Botanical Studies (2009) 50: taxonomy Reevaluating the taxonomic status of Ceriops australis (Rhizophoraceae) based on morphological and molec

Botanical Studies (2009) 50: 89-100. taxonomy Reevaluating the taxonomic status of Ceriops australis (Rhizophoraceae) based on morphological and molecular evidence Chiou-Rong SHEUE 1, *, Yuen-Po YANG 2, Ho-Yi LIU 2, Fu-Shan CHOU 3, Hsiu-Chin CHANG 1, Peter SAENGER 4, Christopher P. MANGION 5, Glenn WIGHTMAN 5, Jean W. H. YONG 6, and Chi-Chu TSAI 7, ** 1 Department of Biological Resources, National Chiayi University, 300 Syuefu Rd., Chiayi 600, Taiwan 2 Department of Biological Sciences, National Sun Yat-sen University, 70 Lien-hai Rd., Kaohsiung 804, Taiwan 3 Liouguei Research Center, Taiwan Forestry Research Institute, 198 Chunghsing Village, Liouguei 844, Kaoshiung County, Taiwan 4 School of Environmental Science and Management, Southern Cross University, Lismore NSW 2480, Australia 5 Department of Natural Resources, Environment and the Arts, PO Box 496, Palmerston NT 0831, Australia 6 Natural Sciences, National Institute of Education, 1 Nanyang Walk, Nanyang Technological University, 637616 Singapore 7 Kaohsiung District Agricultural Improvement Station, 2-6 Dehe Rd., Changihih Township, Pingtung County 908, Taiwan (Received December 24, 2007; Accepted August 19, 2008) ABSTRACT. Ceriops australis (White) Ballment, Smith & Stoddart, a member of the mangrove family Rhizophoraceae, was originally recognized as C. tagal var. australis White but was raised to species rank based solely on isozyme features and the only distinctive morphological feature of the hypocotyl. Therefore, it was considered a sibling species of C. tagal (Perr.) C. B. Rob. The goal of this study was to test the previous assessment that C. australis and C. tagal differ consistently only in hypocotyl morphology, in order to reevaluate the taxonomic status and to establish its geographic range. Principal components analysis was employed to analyze 29 morphological characters of herbarium specimens from Australia, Madagascar, and Sumatra tentatively identified as C. australis and C. tagal, and two well differentiated distinct taxa were recognized. In addition, both of the detailed morphological features based on fresh and herbarium materials and the intron sequences of trnl gene from plastid DNA support this conclusion. This finding disagrees with previous assessment and supports the current taxonomic status of C. australis. Here, a key to these two species is provided, and a revised distribution range of C. australis is established. This is the initial report of C. australis occurrence in a part of Indonesia, in addition to areas of Australia and Papua New Guinea. Keywords: Australia; Ceriops tagal; Ceriops; Distribution; Indonesia; Mangroves; Papua New Guinea; Plastid DNA; Principal components analysis. INTRODUCTION Ceriops Arn. is one of the mangrove genera in the family Rhizophoraceae, with a widespread geographical range from eastern Africa throughout tropical Asia, and northern Australia to Melanesia, and through Micronesia north to southern China (Tomlinson, 1986). It typically grows in the inner mangroves, often forming pure stands on better drained sites or becoming stunted in exposed and highly saline sites, within the reach of occasional tides (Hou, 1958). *Corresponding author: E-mail: crsheue@mail.ncyu.edu.tw; Tel: +886-05-2717827; Fax: +886-05-2760164. **Co-Corresponding author. E-mail: tsaicc9017@yahoo. com.tw; Tel: +886-08-7746735; Fax: +886-08-7229466. The last revision of the genus Ceriops was done by Hou (1958), with two species recognized: C. tagal (Perr.) C. B. Rob. and C. decandra (Griff.) Ding Hou. Some 20 additional names, including several infraspecific names were synonymized for them, but a variety name C. tagal (Perr.) C. B. Rob. var. australis C. T. White named by White (1926) was not listed. White (1926) noticed a form of C. tagal in which the propagules had smooth, terete hypocotyls rather than the angled or ribbed hypocotyls typical of C. tagal from Australia and Papua New Guinea. He initially intended to describe this form as a new species distinct from C. tagal, based on the less distinctly veined, and more inclined to recurved leaves (White, 1926). After examining additional specimens, however, he found those differences between the new form and C. tagal were not constant

90 Botanical Studies, Vol. 50, 2009 except for the hypocotyl morphology. Thus, White described the form as a variety, C. tagal var. australis. Based on the analysis of starch gel electrophoresis of isozymes from C. tagal var. tagal, C. tagal var. australis and C. decandra in northern Australia, Ballment et al. (1988) found a uniform genetic structure within each taxon and a high level of genetic divergence among taxa. For each taxon having a distinct isozyme profile and the evidence of reproductive isolation, the authors proposed three distinctive species and hence raised White s variety to specific rank as C. australis, despite the fact that the extent of divergence in morphological characters other than propagule morphology remained unclear (Ballment et al., 1988). Ceriops australis was then claimed as a sibling species of C. tagal (Ballment et al., 1988). Due to the confusion regarding diagnostic characters, it is still unknown how far north of Australia C. australis extends (Duke, 2006). Misidentification of these two morphologically similar taxa has been quite common in herbaria (Sheue, personal observation). Making a field identification of C. australis is very difficult at any time other than the fruiting stage with a hypocotyl. Thus, a detailed study of these two morphologically similar species is vital. The goals of this study, therefore, are to detect the differences between C. australis and C. tagal, based on a broad and detailed morphological assessment aided by molecular data, and to establish the geographic distribution range of C. australis. Here we apply principal components analysis to morphometric data obtained from herbarium specimens and investigate the DNA features of the trnl intron of cpdna. We also use detailed characters from fresh and herbarium materials to reevaluate the taxonomic status of these species. The results will be useful for field work identification and herbarium examination, conservation, and for clarification of the relationship between these species. MATERIALS AND METHODS Herbarium specimens and morphometric analysis As it has been reported that only viviparous seedlings could be used to differentiate Ceriops australis from C. tagal, the former having terete (smooth) hypocotyls and the latter having ridged hypocotyls (White, 1926; Ballment et al., 1988), specimens of branches with both vegetative and reproductive features, including viviparous seedlings, were examined for this morphometric study. Fifteen specimens from the Northern Territory (OTUs 1-8) and Queensland (OTUs 9-15) of Australia, tentatively identified as C. australis, and 15 herbarium specimens tentatively identified as C. tagal representing populations of Northern Territory from Australia (OTUs 16-22), Madagascar (OTUs 23-26), and Sumatra (OTUs 27-30) were used in the morphometric study (Appendix). Principal components analysis (PCA) was conducted to analyze 29 morphological characters (25 quantitative and 4 binary characters, Table 1), and the PC-ORD package (McCune and Mefford, 1999) was used to analyze character variable matrices. The possible differentiated characters identified in this analysis will be used to detect diagnostic features in the following analysis. Fresh plant materials for morphological characterization Fresh plant materials of C. australis and C. tagal were sampled from Cape York, Cairns, and Cardwell of northeastern Queensland and from the Darwin area of the Northern Territory of Australia during 2005 to 2007 for morphological characters investigation and molecular study. Three branches from each of five individuals in a population were collected. Characters of fresh materials were investigated by a Leica MZ75 stereoscope and photographed with an Olympus C7070 digital camera. Voucher specimens were deposited at the Herbarium of the Department of Biological Resources, National Chiayi University (CHIA). H e r b a r i u m s p e c i m e n s fo r d e t e r m i n i n g distribution range The loaned specimens (Appendix) from herbaria BM, DNA, GH, K, L and MO were identified as C. australis or C. tagal through the following two steps. In the first step, specimens with viviparous seedlings attached on the shoots were determined and used for getting diagnostic features for identification. In the second step, specimens lacking hypocotyls were identified according to the diagnostic features obtained from the previous first step. Each specimen was carefully examined at least thrice. In addition, a few specimens of C. australis examined from Herbaria BO and CAL were incorporated in the results. Molecular evidence Materials. Populations of C. australis and C. tagal were mainly sampled at five sites on the northeast Queensland coast and north Northern Territory coast in Australia during the period from 2003 to 2007. In addition, C. tagal collected from Singapore and India and C. decandra collected from India were analyzed together (Table 2). Three to five leaves were taken from each individual and stored with silica gel in zip-lock plastic bags until DNA isolation. Voucher specimens were deposited at the Herbarium of National Chiayi University (CHIA). DNA extraction. Using the cetyltrimethylammonium bromide (CTAB) method described previously (Doyle and Doyle, 1987), total DNA was extracted from fresh etiolated leaves. Ethanol-precipitated DNA was dissolved in TE (Tris-EDTA) buffer and stored at -20ºC. Qiagen (Valencia, CA, USA) columns were used to clean the DNA samples, which were difficult to amplify by PCR. The approximate DNA yields were then determined using a spectrophotometer (model U-2001, Hitachi).

SHEUE et al. Reevaluating the taxonomic status of Ceriops australis 91 PCR amplification and electrophoresis. The protocols for PCR were as follows. A 50-µl mixture contained 40 mm Tricine-KOH (ph 8.7), 15 mm KOAc, 3.5 mm Mg (OAc) 2, 3.75 µg/ml BSA, 0.005% Tween 20, 0.005% Nonidet-P40, four dntps (0.2 mm each), primers (0.5 µm each), 2.5 units of Advantage 2 DNA polymerase (Clontech), 10 ng genomic DNA, plus a 50-µl of mineral oil. Amplification reactions were carried out in a dry-block with two-step thermal cycles (Biometra). The universal primers for amplifying the trnl intron of chloroplast DNA were the same as described by Taberlet et al. (1991). The first step of PCR reaction conditions for the trnl intron were: incubation at 94ºC for 3 min, 10 cycles of denaturation at 94ºC for 30 s, annealing at 68ºC for 10 s, and extension at 72ºC for 45 s. The second step was carried out with 30 cycles of denaturation at 94ºC for 30 s, annealing at 66ºC for 10 s, extension at 72ºC for 45 s, and a final extension for 5 min at 72ºC. The PCR products were analyzed by agarose gel electrophoresis (1.0%, w/v in TBE), stained with 0.5 µg/ml ethidium bromide, and photographed under UV light exposure. DNA recovery and sequencing. The PCR products in this study were recovered using glassmilk (BIO 101, California) and directly sequenced following the method of dideoxy chain-termination using an ABI377 automated sequencer with the Ready Reaction Kit (PE Biosystems, California) of the BigDye Terminator Cycle Sequencing. Primers for sequencing were the same as those used for PCR. Each sample was sequenced two or three times to ensure the accuracy of the sequences. The reactions were performed following the recommendation of the manufacturers. These reactions were performed based on the recommendations of the manufacturer. Data analyses. DNA sequence alignment was conducted using the program Clustal W multiple alignment in BioEdit (Hall, 1999). Genetic relationships were then determined using the program MEGA version 2.1 (Kumar et al., 2001). The genetic distance matrix was calculated by the two-parameter method of Kimura (1980) and then used to construct the phylogenetic trees using the Neighborjoining (NJ) method (Saitou and Nei, 1987). Maximum parsimony (MP) analyses (Fitch, 1971) were done using code modified from the Close-Neighbor-Interchange (CNI) algorithm (Rzhetsky and Nei, 1992) in MEGA version 2.1 (Kumar et al., 2001). Bootstrapping (1000 replicates) was carried out to estimate the support for both NJ and MP topologies (Felsenstein, 1985; Hillis and Bull, 1993). The strict consensus parsimonious tree was then constructed using the program MEGA version 2.1 (Kumar et al., 2001). RESULTS Morphometric analysis We performed a principal coordinate analysis, and the result is shown in Figure 1. These two species are well separated, and 51.8% of the variation can be explained by the first two principal coordinates. Only the first ordination axis was considered (Table 1). For components, the ten highest eigenvector values belonged to reproductive characters, except leaf length; accordingly, these include surface of hypocotyl (HS), length of fruit (FL), length of hypocotyl (HL), length of calyx lobe/ width of calyx lobe (CLL/CLW), width of calyx lobe (CLW), thickness of the middle part of calyx lobe (CLT), width of hypocotyl (HW), width of fruit (FW), length of style (STL) and leaf length (LL). The highest three eigenvector values of vegetative characters were leaf length (LL), stipule length at the naturally expanded stage (SL) and leaf width (LW). These characteristic variables represented the relative contribution of the first component in explaining the total variation within the dataset. Each two selected diagnostic characters of organs belonging to leaf (LL, SL), flower (CLW, STL), and fruit (HS, FL) are suggested for use in identification and are shown in Figure 1. Morphological features Ceriops australis and C. tagal have very similar morphological characteristics, including a grey-white trunk and stem with buttressed base, elliptic-obovate leaves with reflexed margins, and small flowers with white petals (Figure 2). The most distinctive basis upon which to differentiate the two species is the viviparous seedling (hypocotyl), as reported before. Based on field experience, C. tagal usually has dark green and elliptic to obovate leaves and longer stipules (the expanded stipules usually longer than 1.5 cm) than C. australis, which has more yellow-green and obovate leaves and shorter stipules (the expanded stipules usually less than 1.2 cm) (Figure 3; Table 1). Figure 1. PCA ordination diagram of OTUs and prominent variables. OUTs 1-15: Ceriops australis; OUTs 16-30: C. tagal. Each two selected diagnostic characters belonged to organs of leaf (LL, SL), flower (CLW, STL) and fruit (FL, HS) for differentiating the two species are suggested. Abbreviations: CLW: Width of the base of calyx lobe; FL: length of fruit; HS: surface of hypocotyl; LL: leaf length; SL: stipule length at the naturally expanded stage; STL: length of style.

92 Botanical Studies, Vol. 50, 2009 Table 1. A List of the selected 29 morphological characters, examined from each 15 herbarium specimens of Ceriops australis and C. tagal from Madagascar, Sumatra and Australia for principal components analysis. Characters 1-11 are leaf characters, 12-22 are flower characters, 23-29 are fruit and hypocotyl characters. No. Character (unit) or (character state) Character abbreviation C. australis Mean±std or character state C. tagal Mean±std or character state Eigenvector value of axis 1 1 Leaf blade length (mm) LL 55.7±5.3 68.5±8.2-0.2169 2 Leaf width (mm) LW 27.5±3.4 32.3±5.1-0.1792 3 The length between the maximum width of leaf to leaf apex (mm) LWmax 22.9±3.6 27.7±5.1-0.1680 4 Petiole length (mm) PL 18.4±4.4 19.1±4.5-0.0326 5 Leaf length/ leaf width LL/LW 2.05±0.20 2.13±0.23-0.0278 6 The ratio of leaf length to the length between the maximum width of leaf to leaf apex LL/LWmax 2.5±0.3 2.5±0.3-0.0058 7 Leaf length/ petiole length LL/PL 3.2±0.8 3.7±0.6-0.1105 8 Leaf apex LAE 0 = 6 0 = 12-0.1260 with (0) or without emarginated (1) 1= 9 1 = 3 9 The degree of acute of leaf apex LAA 61.7±5.0 63.1±7.8-0.0471 10 Number of lateral vein V 7.3±0.8 8.0±0.7-0.1334 11 Stipule length of naturally expanded (mm) SL 11.8±1.4 16.1±2.7-0.2060 12 Length of calyx lobe (mm) CLL 4.3±0.3 4.1±0.4 0.0784 13 Thickness of the middle part of calyx lobe (mm) CLT 0.24±0.04 0.35±0.05-0.2402 14 Width of the base of calyx lobe (mm) CLW 1.4±0.2 2.0±0.1-0.2530 15 Length of calyx lobe/width of calyx lobe CLL/CLW 3.0±0.4 2.1±0.2 0.2538 16 Length of petal (not included bristle) (mm) PL 2.8±0.3 3.04±0.19-0.1519 17 Number of bristles of each petal apex B 3.5±0.7 3.0±0.0 0.1228 18 Length of bristle (mm) BL 0.77±0.15 0.61±0.11 0.1554 19 Length of enlarged part of bristle (mm) BHL 0.23±0.06 0.31±0.05-0.1902 20 Width of enlarged part of bristle (mm) BHW 0.08±0.01 0.13±0.04-0.2180 21 Trichome on abaxial surface of petal PT 0 = 15 0 = 6 0.1991 with (0) or without (1) 1 = 9 22 Length of style (mm) STL 3.07±0.39 2.08±0.49 0.2259 23 Length of fruit (mm) FL 11.6±1.39 19.3±1.62-0.2645 24 Width of fruit (mm) FW 6.2±0.7 9.2±1.28-0.2258 25 Length of fruit/ width of fruit FL/FW 1.9±0.3 2.1±0.2-0.1377 26 Persistent calyx lobe CLR 0 = 7 0 = 15-0.1651 reflex (0) or patent (1) 1 = 8 27 Length of hypocotyl (mm) HL 106.6±23.0 241.3±38.6-0.2596 28 Width of hypocotyl (mm) HW 3.2±0.6 5.9±1.11-0.2393 29 Surface of hypocotyl HS -0.2818 smooth (0) or ridged (1) 0 = 15 1 = 15

SHEUE et al. Reevaluating the taxonomic status of Ceriops australis 93 It was evident that features like calyx lobe, petal morphology, style (Figure 4), fruit length, and hypocotyl surface (Figure 3C) could aid the differentiation of these sibling species. Ceriops australis has longer flowers (Figure 4A), narrower and longer calyx lobes (Figure 4C; Table 1), longer petals (Figure 4D-E) and longer styles (Figure 4F) than C. tagal (Figure 4B, C, E-F). Three to five more slender clavate appendages were commonly found on the petal apex of C. australis, but only three such appendages (more short) were observed on C. tagal (Figure 4D-E; Table 1). DNA evidence Sequence alignment and characteristics. PCR products from each sample studied were directly sequenced. The accession numbers of those plastid DNA sequences Table 2. A list of molecular study for 14 accessions of the Ceriops australis and 15 accessions of C. tagal, as well as three outgroup accessions of C. decandra, and their different geographical distributions. Abb. Taxon Collection location Accessions No. Rh-13 C. australis Moreton Bay, QLD, Australia (AU) EF118948 Rh-70 C. australis Cairns, QLD, Australia (AU) EF118971 Rh-71 C. australis Darwin, NT, Australia (AU) EF118949 Rh-72 C. australis Darwin, NT, Australia (AU) EF118950 Rh-73 C. australis Darwin, NT, Australia (AU) EF118951 Rh-113 C. australis Cardwell, QLD, Australia (AU) EF673713 Rh-114 C. australis Cardwell, QLD, Australia (AU) EF673714 Rh-115 C. australis Cardwell, QLD, Australia (AU) EF673715 Rh-120 C. australis Cardwell, QLD, Australia (AU) EF673717 Rh-132 C. australis Cardwell, QLD, Australia (AU) EF673721 Rh-133 C. australis Cardwell, QLD, Australia (AU) EF673722 Rh-134 C. australis Cardwell, QLD, Australia (AU) EF673723 Rh-135 C. australis Cardwell, QLD, Australia (AU) EF673724 Rh-136 C. australis Cardwell, QLD, Australia (AU) EF673725 Rh-31 C. tagal West Sundarbans, India (IN) EF118987 Rh-32 C. tagal West Sundarbans, India (IN) EF118964 Rh-33 C. tagal West Sundarbans, India (IN) EF118965 Rh-65 C. tagal Cairns, QLD, Australia (AU) EF118966 Rh-85 C. tagal Cairns, QLD, Australia (AU) EF118986 Rh-86 C. tagal Cairns, QLD, Australia (AU) EF118988 Rh-66 C. tagal Darwin, NT, Australia (AU) EF118967 Rh-67 C. tagal Darwin, NT, Australia (AU) EF118968 Rh-68 C. tagal Cape York, QLD, Australia (AU) EF118969 Rh-69 C. tagal Cape York, QLD, Australia (AU) EF118970 Rh-82 C. tagal Pulau Ubin, Singapore (SING) EF118972 Rh-116 C. tagal Cardwell, QLD, Australia (AU) EF673716 Rh-121 C. tagal Cardwell, QLD, Australia (AU) EF673718 Rh-127 C. tagal Cardwell, QLD, Australia (AU) EF673719 Rh-131 C. tagal Cardwell, QLD, Australia (AU) EF673720 Rh-26 C. decandra Pichavarum, India (IN) EF118952 Rh-28 C. decandra West Sundarbans, India (IN) EF118953 Rh-29 C. decandra West Sundarbans, India (IN) EF118954 Abbreviations: AU: Australia, IN: India; QLD: Queensland; NT: Northern Territory; SING: Singapore.

94 Botanical Studies, Vol. 50, 2009 from the 14 accessions of C. australis and 15 accessions of C. tagal plus three outgroup accessions are shown in Table 2. Those sequences were aligned and resulted in 606 characters, from which 13 were variable sites. The sequence alignment was submitted to TreeBase (Submission ID: SN4033). Each variable site was a potentially informative parsimony site. Neither C. australis nor C. tagal showed any sequence variation at the species level. The genetic distance between C. australis and C. tagal was 0.003 using the 2-parameter method of Kimura (1980). Two stable transitions were found within this DNA region between C. tagal and C. australis (data not shown). Phylogeny reconstruction. The phylogenetic tree for the intron of trnl used characters that were equally weighted. Based on the MP method, the analysis yielded 270 equally parsimonious trees with a length of 13 steps, a consistency index (CI) of 1.0, and a retention index (RI) of 1.0. The strict consensus tree is shown in Figure 5. More than 50% of the bootstrap values are shown below/above the supported branches for MP tree. The NJ tree and the MP strict consensus tree constructed from plastid DNA data were highly congruent (Figure 5, MP tree presented only). Based on the phylogenetic tree, accessions of C. australis formed a clade supported by a 69% bootstrap value, and accessions of C. tagal formed a clade supported by a 72% bootstrap value. Molecular data also supported the distinctness of C. australis and C. tagal, even in the sympatric populations of Queensland and Darwin, Australia. Distribution range of C. australis The whole distributional range of C. australis includes eastern (Moreton Bay) and northern Queensland (Cape York, Nassau River), the coast of the Northern Territory, through northern and northwestern Western Australia (to the Ashburton River), the southern part of Papua New Guinea (Port Moresby, Daru Island), through Timor, Flores, Sumbawa, Java and Pulau Bilinton, close to Sumatra, Indonesia (Figure 6). This is the first report that C. australis occurs in parts of Indonesia. According to the examination of herbarium specimens, C. australis has a much wider distribution range than C. tagal in Australia, although C. tagal is widely distributed from East Africa through India and Asia to New Caledonia. However, C. tagal is only found in northeastern and northern Queensland, through Cape York, Arnhem Land, and Melville Island in Australia. There are only about five colonies with a few individuals of C. tagal growing closely with C. australis found in the Darwin area (Sandy Creek) in the Northern Territory according to our field survey. DISCUSSION In this study, both of the morphological features revealed by PCA and molecular evidence demonstrate that C. australis should be recognized as distinct from C. tagal, rather than as a sibling species only slightly different Figure 2. Habitats of Ceriops australis (A-C) and C. tagal (D). A, Close-up of C. australis with flowers and viviparous seedlings with smooth surface; B, The grey-white bark with buttress base of C. australis at Cairns, Queensland; C, Flowers of C. australis. Note the evident long style; D, C. tagal with viviparous seedlings with ridges.

SHEUE et al. Reevaluating the taxonomic status of Ceriops australis 95 Figure 3. Characters of leaves, fruits and hypocotyls of Ceriops australis and C. tagal. A, Leaves of C. australis tend to be more obovate in shape and stipules at naturally expanded stage are usually less than 1.2 cm; B, Leaves of C. tagal are more oblong in shape, and stipules at naturally expanded stage are usually longer than 1.4 cm; C, The fruit is smaller and the hypocotyl is shorter and ridgefree of C. australis (A & a); while the fruit is larger and the hypocotyl is longer and ridged of C. tagal (B & b). Abbreviations: NT: Northern Territory, QLD: Queensland, WA: Western Australia. in hypocotyl feature and genetic structure as proposed by White (1926) and Ballment et al. (1988). However, for a practical application of the concept of species, it is necessary to provide some diagnostic characters. According to the morphometric results obtained in this study, the most distinctive characters differentiating the two species are reproductive features. The diagnostic characters of style length (STL) and width of calyx lobe (CLW) are recommended for the plants with flowers; those of hypocotyl surface (HS) and fruit length (FL) are recommended for plants with fruits. Nevertheless, we suggest that the features of leaf length (LL) and stipule length of the naturally expanded stage (SL) could also aid the identification of plants in the field without flower or fruit. Based on the results of morphological features and PCA, a key to differentiating the populations of C. australis and C. tagal is here provided: Key to C. australis and C. tagal 1a. Leaf blade usually shorter than 6.5 cm in length; stipule less than 1.2 cm long at the naturally expanded stage; base of calyx lobe 12-15 mm in width; style 2.7-4 mm in length; fruit 9-14 mm long; hypocotyl terete (without longitudinal ridges), 5-12 cm in length...c. australis 1b. Leaf blade usually longer than 6.5 cm in length; stipule longer than 1.4 cm long at naturally expanded stage; base of calyx lobe 18-25 mm in width; style 1.5 mm in length (-3.5 of populations from Darwin area, Northern Territory of Australia); fruit 18-25 mm long; hypocotyl angular (with longitudinal ridges), 15-35 cm in length...c. tagal According to Wightman (2006), populations of C. tagal in Northern Territory generally have elliptic leaves and relative shorter petiole length (usually less than 1/4 of the blade length) than the mostly obovate leaves and relative longer petiole length (generally reaching 1/3 or more of the blade length) of C. australis. Based on the observation of this study, we agreed with Wightman s statement and found that the populations of C. australis in Western Australia have the most typical obovate and smaller leaves than other populations. It is likely that C. australis is the only one species of this genus occurring in Western Australia, which results in much less opportunity to have gene flow with other species of Ceriops, if compared to the other sympatric populations of Ceriops. In addition, we noted some of the detailed differences between these two taxa, including the number of colleters inside the adaxial base of the stipule (Sheue, 2003), the thickness of the middle of the calyx lobe, and the number and shape of the clavate appendages on the petal apex.

96 Botanical Studies, Vol. 50, 2009 However, to observe these delicate features a hand lens (10X) or a stereoscope may be needed. It is notable that the observation of herbarium specimens in this study revealed no evident morphological variations of C. tagal between the populations of Madagascar and Sumatra and those from northern Australia. The low levels of morphological variation across a big geographic range of C. tagal noted by this study were consistent with the inter simple sequence repeat (ISSR) markers of C. tagal studied in Asia (Ge and Sun, 2001) and the trnl intron sequences of plastid DNA from different locations of C. tagal in this study. Correct information for identification is essential to getting an accurate biogeographic description. Since the confusion in diagnostic characters applies to these two taxa in Australia, obtaining accurate information on their distribution ranges is not easy. This is perhaps why Australia s Virtual Herbarium (AVH) could not supply the correct information for these two taxa (http://www. anbg.gov.au/cgi-bin/avhxml.cgi). In terms of AVH Mapper, C. australis only occurs in the Nothern Territory and northeastern Western Australia while C. tagal has a much wider distribution range from Queensland, through the Northern Territory to Western Australia. Based on a detailed examination of herbarium specimens in this study, we have reconstructed the geographic range of C. australis and C. tagal in Australia. The dominant species of Ceriops in Australia is C. australis. It ranges from Western Australia and the Northern Territory to Queensland. This result is consistent with the report of Duke (2006). In Papua New Guinea, C. australis has only been observed from Port Moresby and Idlers Bay. This was consistent with the observations of White (1926), McCusker (1984), and Wells (1983). It is quite interesting that C. australis occurs in Timor, Flores, Sumbawa, Java and Pulau Bilinton of Indonesia. The first collector of C. australis from Indonesia may have been Teijsmann in 1875 (specimens found at BO herbarium, Sheue, personal observation). Due to the limited herbarium specimens available from Indonesia, an extensive field survey for C. australis from the nearby islands of Indonesia would be useful. This information would be valuable for mangrove conservation and the study of phytogeography and dispersal ecology. Figure 4. Flower morphology of of Ceriops australis and C. tagal. A, Lateral view of a flower of C. australis; B, Lateral view of a flower of C. tagal; C, Calyx lobes with abaxial and adaxial sides of C. australis (left) and C. tagal (right); D, Petals of C. australis, with 3-5 more slender clavate appendages; E, Petal of C. tagal, usually with 3 short clavate appendages; F, Lateral view of detached flowers of C. australis (left) and C. tagal (right) showing calyx lobe, anther and style. Scale bars: A-B = 5 mm, C-F = 1 mm.

SHEUE et al. Reevaluating the taxonomic status of Ceriops australis 97 Figure 5. The strict consensus parsimonious tree of 14 accessions of Ceriops australis and 15 accessions of C. tagal plus three outgroup accessions of C. decandra derived from the trnl intron sequence. Bootstrap values > 50% are shown on each branch. Figure 6. The distribution range of Ceriops australis and the sympatric localities of C. tagal in Australia. The arrows indicate the new localities of C. australis in Indonesia first reported in this study.

98 Botanical Studies, Vol. 50, 2009 These two species are sympatric in Papua New Guinea and northern Queensland (White, 1926; McMillan, 1986), and both occur on the northern coast of Northern Territory (Wells, 1983). No intermediate forms between them has been recorded, as reported by McCusker (1984). After the examination of numerous herbarium specimens and limited fresh materials collected from Darwin area, we found that several characters of flowers of C. tagal collected from Northern Territory and Papua New Guinea are closer to those of C. australis. Namely, the populations of C. tagal from the Northern Territory and Papua New Guinea have narrower and oblong calyx lobes, longer clavate appendages on the petal apex and longer styles than C. tagal from other populations in the world. However, the characters of fruit and propagule of C. tagal from this area resemble those of other global populations of C. tagal. We assume that a possible hybridization between these two taxa may have occurred. According to Duke et al. (1984), the major flowering season of the populations from northeastern Australia are November and January to March for C. australis and C. tagal, respectively. Based on the observation of herbarium specimens, a broader period of flowering season for both species could be inferred. The possible overlap of flowering season may increase the opportunity of hybridization between these two species. An anecdotal report notes that hybridization has occurred, and some trees with both types of propagules from the Murray River, Admiralty Island, and Pigeon Island in northeastern Queensland have been observed (Ballment et al., 1988). However, except for the flower variation in the Northern Territory and Papua New Guinea previously mentioned, we did not find such intermediate forms in this study. A further study to compare the morphological and genetic variations of populations of C. tagal in the Northern Territory, Australia, and Papua New Guinea and the other populations from the world should be useful and interesting. Moreover, the factors influencing the sympatric populations of C. australis and C. tagal in the Northern Territory and Queensland may be worth exploring, in order to reveal why a possible hybridization only occurs in the Northern Territory, but not in northern Queensland. Acknowledgements. The authors thank T. Lammers and M. S. B Ku for improving the manuscript, D. Foster and V. Sarafis for helping to collect some materials in Australia, K. N. Ganeshaiah for helping with the distribution map, and the following herbaria for permission to study and/or borrow specimens: BM, BO, CAL, DNA, GH, K, L, and MO. This study was supported by the National Science Council (NSC94-2311-B-020-001-) and by National Chiayi University (NCYU 96T001-02-06-018) of Taiwan. LITERATURE CITED Ballment, E.R., T.J. Smith III, and J.A. Stoddart. 1988. Sibling species in the mangrove genus Ceriops (Rhizophoraceae), detected using biochemical genetics. Aust. Syst. Bot. 1: 391-397. Doyle, J. and J. Doyle. 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochem. Bull. 19: 11-15. Duke, N.C. 2006. Australia s Mangroves: the Authoritative Guide to Australia s Mangrove Plants. University of Queensland, Queensland. Duke, N.C., J.S. Bunt, and W.T. Williams. 1984. Observations on the floral and vegetative phenologies of north-eastern Australian mangroves. Aust. J. Bot. 32: 87-99. Felsenstein, J. 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783-791. Fitch, W.M. 1971. Towards defining the course of evolution: minimum change for a specific tree topology. Syst. Zool. 20: 406-416. Ge, X.J. and M. Sun. 2001. Population genetic structure of Ceriops tagal (Rhizophoraceae) in Thailand and China. Wetl. Ecol. Manag. 9: 203-209. Hall, T.A. 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/ NT. Nucl. Acids Symp. Ser. 41: 95-98. Hillis, D.M. and J.J. Bull. 1993. An empirical test of bootstrapping as a method for assessing confidence in phylogenetic analysis. Syst. Biol. 42: 182-192. Hou, D. 1958. Rhizophoraceae. In C.G.G.J. van Steenis (ed.), Flora Malesiana, ser. 1, vol. 5. Noordhoff- Kolff N. V., Djakarta, pp. 429-473. Kimura, M. 1980. A simple method for estimating evolutionary rates of base substitution through comparative studies of nucleotide sequences. Mol. Evol. 16: 111-120. Kumar, S., K. Tamura, I.B. Jakobsen, and M. Nei. 2001. MEGA 2.1: Molecular Evolutionary Genetics Analysis software. Arizona State University, Tempe, AZ. McCune, B. and M.J. Mefford. 1999. Multivariate Analysis of Ecological Data, Version 4. MjM Software Design, Gleneden Beach, Oregon, USA. McCusker, A. 1984. Rhizophoraceae. In A.S. George, B.G. Briggs, B.A. Barlow, H. Eichler, L. Pedley, J.H. Ross, D.E. Symon, P.G. Wilson, and A. McCusker (eds.), Flora of Australia Vol. 22: Rhizophorales to Celastrales, Australian Government Publishing Service, Canberra, pp. 1-11. McMillan, C. 1986. Isozyme patterns among populations of black mangroves, Avicennia germinans, from the Gulf of Mexico-Caribbean and Pacific Panama. Contrib. Mar. Sci. 29: 17-25. Rzhetsky, A. and M. Nei. 1992. Statistical properties of the ordinary least-squares, generalized least-squares, and minimum-evolution methods of phylogenetic inference. J. Mol. Evol. 35: 367-375. Saitou, N. and M. Nei. 1987. The Neighbor-joining method: a new method for reconstructing phylogenetic trees. Molec. Biol. Evol. 4: 406-425. Sheue, C.R. 2003. Comparative Morphology and Anatomy of

SHEUE et al. Reevaluating the taxonomic status of Ceriops australis 99 the Eastern Mangrove Rhizophoraceae. Ph. D. Dissertation. National Sun Yat-sen University, Kaohsiung, Taiwan. Taberlet, P., L. Gielly, G. Pautou, and J. Bouvet. 1991. Universal primers for amplification of three non-coding regions of chloroplast DNA. Plant Mol. Biol. 17: 1105-1109. Tomlinson, B.P. 1986. The Botony of Mangroves. Cambridge Univ. Press, Cambridge. Wells, A.G. 1983. Distribution of mangrove species in Australia. In H.J. Teas (ed.), Tasks for Vegetation Science, 8: Biology and Ecology of Mangroves. Dr W. Junk Publishers, The Hague, The Netherlands, pp. 57-76. White, C.T. 1926. A variety of Ceriops tagal C. B. Rob. (=C. candollean W. and A.). J. Bot. Lond. 64: 220-221. Wightman, G. 2006. Mangroves of the Northern Territory, Australia: identification and traditional use. Department of Natural Resources, Environment and the Arts, and Greening Australia, Northern Territory, Australia. 以形態及分子特徵再評估南方細蕊紅樹 ( 紅樹科 ) 的分類位階 1 2 2 3 1 PeterSAENGER 4 ChristopherP.MANGION 5 GlennWIGHTMAN 5 JeanW.H.YONG 6 7 1 2 3 4 SchoolofEnvironmentalScienceandManagement, SouthernCrossUniversity,LismoreNSW,Australia 5 DepartmentofNaturalResources, EnvironmentandtheArts,PalmerstonNT,Australia 6 NaturalSciences,NationalInstituteofEducation, NanyangTechnologicalUniversity,Singapore 7 White Ballment 逹 29 PCA DNA trnl intron 關鍵詞 : DNA

100 Botanical Studies, Vol. 50, 2009 Appendix. Specimens list (herbarium acronyms follow Index Herbariorum, available at http://sweetgum.nybg.org/ih/). Specimens list for morphometric analysis for principal components analysis. Ceriops australis: AUSTRALIA: Northern Territory: Bowman & Wilson 263 (DNA), Dunlop 3629 (DNA), Dunlop & Munns 7512 (L), Egan 2821 (DNA), Forster & Russell-Smith PIF5920 (DNA), Henshall 857 (DNA), Latz 3192 (DNA, L, MO), Must 1649 (DNA), Russell-Smith & Lucas 5884 (DNA). Queensland: Smith 4825 (L), 12441 (GH, L) Stoddart 4536, 4699, 4761, 4903 (L, MO), 4992 (L). Ceriops tagal: AUSTRALIA: Northern Territory: Brock 116 (DNA), Dunlop 3899 (DNA), Dunlop & Wightman 9709 (DNA), J. & Eurell 78/20 (DNA, MO), Scarlett 164 (DNA), Wightman 458, 786 (DNA). MADAGASCAR: Birkinshaw & Jules 13 (MO), Darcy & Rakotozafy 15470 (MO), Dorr & Koenders 3063 (GH, MO), Rahajasoa 356 (MO). SUMATRA: Iwatsuki et al. S1319 (MO), S1321 (L), Schmad 146 (L), Teijsmamn & Miquee s. n. [no date] (K). Specimens examined for revising the distribution range of C. australis and the sympatric localities of C. tagal in Australia. Ceriops australis: AUSTRALIA: New Holland: Banks & Solander s. n. [1770] (BM), Queensland: Blake 14127 (MO), Clarkson 2016 (MO), 3875 (DNA, MO), Cribb & Newton s. n. [1950] (BM), Dietrich s. n. [1863-65], 657 (MO), Durrington (L), Everist 7881A (L), Fosberg 61833 (MO), Macnae s. n. [1962], Mrs. Stephenson 569 (BM), Neldner & Clarkson 2993 (DNA), Smith 4825, 11435 (L), 12441 (GH, L), Stoddart AQ14784 (K), 4510 (MO), 4527, 4536, 4699, 4761, 4786, 4903 (L, MO), 4992 (L), Webster & Hildreth 15005 (GH), White s. n. [1915] (BM), 3372A (K, type), 3373A (GH); Northern Territory: Bardsley s.n. [1985] (DNA), Barlow 506 (DNA), Blake 17050 (K, GH), Bowman & Wilson 263 (DNA), Brennan 2619 (DNA), Brooker 3258 (DNA), Byrnes NB275 (DNA), Byrnes & Maconochie 1077 (DNA), Calliss 63 (DNA), Chippendale s. n. [1961], 8180 (DNA), Clark 948 (DNA), Cowie 5183 (DNA), Cowie & Dunlop 4131, 7926 (DNA), Dunlop 1869 (DNA), 2782 (DNA, MO), 3984 (DNA), Dunlop & Leach 8062 (DNA), Dunlop & Munns 7512 (L), Dunlop & Wightman 9203 (DNA), Egan 2391, 2821 (DNA), Forster & Russell-Smith PIF5920 (DNA), Gill s. n. [1970] (GH), Henry 88 (DNA), Henshall 857 (DNA), Hodder s. n. [1971] (K), D4044 (DNA), Latz 3192 (DNA, L, MO), 3390 (DNA), Leach 3993, 4231 (DNA), Leach & Cowie 3641 (DNA), Martensz & Schodde AE737 (DNA), McKean B142, 974 (DNA), Michell & Ingraham 27 (DNA), Must 884, 1310 (DNA), 1348, 1649 (DNA, MO), Nelson 1078 (DNA), Rankin 1171, 1248, 1380, 2220 (DNA), Ridpath Mck B7 (DNA), Russell-Smith 8918 (DNA), Russell-Smith & Lucas 4375, 5642, 5884, 8368 (DNA), Scarlett 163 (DNA), Shaw & Dunlop 3629 (DNA, MO), Smith 1030, Specht 591 (GH), Story 8337 (DNA), Thomson 661, 1878, 2621, 2661 (DNA), van Kerckhof 29, 33, 39 (DNA), Waddy 560 (DNA), Wells s. n. [1975, 1978] (DNA), Wheelwright DW8 (DNA), 24 (DNA), Wightman 475, 488, 504, 506, 520, 543, 619, 673, 701, 814, 1070, 1544, 1663, 2290, 2389, 2453, 2472 (DNA), 4603 (DNA, MO), 6162, 6652 (DNA), Wightman & Dunlop 551, 563 (DNA), Wightman & Giulian 2926 (DNA), Wightman & Smith 3531, 4523 (DNA), Williams 350 (DNA), Williams & Wightman 135 (DNA), Wilson 790 (DNA); Western Australia: Croat 52316A (MO), Cunningham 235 (K), Fstyguold s. n. [1906] (BM), George 12724 (DNA), 14829 (K), Hartley 14587 (DNA), Mitchell 5949 (DNA), Morrison s. n. [1950], Paijmans 2469 (DNA), Perry 2548 (DNA), Wightman 7111 (DNA); PAPUA NEW GUINEA: Central District: near Barune: Frodin UPNG4444 (K); Fairfax Harbour: Gillison NGF22159 (GH); Kairuku subdistriction: Darbyshire 773 (K); near Lae Lae: Schodde 2681 (GH); Kappa Kappa Papua: Brass 786 (BM); Port Moresby: Frodin & Millar UPNG562 (L). INDONESIA: Timor: Anonymous s. n. [1923] (BO), L. v. d. Pijl 820 (BO); Flores: M. Kew s. n. [1905] (BO); Subawa: Kostermans & Wirawan 348 (BO); Java: Teijsmann s. n. [no date] (CAL); Bilinton Island: Teijsmann s. n. [1875] (BO). Ceriops tagal: AUSTRALIA: no data: Leschenault s. n. [1802] (BM), Queensland: Stoddart 4086, 4113, 4154 (MO), 4317 (MO, L), 4385 (MO), 4634 (MO, L), 4637 (MO), Smith 11409 (GH), 11618 (K), 11619 (L), 12445 (GH), 12521 (L), Smith & Webb 3243 (L), Mrs. Stephenson 486, 545, 570, 606 (BM), Thom 4168, 4170, 4171, 4172 (MO), Clarkson 3387 (MO). Northern Territory: Bardsley 15 (DNA), Brennan 4563, 2627, 2877 (DNA), Brock 116 (DNA), Byrnes & Maconochie NB1078 (DNA), Cowie 3397, 5140, 6924 (DNA), Dunlop 3899 (DNA), Dunlop & Wightman 6541, 9709, 9739 (DNA), Egan 2713 (DNA), Eurell s.n. (MO), s. n. [1978] (GH), J. & Eurell 78/ 20 (DNA, MO), Kerrigan & Risler 57 (DNA), Scarlett 164 (DNA), Stocker GS79 (DNA), Wells s. n. [1975] (DNA), Wightman 458, 786, 1970, 4113, 4185, 4457, 6506 (DNA), Wightman & Giuliana 2993 (DNA), Wightman & Smith 3531 (DNA).