A morphological review of the Cuora flavomarginata complex (Testudines: Geoemydidae)

A morphological review of the Cuora flavomarginata complex (Testudines: Geoemydidae)
of 7
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  A morphological review of the  Cuora flavomarginata  complex(Testudines: Geoemydidae) Carl H. Ernst*, Arndt F. Laemmerzahl, and Jeffrey E. Lovich (CHE, AFL) Department of Environmental Science and Policy, George Mason University,Fairfax, Virginia 22030-4444, U.S.A. (current address of CHE: Division of Amphibians andReptiles, MRC 162, Smithsonian Institution P.O. Box 37012, Washington, D.C. 20013-7012,U.S.A.), e-mail:,;(JEL) U.S. Geological Survey, Southwest Biological Science Center, 2255 North Gemini Drive,Flagstaff, Arizona 86004, U.S.A., e-mail: Abstract.—  A reevaluation of the morphometric and color patterndifferences within the Asiatic box turtle,  Cuora flavomarginata  sensu latu,was conducted in view of determining the taxonomic position of the threecurrently recognized subspecies:  C. f. flavomarginata  (Taiwan),  C. f. sinensis (southern mainland China), and  C. f. evelynae  (Ryukyu Islands, Japan).Recent analyses indicate that the allopatric population of   C. f. evelynae  isthe most divergent of the three taxa and shares little possibility for geneexchange with the other two populations. In contrast, the populations of   C. f. flavomarginata  and  C. f. sinensis  share many characters. We recommendthe recognition of the Ryukyu population as a full species,  C. evelynae .The Asian box turtle,  Cuora flavomar- ginata , is currently considered polytypic,consisting of three subspecies:  Cuora flavomarginata flavomarginata  (Gray,1863) of Taiwan;  C. f. sinensis  (Hsu¨,1930) of southern mainland China (Fonget al. 2002); and  C. f. evelynae  Ernst &Lovich, 1990, of the Japanese RyukyuIslands (Ernst et al. 2000, Hennen 2004,Schilde 2005, Bonin et al. 2006, Fritz &Havasˇ 2007). The validity of   C. f. sinensis and the taxonomic rank of   C. f. evelynae ,both of which were described as fullspecies, have been questioned (McCord& Iverson 1991, Zhao & Adler 1993, Fritz& Obst 1999, Yasukawa & Ota 1999,Fritz & Havasˇ 2007).The goal of this study is to examinemorphometric and color pattern differ-ences between the three populations toreevaluate the taxonomic status of threegeographic disjunct populations. We nowpossess significantly larger samples of thethree separate populations of   C. flavo-marginata  sensu latu than in our formeranalysis (Ernst & Lovich 1990). Ourresults and those of other studies arepresented below.Materials and MethodsA total of 125 turtles were examined(Ryukyu Islands, 38; Taiwan, 50; south-ern mainland China, 37; see Appendix 1).Notes and drawings were made of head,neck, limb, carapacial and plastral pat-terns of each specimen. Sexes weredetermined by the characters given byErnst & Barbour (1989) and Ernst et al.(2000). Straight-line measurements of each specimen were taken with dialcalipers accurate to 0.1 mm. Variablesincluded: the greatest carapace length(CL, not midline), carapace width (CW)and depth (CH) at the level of the seamseparating vertebral scutes 2 and 3, *Corresponding author. PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON121(3):391–397. 2008.  marginal width (MW, the differencebetween the CW and the width acrossthe carapace taken between the pointsof juncture of the marginal scutes andpleurals at the level of the seam betweenvertebrals 2 and 3), greatest plastronlength (PL), greatest width and length of both plastral lobes (APW, APL, PPW,PPL), greatest length of the right shellbridge (BL), greatest width and length of the cervical scute (CSW, CSL) and allvertebrals (V 1 W, V 1 L, etc.), and themedial seam lengths of all plastral scutes(Gul., Hum., Pect., Abd., Fem., An.).Shell proportions were expressed asratios of one measurement to another.Several ratios proved useful: width/lengthof cervical scute, width/length of designat-ed vertebrals, MW/CL, CH/CL, CH/CW,length of light-colored blotch at its medialheight on pleural scute 2/total length of pleural scute 2 at the same point (PBL/PSL), length of the light-colored blotch atits medial height on pleural 2/carapacelength (PBL/CL), PL/CL, BL/PL, APL/PL, APW/PL, APW/APL, PPL/PL, PPW/PL, and PPW/PPL. The number of rowsof large scales at the lateral edge of theantebrachium between the knuckle jointof digit 5 and the first horizontal skin foldproximal to the elbow (FLSR, mistakenlydescribed in Ernst & Lovich 1990 as thescale rows between the claw of digit 5 andthefirsthorizontalskinfold)werecountedand recorded.Only adult turtles (those that displayedsecondary sexual characteristics) wereused in the analyses. Data of both sexeswere combined for further analysis as nocharacters displayed significant differ-ences between females and males (adjust-ed for region). Effects due to sexualdimorphism were also minimized sinceturtles on the Ryukyus and Taiwanappear to have similar levels of sizedimorphism (Gibbons & Lovich 1990).Statistical techniques were executed usingSAS package 8.2. Levels of significancewere set a priori at  a 5 0.05.Continuous variables were checked fornormality using q-q plots (Conover 1980).As a result, most were judged to beapproximately normal. These variableswere then used in a stepwise discriminantfunction analysis, which identified thevariables used in subsequent analyses.Because FLSR was deemed significant-ly different in Ernst & Lovich (1990) butnot used by McCord & Iverson (1991),this trait was analyzed separately using aKruskal-Wallis test. Because sample dif-ferences were highly significant, multiplecomparisons (Conover 1980) were con-ducted to look for significant differencesamong all three populations.ResultsThe results of the discriminant functionanalysis (resubstitution) clearly showedthat the three populations were morpho-logically distinct, and a follow up cross-validation analysis confirmed this. Thefirst two canonical variables (from theresubstitution results) are plotted inFig. 1 and show the three populations tobe basically separated, with very minoroverlap between the Ryukyu and Chinapopulations. Only one outlier causes themajor polygon overlap between the Chinaand Ryukyu Island populations. Thisspecimen has not been identified butmay have erroneous locality data associ-ated with it; even with this specimenincluded, the overlap is no more thanexpected in such a genetically relatedcomplex. Some overlap, possibly due tomore recent gene flow, occurs betweenTaiwan and China. This result is alsosupported by a MANOVA performed aspart of the discriminant function analysis.Univariate results generally confirm theresults of the discriminant analysis, al-though APW and CH showed no signif-icance. This result is not unexpected asmultivariate statistics often show signifi-cant differences when univariate statisticson the same data do not (Johnson & 392 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON  Wichern 1998). The two most importantunivariate characters were PBL/PSL(Ryukyu Islands:  X  ¯   5 64.9 [58.0–73.7] % ;Taiwan:  X  ¯   5  43.0 [24.0–69.6]  % ; China: X  ¯   5  38.7 [31.9–51.1]  % ) and FLSR(Taiwan:  X  ¯  5 9.7 [6–13]; Ryukyu Islands: X  ¯  5 8.8 [6.5–11]; China:  X  ¯  5 8.3 [6–10]).Morphological and pattern data werecompared by locality (mainland China,Taiwan, and the Ryukyu Islands) todetermine if geographic variation occurs,and to what extent. Univariate test resultsindicate that PBL/PSL and FLSR weresignificantly different for the combinedand each individual sex samples (  p  5 0.001); in addition, Abd (  p  5  0.01) andPPL (  p  5  0.004) were also significantlydifferent in females, but not for males orfor combined sex data. A NPAR1WAYprocedure (Table 1) yielded Wilcoxonscores (rank sums) for FLSR (Kruskal-Wallis Test chi-square 18.07,  p 5 0.0001)that differed significantly. Results of the discriminant analysis are shown inFig. 1.Classification by cross validation of combined data from both sexes for thethree geographic areas is presented inTable 2. For the Ryukyu sample, 100 % of males and 77.8 %  of females were classi-fied correctly (females, 22.2 %  as main-land China). Data from mainland Chinayielded a correct female identification of 57.1 %  (42.9 %  as Taiwan), and 85.2 %  of Taiwanese females were identified cor-rectly (11.1 % as mainland China, 3.7 % asRyukyu Islands). Male data from bothmainland China and Taiwan yielded71.4 %  correct identifications and 28.6 % as from the Ryukyu Islands. Fig. 1. Discriminant function analysis (resubstitution) using the first two canonical variables of individuals (sexes combined) of   Cuora  populations from southern mainland China (C), the Ryukyu Islands(R), and Taiwan (T). VOLUME 121, NUMBER 3 393  DiscussionDiscriminant function analysis withcross validation of combined sex data ispresented in Table 2. The same test of thesame data set but testing each sexseparately, properly classified 100 %  of males and 77.8 %  of females from theRyukyu Islands (presently,  Cuora f. eve-lynae ); 71.4 %  of males, and 57.1 %  of females from southern mainland China( C. f. sinensis ); and 71.4 %  of males and85 % of females from Taiwan ( C. f. flavo-marginata ). Males are less variable thanfemales.The populations of   C. flavomarginata from mainland China and Taiwan shareenough similarities among the charactersto have both males and females incor-rectly assigned to the other population orthe Ryukyus (however, only 3.7 %  of Taiwanese females and no females fromsouthern mainland China were identifiedas from the Ryukyu population). Themorphological overlap in the Chinese andTaiwanese populations might result fromproximity and relatively recent geneexchange, and strengthens the interpreta-tion that the two populations represent asingle subspecies,  C. f. flavomarginata (Zhao & Adler 1993, Fritz & Obst 1999,Yasukawa & Ota 1999, Fritz & Havasˇ2007). In contrast, the more distantRyukyu population appears less variable,and it is widely separated morphological-ly from the other two populations.A plot of the discriminant analysis(resubstitution) using the first two canon-ical variables of the total adult samplefrom the three populations (Fig. 1) indi-cates a sharing of characters betweenmainland China and Taiwan turtles, butessentially none, except one distant Chi-nese outlier (see above), between theRyukyu population and mainland China,and no character overlap between turtlesfrom the Ryukyus and Taiwan.These results support the concept thatthe populations from mainland Chinaand Taiwan represent one taxon, andthat  C. f. sinensis  (Hsu¨, 1930) is a juniorsynonym of   C. f. flavomarginata  (Gray,1863), as recognized by McCord &Iverson (1991), Zhao & Adler (1993),Fritz & Obst (1999), Yasukawa & Ota(1999), and Fritz & Havasˇ (2007).The taxonomic status of the RyukyuIslands population,  C. f. evelynae , isanother matter. McCord & Iverson(1991) thought elevation of the Ryukyupopulation to full specific status by Ernst& Lovich (1990) was ‘‘unwarranted,’’ andthat it ‘‘perhaps’’ represented a subspeciesof   C. flavomarginata . They did not Table 1.—Wilcoxon scores (rank sums) for the character FLSR classified by population (Ryukyu Islands,southern mainland China, and Taiwan. Population ( n ) Sum of scores Expected under HO Standard deviation under HO Mean score Ryukyu (28) 1114.0 1190.0 102.4 39.8China (21) 561.5 892.5 94.1 26.7Taiwan (35) 1894.5 1487.5 107.1 54.1Table 2.—Classification by cross validation of individual turtles from the Ryukyu Islands, southernmainland China, and Taiwan. Population ( n )Taxon frequencyRyukyu China Taiwan Ryukyu (28) 25 (89.3 % ) 0 3 (10.7 % )China (21) 0 17 (81.0 % ) 4 (19.1 % )Taiwan (35) 2 (5.9 % ) 7 (28.9 % ) 25 (73.5 % ) 394 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON  include in their analysis all charactersused in the srcinal diagnosis by Ernst &Lovich (1990). They also criticized Ernst& Lovich (1990) for including juveniles intheir srcinal analyses; our present anal-yses use only adult  Cuora . Subsequentstudy of morphological variation within Cuoraflavomarginata byYasukawa(1998)and Yasukawa & Ota (1999) demonstratethat  C. f. evelynae  is the most divergentof the three described subspecies.A study of variations in 882 base posi-tions of mitochondrial DNA (mtDNA)12S and ribosomal RNA (rRNA) 16Sgenes of Asian  Cuora  by Honda et al.(2002) also revealed that Ryukyu  C. f.evelynae  are genetically distinct from theTaiwan population of   C. flavomarginata .Both groups were shown to be monophy-letic in three separate cladistic analyses: aneighbor-joining dendogram derivedfromthe distance matrix from mtDNA 12S andrRNA 16S sequence data, a maximum-likelihood dendogram, and a maximumparsimony dendogram using heuristicoption. In addition, two additional recentstudies also reported genetic differencesbetween the populations from China andRyukyus. Spinks et al. (2004), use proteincoding cytochrome b, mtDNA, 12 s ribo-somal RNA, and a , 1 kb intron from theR35 neural transmittter gene. Spinks &Shaffer (2006), although concentrating onother species of   Cuora , use cytochrome coxidase subunit I mtDNA, the nicotin-amide adenine dinucleotide dehydroge-nase subunit 4 gene, plus the flankinghistidine andserine tRNAsand partof theleucine tRNA in their sequencing. Unfor-tunately, the turtle sample sizes in all threeof these studies are small, but differencesbetween the populations tested wereindicated in every test.Our data clearly strengthen the obser-vations of Yasukawa (1998), Yasukawa &Ota (1999), Honda et al. (2002), Spinks etal. (2004), and Spinks & Shaffer (2006)and further indicate that the  Cuora  on theRyukyu Islands should be considered aseparate species. In addition, the Ryukyuturtles are geographically separated (allo-patric) and have been at least since themid-Pleistocene (Inger 1947, Ota 1998,2003; Takahashi et al. 2003). Currently,natural gene exchange is highly unlikelywith either the Taiwan or mainland Chinapopulations of   C. flavomarginata , addingto the argument for specific status.Discussions about what constitutes a‘‘species’’ flourished in scientific journalsduring the last two decades, occasionallymired in both theoretical and operationalconceptualizations and criteria (Frost &Hillis 1990). Alternative species explana-tions have been proposed, including thebiological, evolutionary, phylogenetic andunified species concepts, all of whichdefine species as separately evolvingmetapopulation lineages (De Queiroz2005, 2007). As reiterated by Berry et al.(2002), the species is the fundamental unitof evolution, forged through separation of ancestral lineages by mechanisms includ-ing vicariance, dispersal, ecology, or be-havior. We made no attempt in our analy-sis to recover the phylogenetic history of this complex. However, our morphologi-cal analyses once again (Ernst & Lovich1990) demonstrate the existence of inter-nally cohesive, diagnosable taxonomicunits with actual or potential gene flowoccurring between some but not othermetapopulations: a characteristic of theevolutionary species concept (Berry et al.2002). The congruence of our findingswith those of independent researchersusing molecular techniques providesadditional strong support for recognitionof   C. evelynae  as a separate species.Similar to Frost & Hillis (1990), we donot believe that recognition of subspeciesin this complex advances our understand-ing of variation in the  Cuora flavomargi-nata  complex. We, therefore, recommendthat  Cuora flavomarginata evelynae  Ernst& Lovich, 1990, be reinstated as a fullspecies,  Cuora evelynae  Ernst & Lovich,1990. VOLUME 121, NUMBER 3 395
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks