MR. G. BENTHAM ON AUSTRALIAN PROTEACELE. 63 
(when several) iit one double row, the obliquity or eccentricity of 
the ovary and ovules, and other minor circumstances have left in 
my mind the conviction that both are essentially monocarpellary. 
The division of the end of the style of each carpel oecurs in many 
bicarpellary gynecia, as in Cleonia, Cordia, &c. Composite, 
on the other hand, Paronychiacez, Amarantacez, and many others 
are examples of a really 2-3-carpellary gynecium with a single ovule, 
There are two other genera of Proteaces whose remarkable 
fecundating-apparatus requires special notice, Conospermum and 
Synaphea. In the majority of species in both genera the perianth 
bears externally a somewhat similar shape, a broad upper lobe 
concave over the anthers and remaining erect when the flower 
opens, the three other segments narrower and becoming recurved 
or reflexed, at least above the anthers. In both the anthers are 
peculiar: one in each flower has two perfect cells separated by a 
distinet connectivum ; the two lateral ones have each one perfect 
cell and one barren one, usually reduced to a simple protuberance ; 
and the fourth is quite barren. The perfect cells, instead of being 
at first closed and then opening by a longitudinal slit, are shaped 
like a hemispherieal eup or bowl open on the broad face; but in 
the unripe bud the perfect cell of each lateral anther is closed 
face to face against the adjoining cell of the 2-celled anther, so 
that the four cells of the flower are united into two globes, which, 
the moment they are separated by the expanding of the flower, 
let fall the enclosed pollen, but not upon their own, or rather their 
sister stigma, which is well kept out of their way. In Conosper- 
mum (Plate II. fig. 11) it is the upper anther under the broad 
concave upper perianth-segment which has the two perfect cells ; 
the opposite lowest anther with the lower cells of the lateral 
anthers are barren; the style has lengthened at an early stage 
far beyond the stamens, with a lateralstigma near the eud which 
in the bud is turned towards the upper lobe; but as the flower 
opens, as observed by the late Dr. Graham of Edinburgh, the 
style is elastically bent back, presenting its stigma to the side 
of the barren anthers far away from the pollen-dust let loose 
by the separation of the perfect cells. Synaphea (Plate II. fig. 12) 
required, however, a different contrivance ; for there it is the upper 
anther under the coneave segment and the upper cells of the 
lateral anthers that are barren, the lower ones perfect. The 
style is, as in Conospermum, lengthened far beyond the anthers and 
expanded into a disk, stigmatie on its upper surface, which is bent 
