100 MR. К, Р. GREGORY ON THE 
Conclusion. 
It appears that the individuals of Valeriana dioica may, for convenience, 
be arranged in classes, distinguished from one another by the relative 
development of the male and female reproduetive organs, and to some 
extent by the size of the corolla. Each class eonsists of a number of plants 
which differ slightly from one another, but may be arranged around а 
central form, which may be taken as the type for the group. In each group 
the extreme forms approach very nearly the extreme forms of the neigh- 
bouring group or groups; and, while the central forms of each type are 
easily distinguishable from one another, there is absolutely no sharp line of 
separation between successive groups, although the intermediate forms occur 
in relatively small numbers. The wide range of variation shown by flowers 
of one individual, in certain cases, tends to emphasise this eontinuity of the 
series from type to type. 
Whether the hermaphrodite plants described above may properly be 
regarded as a group, or whether they ought to be regarded simply as a 
serles of transitional forms, is a debateable matter. It would be natural to 
suppose that in a species where the males are separable into two groups, 
aecording to the degree of development of the rudiments of the female 
organs, а corresponding line of division might be found in the sex in which 
these organs are functional.  Müller' class of * Females with distinct 
rudiments of anthers” corresponds much more closely, I think, with certain 
forms which I have classed with the females, than it does with the herm- 
aphrodite group ; yet I think a separation of these forms from the commoner 
type of female would be extremely difficult to effect, and the boundaries of 
the two resulting classes would be even more ill-defined and artificial than 
they are in the males. On the other hand, the eight hermaphrodite plants 
would form a group most disproportionately small in comparison with the 
size of the other groups. 
As regards the relative numbers in which the various forms occur, in the 
wild state at Dernford Fen the males appeared greatly to outnumber the 
females. On the other hand, of the plants obtained by breeding under 
cultivation, 298 were females, 3 hermaphrodite, and 65 males. Of the 298 
females, 132 were obtained from crosses of which all the offspring were 
female. With regard to the apparent preponderance of males at Dernford 
Fen, two cireumstances, both of which introduee a source of error into any 
counts which may be made, must be taken into consideration: first, that 
the flowering pe.iod of the males tends to be slightly earlier than that of the 
females (at any rate when the plants are cultivated in a garden), so that 
the counts might vary substantially, according to the season at which they 
were made ; second, that, owing to the stoloniferons habit of the plant, the 
