Ficaria verna , and other Members of the Ranunctilaceae. 61 
Table IV. 
Eranthis hy emails. 
Number of 
Carpels in Number of Ovules. 
Flower. 
6 
7 
8 
9 
10 
4 
Number of specimens 
— 
8 
4 
— 
— 
66 . 6 % 
33*3 % 
5 
5 
8 
6 
— 
1 
25% 
40 % 
3 o % 
5 % 
6 
5 
1 1 
18 
5 
3 
11 *9 % 
26.1 % 
42.8 % 
n- 9 % 
7 - 2 % 
7 
>> >> 
— 
6 
11 
2 
1 
30 % 
55 % 
5 % 
8 
)) 
2 
4 
4 
6 
— 
p 
12.5% 
25 % 
25 % 
36.5 % 
9 
1 
6 
8 
2 
1 
5*5 % 
33-3 % 
44-4 % 
11 % 
5-5 % 
Totals 
r 3 
43 
5 1 
15 
6 
128 
II. Ficaria verna, L. 
The flower of this species (cf. Clos, 1852 ; Baillon, 1863, p. 4 ; Payer, 
1857, p. 254) usually consists of a calyx of three members, exhibiting 
a one-third spiral, followed by eight or seven coloured perianth segments 
bearing nectaries. The androecium is most commonly composed of from 
eighteen to twenty-five stamens which show an obvious tendency to dehisce 
in whorls of three, though the successive whorls exhibit lateral displace- 
ment (cf. Fig. 10, G-H). The gynaeceum generally consists of from nine to 
twenty-four carpels, the lower numbers being often clearly arranged in 
trimerous whorls. 
(a) M eristic variation. 
(1) The calyx . The number of sepals ranges from three to seven, more 
than 400 of the 514 examples examined possessing the former number. 
The arrangement of the extra sepals as shown in Fig. 10, a-h, negatives 
any suggestion that these are transformed petals. It is true that the sepals 
are sometimes coloured, but in no instance do such bear any indication of 
a nectary, which is always present on virescent petals. On the other hand, 
the position of the supernumerary sepals (Ki'-K^') does harmonize with 
the view that they represent fission products of one or more of the three 
members normally present (K1-K3). We thus have exhibited in the calyx 
a phenomenon commonly met with in other members of the group and also 
in the other whorls of this flower. Celakovsky’s dogmatic statement that 
these extra sepals here, and in Anemone hepatica , are modified members of 
the coloured perianth within cannot therefore be upheld. Having regard 
to the correlation generally exhibited between the number of parts in the 
different whorls, it might be regarded as strong evidence against the view 
