FORMS OF FLOWERS IN VALERIANA DIOICA. 97 
recognized, for, since the style undergoes elongation while the flower is open, 
the younger flowers of each type tend to resemble the older flowers of the 
succeeding one. In addition, there is often considerable variation between 
the different flowers of the same plant *, apart from any differences due to 
the age of the flower. In practice, therefore, each plant was recorded as 
belonging to the group indieated by the structure of that flower which had 
the longest style, a large number of flowers from each plant being examined 
in each case, especially in the short-styled plants. It may be mentioned here 
that flowers of the type © (figs. 29, 35) were most frequently met with on 
plants which bore other flowers conforming to one or both of the long-styled 
types, although in one or two cases such flowers were found on plants which 
otherwise bore only flowers of the short-styled types. 
While there is no doubt that the majority of the males ean be separated 
into two groups, according to the length of the style, if care be taken to 
dissect several fully mature flowers, there is equally no doubt of the existence 
of a series of connecting forms. I do not wish to suggest that there is any 
true line of separation whatever between my groups A & В, and D & E, 
which were used merely for convenience in recording types intermediate 
between those described by Müller. That such types were of not infrequent 
occurrence is shown by the fact that of 60+ male plants recorded under this 
system, | 
21 were placed in Class А, 
4 were recorded as intermediate between A A D, 
8 were placed in Class В, 
4 were placed in Class C, 
were placed in Class D, 
17 were placed in Class E. 
©» 
Not only are there connecting forms between the various types of male 
flower, but the long-styled males, in their extreme forms, approach the 
hermaphrodite type. Thus, one of the plants found in the wild state, which 
was classed as a long-styled male, had a very well developed style with 
distinct (stigmatic) lobes at its apex. Two other plants, offspring of a cross 
[Female x Long-styled male], in some of their flowers approached so nearly 
the hermaphrodite type, that they were distinguishable from the plant shown 
in figures 15-18 (which bore fruit) only in their somewhat smaller ovary and 
* One very extreme case of this 18 referred to later (p. 98.) 
T All the offspring of one cross are excluded from these figures, because continued 
examination of some of the plants has shown that such extreme variation occurs in the 
different flowers of the same plants, that the accuracy of the classes recorded for the remainder 
is laid open to doubt (see figures 32-37 ; and the description on pp. 98, 99). 
