44 C. Worsdell — 
different origin from true petals. He ascribes to Actaea , along 
with Cimicifuga, the universal possession of staminodes. Celakovsky* 
from whose works all that this article contains is gleaned, is of 
opinion that the sharp distinction made by Prantl cannot well be 
maintained; for instance, the tiny petals of some species of Actaea 
( Euactaea Prantl) are without nectaries, while others ( Cimicifuga ) 
produce them; the same holds good for Coptis. Further, the fact 
that many petals not possessing nectaries, and to which Prantl 
assigned the character of staminodes, exhibit rudiments of anthers, 
while the petals of Adonis do not is of no decisive importance, for 
such transitional forms between stamens and petals do not occur in 
all species of those genera to which staminodes are ascribed, as is 
the case in our own Actaea spicata (Fig. 3). It is, therefore, far 
from clear why these latter should have been derived from stamens 
and the petals of Adonis enjoy a different origin. Other important 
facts, for which there is no room here, tend to render Prantl’s view 
improbable. And we have only to turn to any “double” flower in 
the garden to observe the gradual transitions taking place under our 
very eyes, between stamens and true petals to be convinced that 
staminodes and petals have precisely the same origin, and that what 
to-day occurs abnormally may have been phylogenetically the normal 
process. 
We have hitherto assumed that staminodes are of staminal 
origin; the real raison d' etre of this view lies in the fact that 
nectaries usually exhibit a more or less distinct 2-lipped structure, 
the upper (inner) lip being homologous with the upturned based lobe 
of a peltate anther ,’ while the non-secretory staminodes very fre¬ 
quently bear rudiments or portions of pollen-sacs. 
Without entering further into the question we may conclude 
with Prantl that nectaries and other staminodes are of staminal 
origin; but it seems quite permissible to proceed farther and add 
that in the same way as the petals of Ranunculus and Callianthemum 
have been derived from the expansion and enlargement of the outer 
or lower lip of the nectaries ( e.g ., of Helleborus (Fig. 4)—so also 
the petals of Adonis have been derived from small pollen-bearing 
or sterile staminodes and, therefore, ultimately from stamens. 
A very pretty piece of evidence in favour of the view just stated— 
that petals are of staminal origin—can be adduced from the behaviour 
of the flowers in the Clematis-group. The perianth of most species of 
this genus is single and of the nature of a calyx ; within, the 
numerous stamens are spirally arranged. In Clematis integrifolia , 
1 This is shewn by Celakovsky as a result of his investigations 
into the origin of the perianth of Narcissus as seen in the 
phenomena presented in many double flowers of that plant. 
