INFERTILE HYBRIDS. 187 
Microscopic inspection, however, showed that the pollen grains were 
shrivelled and useless. An occasional grain was found to be plump, but 
not quite normal. Nectar was secreted in considerable quantity. Bees 
visited the flowers busily, and they (including the humble-bees) had no 
difficulty in reaching the nectar by entering the mouth. The corolla was 
not pierced. If the pollen had been good, one would have expected seed 
to beset; but, although the ovaries swelled and gave promise of seed, none 
was got. Fasciation did not occur. 
From the above descriptions it is obvious that in crossing with the 
plants in question the reciprocals are not alike. In form the flowers in 
each case take more after the seed-parent. It is a matter of very con- 
siderable interest to find the purple reappearing in those haying D. lutea 
as pollen-parent, pointing to the fact of that colour being latent in the 
white foxglove. It is of further interest to notice that the flowers of this 
series are evidently the more nearly normal, although the pollen was taken 
from the smaller-flowered species. It does not appear that there is any 
reason to suppose that the pollen of D. lutea is more potent than that of 
D. purpurea alba because the former is a perennial plant and the latter 
a biennial one. 
New series of the same reciprocal hybrids are at present in flower. 
Again the differences between the reciprocals are noticeable, the plants 
having D. lutea as pollen-parent bearing larger and more foxglove-like 
flowers. In this lot, however, there is more variation amongst the 
individual plants in respect of colour. Some are decidedly reminiscent of 
the purple foxglove, others are equally so of the yellow parent. As before, 
in all the plants the four stamens are all present and apparently well- 
developed. 
In the reciprocal set there is also a much less rigid adherence to a 
type. The flowers are always narrower than the above, but many are 
wider than was ever found in the former corresponding series; in fact 
some are a very near approach to the smaller-flowered examples of those 
of the other series (D. purpurea alba x D. lutea). Not only so, a good 
many of them are well suffused with light rose-purple. Further, the 
stamens are better developed, although abortion is common enough. 
The plants now in flower, while still demonstrating the fact that 
reciprocal crossing does not give similar results, show that, under certain 
conditions nct discovered, the distinctions between the reciprocal crosses 
may sometimes be less marked. The white foxgloves used in the experi- 
ments were alike in being destitute of coloured spots. It should be noted 
that a few faint brownish-red markings often appear at the mouth of the 
tube of D. lutea, at the angles of the corolla lobes. Corresponding colora- 
tion has not been seen in the hybrids. 
In summing up the observations on the three series of crosses having 
D. purpurea alba as pollen-parent, we observe as outstanding features 
that the first series was characterised by great vigour, very close 
approximation to a type, and an almost universal development of 
fasciation of a very pronounced kind. The second series was marked by 
also keeping close to type, but no fasciation appeared. The third series 
is decidedly more variable than either of the above in the form and 
colour of the flower, and there is no development of fasciation. 
