TAKEDA—JAPANESE PRIMULAS. 85 
are nearly or even more than five times longer in P. saxatilis. 
This character is not always clearly shown in herbarium specimens 
of young stage, but seems to be well marked in cultivated in- 
dividuals of P. saxatilis. In consequence of scarceness of trust- 
worthy specimens of P. cortusoides, I have been unable to settle 
the question whether the pedicels of P. cortusoides always 
remain short even in the fruit-bearing stage. If this character 
fails to distinguish those two plants, there would exist no other 
point of specific distinction between them, excepting the more 
globose calyx-base of P. saxatilis.. A careful comparison of 
living specimens is very much desired. 
As far as we are aware, P. saxatilis is distributed over Eastern 
Asia, with extension from the Altai through Amurland as far 
as Korea, whereas P. cortusoides is found in Western Siberia 
extending from the Urals to the Altai. 
Turning to P. Sieboldii, we can easily distinguish this species 
from the other two by the almost glabrous and markedly accres- 
cent calyx with patent lobes. This holds good in wild as well as 
in cultivated forms of this species. The patent nature of the 
calyx-lobes was noticed by Turczaninow as long as seventy-five 
ears ago, when he proposed to call the plant P. patens. Later, 
when publishing its description, he unfortunately changed 
the name to P. cortusoides var. patens. Therefore we are 
unable to use the suggestive name patens, because in 1873 E. 
Morren described a garden variety of this species under the 
name of P. Sieboldii, and this name, in spite of its being somewhat 
unsuitable for a species occurring wild, must be adopted. 
Now I propose to distinguish two forms with the names «. hortensis 
and 3. spontanea, as I have indicated above. 
The garden variety bears flowers of larger size and of colours 
varying from pure white to deep crimson and almost purple. 
Some of them have the corolla inside pink or white and outside 
crimson or pale mauve. In some varieties the corolla tube is 
deep crimson. The corolla-lobes are sometimes very broad, and 
sometimes considerably narrowed towards the base. 
The wild form also shows to some extent plasticity in colour 
and shape of the corolla-lobes. 
The inflorescence is usually a simple umbel, but occasionally 
the axis grows out, producing another tier of flowers. ge 
I may perhaps mention here that P. cortusoides var. lichian- 
gensis, Forrest,! is not a variety of P. cortusoides, but belongs 
to a distinct species. It is distinguished above all by the sinuate 
leaf, thick lanate indumentum on the petiole and scape, and the 
larger bell-shaped calyx. This species, if it is different from P. poly- 
neura, Franchet, should retain the name P. lichiangensis, Forrest. 
P. Sieboldii is distributed in Japan over Yezo, Hont6, and 
1G. Forrest in Notes, Roy. Bot. Gard., Edin. (1908), p. 217. 
