduced and more 
rudimentary than that of Orobanche; and thus it is not sur- 
prising if in the embryonic condition of the plant a more reduced 
structure is also met with. 
We have seen, however, that the development of the ovule 
is essentially similar at every stage to that of the ovule in 
Orobanche. 
The main part of this investigation was carried out in the 
botanical laboratory of the Royal College of Science, South 
‘Kensington. It was eventually finished and completed in the 
