Botanical Studies (2012) 53: 275-282.

SYSTEMATICS

A taxonomic study of Camellia brevistyla and C. tenuiflora (Theaceae) based on phenetic analyses

Mong-Huai SU¹, Chang-Fu HSIEH², Jenn-Che WANG³, and Chih-Hua TSOU⁴ *

¹ Department of Forestry and Nature Conservation, Chinese Culture University, Taipei 111, Taiwan
²Institute of Ecology and Evolutionary Biology, National Taiwan University, Taipei 106, Taiwan
³Department of Life Science, National Taiwan Normal University, Taipei 116, Taiwan
⁴Institute of Plant and Microbial Biology, Academia Sinica, Taipei 115, Taiwan

(Received January 10, 2011; Accepted December 1, 2011)

ABSTRACT. Camellia brevistyla and C. tenuiflora, two ecologically and economically important species of Taiwan, have long been confused due to their morphological similarities. This paper aims to clarify their taxonomic relationship based on phenetic analyses of morphological data. We characterized the patterns of morphological variations of C. brevistyla and C. tenuiflora, along with C. hengchunensis and C. kissi, throughout their geographic ranges in Taiwan. The ordination plots obtained from non-linear principal analyses showed considerable overlap among the samples of C. brevistyla and C. tenuiflora, and none of the characters applied was able to distinguish the two species apart. Further assessment of morphological characters revealed considerable morphological variations across the entire geographic range. Pearson's correlation coefficients for the data suggested the occurrence of a continuum of morphological variation among the C. brevistyla-C. tenuiflora complex. It is concluded that C. brevistyla and C. tenuiflora belong to a single highly-variable species and the taxonomic treatment of C. tenuiflora as a synonym of C. brevistyla is supported.

Keywords: Camellia brevistyla; Camellia tenuiflora; Phenetics; Taiwan; Taxonomy.

INTRODUCTION

Camellia are the largest and the type genus of Theaceae, with the number of species varying from 110 to 300 depending on different taxonomic treatments (Chang and Ren, 1998; Ming, 2000; Ming and Bartholomew, 2007). In Taiwan, 12 native species of Camellia were described in the second edition of the Flora of Taiwan (Hsieh et al., 1996). Thereafter, C. buisanensis Sasaki was moved to the genus Pyrenaria based on its morphological characters (Su et al., 2004) and C. kissi Wallich was newly recorded (Su et al., 2008). Among these Taiwanese Camellia species, the taxonomic ambiguity on the highly similar C. brevistyla (Hayata) Cohen-Stuart and C. tenuiflora (Hayata) Cohen-Stuart appears as one of the major taxonomic problems.

The type specimen of C. brevistyla was sourced from Arizan and Tozan, Chiayi County, south-central Taiwan (Hayata, 1908; in TAIF) and that of C. tenuiflora was from Wantan, Taipei County, northern Taiwan (Hayata, 1911; in TI) (Figure 1); but no comparison on the two taxa was provided in Hayata's publications. Due to their high similarities, the taxonomic treatments on C. brevistyla and C. tenuiflora have been controversial for a long time. Most studies accepted them as two independent species (see literature in Table 1), but Ming and Bartholomew (2007)

*Corresponding author: E-mail: chtsou@gate.sinica.edu.tw.

merged them under the name of C. brevistyla. Several features on distinguishing C. brevistyla from C. tenuiflora were proposed, with leaf morphology and style number as key criteria (Table 1); however, serious problems occur while applying these criteria in the identification in the field. For example, the most dominant Camellia species in Chatienshan Nature Reserve (northern Taiwan) has been identified as C. brevistyla (Su and Wang, 1988; Lu et al., 1998) and C. tenuiflora (Tsoong and Chang, 1954; Hsieh et al., 1987). The same problem exists in the identification of the most dominant Camellia species in Hoping Logging Road in eastern Taiwan (e.g., C. brevistyla, S. Y. Lu 12703, TAIF; C. tenuiflora, M. F. Kao 260, TAIF; C. tenuiflora, J. C. Wang et al. 8620, TNU; Appendix 1).

Camellia brevistyla and C. tenuiflora are widely distributed in Taiwan based on specimens from various herbaria (Figure 1) and form important components in those broadleaved forests (Hsieh et al., 1987; Yang, 1997; Chen, 2005). These two species are also cultivated for oil production and horticultural purposes (Wang and Lin, 1990; Wang et al., 1994; Ming, 2000). Therefore, detailed taxonomic studies are essential not only for the identification of these species in the field, but also for the conservation and management of natural resources (Meiri and Mace, 2007). In this study, the taxonomic clarity of C. brevistyla and C. tenuiflora is addressed and the patterns of morphological variation of C. brevistyla

276

Botanical Studies, Vol. 53, 2012

and C. tenuiflora throughout their geographic ranges in Taiwan are characterized. Two native sibling species, C. hengchunensis Chang and C. kissi from the section Paracamellia (Ming, 2000; Vijayan et al., 2009) which C. brevistyla and C. tenuiflora belong to, are also compared. Many previous studies show that plants usually exhibit a morphological plasticity in response to environmental

conditions (Thompson, 1991; Ro9as et al., 1997) and that frequently causes taxonomic problems. Since C. brevistyla and C. tenuiflora are widely distributed, a test of the correlation between morphological variation and geographical location of the two taxa is also conducted in this study.

MATERIALS AND METHODS

Materials

For scoring morphological characters, specimens deposited in the herbaria were examined (Appendix 1). Each collection was designated as an operational taxonomic unit (OTU). Duplicate specimens were treated as the same OTU. Scientific names given on the labels of specimens were tentatively used. A total of 149 specimens were examined, but those without sufficient measureable or label information were excluded from the data analyses. Finally, 74 specimens (OTUs), including 24 specimens of C. brevistyla, six of C. hengchunensis, four of C. kissi and 40 of C. tenuiflora, were included (Appendix 1). The specimens of C. brevistyla and C. tenuiflora represent their distribution range in Taiwan (Figure 1). Although type specimens of C. brevistyla and C. tenuiflora were not included in the study because they did not provide enough measurable traits, specimens from the type locality were used (e.g. C. brevistyla, M. H. Su 679 and 681, TAI; C. tenuiflora, P. F. Lu 7175, TNU; Appendix 1). The small sample sizes of C. hengchunensis and C. kissi were due to their limited distribution and small population sizes in Taiwan (Figure 1).

Characters analysis

Thirty-three characters were chosen on the basis of the variation mentioned in the literature and our own judgement: 12 leaf characters, 16 floral characters and five fruit characters (Tables 2-4). For each specimen, five mature and healthy-looking leaves were scored and averaged. The measurements of floral and fruit characters were averaged from one to three replicates, depending on the condition of the specimen.

Figure 1. Distribution of the populations of Camellia brevistyla and C. tenuiflora in Taiwan. Solid circle, C. brevistyla sampled. Open circle, C. brevistyla not sampled. Solid star, C. tenuiflora sampled. Open star, C. tenuiflora not sampled. Solid triangle, C. kissi. Open triangle, C. hengchunensis. Arrow marked with B, collection site of type specimen of C. brevistyla. Arrow marked with T, collection site of type specimen of C. tenuiflora. Data were sourced from HAST, PPI, TAI, TAIF, TNM, and TNU.


Table 1. A comparison of the useful taxonomical features of Camellia brevistyla and C. tenuiflora cited in previous reports.
Source	C. brevistyla	C. tenuiflora
Yamamoto and Mori, 1934 Secondary veins inconspicuous.		Secondary veins flat or stressed above, conspicuous beneath.
Keng, 1950	Styles 4, 1 mm long, connate at base.	Styles 3, 3 mm long, free at upper portion or entirely connate at the base.
Liu and Lu, 1967	Secondary veins inconspicuous.	Secondary veins 5-8 pairs, conspicuous, slightly elevated on both surfaces.
Li, 1976	Styles 1-1.5 mm long, connate at base.	Styles 3 mm long, free at upper portion or entirely connate.
Ying, 1995	Leaf apex obtuse; styles 3-fid.	Leaf apex acute; styles shallowly 3-fid or entirely connate.
Hsieh et al., 1996	Styles 4, 1-1.5 mm long; leaves rhombic-oblong, midribs pubescent above.	Styles 3, more than 2.5 mm long; leaves obovate-elliptic, midribs almost glabrous above.

SU et al. ― Taxonomy of Camellia brevistyla and C. tenuiflora

277

Table 2. Loadings of the 12 leaf characters on the first three axes of non-linear principal component analysis (NLPCA). Eigenvalues, Cronbach's a, percentage and cumulated percentage of variance accounted for are given for each axis.

a = 0.88, Table 2). The second axis was close to the level of reliability of the first (Cronbach's a = 0.64) and explained a further 20.23% of the variance. The ordination plot based on the first two axes showed three distinct groups, designated as Gr1-1, Gr1-2 and Gr1-3 (Figure 2a). Gr1-1 was composed of C. kissi and Gr1-2 of C. hengchunensis, while Gr1-3 was a mixture of C. brevistyla and C. tenuiflora. The leaf characters with high loadings on the first axis were length, width, thickness, serration, midrib pubescence, secondary vein angle, petiole length and petiole pubescence (Table 2), which separated C. hengchunensis (Gr1-2) from the other three taxa. The characters that contributed most to the second axis were leaf length and apex shape (Table 2). Camellia kissi (Gr1-1) can be distinguished easily from the C. brevistyla-C. tenuiflora mixture (Gr1-3) along this axis.

The first three axes of NLPCA using floral characters explained less than half of the total variance (49.62%). The first axis was reliable (Cronbach's a=0.81, Table 3) but accounted for only 26.03% of the variance. The characters with high loadings on the first axis were petal length, petal width, sepal length, sepal width, style length and style pubescence (Table 3). However, variation of those characters was continuous and overlapped among OTUs from

Table 3. Loadings of the 16 floral characters on the first three axes of non-linear principal component analysis (NLPCA). Eigenvalues, Cronbach's a, percentage and cumulated percentage of variance accounted for are given for each axis.


	NLPCA axis
Character	1 2 3
Leaf length	0.67	0.68	-0.06
Leaf width	0.61	0.48	0.12
Leaf thickness	0.83	-0.38	-0.05
Leaf shape	0.24	-0.53	0.59
Leaf apex shape	-0.10	-0.80	-0.03
Leaf serration	-0.85	0.37	0.12
Midrib pubescence	0.92	-0.32	-0.11
Pairs of secondary veins	0.36	0.36	0.34
Secondary vein angle	-0.62	-0.12	0.28
Secondary vein prominence	-0.46	-0.18	-0.69
Petiole length	0.62	0.39	-0.17
Petiole pubescence	0.92	-0.32	-0.11
Eigenvalue	5.09	2.43	1.10
Cronbach's a	0.88	0.64	0.10
Variance explained (%)	42.38	20.23	9.19
Cumulative variance explained (%)	42.38	62.61	71.80

Data analysis

To analyze the phenological features of the specimens, the data were rearranged into three combinations of leaf characters (74 OTUs x 12 characters), floral characters (48 OTUs x 16 characters) and fruit characters (34 OTUs x 5 characters).

To group the OTUs and assess the pattern of morphological variation, nonlinear principal components analysis (NLPCA, de Leeuw, 1982) was performed. To assess the reliability of each component (NLPCA axis), Cronbach's a (Cronbach, 1951) was computed and a value of 0.70 or greater was considered to indicate the reliability (Bland and Altman, 1997). To investigate the relationship between geographic location and morphological variation among specimens of C. brevistyla and C. tenuiflora, an analysis of the Pearson correlation was conducted. Each geographic locality, obtained from the specimen labels, was defined by longitude, latitude and altitude. Total morphological variation was represented by the OTU scores on the first two NLPCA axes. All analyses in this study were performed using SPSS software v11.0 (SPSS Inc.).

RESULTS

The first three NLPCA axes generated in the analysis based on leaf characters accounted for 71.8% of the total variance; the first axis accounted for 42.38% of the total variance and was highly reliable (Cronbach's


	NLPCA axis
Character	1 2 3
Petal number per flower	-0.06	-0.11	-0.13
Petal length	0.81	-0.03	0.03
Petal width	0.75	-0.16	-0.09
Petal pubescence	0.01	0.48	-0.74
Petal emargination	-0.06	-0.10	-0.23
Sepal number per flower	0.35	0.51	0.21
Sepal length	0.73	0.10	-0.07
Sepal width	0.78	0.04	-0.02
Sepal pubescence	-0.11	-0.47	0.77
Filament length	0.44	-0.52	-0.09
United filament tube length	-0.55	-0.43	0.06
Style length	0.81	-0.08	0.08
Style number per flower	0.03	0.62	0.59
United style length	0.13	-0.56	-0.38
Style pubescence	-0.64	0.16	-0.17
Ovary length	0.31	0.03	0.08
Eigenvalue	4.16	1.96	1.82
Cronbach's a	0.81	0.52	0.48
Variance explained (%)	26.03	12.23	11.36
Cumulative variance explained (%)	26.03	38.26	49.62

278

Botanical Studies, Vol. 53, 2012

Table 4. Loadings ot the five truit characters on the first three axes of non-linear principal component analysis (NLPCA). Eigenvalues, Cronbach's a, percentage and cumulated percentage of variance accounted for are given for each axis.


	NLPCA axis
Character	1 2 3
Fruit shape	0.72	0.55	-0.02
Fruit length	0.95	-0.13	-0.23
Fruit width	0.67	-0.64	-0.32
Fruit pedicel length	0.58	0.62	0.31
Fruit pubescence	0.36	-0.58	0.72
Eigenvalue	2.33	1.44	0.77
Cronbach's a	0.71	0.39	-0.37
Variance explained (%)	46.59	28.89	15.44

Cumulative variance explained (%)	46.59	75.48	90.92

C. hengchunensis, C. kissi and the species complex of C. brevistyla and C. tenuiflora. No clear distinction could be found among the previously-recognized groups along the first axis (Figure 2b).

In NLPCA based on fruit characters (Table 4), the first axis accounted for nearly half of the total variance (46.59%) and was the only reliable axis (Cronbach's a =0.71). Three distinct groups, Gr2-1, Gr2-2 and Gr2-3, were recognized in the ordination plot based on the first two axes (Figure 2c). Gr2-1 was composed of C. kissi and Gr2-2 of C. hengchunensis, while Gr2-3 was a mixture of C. brevistyla and C. tenuiflora. Fruit shape, length, width and pedicel length were the most important for determining the scores of OTUs on the first axis, whereas fruit width, pedicel length and pubescence were the most important for determining the second axis (Table 4).

The test of correlations between geographic location and morphological variation among specimens of C. brevistyla and C. tenuiflora (Table 5) revealed that variation of leaf characters was significantly correlated with longitude (the first and second axes of NLPCA) and latitude (the second axis of NLPCA), but not correlated with altitude. Neither the test for floral characters nor that for fruit characters showed any significant correlation.

DISCUSSION

The phenetic analyses showed that samples of C. brevistyla and C. tenuiflora were mixed in a group in all the ordination plots. The morphological characters traditionally used to distinguish them apart include leaf shape, leaf apex shape, secondary vein prominence, style number and style length (Table 1). However, our analyses based on every single character and all kinds of combinations of multiple characters failed to show any discrete clusters within the C. brevistyla-C. tenuiflora complex.

Figure 2. Ordination plots of the first two axes of NLPCA based on leaf (a), floral (b) and fruit (c) characters. Open circle, Camellia brevistyla. Solid circle, C. tenuiflora. Solid star, C. hengchun-ensis. Solid triangle, C. kissi.

SU et al. ― Taxonomy of Camellia brevistyla and C. tenuiflora

279

Table 5. Pearson's correlation coefficients between geographic locations and scores of the first two NLPCA axes based on leaf, flower and fruit characters.


		Leaf (n=64)			Flower (n=40)			Fruit (n=28)
	LG	LT	ALT	LG	LT	ALT	LG	LT	ALT
NLPCA axis 1	-0.26*	-0.23	0.07	0.22	0.15	0.12	-0.02	-0.03	0.05
NLPCA axis 2	-0.31*	-0.44**	0.11	0.12	0.20	0.12	0.04	-0.09	-0.05

LG: longitude; LT: latitude; ALT: altitude. Correlation is significant at the 0.05 (*) or 0.01 (**) level using a 2-tailed test.

According to Keng (1950) and Hsieh et al. (1996), C. brevistyla has a style less than 1.5 mm long, whereas C. tenuiflora has a style longer than 2.5 mm; however, half (20/40) of the specimens examined have a style between 1.5 and 2.5 mm long. The four-styled feature was also used as a diagnostic character for C. brevistyla when the species was firstly described (Hayata, 1908); but in fact, the number of styles of the specimens scored in this study ranged from one to four: 28 out of 40 specimens with three styles, and only one specimen (A. T. Hsieh and M. T. Kao 2, Appendix 1) with four styles. A flower with three styles is a common character of species in section Paracamellia (Sealy, 1958); thus the four-styled state seems to be a rare case and unsuitable for distinguishing C. brevistyla from C. tenuiflora. The taxonomic importance of secondary vein prominence, which was emphasized in some literatures (Yamamoto and Mori, 1934; Liu and Lu, 1967) and adopted by some field investigators, displayed only a low loading in the NLPCA results (Table 2). Leaf shape and leaf apex shape displayed higher loadings on the second NLPCA axis; however, these characters are effective for the identification of C. kissi rather than for resolving the C. brevistyla-C. tenuiflora complex.

The correlation test showed that leaf variation in the C. brevistyla-C. tenuiflora complex is associated with geographic location (Table 5). Many studies have revealed such a relationship, which has been interpreted as a phenotypic response (adaptation) to the environmental gradient (Jonas and Geber, 1999; Santamaria et al., 2003; Herrera, 2005) and claimed to cause taxonomic confusions (Ellison et al., 2004). C. brevistyla was first described based on a specimen from Chiayi County (south-central Taiwan) and C. tenuiflora based on a specimen from Taipei County (near northern tip of Taiwan). Since type specimens of these two species were obtained from two extremes of the geographical range in Taiwan, our analyses suggest that it could be one of the reasons why C. brevistyla and C. tenuiflora have long been treated as two species. However, the patterns of morphological variations were found to be continuous along the geographical gradient. While the variation throughout the range of the complex has been considered, no proper delimitation of C. brevistyla and C. tenuiflora could be made.

Based on leaf and fruit characters, samples of C. hengchunensis and C. kissi were clearly separated from each other in the ordination plots, and both were also clearly distinct from the C. brevistyla-C. tenuiflora

complex. The results suggest that C. hengchunensis and C. kissi should be treated as distinct species, respectively, and thus are not in agreement with the view that C. hengchunensis should be reduced to a synonym of C. brevistyla (Ming and Bartholomew, 2007). Useful characters to identify C. hengchunensis are the glabrous midrib, acute or obtuse leaf apex, thick leaf blades, nearly entire margins, glabrous petioles, and oblong capsules. Camellia kissi can be identified by its caudate leaf apex, longer leaf blades (> 5.5 cm), villose midrib and pyriform capsules, as mentioned by Su et al. (2008).

In conclusion, the present phenetic study does not support the treatment of C. tenuiflora and C. brevistyla as two distinct species. The treatment of C. tenuiflora as a synonym of C. brevistyla as earlier proposed by Ming and Bartholomew (2007) is supported.

Acknowledgments. We thank the directors and crews of HAST, PPI, TAI, TAIF, TNM, and TNU for their helps while we were studying specimens there. Two reviewers are also acknowledged for providing helpful comments. This study was supported by grants NSC99-2321-B-034-001 from the National Science Council.

LITERATURE CITED

Bland, J.M. and D.G. Altaian. 1997. Statistics notes: Cronbach's alpha. Brit. Med. J. 314: 572.

Chang, H.T. and S.X. Ren. 1998. Theaceae (1): Theoideae. In H.T. Chang (ed.), Flora Reipublicae Popularis Sinicae to-mus 49. Science Press, Beijing, pp. 22-33.

Chen, T.Y. 2005. A Survey on Plant Resources of Beech Forest in Tungshan Area (2/2). Forestry Bureau, Council of Agriculture, Executive Yuan, Taipei.

Cronbach, L.J. 1951. Coefficient alpha and the internal structure of tests. Psychometrika 16: 297-334.

de Leeuw, J. 1982. Nonlinear principal components analysis. In H. Caussinus, P. Ettinger, and R. Tomassone (eds.), COMPSTAT Proceedings in Computational Statistics. Physica Verlag, Vienna, Austria, pp. 77-89.

Ellison, A.M., H.L. Buckley, T.E. Miller, and N.J. Gotelli.

2004. Morphological variation in Sarracenia purpurea (Sarraceniaceae): geographic, environmental, and taxonomic correlates. Am. J. Bot. 91: 1930-1935.

Hayata, B. 1908. Flora montana Formosae. J. Coll. Sci. Imp.

280

Botanical Studies, Vol. 53, 2012

Univ. Tokyo 25: 63.

Hayata, B. 1911. Materials for a flora of Formosa. J. Coll. Sci. Imp. Univ. Tokyo 30: 46-47.

Herrera, J. 2005. Flower size variation in Rosmarinus officinalis: individuals, populations and habitats. Ann. Bot. 95: 431437.

Hsieh, C.F., L.K. Ling, and K.C. Yang. 1996. Camellia, Theaceae. In D.E. Boufford, C.F. Hsieh, T.C. Huang, H. Ohashi, and Y.P. Yang (eds.), Flora of Taiwan 2nd. ed., Vol. 2. Editorial Committee of Flora of Taiwan, Taipei, pp. 662693.

Hsieh, C.F., W.S. Tang, Y.F. Lin, Y.C. Lin, Z. S. Chen, K.C. Lin, and C.M. Chang. 1987. Developing an Ecological Data System for Nature Preserves (II). Council of Agriculture, Executive Yuan, Taipei.

Jonas, C.S. and M.A. Geber. 1999. Variation among populations of Clarkia unguiculata (Onagraceae) along altitudinal and latitudinal gradients. Am. J. Bot. 86: 333-343.

Keng, H. 1950. The Theaceae of Taiwan. Taiwania 1: 223-268.

Li, H.L. 1976. Theaceae. In H.L. Li, T.S. Liu, T.C. Huang, T.

Koyama, and C.E. DeVol (eds.), Flora of Taiwan, Vol. 2. Epoch Publ. Co., Taipei, Taiwan, pp. 592-619.

Liu, T.S. and F.Y. Lu. 1967. Studies in the Taiwan Theaceae based on the morphological characters of leaves. Tech. Bull. Exp. For. Nat. Taiwan Univ. 52: 1-55.

Lu, J.C., C.H. Ou, and C.A. Chiou. 1998. Studies on the

vegetation of Chatienshan Nature Reserve (II) - the population constitution of Taiwan beech. Bull. Exp. For.

Natl. Chung Hsin Univ. 20: 79-92.

Meiri, S. and G.M. Mace. 2007. New taxonomy and the origin

of species. PLoS Biol. 5: 1385-1386.

Ming, T.L. 2000. Monograph of the Genus Camellia. Yunnan Science and Technology Press, Kunming.

Ming, T.L. and B. Bartholomew. 2007. Theaceae. In Z.Y. Wu, P.H. Raven, and D.Y. Hong (eds.), Flora of China, Vol. 12 (Hippocastanaceae through Theaceae). Science Press and Missouri Botanical Garden Press, Beijing, St. Louis, pp.

366-478.

Ro9as, G., C.F. Barros, and F.R. Scarano. 1997. Leaf anatomy plasticity of Alchornea triplinervia (Euphorbiaceae) under

distinct light regimes in a Brazilian montane Atlantic rain

forest. Trees Struct. Funct. 11: 469-473.

Santamaria, L., J. Figuerola, J.J. Pilon, M. Mjelde, A.J. Green,

T.d. Boer, R.A. King, and R.J. Gornall. 2003. Plant performance across latitude: the role of plasticity and local adaptation in an aquatic plant. Ecology 84: 2454-2461.

Sealy, J. 1958. A revision of the genus Camellia. Royal Horticultural Society, London.

Su, H.J. and L.J. Wang. 1988. Forest vegetations in the upper Nanshihchi watershed of northern Taiwan. Q. J. Exp. For.

NTU 2: 89-100.

Su, M.H., S.Z. Yang, and C.F. Hsieh. 2004. The identity of Camellia buisanensis Sasaki (Theaceae). Taiwania 49: 201208.

Su, M.H., C.F. Hsieh, C.H. Tsou, and J.C. Wang. 2008. Camellia kissi Wallich (Theaceae): a newly recorded tree from Taiwan. Taiwan J. For. Sci. 23: 301-307.

Thompson, J.D. 1991. Phenotypic plasticity as a component of evolutionary change. Trends Ecol. Evol. 6: 246-249.

Tsoong, P.C. and L.M. Chang. 1954. A preliminary survey of the vegetation of the South Cha-Tian Mountains. Bull. Taiwan For. Res. Inst. New Ser. 41: 1-36.

Vijayan, K., W.J. Zhang, and C.H. Tsou. 2009. Molecular taxonomy of Camellia (Theaceae) inferred from nrITS sequences. Am. J. Bot. 96: 1348-1360.

Wang, C.L. and Y.H. Lin. 1990. The extraction and analysis of oils from selected species of oil-tea Camellia in Taiwan. Bull. Taiwan. For. Res. Inst. 5: 11-15.

Wang, C.L., H.W. Yin, and W.Y. Liu. 1994. Stability examination of oil-tea oil and analysis of oil's tocopherol and sterol components. Bull. Taiwan For. Res. Inst. New Ser. 9: 73-86.

Yamamoto, Y. and K. Mori. 1934. Determination of the Formo-san species of Ternstroemiaceae (Theaceae) based on the morphological characters of leaves. Sylvia 5: 25-49.

Yang, J.C. 1997. Study on the streamside vegetation of Tonghou Stream in Wulai, Taipei County. Taiwan J. For. Sci. 12: 335346.

Ying, S.S. 1995. Systematic study on Theaceae of Taiwan. Mem. Coll. Agric. Nat. Taiwan Univ. 35: 131-170.

SU et al. ― Taxonomy of Camellia brevistyla and C. tenuiflora

281

Appendix 1. List of voucher specimens used in the phenetic analyses. Counts of specimens collected in the same county are included in parentheses.


	County	Collection information
	C. brevistyla
	Taipei (4)	Erhkoshan, 8 Nov. 1998, S. S. Li 174 (TAIF).
		Lopeishan, 1,500 m, 27 Aug. 1935, T. Suzuki ST18965 (TAI).
		Peichatienshan, 600-1,400 m, 28 Sep. 1984, R. T. Li 3535 (TAI).
		Sintien Stream, 23 Nov. 1932, T. Suzuki 7829 (TAI).
	Ilan (2)	Songluo Lake, 1240 m, 20, Jul. 2000, S. W. Chung 2677 (TAIF); same date, C. C. Chen 7460 (TAIF).
	Chiayi (3)	Menyue, 2,300 m, 1 Nov. 1918, E. H. Wilson s. n. (TAIF).
		Tozan, 2,100 m, 26 Oct. 2005,M. H. Su 679, 681 (TAI).
	Hualien (7)	Hopping Logging Road, 57 km, 2,100 m, 3 Aug. 1983, S. Y. Lu 12703 (TAIF).
		Lainshan, 1,850-1,900 m, 10 Aug. 1956, Liu et al. 150 (TAI); 2100-2330 m, 15 Aug. 1956, Liu et al.
		242 (TAI).
		Paliwanshan, 900 m, 21 Dec. 2004, M. H. Su 587, 588, 589 (TAI).
		Pilu, 2300 m, 20 Sep. 1985, S. Y. Lu 17013 (TAIF).
	Kaohsiung (1)	Sasi Logging Road, 1,640 m, 19 Nov. 1983, S. Y. Lu s. n. (TAIF).
	Pingtung (4)	Fanshan, 31 Jan. 1997, S. Y. Lu 25226 (TAIF).
		Chachayalaishan, 900 m, 27 Dec. 1995, S. Y. Lu 24940 (TAIF).
		Lilongshan, 1,000 m, date unknown, H. M. Chang s. n. (TAI); 12 Jul. 1993, T. T. Chen et al. 1544
		(TAIF).
	Taitung (3)	Kwetien, 560 m, 11 Nov. 2003, M. H. Su 439, 441 (TAI).
		Tawu, 28 Sep. 1997, K. C. Yang 5106 (TAIF).
	C. tenuiflora
	Keelung (2)	Chingjen Lake, 200 m, 30 Oct. 2003, Y. P. Cheng 4232 (TAIF); 150 m, 28 Aug. 1993, W. P. Leu et 1988 (TNU).

	Taipei (17)	Chaitsushan, 18 Oct. 1938, T. Nakamura 2217 (TAI).
		Erhkoshan, 600 m, 23 Nov. 1996, C. H. Chen 1878 (TNU).
		Houtung, 2 Jan. 1996, S. Y. Lu 24944 (TAIF).
		Huangtitien, 10 Sep. 1985, K. C. Yang 1612 (TAI); 500-550 m, 22 Aug. 1996, C. K. Liou et al. 355
		(TNU); 300-500 m, 22 Dec. 1995, J. C. Wang 9726 (TNU); 500-570 m, 4 Oct. 1996, J. C. Wang et
		al. 10196 (TNU); 450-500 m, 16 Aug. 2003, S. C. Liu & H. M. Chang 1450 (TNU).
		Kuolai, 25 Dec. 1988, Y. P. Chiang s. n. (TNU).
		Lusytarn, 21 Feb. 1983, S. Y. Lu 12521 (TAIF).
		Mon Lake, 600 m, 23 Dec. 2003, M. H. Su 488 (TAI).
		NTU campus, 1939 (date unknown), K. Mori s. n. (TAI).
		Panpinshan, 510-550 m, 11 Aug. 1996, C. K. Liou et al. 281 (TNU).
		Peichatienshan, date unknown, H. M. Chang s. n. (TAI).
		Taipei, 9 Dec. 1927, Sasaki s. n. (TAI).
		Tongho, 23 Sep. 1979, C. L. Huang 834 (TAI).
		Tsankuangliaoshan, 738 m, 3 Jun. 1998, Y. P. Cheng 2290 (TAIF).
	C. tenuiflora
	Taipei (5)	Wantan, 450 m, 31 Dec. 2003, P. F. Lu 7175 (TNU).
		Wu-er-cha-hu-shan, 300 m, 7 Nov. 2005, M. H. Su 682 (TAI)
		Yangmingshan, 13 Nov. 1924, Sasaki s. n. (TAI); 2 Nov. 1928, Sasaki s. n. (TAI); 15 Dec. 1934,
		Sasaki s. n. (TAI).
	Ilan (2)	Taipingshan, date unknown, C. T. Huang s. n. (TAI)
		Yu-en-shan, 700 m, 8 Oct. 2003, M. H. Su 326 (TAI)


	282	Botanical Studies, Vol. 53, 2012
	Appendix 1. (Continued)
	County	Collection information
	Taoyuan (1)	Wutukushan, 100-300 m, 11 Sep. 2002, C. C. Chen s. n. (TNU).
	Hsinchu (5)	Hsinchu, Oct. 1913, S. Sasaki 19699 (TAI),
		Litungshan, 1,940 m, 19 Sep. 2000, S. W. Chung 3077 (TAIF); 1,914 m, 7 Dec. 2002, Y. Y. Huang
		1261 (TAI); 1,540-1,913 m, 28 Aug. 2003, S. C. Liu 1512 (TNU).
		Peipu, Dec. 1910, Kawakami s. n. (TAIF)
	Hualien (4)	Ertzishan, 2,100 m, 28 Jul. 1984, S. Y. Lu 15010 (TAIF).
		Hoping Logging Road, 1,800-2,100 m, 21 Sep. 1991, M. F. Kao 260 (TAIF); 1,800 m, 28 Jul. 1993, J.
		C. Wang et al. 8620 (TNU).
		Tayulin, 2,047 m, 4 Dec. 1997, C. K. Liou et al. 918 (TAIF).
	Nantou (2)	Noko, 11 Dec. 1930, Sasaki s. n. (TAI)
		Weisun, 27 Feb. 1955, A. T. Hsieh & M. T. Kao 2 (TAI)
	Pingtung (1)	Peililongshan, 800 m, 27 Jan. 1994, T. T. Chen et al. 3062 (TAIF).
	Taitung (1)	Tuyayuanshan, 1,230 m, 30 Apr. 2005, M. H. Su 619 (TAI).
	C. hengchunensis
	Pingtung (6)	Nanjen Lake, 100 m, 30 May 1991, C. H. Tsou & K. H. Wang 482 (PPI).
		Nanjenshan, Oct. 1919, E. Matuda s. n. (TAI); 15 Dec. 1977, C. I. Peng 3464 (TAI); 320 m, 5 Jan.
		1984, S. Y. Lu 13900 (TAIF); 11 Jan. 1980, C. E. Chang s. n. (PPI).
		Tsufonbi, 370 m, 6 Jan. 1984, S. Y. Lu 13877 (TAIF).
	C. kissi
	Taitung (4)	Chuifenshan, 1,200 m, 29 Sep. 2005, M. H. Su 662, 663, 664, 665 (TAI).

使用數值分類方法探討短柱山茶與細葉山茶之分類問題

蘇夢淮¹ 謝長富² 王震哲³ 鄒稚華⁴

¹中國文化大學森林暨自然保育學系
²台灣大學生態學與演化生物學研究所
³台灣師範大學生命科學系
⁴中央研究院植物暨微生物學研究所

短柱山茶與細葉山茶是兩個在生態與經濟上都佔有重要地位的原生山茶屬植物，但因爲形態上的近
似度高而經常有難以區分的情況。本硏究旨在使用以形態特徵爲基礎的數値分類方法來探討兩者之間的
關係。我們測量了短柱山茶、細葉山茶、恆春山茶與落瓣油茶的標本，標本來源涵蓋了台灣全島。非線
性主成份分析結果指出，短柱山茶與細葉山茶形成一個無法分割的混合群，顯示兩者的形態變異極大且

有所重疊。對此混合群'進一歩以Pearson相關係數進行分析'發現其形態變異與地理位置呈現顯著相

關。綜合上述，本硏究認爲短柱山茶與細葉山茶並無法以形態進行確切的分類，本結果支持先前將細葉
山茶倂入短柱山茶之分類處理。

關鍵詞：短柱山茶；細葉山茶；數值分類；台灣。



	280	Botanical Studies, Vol. 53, 2012

	Univ. Tokyo 25: 63. Hayata, B. 1911. Materials for a flora of Formosa. J. Coll. Sci. Imp. Univ. Tokyo 30: 46-47. Herrera, J. 2005. Flower size variation in Rosmarinus officinalis: individuals, populations and habitats. Ann. Bot. 95: 431437. Hsieh, C.F., L.K. Ling, and K.C. Yang. 1996. Camellia, Theaceae. In D.E. Boufford, C.F. Hsieh, T.C. Huang, H. Ohashi, and Y.P. Yang (eds.), Flora of Taiwan 2nd. ed., Vol. 2. Editorial Committee of Flora of Taiwan, Taipei, pp. 662693. Hsieh, C.F., W.S. Tang, Y.F. Lin, Y.C. Lin, Z. S. Chen, K.C. Lin, and C.M. Chang. 1987. Developing an Ecological Data System for Nature Preserves (II). Council of Agriculture, Executive Yuan, Taipei. Jonas, C.S. and M.A. Geber. 1999. Variation among populations of Clarkia unguiculata (Onagraceae) along altitudinal and latitudinal gradients. Am. J. Bot. 86: 333-343. Keng, H. 1950. The Theaceae of Taiwan. Taiwania 1: 223-268. Li, H.L. 1976. Theaceae. In H.L. Li, T.S. Liu, T.C. Huang, T. Koyama, and C.E. DeVol (eds.), Flora of Taiwan, Vol. 2. Epoch Publ. Co., Taipei, Taiwan, pp. 592-619. Liu, T.S. and F.Y. Lu. 1967. Studies in the Taiwan Theaceae based on the morphological characters of leaves. Tech. Bull. Exp. For. Nat. Taiwan Univ. 52: 1-55. Lu, J.C., C.H. Ou, and C.A. Chiou. 1998. Studies on the vegetation of Chatienshan Nature Reserve (II) - the population constitution of Taiwan beech. Bull. Exp. For. Natl. Chung Hsin Univ. 20: 79-92. Meiri, S. and G.M. Mace. 2007. New taxonomy and the origin of species. PLoS Biol. 5: 1385-1386. Ming, T.L. 2000. Monograph of the Genus Camellia. Yunnan Science and Technology Press, Kunming. Ming, T.L. and B. Bartholomew. 2007. Theaceae. In Z.Y. Wu, P.H. Raven, and D.Y. Hong (eds.), Flora of China, Vol. 12 (Hippocastanaceae through Theaceae). Science Press and Missouri Botanical Garden Press, Beijing, St. Louis, pp. 366-478. Ro9as, G., C.F. Barros, and F.R. Scarano. 1997. Leaf anatomy plasticity of Alchornea triplinervia (Euphorbiaceae) under	distinct light regimes in a Brazilian montane Atlantic rain forest. Trees Struct. Funct. 11: 469-473. Santamaria, L., J. Figuerola, J.J. Pilon, M. Mjelde, A.J. Green, T.d. Boer, R.A. King, and R.J. Gornall. 2003. Plant performance across latitude: the role of plasticity and local adaptation in an aquatic plant. Ecology 84: 2454-2461. Sealy, J. 1958. A revision of the genus Camellia. Royal Horticultural Society, London. Su, H.J. and L.J. Wang. 1988. Forest vegetations in the upper Nanshihchi watershed of northern Taiwan. Q. J. Exp. For. NTU 2: 89-100. Su, M.H., S.Z. Yang, and C.F. Hsieh. 2004. The identity of Camellia buisanensis Sasaki (Theaceae). Taiwania 49: 201208. Su, M.H., C.F. Hsieh, C.H. Tsou, and J.C. Wang. 2008. Camellia kissi Wallich (Theaceae): a newly recorded tree from Taiwan. Taiwan J. For. Sci. 23: 301-307. Thompson, J.D. 1991. Phenotypic plasticity as a component of evolutionary change. Trends Ecol. Evol. 6: 246-249. Tsoong, P.C. and L.M. Chang. 1954. A preliminary survey of the vegetation of the South Cha-Tian Mountains. Bull. Taiwan For. Res. Inst. New Ser. 41: 1-36. Vijayan, K., W.J. Zhang, and C.H. Tsou. 2009. Molecular taxonomy of Camellia (Theaceae) inferred from nrITS sequences. Am. J. Bot. 96: 1348-1360. Wang, C.L. and Y.H. Lin. 1990. The extraction and analysis of oils from selected species of oil-tea Camellia in Taiwan. Bull. Taiwan. For. Res. Inst. 5: 11-15. Wang, C.L., H.W. Yin, and W.Y. Liu. 1994. Stability examination of oil-tea oil and analysis of oil's tocopherol and sterol components. Bull. Taiwan For. Res. Inst. New Ser. 9: 73-86. Yamamoto, Y. and K. Mori. 1934. Determination of the Formo-san species of Ternstroemiaceae (Theaceae) based on the morphological characters of leaves. Sylvia 5: 25-49. Yang, J.C. 1997. Study on the streamside vegetation of Tonghou Stream in Wulai, Taipei County. Taiwan J. For. Sci. 12: 335346. Ying, S.S. 1995. Systematic study on Theaceae of Taiwan. Mem. Coll. Agric. Nat. Taiwan Univ. 35: 131-170.