EDGEWORTHIA, ERYTHRJEÉA, AND EUPHORBIA. 71 
since they bear an obvious resemblancé to the upper edge of 
the convolutions seen in an early stage. 
If this hypothesis be found to be true, it bears strongly 
on Orchidez, since it shews that carpellary lamine may 
exist without any obvious style; hence the 9 ribs may be 
really abortive carpella. 
This is a question of great importance. 
If it be remembered, that the apex of tbe leaf is the part 
first: formed, it appears impossible but that there must be 
some traces appreciable of the original number of carpellary 
leaves formed. Consequently, Orchidez must have 3 stig- 
mata at least. 
In Edgeworthia buxifolia three STE exist, yet all are 
united by lines of stigmatic tissue; hence the usual"stigma 
of Orchidez may consist of three confluent ones. 
In Erythrea nothing can be more satisfactory than the 
manner in which the margins are involute, or the evident 
origin of the ovula from their involute margins. 
Between the edges of the leaves, and the first part of the 
involuted lamina, cellular tissue is interposed; but it has 
nothing whatever to do with the ovula. Nothing can be more 
distinet than the fact, that the pistillum in this consists of 
two concave or sub.involute leaves with thickened margins. 
The pistillum is of latish formation, appearing after the 
anthers, or even perhaps the petals. 
Pl. 29.—Fig. 3. a, very early, like a bilabiate capsule; b, 
opaque from air; c, placenta—e, continuation of placente 
with stigmata. 
Why is it, that the whole inflected part of this leaf, does 
not bear ovula ? 
In Euphorbia the three styles are bipartite, each has one 
vessel which branches at the fork ; there is a decided obli- 
-quity in the stigmata, and a connivance towards each other, 
which shews that they are not really independent organs. 
No furrow exists along the centre. 
Pl. 30.—Fig. 5. a, e;—5, stigmatic surface. 
The existence of the groove along the centre of very 
many linear stigmata, and its absence along the lobes into 
