Inheritance of Mutation in the Common Foxglove. 55 
from the type plants notwithstanding this grading, for in most cases 
the distinction between variety and type is absolute, all the flowers 
on the spike being deformed in some degree in the one case, and 
normal in the other. Even when this is not the case there is still 
a very wide gap between the heptandrous individual which eventually 
produces a small number of normal flowers and the type plant 
which shows a trace of the abnormality in some of the earliest 
flowers. No true intermediates were observed in any of the matings. 
If they occur, they are evidently rare. Within the heptandra class 
we may conveniently group the several grades as follows:— 
1. Extreme lieptandra form. Plants in which all, or almost 
all, the flowers are without petal-like structures. 
2. Heptandra form proper (fig. 12). Plants in which only the 
flowers in the lower region of the spike and those at the extreme top 
appear to have no corolla. In the intervening region the upper lip 
appears as a conspicuous petaloid structure which is usually deeply 
bifid. By far the larger number of heptandra plants belong to this 
group. 
3. Intermediate heptandra form. Plants in which all, or 
nearly all, the flowers are more or less petaloid. The lowest 
flowers, in which an upper lip is often already well-developed, are 
succeeded by others in which a lower lip also gradually makes its 
appearance, and finally tubular or even quite normal flowers may be 
found towards the top. 
As previously stated the only seed obtained in 1906 was 
collected from exposed flowers on a single individual. It was there¬ 
fore to be expected that much of this seed would have resulted from 
crossing with neighbouring type plants, but that some would be due 
to self-fertilisation. Some 240 plants raised from this seed were 
flowered in 1908 of which more than 100 showed the heptandra 
characters, while the remainder resembled the type. In this case 
the precise numbers have no value as the fertilisation of the seed- 
parent was not controlled, but from the appearance of so many 
heptandra offspring it was scarcely doubtful that, when self-fertilised, 
this form would be found to breed true; the results in 1910 have 
shown that such is the case. Six of the heptandra individuals 
flowering in 1908 were self-fertilised and all the 256 offspring were 
heptandra} The numbers obtained in each case were as follows:— 
1 A single type plant was found in the rows, but that this was a 
rogue was evident from the fact that it possessed the smooth 
stem occurring in certain other families, but not otherwise 
found in the six families above mentioned, the 256 individuals 
composing them all showing the normal degree of hoariness 
(see p. 60). 
