Rubiales. 
235 
Adoxa . We must dismiss the monotypic Adoxa with little 
more than a passing reference. Like many other forms whose 
characters suggest reduction in descent (Adoxa is essentially 
geophilous in habit), the affinities of this curious genus are obscure, 
and this largely owing to the doubtful nature of the perianth. If 
the outer whorl of the latter be a true calyx, then Adoxa should 
properly be classed with epigynous Sympetalae ; and its multilocular 
multistylar ovary with one pendulous ovule in each loculus maturing 
to a berry-like fruit, coupled with its capitate inflorescence of regular 
flowers, point to the probability of its umbellifloral ancestry and 
consequent affinity with the descendants of the Rubialian Stock. 
Its place is so indicated in the diagram, p. 230; and its extreme 
specialization is seen in the definiteness of the number of flowers 
in the inflorescence, which is most usually five, in the regularly 
pentamerous arrangement of the lateral flowers and tetramerous 
terminal flower, and in the constantly bipartite stamens. 
* * * t- 
This completes our review of the Rubiales. We have seen 
that their origin is readily referable to a well-defined archichlamy- 
deous ancestry represented by Umbelliflorae. The latter has been 
regarded as an offshoot from the calycifloral plexus typified in the 
heterogeneous Rosales. The other cohort of Inferae, Campanulatae, 
will form the subject of the next chapter. 
[To be continued J. 
