72 JOURNAL OF THE ROYAL HORTICULTURAL SOCIETY. 
which, besides the hybrid type, the paternal and maternal characters 
appear more or less sharply defined. For my purpose, however, it was 
specially desirable to observe the behaviour of the subsequent generations. 
When in 1892 I wished to effect the first cross, I sowed to that end 
seed of Lychnis diwrna obtained by exchange in the usual way from 
botanical gardens. The seed-bed, about one metre square, yielded a brightly 
coloured mixture of various forms. Certainly all the plants were normally 
pubescent, but between L. diwrna and individuals similar to L. vespertina 
there were almost all grades and combinations, both as regards habit of 
growth, form of leaf, and colour of flower, as well as the formation of 
flower stalks in the first year, or only leafy rosettes. 
During the flowering I selected three female plants, which appeared 
to me to present the pure characteristics of Z. diwrna. They had the 
normal pubescence and dark-red flowers. All the rest were weeded out, 
and from these three examples the flowers, which were already open or 
had faded, were removed. 
I then fertilised these three plants with pollen from my Lychnis 
vespertina glabra, and they yielded plenty of seed. 
The following year I sowed from this seed a bed of about two square 
metres, and had at the flowering period something over 200 examples. 
I sowed some more also in the greenhouse, for control and pot culture. 
This first generation of hybrids, up to the time of flowering, appeared 
very uniform. All the plants were hairy and had the characters of 
Lychnis diurna. Closer examination, however, during the flowering 
period, afforded opportunity for some not unimportant observations. 
In his excellent work upon ‘‘ The Minute Structure of Plant Hybrids,” 
Mr. Macfarlane has carefully examined how hybrids are constituted with 
regard to anatomical characters which are absent in one of the parents 
but present in the other.* He treats of these cases together under the 
name of Unisexual Heredity, and deduces the rule that they “ are handed 
down, though reduced by half.’ Lychnis vespertina and L. diwrna have 
both, upon all parts of the plant, long, soft, many-celled, sharply pointed 
hairs, among which there are scattered glandular hairs, rare upon the 
stems and more numerous upon the flowers. The hybrid between 
L. diurna and the hairless form of LZ. vespertina has shorter and blunt- 
ended hairs and a lesser number of glandular hairs, especially on the 
stalks. 
Of my hybrids, in the first year about two-thirds produced flower 
stalks, whilst the remainder only formed rosettes of basal leaves. It is 
well known that ZL. diwrna is a perennial species, whilst L. vespertina is 
an annual or biennial. But then ZL. diurna also frequently flowers the 
first year, as did the mother plants, which germinated and flowered in 
1892, when they served for my crossing experiment. 
I have especially directed my attention to malformations. It is said, 
and pretty generally acknowledged, that hybrids show a greater tendency 
towards malformations than do the parental forms. In my seed-beds they 
were not rare. In the first place, as regards the number of the pistils, 
Gaertner (/.c. pp. 842 and 531) foundin L. diurna x vespertina occasion- 
* J. Muirhead Macfarlane, ‘The Minute Structure of Plant Hybrids,’ Trans. 
Ioy. Soc. Edin. vol. xxxvii. Part 1, No. 14, 1892, p. 273. 
