924 CORRESPONDENCE. 
'eonfluent, be there ovuligerous ; the supposition is as admissible as that the 
ovuligerous margins meet in the axis, and it would be no easy task to offer a 
proof against the existence of the fact: it is clear that as regards the develop- 
ment of the fruit, the growth would be precisely the same under either hypo- 
thesis; in both cases the cells, as well as the placente, would in like manner 
stand opposite the lobes of the stigma, the dissepiments would equally corre- 
spond, and the dehiscence be the same, either from the resilience of the con- 
fluent margins, or the rupture along the median nervure of the carpels ; but 
which of the two actually occurs in particular cases, o s of 
judging, except by analogy. Now analogy shows that both conditions can 
leaves ; it is known as the ordinary rule, that the stamens are placed with their 
backs exteriorly, with their faces turned towards the axis, or in other words 
introrse ; but we know they are occasionally extrorse, with their backs turned 
towards the centre. If, then, the staminal normal leaflets be both introrse and 
extrorse, it is.reasonable to suppose that the carpellary leaflets may be subject 
to the same conditions, and it cannot be said that I have violated the law of 
probability in applying this explanation in the case of Bignoniacee, where the 
circumstances attendant on the growth of the fruit cannot be accounted for on 
any other hypothesis. We have indeed something like proof of an analogous 
ence in the Brazilian genus Vasconcella, which has a 5-locular ovary with 
5 sterile complete dissepiments, intermediate with which are as many parietal 
placentæ ; also in Cucurbitacee, where 3 or 5 extrorsely-formed carpels, with 
-are sometimes placed back to back, with their 
Vil WUC. 
Mr. Brown demonstrated that in Orchidacee. 
rule : 
alternate with the lobes of the stigma; this is ae aa m onum 
lridacee. This anomaly was accounted for by 
genious but not quite satisfactory hypothesis, 
normally bifid, and that the lobes of the adj 
they thus appear as if alternate, 
the placentze. 
Con 
cerning Montiea and Reyesia, the evidence is certain that they belong to 
