74 JOURNAL OF THE ROYAL HORTICULTURAL SOCIETY. 
experiments I have subsequently been able to test my hybridising 
material in this respect, and found the assumption, above given to be 
confirmed ; unfortunately, in this case it was no longer possible. 
In the horticultural practice of hybridising it is a rule to choose forms 
of which one at least is very variable, and hence arises the known multi- 
formity of the hybrids. 
With regard to the flowers and other important characters, my 
hybrids resembled the true Lychnis diurna. 
IT come now to the third generation, which I cultivated in 1894. For 
this I used seed of the dark-red examples of 1893, which I had fertilised 
with pollen from equally dark-coloured flowers, taking care to exclude 
insect visits. 
But whilst in 1898 all the hybrids had been hairy, this was no longer 
the case in 1894. Only about three-fourths were hairy, the rest hairless. 
T had 99 hairy and 54 hairless, in all 155 plants, and counted them in 
July at the commencement of flowering. The character of the grand- 
father, the transfer of which I had had in view, was therefore once again 
visible. 
Both among the hairy and among the hairless plants there were red- 
flowered and white-flowered and _ broad-leaved and _ narrow-leaved 
examples; the broad-leaved had the habit of L. diwrna, the narrow- 
leaved that of LZ. vespertina. Also, as regards the corolla and the calyx, 
there was a similar diversity of form. The parental characters were, 
in all imaginable combinations and grades, to be found in the bed of 
seedlings. 
As confirmation of the above, as regards the inheritance of malforma-_ 
tions, I found, for instance, a flower with two corollas in one calyx. 
Among the vari-coloured mixture, I now sought out my Lychnis 
diurna glabra, selecting some male and female examples, which latter 
I fertilised with the former, excluding insect visits. They were entirely 
hairless broad-leaved plants with the flowers of L. diwrna and with dark- 
red petals—characteristics of the grandfather, with the exception of the 
entire hairlessness, I failed to find. 
The seed reproduced the desired form in the following year (1895) 
quite truly—at least as regards the absence of hairs and the other 
constant characteristics of my starting plants. Only the colour of the 
flowers remained variable. Out of 206 plants there were 13 white- 
flowered, or about 6 per cent. The rest were partly purple, partly dark 
red. 
As regards the pubescence, I have made the following experiments. 
As the bed was in the vicinity of that in which the cuiture of the previous 
year had been made, its integrity, owing to possible subsequent 
germination of older seed, was not certain. I sowed, therefore, a portion 
of seed in pans with sterilised soil and raised 390 young plants, all of 
which were hairless. The transmitted character could therefore be 
regarded as constant. 
In the two following years (1896 and 1897) I have carried the new 
form through two further generations, choosing always the dark-red 
individuals as pollen and seed parents. In 1897 I had a bed of four 
square metres, containing about 1,350 plants. Pubescent plants occurred 
