natural Family of Floats called Composites. 141 
hypothesis not meant to be extended to all—this part, even in its 
simplest state, may be considered as formed of a series of modified 
stamina: Or, merely to state the facts from which the conjecture 
originates, that there are certain families in some of whose genera 
this organ exists in its simplest form, that of an undivided fleshy 
ring; while in other genera of the same families it consists of 
several distinct bodies alternating with the stamina, and in some 
cases putting on the appearance of barren filaments. 
This hypothesis is chiefly applicable to families in which the 
number of stamina is equal to the divisions of one floral envelope 
only, the nectarium being supposed to be formed of the second 
series: but it receives its principal support from Scitammece*, where 
the glandular bodies belong actually to the same series with the 
perfect stamen. 
I am aware at the same time of several objections to its gene¬ 
ralization. Thus, the nectarium or glandular disk exists in fami¬ 
lies where, though the stamina are definite, the}’' are equal in num¬ 
ber to the divisions of calyx and corolla united ; and moreover, 
in such families where it consists of distinct parts, these parts are 
placed where an addition to the number of stamina is least likely 
to take place, as in Crassulacece. Here, however, as in many 
other cases, the divisions of the disk are opposite to the ovaria; 
they may therefore be supposed more intimately connected with 
the pistilla than with the stamina; an opinion which is I believe 
held, though not yet published, by the ingenious M. Decandolle 
with respect to Fanunculacece. In support of this opinion it may 
be noticed that in Pcconia Moutan , where the disk or urceolus is in 
the state of the greatest development, when a multiplication of the 
pistilla takes place, which in the double-flowered varieties of this 
* See Flinders’s Voyage to Terra Australis, ii. p. 574. 
species 
