6 ON THE MOllPHOLOGY OF CUUCIFEK^. 
stamen, which is obviously alternate with two petals, can become two 
in apposition with each other, it seems very probable that the two pairs 
of long stamens (each stamen being in close apposition with the other 
at its base) represent only two single ones^ and agreeably with this 
explanation I find, on examining MattJdolaj Iberis, and Lepidium, that 
each pair of long stamens is between two petals, and that in the latter 
genns the claws of the petals do not conceal any portion of the two 
filaments. If, then, it is admitted that the two pairs of long stamens 
represent only two single ones, the structure of the flower will be tetra- 
merous throughout, with the peculiarity of having a decided tendency 
in the sepals, stamens, and carpels when the latter are four, to be in 
pairs, two within two, like the 4-sepaled calyx of Epimeditim, and the 
position of each will then be quite normal. 
As thus understood, the calyx consists of four sepals, two within 
two, and the corolla of four petals alternate w^ith them, next come the 
two short stamens alternate with petals, and opposite the external pair 
of sepals, and then more internally the two pairs of long stamens (each 
pair representing one stamen) alternate with the petals and opposite 
the internal sepals. The two carpels would then be alternate with the 
last two stamens {i, e. the two pairs of long stamens) which, as far as 
my observation goes, is always the case. There is also in monstrous 
flowers a tendency in the carpels to develope by pairs, as, for instance, 
the second pair may become elevated on a kind of gynophore within 
the first pair, which have separated at their margins. 
The glandular receptacles of the short stamens Mr. Smith has espe- 
cially noticed (fig. 4, p. 270), and these, together with the six glands, 
I regard as showing a decided tendency in Gmclferm to an increase of 
stamens, and, I believe, to become polyandrous, like Capparhkm and 
VapateracecB^ I, however, take to some extent a different view of their 
nature. The receptacles of the short stamens being much larger than 
the glands, appears to indicate a tendency to the production of a larger 
number of stamens on the sides of the flower (i. e. laterally, opposite 
the tAvo carpels which are right and left the axis), which would then 
agree with Reseda, among the species of which there is a distinct ten- 
dency of the stamens to develope on the sides of the flower more than 
either directly anterior or posterior .- in R, odorata there are no stamens 
anterior. 
I believe there is some variety in the number and form of the glands. 
