( 681 ) 
dwarf-form, which already occurred at the beginning of his experi- 
ments in 1888. It has since shown itself each time that a sufficiently 
large number of plants has been worked with. 
Dr Vries named this plant Oenothera Lamarckiana nanella, or 
rather, briefly, Oenothera nanella, because, although the zana-charac- 
teristic can occur in species of the most diverse orders, yet this Oeno- 
thera with regard to its constancy on sowing can in no respect be 
distinguished from an elementary species. 
Not only from Oenothera Lamarckiana, but also from Oenothera 
lacvifolia, O. scintillans, O. leptocarpa, and from hybrids of Oeno- 
thera Lamarckiana, with the new species, Oenothera nanella appeared 
quite uniformly. On the average '/, °/, of the plants showed the 
dwarf-type. 
The nanella can already be recognized as a seedling by the two 
first leaves, which are broad and have a short petiole. Then follow 
2—4 leaves with long petioles, which resemble more the Lamarchiana- 
type; DE Vries regards their appearance as an atavism. 
As a rule Oenothera nanella is an annual. In this case the ascend- 
ing stem grows out at once; in plants which will be biennials, 
several broad radical leaves with short petiole are developed, so that 
the plant hibernates with a dwarf rosette. 
The fully grown stem has remarkably short internodes. This, in 
addition to the broad shape of the leaves, gives the plant a verv 
squat appearance. 
In spite of the small size of the plant, its flowers and fruits are 
hardly smaller than those of Lamarckiana It sometimes happens 
that a plant bears these flowers when the shoot is no more than 
10 em. high. | 
This very characteristie form showed itself through mutation about 
400 times in 80.000 plants. Its constancy was carefully and repeat- 
edly tested. . 
In 1893 some nanellas, of which the ancestors had arisen in 1889 
as mutants from Oenothera laevi folia, were pollinated with their own 
pollen. The seed yielded 440 plants, all of which bore the nanella- 
characteristic. 
In 1895 twenty nanellas were treated in the same way, they 
themselves having occurred as mutants from Zamarchiana. They 
yielded 2463 descendants: all were nanellas. 
The experiment was performed in 1896 with 18000 plants and 
the same favourable result was cbtained. Three seedlings showed 
simultaneously with the nanella characteristics, those of Oenothera 
oblonga and one those of 0. elliptica. 
