SHORT COMMUNICATION

HORTSCIENCE 42(2):399–402. 2007. not known. Evaluation and comparison of accessions in different seasons will be helpful in developing improved cultivars and in Variation in Morphological Traits cultivar identification. The objectives of this study were to 1) among Loropetalum chinense var. compare leaf and flower traits of various accessions in spring, summer and autumn; rubrum Accessions 2) characterize phenotypic diversity among these accessions using cluster analysis; and Zhiyi Bao1 and Bo Chen 3) select some accessions with good charac- School of Landscape Architecture, Zhejiang Forestry University, 311300 ters to use in the landscape. Linan, China, and Department of Horticulture, Zhejiang University, 268, Kaixuan Road, Hangzhou, Zhejiang Province, China 310029 Materials and Methods Hua Zhang Seedlings were collected from Liuyang City, Hunan Province, China, and planted in School of Tourism and Health, Zhejiang Forestry University, 311300 Linan, the L. chinense var. rubrum Garden at the China Academic Institution of Landscape Garden- Additional index words. data processing system, cluster analysis, phenotypic diversity ing of Changsha, the capital city of Hunan Province, in 1998. Field tests were estab- Abstract. Loropetalum chinense var. rubrum accessions in China have not been lished in spring 2001. Twenty-one accessions adequately characterized for their morphological traits. Such characterization would were selected from the garden according to be helpful in the development of improved cultivars and in cultivar classification. In this their different appearances of flowers in study, the morphological traits of 23 accessions were evaluated in spring, summer, and spring, and were numbered from 1 to 21. autumn to determine their phenotypic diversity. Cluster analysis with average distance There were two that flowered heavily in was performed for the main traits of leaf and flower using data processing system autumn of 2001, so they were added as software. The morphological investigation indicated that the number of flowering nos. 26 and 27. The dates of investigation in accessions and the flower number of L. chinense var. rubrum in spring were more than spring, summer, and autumn were from 1 to those in autumn. Only one accession (no. 13) yielded several flowers in summer. Some 10 Apr., from 3 to 9 Aug., and from 2 to 6 accessions had the same or similar color of leaf or flower in spring and autumn. Two Nov. respectively. Young leaf color, mature accessions had the same flower color in spring and autumn, whereas others showed leaf color, underside color of young leaf, leaf slightly different colors in spring and autumn. The 23 entries were grouped into four shape, leaf length, leaf width, leaf length-to- clusters in spring and five clusters in autumn based on multivariate analysis of nine width ratio, flower number, flower color, classification variables. Each cluster had some specific characteristics of its own. petal length, petal width, petal length-to- Generally, the cluster formed first because of the similarity in leaf color. The accessions width ratio, floret number, and petal number with similar flower color formed subclusters within a cluster. These accessions are an were measured in every investigation. Also, important resource for the establishment of a core collection of L. chinense var. rubrum in spring shoot length and offshoot angle were the world. Several accessions with good qualities were selected and should be further measured in the summer investigation, and tested for horticultural merit. summer shoot length was measured in the autumn investigation. Ten observations were collected for each variable from each plant and means were calculated. The representa- Loropetalum chinense var. rubrum is an City, Japan, by the former U.S. National tive branches were cut from different parts of evergreen, purple-leafed, medium-size shrub Arboretum Director, John L. Creech, in the plant and carried to the laboratory to from Hunan Province, China (Hou et al., September 1989 at the request of Sylvester determine leaf and flower color using the 2003). It is a member of the witch hazel G. March (Gawel et al., 1996). The plants Royal Horticultural Society Color Chart family (Hamamelidaceae) and is defined as a rapidly entered the commercial market. In an under northward scattered light (Royal Hor- variety of L. chinense (Creech, 1960). The attempt to give some order to the confusion of ticultural Society and Flower Council of gracefully layered branches are laden during multiple accessions and clones, the U.S. Holland, 1986). spring and sporadically throughout the year National Arboretum named the purple- The representative traits were subjected to with pendulous, bright-pink, spiderlike flowers. foliage form ‘Burgundy’, and the green- variance and cluster analyses with data pro- It is fast growing and tolerant to diseases foliage form ‘Blush’. More than 15 different cessing system software (Tang and Feng, and insects. Loropetalum chinense var. introductions of L. chinense var. rubrum have 2002). Multivariate analysis has been used in rubrum is versatile and can be grown in mild been established in North America since the many plant species to group accessions and coastal climates in light shade to full sun, and late 1980s (Hou et al., 2004). Selection and cultivars into clusters (Berdahl et al., 1999; can tolerate the more extreme winter and registration of cultivars has been carried out Naghavi et al., 2005; Steiner et al., 1998). summer temperatures associated with other since then. The main traits of leaf and flower—young regions. As an important ornamental plant, In recent years, with rapid introduction leaf color, underside color of young leaf, L. chinense var. rubrum is very popular in and commercialization of L. chinense var. mature leaf color, leaf length-to-width ratio, Chinese cities and has been introduced into rubrum, much confusion concerning the flower number, flower color, petal length-to- many foreign countries like America and identity of the cultivars has arisen. Although width ratio, floret number, and petal number— Japan. For example, two selections of the many investigations have been carried out by were used for classifying accessions into hot-pink/purple-flowered L. chinense var. Chinese researchers (Hou et al., 2003; Huang clusters in the current study. All qualitative rubrum were introduced into the United et al., 2004; Liu and Zhang, 2001; Song et al., data were transformed into numbers, and States from Nihonkaki Nursery, Kawaguchi 1982; Tang and Zhou, 2001; Tang et al., cluster analysis was performed using the 2003), the study of L. chinense var. rubrum in following algorithm: 1) identify the mini- China is still in its primary stage, and acces- mum distance between any two taxa, 2) Received for publication 3 Sept. 2006. Accepted sions have not been adequately characterized combine these two taxa as a single pair, 3) for publication 30 Oct. 2006. for morphological traits. The extent of phe- recalculate the average distance between this 1To whom reprint requests should be addressed; notypic diversity among L. chinense var. pair and all other taxa to form a new matrix, e-mail [email protected]. rubrum accessions throughout the world is and 4) identify the closest pair in the new

HORTSCIENCE VOL. 42(2) APRIL 2007 399 matrix and so on, until the last two clusters are joined.

Results and Discussion Data from spring observations

Leaf color, size, and shape. New spring Floret no. Petal no. growth emerged crimson to light red. Most of the colors belonged to Greyed-Purple Group (183–187), with others in the Yellow-Green

Group (144–154), Greyed-Orange Group y (163–177), or Greyed-Red Group (178– 182). The leaf underside colors ranged from ratio red to gray-red to gray-purple. The mature leaf color, measured in spring, ranged from Petal length-to-width brown to gray-purple to green to yellow- green. Mature leaf length ranged from 2.06 to 5.52 cm, and leaf width ranged from 1.28 to Petal 3.44 cm. The leaf length-to-width ratio width (cm) ranged from 2.24 to 1.40 (Table 1). Flower number and color. All accessions data available. flowered in spring except nos. 16, 19, and 20.

All flower colors were in the Red-Purple Petal

Group (57–74) with variation in the intensity length (cm) (Table 1). Petal size and number, and floret number. color

Petal length ranged from 1.34 to 2.40 cm, and Flower

petal width ranged from 0.13 to 0.22 cm. The x petal length-to-width ratio ranged from 8.02 to 16.48. Most accessions had four petals per floret, with several accessions containing occasional five petals per floret. There were Flower no. generally five to eight florets per inflores- cence (Table 1). y Data from summer observations The young leaf color of accession nos. 10, ratio 11, and 12 was gray-purple (183A, 183A, and

183B respectively). The young leaf color of Leaf length-to-width accessions nos. 5, 6, 7, 8, 9, 13, 14, 15, 16, 17, and 18 was gray-orange (163–177). The young leaf color of accession no. 1 was different from others, which turned green (cm) soon after new growth and was yellow-green Leaf width (146A). The young leaves of accession no. 4 emerged earlier, some of which were bigger accessions in spring evaluated at Changsha, Hunan, China. than mature leaves. Its color was green (cm)

(137B). But its sprouts were still red, and Leaf length turned green quickly under high temperatures rubrum in summer. All mature leaves in the investi- var. gation in summer emerged in spring, and turned green. Their colors were yellow-green

(147A or 146A). Only one accession (no. 13) Leaf shape yielded several flowers in summer. There were four to five light-pink (62A) flowers at each branch tip. z Loropetalum chinense Data from fall observations Summer shoot length. Summer shoot of young leaf length ranged from 8.48 to 18.48 cm (average, Underside color 10.0–16.0 cm; Table 2). Generally, the length of summer shoots was greater than that of spring shoots. Leaf color, size, and shape. There were a color few gray-purple young leaves at the branch Mature leaf tips of every plant. Fewer young leaves emerged in autumn than in spring. The color

of new growth was not persistent, because all color

summer leaves had turned green by fall. Leaf Young leaf length ranged from 2.06 to 5.60 cm, and leaf 123 144B4 152B5 152B6 184B 146B7 183A 147A8 187A 147B9 183B 146A 187A 147A 145A 145A 183C 146A 145A 146A Elliptic 185D 147A Elliptic 182B 146A Elliptic 183C Elliptic 183D Elliptic 183B Elliptic 183D 3.44 Broadly 3.54 elliptic Elliptic 3.05 Ovate 3.24 2.94 2.0 3.46 1.56 5.52 1.91 2.08 1.76 3.74 2.38 1.72 bc 3.2 2.56 2.24 a 1.60 cdef 1.80 1.56 cdef 1.66 cde 1.46 ef 1.84 1.72 bc ++ ++ ++ 2.10 a ++ +++ 1.40 f 57D ++ 62A 68B + 60B 68B 1.83 1.65 +++ 63B 1.44 1.34 2.03 ++ 67B 0.17 1.69 61B 0.16 0.13 0.17 0.19 2.00 67C 2.30 0.18 11.08 cd 1.54 10.30 10.74 cd 0.14 cd 10.86 0.22 cd 8.02 e 9.38 0.19 de 6–8 6–8 13.94 7–10 b 6–7 10.52 cd 5–7 4 5–8 4 8.06 4 e 4 4–8 5–8 4 4 7–9 4 4 4 Different lowercase letters denote significant+, differences ++, at +++ the denote 0.05 ‘‘less,’’ ‘‘more,’’ or level ‘‘many’’ according respectively. to analysis of variance. ‘‘–’’ in the column of leaf color indicates the data cannot be measured because the leaf color is in red and green. ‘‘—’’ in the other columns indicates no Table 1. Origin and major attributes of 21 z y x No. width ranged from 1.25 to 2.65 cm. The 101112 187A13 187A14 187A15 187A 200A16 183C 147A17 187A 200A18 187A 147A19 183C 147A20 – 176A 147A21 – 187A 147A – 184B 147A – 178A 137A Ovate – 187A Elliptic – 187A Elliptic – 185D 137A Elliptic Elliptic – 185D Elliptic Elliptic 4.74 185C Elliptic 3.42 3.66 – Elliptic Elliptic 3.54 Elliptic 3.44 3.44 2.16 3.70 2.1 3.46 Elliptic 2.68 2.44 2.14 2.08 2.72 1.40 f 1.94 2.06 1.58 cdef 2.18 1.72 1.70 bcd 1.46 ef 1.38 2.66 1.88 1.62 cde 1.28 1.88 b 1.60 ++ cdef 1.54 cdef + 1.86 + 1.50 def 1.48 ef ++ 1.62 cde 60A + ++ 59C — ++ 1.46 67A ef 1.76 62A — 64D 2.40 60B ++ — 2.15 64C — 2.30 0.16 1.61 — 1.68 0.15 ++ 68A 1.89 — 0.14 — 0.20 0.14 11.30 2.13 — 0.16 c 68A 0.14 — 16.48 a 15.02 — ab 11.76 2.39 c 0.19 11.74 c — 10.54 cd 5–7 13.72 — b 5–7 0.21 5–7 4–7 11.32 4–5 c — 7–8 4–6 4–5 6–8 — 4 4–5 11.50 — c 4–5 4–5 5–7 4–5 — — 4–5 3–9 — — — — —

400 HORTSCIENCE VOL. 42(2) APRIL 2007 maximum leaf length-to-width ratio was 2.29 cm (Table 2). Flower number and color. Only a part of the accessions flowered in autumn. The flower number of all flowered accessions in autumn was less than that in spring. Flower colors were in the Red-Purple Group (Table Floret no. Petal no. 2).

z Petal size and number, and floret number. Petal length ranged from 0.87 to 2.06 cm, and petal width ranged from 0.14 to 0.23 cm. Most accessions had four to five petals per

ratio of petal floret. There were generally four to seven Length-to-width florets per inflorescence, with some acces- sions containing occasional 9 to 10 florets per inflorescence (Table 2). (cm) Petal width Cluster analysis According to the data obtained in spring,

data available. the entries were grouped into four clusters at

(cm) the threshold of 3.80 (Fig. 1). Cluster I

Petal length included four subclusters. Subcluster I-1 included accession nos. 1, 3, 18, and 21. Subcluster I-2 included accession nos. 5,

color 7, 13, 14, and 15. Subcluster I-3 included Flower accession nos. 4 and 17. Subcluster I-4 x included accession nos. 6, 10, 11, 26, and 27. Cluster II included accession nos. 2, 8, and 9. Cluster III only included accession no.

Flower no. 12. Cluster IV included accession nos. 16, 19, and 20.

z According to the data obtained in autumn, the entries were grouped into five clusters (Fig. 2). Cluster I included accession nos. 1, 3, 5, 8, 10, 11, 12, and 16. Cluster II included ratio of leaf accession nos. 6, 14, 18, and 9. Cluster III Length-to- width only included accession no. 4. Cluster IV included accession nos. 7, 17, 26, and 27. Cluster V included accession nos. 13 and 15.

(cm) Results from this analysis showed that the

Leaf width cluster grouped first as a result of the simi- larity in leaf color. The accessions with (cm) accessions in autumn evaluated at Changsha, Hunan, China. Leaf length y rubrum var. Under- side color of young leaf Loropetalum chinense color Mature leaf y color Young leaf z

Fig. 1. Dendrogram of 21 Loropetalum chinense length var. rubrum accessions in spring, with two Summer shoot additional accessions (nos. 26 and 27) in autumn obtained by analysis of nine morpho- 134 11.38 ab56 11.65 — ab7 10.98 ab 146C8 18.48 a9 17.55 a 187A 11.93 ab 146C 147A 15.53 – ab 187A 147A 183A 147B – 183C 147A 146A — 146B 146A — 146A — — — 183D 2.35 — 2.90 — — 1.80 3.21 5.05 3.77 1.85 3.07 1.30 2.74 d — 2.06 2.30 2.63 1.58 bc 2.20 1.74 1.55 2.20 1.45 bcd a bcd — + 1.39 + cd 1.60 bc +++ 63A — — 60B — 64C 0.87 — +++ — — 1.64 — 1.84 60C — 0.20 — — 0.15 — 2.06 0.20 — 4.40 f — 11.50 b — 10.20 0.20 bcd — — 4–6 3–6 6–7 10.21 bcd — — — 4 4–5 — 5–8 4 — — 4–5 — — — — — — ‘‘—’’ in the column of+, leaf ++, color +++ indicates denote the ‘‘less,’’ data ‘‘more,’’ or cannot ‘‘many’’ respectively. be measured because the leaf color is in red and green. ‘‘—’’ in the other columns indicates no Different lowercase letters denote significant difference at the 0.05 level according to analysis of variance. Table 2. Origin and major attributes of 19 z y x No. 101112 17.80 a13 15.38 ab14 11.95 ab15 15.58 ab16 183C 187A 8.4817 b 11.18 ab 187A18 15.63 ab 183A26 10.55 146A ab27 147A 12.63 178A ab 200A 147A 10.38 ab 187A 147B 11.25 ab 187A 183A 147B 147A — 187A 147A — 186A 187A 147A 146A — 147B 147A — — 183D 5.60 2.94 183D 3.67logical — 183D 2.70 185D 2.65 1.76 2.43 characterization 2.78 2.06 3.92 2.30 1.69 2.14 2.30 3.89 a 1.67 1.51 1.70 b 1.25 bcd 1.72 2.95 1.55 1.61variables. bc 1.66 2.30 1.64 1.66 2.29 bc bc 1.90 a — — 1.50 bcd — ++ 1.39 1.68 cd b 1.55 — bcd — — — + + + 62A — ++ — — — 60D + + 1.82 64C 60B 60B 1.79 — 2.00 — 64C 1.77 61A — 0.14 — 1.93 1.65 1.96 0.21 0.15 0.18 13.69 — a — — 0.19 — 0.23 0.18 8.67 13.91 de a 9.81 cd 5–9 10.22 bcd — — 11.16 — bc 7.11 — e 5–9 6–8 4–6 4–6 6–7 — — 4–5 8–10 — — 4–5 4–5 6–9 4 4 — 4–5

HORTSCIENCE VOL. 42(2) APRIL 2007 401 Several accessions with good qualities Hou, B.X., F. Liu, W.P. Li, X.M. Wang, G.F. Yu, Results from this study have implications and Q.A. Song. 2003. List of Loropetalum for future plant collection and application chinense var. rubrum classification system. efforts with L. chinense var. rubrum. After For. Res. 16:430–433. comparing the morphological traits of all Huang, Q.C., L.Y. Jiang, X.H. Peng, and G.S. Hu. accessions in different seasons, four acces- 2004. On resistance of Loropetalum chinense sions with good qualities were selected and var. rubrum. J. Hunan Agri. Univ. (Nat. Sci.) should be further tested in the landscape. 30:37–39. Liu, D.L. and Q. Zhang. 2001. The germplasm Accession no. 8 had abundant, attractive resources of Loropetalum chinense var. rubrum flowers in spring and autumn. Its flowers were and their further exploitation. Wild Plant purplish red. Flowers of accession no.11 were Resources China 20:17–18. a vivid red. Leaf color of accession no. 12 was Naghavi, M.R. and M.R. Jahansouz. 2005. Varia- dark purple, and the leaf underside color was tion in the agronomic and morphological traits red. Its mature leaf color was long-lasting of Iranian chickpea accessions. J. Integrative brownish black. Accession no. 13 yielded only Plant Biol. 47:375–379. several flowers in the summer in this study. Royal Horticultural Society and Flower Council The color and shape of its leaves were of Holland. 1986. The RHS colour chart. Royal attractive. Its mature leaves were dark glossy Horticultural Society, London. green and ovate. The classification and Song, P.L., C.Y. Peng, and T.X. Zhang. 1982. Tissue nomenclature of these plants are currently in culture and organogenesis of Loropetalum progress. chinense var. rubrum.PlantPhys.Com.4:33. Steiner, J.J., E. Piccioni, M. Falcinelli, and A. Fig. 2. Dendrogram of 19 Loropetalum chinense Literature Cited Liston. 1998. Germplasm diversity among cul- var. rubrum accessions in autumn (accession tivars and the NPGS crimson clover collection. nos. 2, 19, 20, and 21 were dead) obtained by Berdahl, J.D., H.F. Mayland, K.H. Asay, and P.G. Crop Sci. 38:263–271. analysis of nine morphological characterization Jefferson. 1999. Variation in agronomic and Tang, Q.R., Y.Y. Chen, and P.H. Zhou. 2003. variables. morphological traits among Russian wildrye Study on the stability of anthocyanin and pH accessions. Crop Sci. 39:1890–1895. changes in cell sap of leaves in Loropetalum Creech, J.L. 1960. On the distribution of Lorope- chinense var. rubrum. Hunan For. Sci. Technol. similar flower color grouped into subclusters talum chinense. Amer. Hort. Mag. 39:236. 30:24–25. within a cluster; however, there were some Gawel, N.J., G.R. Johnson, and R. Sauve. 1996. Identification of genetic diversity among Tang, Q.Y. and M.G. Feng. 2002. DPS data exceptions. For example, the spring leaf processing system for practical statistics. Sci- colors of accession nos. 2 and 3, both in the Loropetalum chinense var. rubrum introductions. J. Environ. Hort. 14:38–41. ence Press, Beijing. Yellow-Green Group, and their flower colors Hou, B.X., Z.H. Cheng, F. Liu, G.F. Yu, A.Q. Yi, Tang, Q.R. and P.H. Zhou. 2001. Comparisons of are similar, belonging to the Red-Purple and Q.A. Song. 2004. Breeding selection and morphological characteristics and pigment Group. However, in the cluster dendrogram, popularization of new variety of Loropetalum contents among variants of Loropetalum the two accessions were in fact far apart from chinense var. rubrum. J. Chin. Urban For. 2: chinense var. rubrum. J. Hunan Agri. Univ. each other. 33–35. (Nat. Sci.) 27:362–366.

402 HORTSCIENCE VOL. 42(2) APRIL 2007