596 Prain.—A Review of the Genera 
The question remains as to whether Erythrococca , as enlarged by Pax, 
and Athroandra , Hook. f., constitute two distinct genera, or if they may be 
treated as integral parts of one natural genus. As genera in Euphorbiaceae 
go, no great harm could accrue were they kept distinct; their discrimination 
would certainly present no difficulty. But while their separation would be 
a simple matter it does not appear to be necessary. In the first place the 
stipular character, which is perhaps unequivocal so far as the distinction 
between Eiythrococca and Claoxylon is concerned, is not so definite as 
indicating a distinction between Erythrococca and Athroandra . The stipules 
in Athroandra , it is true, remain minute and unmodified. But these organs 
are not, in Athroandra , normal stipules; they are firm, persistent, hyaline 
or scarious bodies that, even in those species which have pubescent twigs 
and petioles, are polished and glabrous. They differ in degree rather than 
in kind from stipules of the Erythrococca type. In all those species which 
can be referred to Erythrococca , as understood by Pax, the stipules are 
accrescent, the filaments are longer than the anthers, the stigmas are 
plumosely laciniate, and the hypogynous scales which alternate with the 
carpels are free and discrete, or very rarely are accompanied by smaller 
additional scales. In all the species which can be referred to Athroandra , 
Hook, fi, the stipules remain unmodified, the anthers are subsessile, the 
stigmas are entire, and the hypogynous scales which alternate with the 
carpels are rarely free, and if they be free are contiguous at their margins 
under the ovary; more usually they are connate in an urceolate disc. 
There is, however, one species with perulate buds, C. panciflorum , Muell. 
arg., which has the stamens, the hypogynous scales, and the stigmas of 
Erythrococca , with the unmodified stipules of Athroandra. On the other 
hand, there is a perulate species, C. membranaceum , Muell. arg., with the 
unmodified stipules, the entire stigmas, and the contiguous hypogynous 
scales of Athroandra , which has anthers borne on long filaments, as in 
Erythrococca. These two species, one a somewhat aberrant Erythrococca , 
the other a somewhat aberrant Athroandra , serve as links between the 
groups to which they severally belong. The intimate relationship of these 
groups is further indicated by a third species with perulate buds, C. poly- 
andrum , Pax and K. Hoffm. (Engl. Bot. Jahrb., xlv. 237). As regards the 
sum of its characters this species resembles C. panciflorum so much that it 
may be looked on as the East African representative of that Angolan 
plant. It combines the anthers with long filaments and the discrete hypo¬ 
gynous scales of an Erythrococca with the unmodified stipules of an Athro¬ 
andra. The stigmas, however, are curiously intermediate ; they are ovate- 
lanceolate and divaricate, and are borne on a distinct, if short, style, exactly 
as they are in one-third of the species of Athroandra. But in place of 
being entire, as they always are in Athroandra, the stigmas of C. poly andrum 
have incised lobulate margins. They thus approach the plumosely laciniate 
