536 
MORPHOLOGY. 
BOOK VI. 
the anthers are mere alterations of the margins of petals, there 
is no difficulty in demonstrating. In Nymphsea the passage 
from one to the other may be distinctly traced. In double 
roses the precise nature of this metamorphosis is shown in a 
very instructive way : if any double rose is examined, it will 
be seen that those petals which are next the stamens contract 
their claw into the form of a filament ; a distortion of the 
upper part, or limb, also takes place ; the two sides become 
membranous, and put on the colour and texture of the anther ; 
and sometimes the perfect lobe of an anther will be found on 
one side of a petal, and the half-formed mis-shapen rudiment 
of another on the opposite side. In Aquilegia vulgaris this 
transformation is still more curious, but equally distinct: the 
petals of that plant consist of a long sessile purple horn or 
bag, with a spreading margin ; while the stamens consist of a 
slender filament, bearing a small, oblong, two-celled, yellow 
anther ; in single and regularly formed flowers, nothing can 
be more unlike than the petals and stamens ; but in double 
flowers the transition is complete : the petals, which first 
begin to change, provide themselves with slender ungues ; the 
next contract their margin, and acquire a still longer unguis ; 
in the next the purple margin disappears entirely; two yellow 
lobes like the cells of the anther takes its place, and the horn, 
diminished in size, no longer proceeds from the base as in the 
genuine petal, but from the apex of the now filiform unguis : 
in the last transition the lobes of the anther are more fully 
formed, and the horn is almost contracted within the dimen- 
sions of the connective, retaining, however, its purple colour : 
the next stage is the perfect stamen. No further evidence can, 
I think, be required of the formation of stamens out of petals; 
if more be wished for, the first double flower that may present 
itself to the observer may be appealed to. The conversion of 
stamens into green leaves is far more uncommon ; this, indeed, 
very rarely occurs. It was seen by Roper in the Campanula 
Rapunculus already referred to ; and Du Petit Thouars found 
the stamens of Brassica napus converted into branches, bear- 
ing verticillate leaves. Thus it appears that the stamens, 
like the petals, calyx, and bracteae, are merely modified 
leaves. 
