597 
Erythrococca and Micrococca . 
type of stigma characteristic of Erythrococca , to which group, in spite of its 
markedly aberrant features, C. polyandrum is most conveniently referred. 
The existence of these intermediate species justifies the treatment of A thro- 
andra as a sub-genus of an Erythrococca somewhat wider even than Pax has 
indicated ; the criterion which distinguishes this enlarged Erythrococca 
from Claoxylon , A. Juss., is that it has perulate buds, while Claoxylon 
has not. 
Throughout this genus, in spite of the differences as regards hypogynous 
scales and stigmas which occur in the species of its two sub-genera, the fruit 
and the seed are remarkably uniform, and the account already given under 
E. aculeata is applicable to the fruit of all its congeners except that in the 
majority we find no hairs on the pericarp, and that in a few the arillus does 
not quite envelop the testa. 
Mueller (DC. Prodr., xv. 2,775) term s the fruit in Claoxylon —-which 
genus, as understood by him, included Claoxylon itself as here understood, 
Micrococca as here understood, and the whole of Erythrococca as here 
understood, except E. aculeata , Benth. — capsular, 2-3-coccous. This charac¬ 
terization is sufficiently general to admit of its employment for all three 
genera. The seed Mueller speaks of as ecarunculate and as covered by 
a coloured epidermis which soon becomes loose. This ‘ epidermis * is the 
arillus ; it never becomes loose, because it shows no trace, at any stage, of 
organic union with the testa. Welwitsch, on the other hand, refers, in the 
case of C. pauciflorum, to a fleshy endocarp (Cat. Afr. PI. Welw., 975) ; on 
examination Welwitsch’s specimens show that this ‘ endocarp * is the arillus ; 
there is no trace of organic union between it and the pericarp. The arillus 
is viscid, as Welwitsch remarks under C. Welwitschianum ; it therefore 
sometimes adheres, now to the seed coat which it overlies, now to the inner 
wall of the pericarp which encloses it. The apparently incompatible inter¬ 
pretations of Mueller and Welwitsch are thus not only readily intelligible 
but easily reconciled. The seeds of C. Welwitschianum and of C. triste are 
said by Welwitsch to have an arillode, but his own specimens show that the 
structure is really an arillus. 
Bentham (Gen. Plant., iii. 309) speaks of the seed of Claoxylon — the 
genus being understood by him as it was by Mueller, except that it does 
not include Micrococca —as ‘ estrophiolate *; he does not mention the arillus. 1 
The fruit of Claoxylon is said by Bentham to break up into 2-valved cocci 
{capsula in coccos 2-valves dissiliens). This description applies only in a 
qualified sense to any species of the sections Adenoclaoxylon or Athroandra , 
which we have now transferred to Erythrococca. The same qualification is 
needed in the case of the section Discoclaoxylon , where the cocci, though 
1 The first author to allude definitely to the fact that the seed in Claoxylon is arillate was 
Kurz (For. Flor. Brit. Burma, ii. 395) : Kurz had, however, the advantage of having collected the fruits 
and examined the seeds of certain species of Claoxylon in the field. 
