Tubiflorce. 
3°5 
folding of one leaf, of by the fusion of more than one, it will be 
reasonably conceived that the most primitive ovary would be 
unseptate— e.g., the typical apocarpous carpel of Ranales. Septation 
is the result of more than one cause; it may he the consequence of 
inturning of the carpellary leaf-margins; of projecting placenta; 
or of the organization of membranes to serve for the special purpose 
of chambering the ovary, as is the case with the secondary septa 
to which we have already made reference ; hut it is doubtful how 
far, if at all, this last condition obtains in the primary septation of 
the ovary. A septate ovary, again, may be the phylogenetic product 
of the union of apocarpous carpels. The problem in each case is 
one of floral development from the primordia to maturity, and any 
conclusions unbased upon actual research in this direction may be 
only worse than useless. 
In certain cases, however, the progression to what we may 
name secondary unilocularity of the ovary may be traced in the 
observation of a series of allied forms as the outcome of degradation 
of septa ; and it is urged that the higher groups characterized by a 
unilocular ovary may not impossibly be descended from stocks with 
septate ovaries in this way, the loss of the septa being sometimes 
correlated with reduction in the number of ovules. A case in point 
(except in the last-named particular) is that of the Primulales, with 
their “ free-central ” placentation, descendants of the caryophylline 
stock, in which the basal placenta which projects into the single 
chamber of the ovary is derived admittedly from an aggregate of 
degraded septa. 
We shall have need to resort to some such tentative consider¬ 
ations as have been briefly sketched above in the study of the 
higher groups which we are now about to approach. In the absence 
of experimental evidence based on observation of the ontogeny of 
floral forms, we must take broad general considerations of biology 
as our guide in these questions of phytogeny and descent; but we 
cannot insist too strongly on the fact that our conclusions through¬ 
out amount to no more than mere suggestions as to the possible 
course of descent along a few main lines, together with some 
indication, perhaps, of the lines along which research might usefully 
be directed. 
We now leave the Transitional Group and proceed to deal, in 
the next chapter, with the progeny of the stock which it represents. 
(i to be continued ). 
