94 MR. R. P. GREGORY ON THE 
the two hermaphrodite plants were no better than the rest in this respect. In 
1905 both plants were badly blighted, in 1906 one of the plants did set four 
seeds, in spite of the attacks of Aphis, but no seedlings were obtained. Аз, 
however, this result is hardly, if at all, worse than that experienced in the 
attempts to obtain seed from the ordinary females which were raised in 
captivity, it does not reflect on the power of the hermaphrodite plants to 
produce good seed under more favourable conditions. 
There is, of course, some variation of flower-structure within the limits of 
this group of plants as defined above. Even so, the number of plants which 
fall within the group is very small. In the course of this work I have 
examined about a thousand plants, of which only eight belonged to the group 
in question. Two plants *, which I have included with the females, nearly 
approached this type in all characters but the production of pollen, while on 
the other side, the line between the hermaphrodite and the long-styled male 
forms is not always easy to draw, since it depends chiefly on the negative 
character of non-production of seed. At the most, however, this uncertainty 
only affects three plants which I have classed with the males. Of the eight 
hermaphrodite plants, five were found in the wild state at Dernford Fen, and 
three have appeared among the offspring obtained by breeding from them. 
The flower of the most typical hermaphrodite form (figs. 14, 14а) has a 
somewhat larger corolla than that of the typical female flower. Neither the 
style nor the stamens project much beyond the corolla; the style bears 
stigmatic lobes at its apex, and the stamens contain pollen, though not always 
in very large quantity. Such plants as this might perhaps be considered to 
represent Miiller’s third type, although his description “ with distinct 
rudiments of anthers" hardly seems to meet the case where good pollen is 
produced. On the whole, I think both his description and his figures of this 
type agree more closely with the more extreme forms which I have classed 
as females. 
Among the hermaphrodite forms one meets with variations from the 
central туре just described, approaching respectively the female and the male 
types. Examples of the former occurred among the offspring of a eross 
between a female plant and one of the wild hermaphrodite plants. From 
this eross, 25 offspring were obtained, of which 15 were female of the usual 
type, and 6 male. Тһе remaining four were provisionally classed as herm- 
aphrodite ; two however failed to give pollen, and had to be placed among 
the females. It should be noted that these two plants (figs. 10-12) showed 
distinet structural similarities to the hermaphrodite type. The flowers were 
rather larger than in the typical female, the staminal rudiments were larger 
and easily visible, and showed a faint reddish tinge, in which they resembled 
the distinetly reddish young anthers of the hermaphrodite and male piants, 
* See also p. 93 and PI. 8. figs. 10, 11. 
