NORMAL STRUCTURE OF CRUCIFEROUS FLOWERS, 271 
The result can be easily arrived at in the following manner :—The 
glands in Cheiranthus are six, in Iberis only four in number 
Two of the glands of Cheiranthus, as before stated, are situated at 
the base of the pod and between the cells. Without doubt, then, these 
two extra glands represent the two absent cells of the pod (a, Fig. 1, 
and Fig. 4). This will reduce the glands to four each, and were they 
situated in the same places in both genera, only ten stamens could be 
arrived at. But it is not so. In Cheiranthus the undeveloped stamens 
are outside the two pairs of long stamens (8,B, Fig. 1), and in Iberis 
they are inside the two short stamens (c,c, Fig. 2). 
We may now reasonably infer that the two pairs of longitudinal 
glands are wholly suppressed in Cheiranthus (Fig. 1), but developed 
in Iberis (Fig. 2 at c,c), whilst the transverse glands do not occur in 
Iberis, but are present in Cheiranthus at B,B, Fig. 1. This is simply 
applying the same rule to the stamens that is applied to the cells of the 
pod, and by supplying from one genus the organs that are deficient 
in the other, and vice versá, we get a 4-celled pod, and sixteen sta- 
mens in two whorls of eight each, as in Fig. 3.* I may add that by 
longitudinal is meant the greater width of the flower-plan from one 
single short stamen to the other; by transverse, the lesser width, or 
from one pair of long stamens to the opposite pair. 
There are four other points in Cruciferous flowers that may be called 
the petals and sepals; and Colonel Madden, in his description of = plant Beas 
ceedings Bot. Soc. Edinb. 1855, p. 43), says ‘that the M are ‘ disposed in tw 
or four sets.’ But upon a a careful re-examination of a number of flowers, I cin 
discover any such arrangement. The stamens, Madera when numerous (never 
owever more than sixteen in any flower I have opened), are Ó into a tuft 
never that two are nearer to nether i wes to the a ing ones, no one is 
really withins ide another 1 the base. n the ves are dinkel (and they 
fall off with the greatest facility = Mcd their scars form a single, irregularly 
waved line, at some distance from the ovary, and surrounded by a slightly glandular 
ring, waved and indented by the cavity i d each flament, This arrangement is 
partieularly evident after the flower is expanded, and the filaments have more 
their natural position. To me arae iti is clear that the whole 
of the stamens ctu in 2 this as in other Crucifera, to a single verticil. This view 
rucifera 
of the case would tend to confirm the most plausible of the modern theories of the 
sortai of Crucifere,—ihat one so clearly expounded by Messrs. Webb and 
M 
ray, 
pros of en rth American Plants.’ " (Hook. Journ. Bot. vii. pp. 353, 
n ia aurea, Nutt., the calyx is nearly equal at the base, and there are 10 
glands, 8 b vp at the base of the c 2 gf rence the base of the 
shorter stamens. (Torrey and Gray, Fl. N 
