Poulton: Abnormal Flowers of the Foxglove. 317 
posterior region of the corolla consisted of two lobes, the 
incision between them being shallower than those between the 
remaining lobes (Fig. II.,1 #,f).. The posteror lobes sometimes 
exhibited a faint spotting. An anterior lobe (a) and a pair of 
lateral lobes (/7) were always present, but in some instances 
the number of segments was greater. In Fig. II. 2, three large 
posterior lobes and a fourth smaller one (’) are represented, 
and the anterior lobe is also subdivided. Four large posterior 
lobes are shown in Fig. II., 3, also an additional anterior lobe. 
Small additional lobes are less common on the posterior than 
upon the anterior and lateral corolla segments. Fig. II., 12-17 
show such lobes on the anterior and lateral petals, and Fig. II., 
18 and 19, represent two posterior lobes from the same flower, 
-each having two small subdivisions. Some of these small 
lobes had their extreme edges curled inwards (Fig. II., 20), and 
their tips were often found to be covered with pollen grains. 
These were at first thought to be rudimentary stamens, but 
-after allowing the material to remain in alcohol for a week or 
two, many of the adhering pollen grains were dislodged, and 
it was found that the small lobes exhibited the ordinary petal 
structure right to their tips. Before the flower is fully opened 
the petals are curled inwards, and are often in contact with 
the anthers, so that it is probable that different parts of the 
corolla become accidentally dusted with pollen. Groups of 
pollen grains were found adhering to various parts of the 
corolla, but at the extreme tips they are more securely en- 
tangled by the small hairs which occur there. If these small 
lobes were rudimentary anthers it would seem reasonable to 
‘expect to find stages of development intermediate between 
these and the ones which are described below. Such inter- 
mediate stages were not observed. Again, would such rudi- 
mentary anthers produce normal pollen grains, as those which 
were found adhering to the tips appeared to be quite normal ? 
It seems clear that these small structures are merely lobes of 
the corolla and exhibit no trace of anther structure. 
The andreecium exhibited a very interesting modification. 
In every flower examined the four well-developed stamens 
characteristic of the Foxglove were present. These were 
readily distinguished by their position and structure. They 
-occur in the usual position, alternating with the corolla lobes 
as shown in Fig. II., 1, 2, and 3s. There was no trace of a 
posterior stamen. The filaments were long, attached to the 
corolla tube near the region where it becomes constricted, and 
carried the anthers far outside the flower (Fig. II. 1-4). The 
anthers were yellow, and faintly spotted. Dehiscence takes 
place in the usual manner. 
In every flower additional stamens occurred, a pair of 
which was attached to the anterior and to the lateral lobes of 
1913 Sept. 1. 
