74 Salisbury . — Variation in Eranthis ky emalis , 
specimens, and coloured sepals, though rare, were also noted. The latter 
however, never show any indication of a nectary and can be readily 
distinguished from virescent petals. In one example a sepal was present 
which bore at the apex a small trilobed lamina. 
(d) Ranunculus auricomus. This species is interesting as it affords 
frequent examples of reversion of petals to stamens. This condition is 
normally found in the so-called var. depauperata. In this form the position 
of the petals is in some specimens occupied by normal stamens, but in others 
by stamens which are much larger than usual or even subpetaloid in character 
(see Fig. 20, A-C). The instructive feature, however, is the fact that, accom- 
panying the reversion of the corolla segments, the sepals usually become 
coloured yellow, whilst in the perfect-flowered form the sepals are quite 
green. 
VII. The Origin of the Perianth. 
There can be little doubt that the solitary inflorescence of Anemone has 
been derived by reduction from a, many-flowered inflorescence. Thus, in 
Anemone narcissiflora two to six flowers arise together in an umbel from 
the involucre of bracts. In A. ranunculoides the number of flowers varies 
from three to one. In A. palmata the flower is normally solitary, but two 
flowers quite frequently occur, whilst in A. nemorosa the single-flowered 
condition has become fixed. In all these cases the flowers arise from an 
involucre of three or rarely four leaves which is obviously homologous in the 
different species. Now whilst in the case of the single-flowered members we 
might a priori imagine these bracts to represent descended perianth leaves 
which had become foliaceous, as Worsdell maintains (1916), such a concep- 
tion is hardly applicable where we find the involucre surrounding a group 
of several flowers. On the other hand, the assumption that the involucre 
consists of foliage leaves is in complete harmony with all the facts and 
equally applicable to the many- or few-flowered inflorescence. Again, if we 
study the numerous species of Anemone we find almost every transition 
from involucral leaves in all respects resembling the foliage leaves (e. g. 
A. nemorosa ), to those in which the involucral leaves are essentially calyx- 
like and closely approximated to the flower (e. g. A. hepatica ). It is, more- 
over, significant that in general where the involucre is situated close to 
a flower its members depart farther from the foliaceous type than where it 
is more remote. 
Now this transition can of course be viewed from either direction, but 
if the bracts forming the involucre are only modified stamens it is scarcely 
conceivable that their ultimate development should present all the indica- 
tions of homology with, rather than analogy to, the foliage leaves. If such 
a striking parallel evolution had actually taken place, it could only indicate 
an identity of function. Yet if we examine species of Anemone with 
a simple type of leaf but in which the involucre is not in close proximity to 
