300 
H. F. Wernham. 
limits of a single family (Celastraceae). 1 In other eases epigyny 
occurs with sufficient frequency to be regarded as the reflection of 
a slight general tendency, as is seen among Centrospermae; in this 
particular example the tendency has impressed itself to a corres¬ 
ponding extent upon the sympetalous descendants Primulales, as, 
witness the genera Maesci, Samolus, etc. We must point out, 
however, that such isolated cases of epigyny as these last admit of 
an alternative interpretation, namely that the inferior position of 
the ovary is here not an ancestral character, but represents a 
tendency initiated in the progeny lather than derived from the 
stock; such, we have suggested in a previous chapter, is probably 
the case with Gesneroideae. 
Each instance must be judged upon its own merits ; but in any 
case the question is largely speculative. Broadly speaking, it would 
seem that, in cases where the tendency to epigyny is approximately 
no stronger ( i.e characterizes no more species) in the sympetalous 
progeny than in the ancestral stock, the character in question may be 
regarded in the progeny as derived from the ancestor. This is the 
case of Primulales and Centrospermae. On the other hand, in cases 
where epigyny characterizes a considerable number of species of 
relatively close and restricted affinity in the progeny, while it occurs 
only very rarely if at all in the ancestral stock, it is to be regarded 
as a tendency initiated in the progeny, and not derived from the 
stock. This is the case of Gesneraceae and Geraniales; and this 
brings us face to face with the problem of the systematic position 
of the Vaccinioideae. For the last-named group may not impossibly 
stand to Ericaceae in the same relation that we have represented 
the epigynous Gesneroideae as standing to Gesneraceae. We must, 
then, digress from our summary at this point, in order to complete 
our study of the epigynous Sympetalae by the consideration of this 
problem. 
VaCCINIOIDE/E. 
The alternative aspect of the question, namely, that the epigyny 
of Vaccinioideae is an ancestral character, has already been 
presented in Chapter III; and in the search for an epigynous 
ancestry we were led to the conclusion that the group in question 
has been derived from a stock different from that of the rest of tha 
Ericales,— i.e., the Rosalian Stock. If this conclusion be the true 
1 Calycanthaceae, a family of Ranales, marks a step on the way 
to epigyny—the flowers are perigynous. 
