The genus Clintonia has four species in North America and one in eastern Asia (Fig．1)．In this paper，the karyotypes and the intraspecific morphological and cytogeographic differentiations of ten populations of C．udensis from China are analysed，and the probable origin area of the genus is also discussed．The conclusions are as follows：

    (1)Based On the chromosome numbers 2n=28 from the North American species，the Japanese and Himalayan groups，the earlier investigators established x=14 as the basic chromosome number of Clintonia，and they thought that there was polyploid in this genus except for aneuploid only in C．borealis (Utech，1975；Utech and Suda，1975)，but a few authors(Sen，1975;Wang et al，1993)pointed out the x=7 basic number of this genus based on 2n=14 in C．udensis from Primorskiy Kray of Russia(Skolovskaya，1966)and Yunlong，Yunnan Province of China(Wang et al，1993)respectively．Our result along with there reports by previous authors(see table 1)shows that at least two ploid levels exist in C．udensis，i.e．2n＝14 and 2n＝28．The diploids are widely distributed from northwest Yunnan of China to Primorskiy Kray of Russia，while the tetraploids are located in northwest Yunnan，Himalayas，Japan，and a narrow area in Mt．Hualongshan of southern Shaanxi(07 population)(Fig．2)．Therefore，the basic chromosome number of Clintonia certainly is x=7 ratherthan x=14，while 2n=28 in North America，Japan and Himalayan area are tetraploid, not diploid， According to the previous data，so far no diploid taxa of Clintonia has been found in North America and Himalayas．So we consider that the primitive type of Clintonia is in eastern Asia，and the secondary diversity center and the present distribution center of this genus are in North America．

    (2)C．udensis is widely distributed in eastern Asia(Fig．2)；it has two cytotypes．The karyotypes for all the diploid populations are remarkablely similar. Taken together, they can be roughly repersented by the formula:K(2n)=14=2V+6J+2v+4j(2NOR+2j). The chromosomes range in length 25.55～12.78 μm, with the ratio of the longest to the shortest 2.0. The karyotype belongs to Stebbins (1971)2A(→2B). For the tetraploid taxa, except for 07 population, karyotypes are also identical and may be roughly symbolized as: K (2n)=28=4V + 12J + 4v+8j ( 2NOR + 6j ). The length of chomosomes is from 27.87 to 13.93 μm, with the ratio of the longest to the shortest 2.0, and thus the karyotype belongs to 2A(→2B). The karyotype of 07 population is similar to those of above tetraploid taxa but also has some differences, especially in the position of satellites and the morphology of 10th and 14th pairs of chromosomes. Its formula is K(2n)=28=4V+12J(2NOR+10J)+4v+ 8j(2n=28=10m+16sm(2SAT)+2st). The ratio of the longest (23.72μm)to the shortest(12.97μm) chromosomes is 1.83. The Karyotype belongs to 2A. And the distribution range of this population is very narrow. We think that it is probably a recent evolutionary event in C. udensis.

      (3)Wether Clintonia in eastern Asia has 1 sp. or 2 spp. or 1 spp. and 1 var. has been debated for a long period. According to our observation, within C. udensis, only the size of seeds is related to its ploidy level, i.e. diploid individuals have smaller seeds and tetraploid ones have larger ones; the colour of seeds is related to its geographic distribution, i.e. the materials from the Himalayas through Yunnan, Sichuan to eastern Qinling Range have pale brown seeds, while those from Mt. Lüliang, Shanxi Province via Hebei, Liaoning, Jilin to Siberia and Japan have dark brown seeds. Some other morphological characters, such as the size of leaves and fruits, inflorescence type and flower numbers between individuals in one locality, even within one populaion have evident variation. Therefore, we consider that evidence(see Table 4 )for separating C. alpina or C. udensis var. alpina from C. udensis is not