70 DEVELOPMENT OF THE FLOWER 
Brassicaceous Cruciferc.—0Of the structure of this group 
the following explanations may be giv 
1. It may be bicarpellary, the serine being produced 
from the midrib of the carpellary leaves. 
This will not explain the difference in structure between 
that which is supposed to be the midrib and the lamina ; 
it will not explain the obvious cessation of the lamina along 
this line, nor the intimate union of the margins of the 
lamina—a union which has altogether the appearance pre- 
sented by the dorsal line of most carpellary leaves. 
It is, besides, contrary to all analogy, although perhaps 
not an improbable structure. Vide, excellent observations 
of Lindley on this point. ; 
2. It may be bicarpellary, the placentz corresponding to 
the margin of the leaves. 
This however will not explain the separation between the 
placenta and the margins, as in the former instance; neither 
will it explain the manifest convexity of the style in a line 
from the placente, its depressed surface along that line 
where it should be most convex ; nor the fact that the high- 
est part of the style terminates this line. Besides, the vena- 
tion is against this view. 
3. It may be quadricarpellary, 2 of the carpella being 
reduced to mere lamine without ovuliferous margins, and 
the others reduced to a midrib, and narrow ovuligerous 
margins, produced upwards into 2 perfect connate styles, 
and linear confluent stigmata. 
This will explain every point of structure, and does not 
assume any thing outrageous. 
It is borne out by the appearance of the exceedingly 
young pistillum, in which 4 involutions are manifest. It 
explains the separation in structure between the laminz, and. 
along the placentigerous lines, although the appearance of 
these with regard to a midrib is less satisfactory than could 
be wished. 
It explains the rim visible at the upper end of the lamina 
between the placentigerous margins, which rim I conceive 
to represent the stigmata of the laminar carpellary leaves, - 
