Bot. Bull. Acad. Sin. (2003) 44: 337-344

Chang et al. — Morphometric analysis of Kalopanax septemlobus

A morphometric analysis of the eastern Asian Kalopanax septemlobus (Thunb.) Koidz. (Araliaceae)

Chin-Sung Chang¹, Hui Kim^2,*, Ho-Sang Kang³, and Don Koo Lee³

¹The Arboretum and Department of Forest Sciences and Products, Agriculture and Life Sciences, Seoul National University, san 56-1, Sillimdong, Gwanak-gu, Seoul, 151-742, Korea

²Korea National Herbarium (KH), Korea National Arboretum, 51-7 Jikdong-ri, Soheul-eup, Pocheon-si, Gyeonggi-do, Korea

³Department of Forest Sciences and Products, Agriculture and Life Sciences, Seoul National University, san 56-1, Sillimdong, Gwanak-gu, Seoul, 151-742, Korea

(Received September 17, 2002; Accepted June 24, 2003)

Abstract. Kalopanax septemlobus (Thunb.) Koidz., a woody species of Araliaceae, exhibits a broad range of morphological variation and occurs throughout much of eastern Asia. A morphometric analysis of herbarium material supplemented with a large sample of field-collected leaves was undertaken to determine if the morphological differentiation found within Kalopanax septemlobus warranted the taxonomic recognition of distinct taxa. One hundred twenty six individuals representing the total geographic range of the species were scored for 20 morphological characters, and the data matrix was subjected to principal components analysis. The results indicated that Kalopanax septemlobus should be recognized as one polymorphic species. Previously recognized infraspecific taxa were not supported to warrant the designation of any taxonomic rank. The observed pattern of variation may be environmentally induced and suggests that the species may exhibit environmental plasticity.

Keywords: Araliaceae; Forms; Kalopanax septemlobus; Kalopanax septemlobus subsp. lutchuensis; Morphometrics; Principal components analysis.

Introduction

Castor Aralia, Kalopanax septemlobus (Thunb.) Koidz. (Araliaceae), is a well-defined species native to eastern Asia, Japan, Korea, China, and eastern coastal Russia (Willis, 1973) (Figure 1). Recent phylogenetic studies of the nuclear ribosom DNA internal transcried spacer (ITS) of Araliaceae have supported the monophyly of Kalopanax and Eleutherococcus and their placement in a broad assemblage comprising taxa largely with compound leaves (Wen et al., 2001).

Within the species, morphological variation is exhibited at many levels, from within a single population to among local and regional entities. Taxonomic treatments of this differentiation by previous workers (Nakai, 1927; Rehder, 1947; Lee, 1980; Hara, 1986) have recognized various numbers of infraspecific taxa at either the varietal or subspecific rank, usually based on minor differences in leaf morphology and trichoms which has made identification difficult. Zabel (in Ohashi, 1994) first noted that some of K. septemlobus was distinguished by tufted hairs on the basal axil of the main nerves, which he recognized as var. magnificus. Similarly, Van Houtte (1874) and Nakai (1927)

attempted to describe specimens with deeply incised leaves, which they designated as var. maximowiczii. However, Hara (1954) presented arguments favoring recognition of Nakai's variety as a form. Authors of regional flora treatments (Lee, 1980; Hoo and Tseng, 1978) have agreed with Nakai and Zabel's circumscriptions for the most part, maintaining the taxa at the varietal level. On the other hand, neither var. maximowiczii nor var. magnificus were

*Corresponding author. Tel: +82-31-540-1082; Fax: +82-31-540-1070; E-mail: huikim@chollian.net

Figure 1. Maximum extent of geographic distribution of Kalopanax septemlobus in eastern Asia.

Botanical Bulletin of Academia Sinica, Vol. 44, 2003

recognized in recent floras of Russia and Japan (Kitamura and Murata, 1984; Charkevicz, 1987). Likewise, Ohwi (1984) did not accept these two varieties in his flora, but he did recognize var. lutchuensis as the southern glabrous type.

There is a clear need for a comprehensive study of this complex that goes beyond the limited work done previously as a small part of a major floristic endeavor thus far. Two major treatments of this complex (Kitamura and Murata, 1984; Charkevicz, 1987 vs. Ohwi, 1984) differed from each other primarily in the importance given to the presence of pubescence and shallowly to deeply lobed leaves. In a recent review, Ohashi (1994) clarified the nomenclature of Kalopanax septemlobus, and proposed two subspecies, including several forms based only on an analysis of Japanese individuals: K. septemlobus subsp. septemlobus for. septemlobus, and for. maximowiczii (Van Houtte) H. Ohashi; and subsp. lutchuensis (Nakai) H. Ohashi. Ohashi (1994) differed from Van Houtte (1874) and Nakai (1927) in considering var. magnificus to represent the same taxon as var. maximowiczii. Also, Ohashi (1994) regarded the previously recognized var. maximowiczii as a highly variable entity with no apparent morphological characteristics to distinguish it from typical material, which led him to reduce it to a form under K. septemblobus subsp. septemlobus. The presence of scarce tufted hairs on the basal axil of the main nerves is characteristic of material assigned to for. septemlobus. These hairs are dense in for. maximowiczii and absent altogether in subsp. lutchuensis. Leaves of var. lutchuensis are usually (3-)5-lobed, thick, and lustrous on the upper surfaces while those of the other taxa are usually 5-7-lobed, relatively thin, and not lustrous. Ohashi (1994) also considered that deeply incised leaves are found in K. septemblobus subsp. septemlobus for. septemlobus and K. septemblobus subsp. septemlobus for. maximowiczii, but not in subsp. lutchuensis.

The purpose of this paper is to analyze the phenetic relationships among the different morphological entities of Kalopanax septemlobus from throughout eastern Asia. Particular attention has been given to determining if the morphological variation among the previously distinguished taxa warrants recognition at the rank of species. For this we have used a large sample of field-collected leaves and comparative methods of quantitative characters to examine patterns of geographical distribution. The level of polymorphism on leaf shape and size was evaluated

within several testable populations to assess the constancy of these characters.

Materials and Methods

Herbarium specimens were selected to represent the entire geographical range of Kalopanax septemlobus complex and to reflect the morphological variability present in the species. Material were borrowed from various herbaria including CNU (Chunnam National University, not listed in Index Herbariorum), PE, SNU, SNUA, and TUS (Holmgren et al., 1990). Additionally data were obtained from pressed specimens collected in the summer and over a four-year period (1998 to 2001) from several populations in South Korea and northern Japan. Voucher specimens were on deposit at SNUA.

Twenty morphological characters (Table 1) were selected for analyses, included those most frequently utilized in keys and diagnoses to distinguish taxa within Kalopanax. Inflorescence and flower features were not measured because these were not previously used for infraspecific delimitation. Measurements were taken with a hand ruler and an ocular micrometer. Illustration of the leaf morphologies was done by drawing the traced outlines of photocopied herbarium specimens (Figure 2).

The OTUs were grouped both geographically and based on a priori identification to varieties (52 specimens from China, 14 from the main islands of Japan, one from Okinawa of Japan, 59 from Korea, and four from Russia) (Appendix). All specimens used in the analyses were at their mature stage of development. In most cases, only one mature leaf per plant was used to represent an individual. In order to assess morphological variation within and among individuals, eighty leaves from one individual tree examined in the field were measured for our initial examination.

Morphological variation was assessed using univariate statistics (mean, maximum, minimum) and multivariate analysis (PCA). Univariate analyses were preformed using the SAS-PC package (version 6.04) developed by SAS Institute (1988). For each geographic group and the individual tree examined, descriptive statistics and character correlation were calculated and normality tests were performed. Samples departing from normality were either log or square-root transformed. One-way analysis of variance, using geographic group as a treatment, was performed on each

Chang et al. — Morphometric analysis of Kalopanax septemlobus

fewer lobes than material assigned to subsp. septemlobus. Glabrous to subglabrous leaves, a trait also associated with var. septemlobus, were frequent but not fixed in any of the regions. No discrete regional groups could be defined by any single measure (Figure 3). Although OTUs from China overlapped with individuals from Korea and Japan, southern Chinese individuals and those from Okinawa shared a close morphological similarity.

In the principal components analysis, the scores of the first three components accounted for 53.70%, 13.84%, and 7.26% of the total variance, respectively, or 74.80% altogether. Ordination of principal components 1 vs. 2 (Figure 4A) revealed clusters of OTUs that did not corresponded to geographic groups. In both plots, several OTUs obtained from an individual tree occupied the central area of the plot and overlapped with OTUs from all geographic groups. The majority of the material from Korea was represented the remainder of the OTUs along the second principal component axis, and some individuals

Figure 2. Diagram of leaf characters measured for numerical analysis. Numbers correspond to character number in Table 1.

character that could be normalized (P £ 0.05). A one-way Kruskal-Wallis test was performed for all non-normal samples.

Principal components analysis (PCA) was performed using the SAS-PC package. PCA was conducted on the matrix of product-moment correlations, obtained from the standardized data, to provide further insight into structure in the data set. This method is well-suited to revealing patterns of continuous variations in a data set (Sneath and Sokal, 1973). Specimens were projected onto the eigenvectors, and a priori assignment to varieties was also plotted in two dimensions for examination.

Results

The results of the univariate analysis were shown in Figure 3, along with the minimum and maximum values for some important characters. Some features traditionally used to differentiate the varieties of Kalopanax septemlobus showed patterns of variation that were inconsistent with the descriptions of the subspecies (Ohashi, 1994). The character of ratio (length/width) to depth of leaf lobe generally used to differentiate subsp. lutchuensis, was common in Korea, northern China and Japan. Analysis of the distribution of these traits indicated that plants in several parts of southern China consistently possessed

Figure 3. Mean and range of values for characters used to differentiate taxa in Kalopanax septemblobus. (A) leaf width; (B) depth of first left lobe; (C) leaf length; (D) width of central lobe; (E) length of first right lobe; (F) length of first left lobe; (G) ratio (leaf legth/width); (H) ratio (depth of lobe / leaf length).

Botanical Bulletin of Academia Sinica, Vol. 44, 2003

from China clustered at the other end of this axis together with the individual from Okinawa at the upper left side of the first axis. It should also be noted that leaves from individual trees were grouped in the center of the axis and spread along first axis. Some samples from the western part of Korea, i.e., the island Deokjeok-do (a coastal type) and from parts of China (Szechuan, Guangdong, Guangxi, Yunnan, Guizhou) were somewhat distinct although their distribution in the PCA overlapped with other clusters of OTUs toward the center of the projection. The small, but relatively shallowly dissected, leaves from southern China would be distinguished from the more dissected leaves of the other populations. The second component corresponded primarily to shape, relative incision size between lobe, and the lengths of the second left and right lobes. However, the specimens with these characters showed a

considerable degree of overlap, observable in frequency diagrams of incision degree [= ratio (depth of lobe/leaf length) (20)] (Figure 5A).

Examination of the relative magnitudes of the character coefficients for each principal component allowed the identification of the relative contribution of each character to the corresponding component (Table 2). The characters that contributed most to the loading on the first axis were leaf width (3), leaf length (7), length of first left lobe (14), length of second left lobe (15), length of first right lobe (16), and length of second right lobe (17). All leaf size characters were highly correlated with the first component axis. The characters that loaded most heavily on the second axis were the depth of left first lobe (5), the depth of first right lobe (6), width of center lobe (9) and the ratio (depth of lobe/leaf length) (20). The character that contributed most to the separation along the third axis was the leaf length/width ratio (18).

In order to examine the relationship between trichome character and cumulative values of variation generated, each of the 126 specimens was scored as either pubescent or glabrous (Figure 4B). That was glabrous individuals were assigned to the var. septemlobus group (N=31) and those with trichome to the var. maximowiczii group (N=95). The results of a PCA discarded the presence of two probable infraspecific groups, one with tufted villose indument at the basal axil of mature leaf surface vs. the other with glabrous to subglabrous abaxial leaf surfaces (var. septemlobus) (Ohashi, 1994). As expected, a lack of

Figure 4. Principal components analysis (PCA) of 20 morphological characters measured on 109 OTUs and 80 leaves from one individual tree. Two extreme leaf shapes are shown. (A) geographic projection by PCA1 and PCA2 (+, China; D, individual tree; ~, Japan; p, Korea; ™, Okinawa; £, Russia); (B) a priori projection. (™, Kalopanax septemblobus var. septemlobus; ¢, Kalopanax septemblobus var. maximowiczii).

Chang et al. — Morphometric analysis of Kalopanax septemlobus

phenetic coherence was evident in the multivariate analysis, and taxa assigned a priori were virtually inseparable in the projection. The analysis of frequency distribution of trichome density demonstrated that this variable showed a normal distribution (Figure 5B).

Discussion

Our initial examination of herbarium specimens suggested a high degree of morphological intergradation among the entities recognized within Kalopanax septemlobus and their inconsistent circumscription among the treatments of various authors. The morphometric analysis indicated that no strong discontinuities existed among the infraspecific taxa recognized in Kalopanax septemlobus. It also demonstrated that taxa within K.

septemlobus could not be delimited with the use of leaf characters. Morphological variation for all characters was continuous with a high degree of overlap across taxa. The difficulties of quantitatively visualizing shape differences could be seen by the fact that the distinction among OTUs was unclear and by the strong indication of similarity among the three geographic groups analyzed.

Comparisons of results of the univariate analyses which compared means and range for the geographic area showed a similarity between individuals assigned to subsp. septemlobus from southern China and those of subsp. lutchuensis from Okinawa. Figures 3 and 5 showed that some characters analyzed had ranges that overlapped completely, but had significantly different means [e.g. depth of first left lobe, length of second right lobe, ratio (length/width)]. Ohashi (1994) delimited taxa on the basis of slight differences in lobe number, and subsp. lutchuensis was distinguished by just one remaining character, leaf texture. Before formally recognizing these differences taxonomically, however, one would like to see evidence of a genetic basis and be assured that differences are not simply due to environmental factors. Geographic distribution should also be taken into consideration along with the analysis of morphological characteristics as a basis for evaluating relationships among taxa (Radford et al., 1974). In fact, variation of these characters was continuous and overlapped among OTUs from southern China and other areas, in particular Okinawa.

No clear distinction could be found between material assigned to the previously recognized var. maximowiczii, and to var. septemlobus, and forms intermediate in leaf morphology between these two taxa suggested that they should be united. The frequencies for the number of trichomes on the basal axil of the main nerves (Figure 5A) showed a normal distribution, suggesting that the degree of pubescence was related to polygenic traits. Although extreme values (glabrous/pubescence) were characteristic of some individuals, none of them could be reliably, consistently, and geographically distinguished based on this character alone. If other traits were to show correlation with a scarcity of trichomes, recognizing a form might be justified. For the two previously recognized taxa sympatric in eastern Asia (K. septemlobus var. septemlobus and var. maximowiczii), we did not confirm the constancy of characters under different environmental conditions, and thus could not determine their taxonomic values. Nevertheless, our data did not support the continued recognition of var. magnificus as distinct from typical material of K. septemlobus.

As noted in PCA, the coastal samples from western Korea appeared to be more variable and distinct compared to collections from other sites. These patterns of variation were consistent with the observation that deeply dissected lobed leaves dominated the coastal specimens. The pattern of variation could be environmentally induced and might reflect a degree of plasticity. However, variation in the coastal type was not sufficiently discontinuous to warrant taxonomic recognition.

Figure 5. Histogram based on the frequencies of (A) the ratio of the first left depth / leaf length ratio and (B) frequencies of trichome density (number of hairs /3 mm²).

Botanical Bulletin of Academia Sinica, Vol. 44, 2003

The results of this study indicate a continuum in leaf morphology among the previously recognized infraspecific taxa. This continuum is especially evident in leaf lobing and pubescence, characters traditionally used to define infraspecific taxa within K. septemlobus. The material studied is not sufficiently differentiated to warrant the designation of subspecies or varieties, and our interpretation is in agreement with the taxonomic treatment of Ohashi (1994), except that we do not maintain subsp. lutchuensis.

Acknowledgements. Our thanks to the Director of PE, Dr. Qin, Hai-ning, and to the directors of CNU, SNU, and TUS for providing loans of specimens for this study. Appreciation is also expressed to Mr. J. I. Jeon for his help in China. This research was supported partly by a grant (PF001302-00) from Plant Diversity Research Center of the 21st Frontier Research Program funded by the Ministry of Science and Technology of the Korean Government and KSF 996-0600-002-2 (Ecology and genetic characters of Kalopanax septemlobus growing in South Korea and Hokkaido, Japan).

Literature Cited

Charkevicz, S.S. 1987. Planatae Vasculares Orientis Extremi Sovietici. Vol. 2. Nauka, Saint Petersburg (in Russian).

Hara, H. 1954. Enumeratio Spermatophytarum Japonicarum 3. Iwanami Shoten, Tokyo.

Hara, H. 1986. Typification of the Japanese plants named by Thunberg. Journ. Jap. Bot. 61: 353-363.

Holmgren, P.K., N.H. Holmgren, and L.C. Barnett. 1990. Index herbariorum. Part I: The herbaria of the world. Bronx, NY: New York Botanical Garden.

Hoo, G. and C.J. Tseng. 1978. Genus Kalopanax. In Flora

Republicae Popularis Sinicae, Beijing: Science Press. 54: 76-80.

Kitamura, S. and G. Murata. 1984. Colored Illustrations of Woody Plants of Japan. Vol. I. Revised Edition. Osaka: Hoikusha Pub. Co. (in Japanese).

Lee, T.B. 1980. Illustrated Flora of Korea. Hyangmun Co., Seoul (in Korean).

Nakai, T. 1927. Flora Sylvatica Koreana. 16. Araliaceae. Forest Experiment Station, Seoul.

Ohashi, H. 1994. Nomenclature of Kalopanax septemlobus (Thunb. ex Murray) Koidzumi and classification of its infraspecific taxa (Araliaceae). Journ. Jap. Bot. 69: 28-31.

Ohwi, J. 1984. Flora of Japan. Washington, D.C.: Smithsonian Institute.

Radford, A.E., W.C. Dickinson, J.R. Massey, and C.R. Bell. 1974. Vascular Plant Systematics. New York: Harper & Row.

Rehder, A. 1947. Bibliography of Cultivated Trees and Shrubs, Hardy in the Cooler Temperate Regions of the Northern Hemisphere. Boston: The Arnold Arboretum of Harvard University.

SAS Institute Inc. 1988. SAS/STAT Guide for Personal Computers, Ver. 6, SAS Institute Inc., Cary, North Carolina.

Sneath, P.H.A. and R.P. Sokal. 1973. Numerical Taxonomy. San Francisco: W.H. Freeman.

Van Houtte, L.B. 1874. Flora des Serres, Ser. 2. 20. Gent, Belgium.

Wen, J., G.M. Plunkett, A.D. Mitchell, and S.J. Wagstaff. 2001. The evolution of Araliaceae: A phylogenetic analysis based on ITS sequences of nuclear ribosomal DNA. Syst. Bot. 26: 144-167.

Willis, J.C. 1973. A Dictionary of the Flowering Plants and Ferns. 8th ed. London: Cambridge University Press.

Chang et al. — Morphometric analysis of Kalopanax septemlobus

Appendix. Origin and accession number for specimens utilized for morphological analysis. All voucher specimens are deposited at Seoul National University, The Arboretum (SNUA) or as otherwise indicated.

CHINA: Jiangxi Prov. Ce-Ming Tan 95712 (TUS); Chekiang anonymous 831 Aug. 1962 (PE); Chekiang PE 29274 (PE); Jiangxi PE 13479 (PE); Jiangxi anonymous 2771 (PE); Hunan L.H. Liu 9949 (PE); Hupei C.F. Li 9498 (PE); Anwei PE 50502 (PE); Kansu anonymous 3662 (PE); s.l. anonymous 587 (PE); Shensi H.W. Kung 2705 (PE); Shensi T.N. Liou 11356 (PE); Liaoning PE 318065 (PE); Szechuan anonymous 3499 (PE); Szechuan PE 105958 (PE); Szechuan anonymous 2713 (PE); Szechuan F.T. Wang 23635 (PE); Guangdong L. Deng 5775 (PE); Guangxi PE 16716 (PE); Yunnnan P.H. Yu 880 (PE); Yunnnan Salwin Valley, Pengta T.T. Yu 23103 (PE); Shandong Y.S. Cui 302 (PE); Liaoning Y. Yabe s.n. Aug. 14. 1917 (PE); Hopei K.M. Liou 287 (PE); Hopei H.F. Chow s.n. ? 1930 (PE); Hopei 15210 (PE); Hopei Tungling C.F. Li s.n. ? 1929 (PE); Hopei Eastern tombs J. Tang 2213 July 1931 (PE); Hopei Eastern tombs K.C. Kuan 2213 (PE); Shensi S.Y. Bao et al. 168 (PE); Shensi anonymous 604 (PE); Shandong Laoshan C.Y. Chiao 2783 (PE); Shandong Laoshan Y.C. Wang 773 (PE); Szechuan ? 2792 (PE); Szechuan W.P. Fang et al. 33045 (PE); Szechuan T.P. Wang 8217 (PE); Szechuan W.J. Zheng 1936 (PE); Szechuan F.T. Wang 23635 (PE); Szechuan F.T. Wang 21716 (PE); Szechuan Wan-Hsien C.T. Hwa 76 (PE); Szechuan W.P. Fang 3406 (PE); Szechuan PE 57574 (PE); Guizhou S.Z. Jiang et al. 32 (PE); Guizhou Y. Tsiang 9074 (PE); Guizhou C.Z. Yu 0621 (PE); Guangdong Y. Tsiang 1466 (PE); Guandong Y. Tsiang 1347 (PE); Yunnnan Changtien Plateau G. Forrest 11256 (PE); Hunan X.Y. Xi 454 (PE); Hupei PE 59176 (PE); Beijing C.S. Chang 3889 (SNUA).

JAPAN: Rikuzen Ohashi s.n. 2. Aug. 1959 (TUS); Kitami Pref. F.G. Meyer et al. 18990 (TUS); Hyogo Pref. N. Kurosaki 20114 (TUS); Fukushima Pref. Ohashi et al. 21895 (TUS); Ooisawa. Ohashi s.n. 4 Aug. 1961 (TUS); Hokaido Kiyomi s.n. 9. Aug. 1991 (TUS); Nagasaki Pref. F.G. Meyer et al. 17202 (TUS); Miyazaki Pref. TWTW17041 (TUS); Yamanishi Pref. TWTW18598 (TUS); Shinano Pref. K.Asano 3354 (TUS); Kanagawa Pref. H.T. Im & T. Kawahara 5278 (JNU); Tokyo Pref. H.T. Im & T. Kawahara 5278 (JNU); Hokkaido H.S. Kang s.n. July 8 2001 (SNUA); OKINAWA Ryukyu TUS4081 (TUS).

KOREA: Jeollanam-do Imjado H.T. Im et al. 38478 (JNU); Jeollanam-do Maehwado H.T. Im et al. 49872-1,2 (JNU); Jeollanam-do Uido H.T. Im & H.H. Hong 988318 (JNU); Gangwon-do Imgye-myon H.T. Im 35861 (JNU); Gyeonggi-do Dukjeokdo ? s.n. Aug. 17 1939 (SNU); Seoul Changkyeongwon ? s.n. July 7 1949 (SNU); Seoul Bukhansan Mori s.n. June 23 1935. (SNU); Seoul Changkyeongwon s.n. July 7 1942 (SNU); SNU9067 (SNU); Gyeongsangbuk-do Ulleungdo SNU9069 (SNU); Gangwon-do SNU9062 (SNU); Jeollanam-do SNU9603 (SNU); Gyeonggi-do Gwangreung SNU9605 (SNU); Gyeonggi-do Gwangreung SNU9606 (SNU); Gyeonggi-do Yongmunsan SNU58146 (SNU); Chungcheongbuk-do Baekunsan H.S. Won 94-4-48 (SNU); Gangwon-do Hyangrobong SNU541 (SNU); Chungcheongbuk-do Songnisan SNU81264 (SNU); Gyeonggi-do Aengjabong SNU62851 (SNU); Gyeonggi-do Yongmunsan S. Ko, Y. Lim & H. Shin 515015 (SNU); Gangwon-do Dongdaesan C.W. Park, et al. 10461 (SNU); Gangwon-do Sangwonsa C.W. Park, et al. 10731 (SNU); Gyeonggi-do Baekunsan S.P. Hong & H.S. Choi 904 (SNU); Chungcheongbuk-do Woraksan SNU68058 (SNU); Gyeongsangbuk-do Uljin-gun SNU51833 (SNU); Gangwon-do Gariwangsan B.R. Yinger et al. 247 (SNUA); Gangwon-do Geonbongsan HR263 (SNUA); Geonggi-do Yeongheungdo SNUA41565 (SNUA); SNUA41566 (SNUA); Mosigae falls T.B. Lee s.n. Aug. 8 1961 (SNUA); Gyeongsangnam-do Changwon-si huikim466 (SNUA); Geonggi-do Gwangneung T.B. Lee & S.J. Park s.n. Oct. 15 1960 (SNUA); Incheon Gangwhado huikim474 (SNUA); Gangwon-do Samcheok-si C.S. Chang 2073 (SNUA); Geonggi-do Yeongheungdo SNUA41572 (SNUA); Incheon Deokjeokdo T.B. Lee s.n. Aug. 6 1952 (SNUA); SNUA41574 (SNUA); Incheon Deokjeokdo T.B. Lee s.n. Aug. 6 1952 (SNUA); SNUA41576 (SNUA); Gangwon-do Taebaeksan T.B. Lee et al. s.n. July 23 1964 (SNUA); Chungcheonbuk-do Minjujisan C.S. Chang 2875 (SNUA); Gangwon-do Jeombongsan JB152 (SNUA); Jeollabuk-do Deogyusan C.S. Chang 2755 (SNUA); Gyeongsangbuk-do Andong-si J.G. Seo s.n. Nov. 18 1983 (SNUA); Gyeonggi-do Suwon T.B. Lee & J.W. Park s.n. Oct. 15 1960 (SNUA); Gyeonggi-do Baekado T.B. Lee s.n. Aug. 4 1952 (SNUA); Incheon Deokjeokdo SNUA41584 (SNUA); Incheon Deokjeokdo SNUA41585 (SNUA); Gyeonggi-do Daebudo T.B. Lee et al. s.n. July 14 1963 (SNUA); Gyeonggi-do Daebudo T.B. Lee s.n. Aug. 6 1952 (SNUA); Gyeonggi-do Daebudo T.B. Lee s.n. July 14 1963 (SNUA); Gyeonggi-do Daebudo T.B. Lee s.n. Aug. 6 1952 (SNUA); Gangwon-do Heungjeong valley H.S. Kang s.n. Oct. 17 1999 (SNUA); Gangwon-do Gachilbong T.B. Lee s.n. July 23 1981 (SNUA).

RUSSIA: Primorsky territory, Vladivostok, C.W. Park et al. s.n. Aug. 11 1995 (SNU); Vladivostok H.S. Kang s.n. Aug. 26 2001 (SNUA); Ussurisk H.S. Kang s.n. Aug. 26 2001 (SNUA); Ussurisk H.S. Kang s.n. Aug. 26 2001 (SNUA).

Botanical Bulletin of Academia Sinica, Vol. 44, 2003