108 MR. BENTHAM ON MALVACE.E 
either in other Malvaces or in Sterculiacee ; but these are generally 
limited to a few genera only, or are too variously combined to war- 
rant the maintenance of the group as a distinct order. Thus, the 
digitate leaflets of the five first genera are unknown in Malvacee 
and Tiliacez, and in Sterculiacee only occur in a very few species 
of Sterculia. The bracteoles in most Bombacex are small and 
inconspicuous, as in Fugosia, &c.; but in the subtribe Duriones 
they are united in an involucre which is often entire, completely 
enclosing the young bud, and bursting irregularly as the calyx 
enlarges. The calyx, sometimes truncate and toothed as in Thes- 
pesia, &c., or more rarely 5-cleft as in most Malvacee and Stercu- 
liacez, is more frequently entire in the young bud, splitting irregu- 
larly into three to five lobes as the flower expands. This is rare in 
Malvace: and Sterculiacee, but occurs in the subtribe Brownlowiee 
of Tiliaceze. In Ochroma, Cheirostemon,and Fremontia the generally 
thick calyx-lobes are more or less expanded on the sides into 
thinner imbricating edges, which is quite exceptional among Co- 
lumnifere. The staminal column, usually more or less Malvaceous, 
is in Hriodendron, Chorisia, Cheirostemon, and .Fremontia excep- 
tionally divided into five lobes, each of which usually bears two 
long linear parallel adnate anthers, which might easily be taken 
at first sight for the parallel cells of single anthers, were it not that 
these are occasionally three instead of two, that the two are often 
not strictly parallel, one being longer or inserted rather higher 
up than the other, and that their real nature is shown by a com- 
parison with the more numerous but similarly adnate anthers of 
Matisia, Quararibea, and Ochroma. As a further evidence of the 
close connexion of Bombacex with Hibiscez, I may observe that 
since the above was read we have received some numbers of the 
* Botanische Zeitung,’ in which Alefeld proposes to remove Gossy- 
pium, Thespesia, and their allies from Hibiscez: to Bombacee. 
Bompax; Linn. 
Bombas differs chiefly from Pachira in its shorter flowers and 
in the dense wool enveloping the seeds within the capsule. In Pa- 
chira humilis, Spruce, and P. Fendleri, Seem., the flowers are longer 
than is usual in Bombax; yet as the capsule is woolly inside as in 
the latter genus, these two species must be transferred to it. The 
small-flowered Eriothecas and the Indian Salmalia, proposed as 
separate genera by Schott, do not appear to be founded on any 
better character than the greater or less degree of union of the 
stamens in pairs, which is variable in the same species; and we 
