is too imperfect to detérmine whether it may have been really — 
our plant or OxyLosrium ellipticum of Hortus Kewensis. 
Caxuistacuys ovata differs but very little from lanceolata, 
except in the form of its leaves and their being more gene- 
rally, though not constantly, ternate ; in Janceolata they grow 
sometimes opposite, sometimes by fours or by threes, and 
not unfrequently without any order. The calyx, and parti- 
cularly the germen in both, are covered with long silky hairs. | 
Perhaps the legumen was not far enough advanced to deter- 
mine it accurately, but, in the early stage, it did not appear 
to me to be properly multilocular ; but, being very fieshy, 
and the seeds partly imbedded in it, the substance of the 
pod runs some way between the seeds, but not so as to make 
a complete septum. The ripe Jegumen is certainly unt 
locular. eee 
Our plant, as well as Canuistacnys lanceolata, was, we 
believe, raised in the garden of the Jate Empress JosepnivE, 
at Malmaison, from seeds brought from New-Holland by 
Captain Baupin. 3 
It is by no means tender, but requires to be protected 
from frost in the winter season.. Propagated by cuttings. 
Flowers in July and August. 
Communicated Jast year by Messrs. Loppicgs and Sons, 
and we have been favoured with flowering specimens of both 
species at the present time, by the same liberal friends. 
> 
