euishable in habit, while the very same character is rejected when 
applied to the separation of an assemblage of genera, all distinctly 
combined by their habit.”. The opinion of Decandolle and others, 
that the apparent leaves of Stellate are in part true leaves and in part 
leaf-like stipules, is thus objected to. If a part of the leaves of each 
whorl of Galium were stipules, they must bear a certain proportion to 
the true leaves; suppose the whorl to consist of two leaves, each will 
have two stipules, and consequently the whole number of parts in the 
whorl must be six, and in al] cases the number must be some power 
of three. The frequent tendency in the whorls to vary from 4 to 6, 
or from 5 to 6, or from 6 to 8, seems to me an incontrovertible proof 
that the apparent leaves of Stellate are true leaves and not a modifi- 
cation of stipules.” 
The weak angular stems and Baoan fruit of Stellate are very 
common in two other tribes retained by Lindley in his Cinchonacee, 
viz. the Spermacocez, and the Anthospermez, and the whole of the 
latter tribe, and many of the Spermacocee, have precisely the habit 
either of some Galia, or more frequently of the Asperule; There re- 
mains therefore the sole question, whether the whorl of leaves in Stel- 
late, consists of two true leaves, and two or more leaf-like stipules, or 
of four or more true leaves without stipules. In the first place it may 
be observed, that in most Dicotyledonex, where the leaves are verticil- 
late, the number of leaves in each whorl bears a regular proportion 
to that of the angles of the stem, either the same number, double the 
number, or half the number; the stem of Stellate is almost always 
quadrangular, and the number of leaves varies from 4 to 10 or 12; 
which is a strong indication that only two or four are true leaves, and 
this is confirmed by the number of cases (especially in Asperula) 
where two opposite leaves are much larger than the others. In the 
next place, comparing Stellate to the nearest Tribes of Lindley’s Cin- 
chonaces \ we observe i in'many: of these, especially in Hedyotidex, Cof- 
,and A , that the stipules are connec- 
ted with the petioles ‘of the leaves into one sheath, often very short, 
from the margin of which arise two opposite leaves, and between the 
leaves on each side, one, two, three, four, or more teeth or setz, these 
teeth or sete are generally, it is true, either membranous or very slen- 
der, but they not unfrequently are to a certain degree foliaceous, and 
in Gaillonia they are sometimes small and setiform, sometimes exactly 
like the true leaves. So in Stellate the apparent leaves are all con- 
nected together at the base in a sheath, very short it is true, but almost 
always evident, and the lateral leaves supply the place of the teeth or 
sete in the other tribes. Thus, when the whorl consists of four, there 
are two real leaves and one stipulary process on each side, as in most 
Anthospermee and many Hedyotidee, Coffeaceze, &c.; where the whorl 
consists of six, there are two stipulary processes on each side, as in 
