i54 
H. F. Wernham. 
families of Bicarpellatae,—a tendency to epigyny. This is indicated 
in many and varied degrees of completeness, and more than 35% of 
the total number of species in the family are concerned ; these belong 
to the sub-family Gesneroideae. The existence of these degrees, 
varying from slight to almost complete epigyny, suggests rather 
strongly that the latter condition represents a tendency peculiar to 
Gesneracere—a secondary tendency towards epigyny, not comparable 
with the general (“ calycifloral ”) tendency to which attention has 
been drawn in our earlier chapters; in other words, Gesneracese 
have availed themselves, so to express it, of the advantages 
associated with epigyny, coupling it with characters typical of a 
highly-evolved member of the Bicarpellatae plexus. This conclusion 
seems to derive some support from the conspicuous absence of any 
such tendency from any other groups of the plexus,—and, indeed, 
from the remoter archichlamydeous stock to which this plexus is 
held to owe its origin. 
How far this line of reasoning may be adopted in the case of 
other groups which differ essentially from their near allies only in 
the relative position of the ovary must be judged on the merits of 
each individual example. The Vaccinioideae provide a case more 
or less in point (chapter III); they will be examined from this 
aspect when we come to deal with the large groups of Sympetalae 
characterised by epigyny. 
The epigyny-line seen in Gesneraceae is realized in the small 
family Columelliaceae, consisting of two or three species only, 
shrubs of western tropical America. The corolla is sub-regular, 
of the Ramondia type, with a short tube. The androecium consists 
of two stamens constantly, but there is reason to suppose that 
each is the fusion-product of a pair of stamens ; in any case the 
resemblance to Gesneraceae is clear. Further, the inferior ovary 
shows transition from the unilocular to the bilocular state, the 
latter resulting from the union of the projecting placentae. The 
ovules are numerous. The systematic position of Columelliaceae has 
been made matter for considerable doubt, and they have been 
variously placed in different systems ; the two modern systems agree 
in associating them with Gesneraceae. They are an old family from 
the point of view of descent; and this, coupled with their shrubby 
habit and their resemblance in so many essential points to Gesner¬ 
aceae—especially to those with a corolla of relatively primitive 
type—favours the suggestion that they may be the sole surviors 
of that branch of the scrophulariaceous stock from which 
Gesneraceae have been derived. 
