行政院國家科學委員會專題研究計畫 成果報告
日行性鱗翅類多重擬態複和群的演化
計畫類別： 個別型計畫
計畫編號： NSC93-2311-B-110-007-
執行期間： 93 年 11 月 01 日至 94 年 07 月 31 日
執行單位： 國立中山大學生物科學系(所)
計畫主持人： 顏聖紘
報告類型： 精簡報告
處理方式： 本計畫可公開查詢
中 華 民 國 94 年 10 月 30 日
行政院國家科學委員會補助專題研究計畫
; 成 果 報 告
□期中進度報告
日行性鱗翅類多重擬態複和群的演化
計畫類別：; 個別型計畫 □ 整合型計畫
計畫編號：NSC 93－2311－B－110－007
執行期間：
93 年 11 月 01 日至 94 年 7 月 31 日
計畫主持人：顏聖紘
共同主持人：
計畫參與人員：
成果報告類型(依經費核定清單規定繳交)：;精簡報告 □完整報告
本成果報告包括以下應繳交之附件：
□赴國外出差或研習心得報告一份
□赴大陸地區出差或研習心得報告一份
□出席國際學術會議心得報告及發表之論文各一份
□國際合作研究計畫國外研究報告書一份
處理方式：除產學合作研究計畫、提升產業技術及人才培育研究計畫、
列管計畫及下列情形者外，得立即公開查詢
□涉及專利或其他智慧財產權，□一年□二年後可公開查詢
執行單位：國立中山大學生物科學系
中 華 民 國 94 年 10 月 30 日
本研究探索茶斑蛾屬-一個具有高度形態多樣性、在東亞高度分化且具極混亂分類歷史之日
行性蛾類群之親緣關係、系統分類以及擬態斑紋之演化。我們使用可見光與紫外光(包含紫
外反射光與螢光)檢查了了這些昆蟲。我們並使用80個成蟲形態形質(包含了41個來自色彩與
斑紋的形質)重建了34個分類群，包含所有被承認的茶斑蛾屬種類與兩種狹翅斑蛾屬種類作
為外群。使用所有資訊所重建之親源關係顯示目前所認知的茶斑蛾屬為一個單系群；而不
同的擬態斑紋具有不同的親緣關係保守性。為了探索斑紋的演化，我們將所有與斑紋相關
之形質抑制並重新建立另一個與原親緣關係假說具顯著性差異的假說。我們使用這個新的
假說來測試穆氏擬態與副貝氏擬態的演化預測。PTP測驗之結果顯示所有與色彩斑紋相關
的形質具有親緣關係保守性，也就是說，茶斑蛾屬的斑紋演化符合穆氏擬態的假說。然而
當我們使用Kishino-Hasegawa測試時，其斑紋演化卻較符合副貝氏擬態的假說。我們因此認
為使用簡單的親緣關係測試驗證擬態模式可能並不實際。
關鍵詞：化學防禦、親緣關係、趨同演化、關聯性演化、副貝氏擬態、副貝氏擬態、多態
性
英文摘要
The present study aims to investigate the phylogeny, systematics and evolution of the mimetic
wing patterns of Eterusia, a day-flying moth genus which exhibits great morphological diversity,
as well as the highest insular differentiation in eastern Asia and which has the most chaotic
taxonomic history in the family Zygaenidae. We examined the wing patterns of the insects
involved using visible and ultraviolet light (both UV reflectance and fluorescence). The
phylogeny of 34 taxa, including all the recognized species of Eterusia plus 2 species of Soritia as
outgroups, was reconstructed based on 80 adult morphological characters, including 41 derived
from colour patterns. Phylogenetic relationships based on the whole data set reveal that (1) the
most current concept of Eterusia is monophyletic, and (2) different types of mimetic patterns
show different levels of phylogenetic conservation. To investigate the evolution of their colour
patterns, we inactivated all the relevant characters and reconstructed another phylogeny, which
was found to differ significantly from the one based on the whole character set in the position of
the Eterusia risa species-group. We used these phylogenetic hypotheses to test evolutionary
predictions based on conventional Müllerian mimicry and quasi-Batesian mimicry dynamics. The
results of PTP tests showed that the colouration characters are phylogenetically conserved, thus
justifying a Müllerian interpretation. However, when comparing the observed topologies with
hypothetical trees constrained to fit perfect Müllerian or quasi-Batesian scenario using the
Kishino-Hasegawa test, the observed phylogenies were more consistent with the phylogenetic
prediction of quasi-Batesian mimicry. Therefore, we consider that applying these two
phylogenetic methods to justify mimicry models may not always be practical.
Keywords: chemical defence, phylogeny, convergent evolution, correlated evolution,
quasi-Batesian mimicry, polymorphism
Phylogeny, systematics and evolution of mimetic wing
patterns of Eterusia moths (Lepidoptera, Zygaenidae,
Chalcosiinae)
S H E N - H O R N Y E N
1 , 2, G A D E N S . R O B I N S O N
3and D O N A L D
L . J . Q U I C K E
1 , 31
Department of Biological Sciences, Imperial College London, Silwood Park Campus, Ascot, Berkshire, U.K.,
2
Department of Biological Sciences, National Sun Yat-Sen University, Kaohsiung, Taiwan,
3
Department of Entomology, The Natural History Museum, London, U.K.
Abstract. The aim of the present study was to investigate the phylogeny, sys-tematics and evolution of the mimetic wing patterns of Eterusia, a day-flying moth genus that exhibits great morphological diversity, as well as the highest insular differentiation in eastern Asia and which has the most chaotic taxonomic history in the family Zygaenidae. We examined the wing patterns of the insects involved using visible and ultraviolet light (both reflectance and fluorescence). The phy-logeny of thirty-four taxa, including all the recognized species of Eterusia plus two species of Soritia as outgroups, was reconstructed based on eighty adult morpho-logical characters, including forty-one derived from colour patterns. Phylogenetic relationships based on the whole dataset revealed that (1) the most current concept of Eterusia is monophyletic, and (2) different types of mimetic pattern show different levels of phylogenetic conservation. To investigate the evolution of their colour patterns we inactivated all the relevant characters and reconstructed another phylogeny, which was found to differ significantly from the one based on the whole character set in the position of the E. risa species group. We used these phylo-genetic hypotheses to test evolutionary predictions based on conventional Mu¨llerian mimicry and quasi-Batesian mimicry dynamics. The results of permutation– tail–probability tests showed that the coloration characters are phylogenetically
conserved, thus justifying a Mu¨llerian interpretation. However, when
comparing the observed topologies with hypothetical trees constrained to fit perfect Mu¨llerian or quasi-Batesian scenarios using the Kishino–Hasegawa test, the observed phylogenies were more consistent with the phylogenetic prediction of quasi-Batesian mimicry. Therefore, we consider that applying these two phylo-genetic methods to justify mimicry models may not always be practical. Finally, the taxonomy of Eterusia is revised. In total, two new species (E. austrochinensis, E. guanxiana), one new subspecies (E. risa palawanica) and four new synonyms (E. lativitta and E. fasciata of E. sublutea, E. coelestina of E. subcyanea,
E. angustipennis gaedeiof E. angustipennis angustipennis) are established.
Introduction
Moths of the subfamily Chalcosiinae Walker, [1865] exhibit the highest diversity in morphology and ecology within the superfamily Zygaenoidea, and even among the non-obtecto-meran apoditrysian Lepidoptera (Yen, 2003c, 2004a; Yen et al., in press). According to current phylogenetic analyses Correspondence: Shen-Horn Yen, Department of Biological
Sciences, National Sun Yat-Sen University, Kaohsiung 804, Taiwan. E-mail: [email protected]
and systematic revisions (Yen, 2003c, 2004a; Yen et al., in press), this subfamily comprises some 370–400 species dis-tributed among eighty-four genera. Except for an isolated genus restricted to the Atlanto-Mediterranean area, all the members of this subfamily are distributed from the Indian subcontinent and Palaearctic eastern Asia, through South-east Asia, to the Melanesian and Micronesian archipelagos, but as yet none is known from Australia and its adjacent islands. Their brilliant coloration and rarity in museum collections have amazed scientists and insect collectors since the 18th century. However, their sexual dimorphism, polymorphism and complex mimetic wing patterns have led to incongruence between character sets and a very misleading classification (Yen, 2003b, c, 2004a; Yen et al., in press).
Among the chalcosiine genera, the genus Eterusia Hope 1841 (type species: E. tricolor Hope 1841) is well known for its complex sexual dimorphism, polymorphism, mimetic wing patterns (Owada & Ta, 2002) and high geographical differentiation in the East Asian Island Arc (Owada, 1989, 1998a, b, 2000a, b, 2001, 2002; Yen, 2003a, 2004b). Several species (e.g. E. aedea (Linnaeus, 1763) and E. risa (Double-day, 1844)) have been reported as defoliating pests of tea trees (Camellia sinensis, Theaceae) in India (Andrews & Tunstall, 1915; Fletcher, 1920; Sevastopulo, 1940; Robinson et al., 2001), Indonesia (Tarmann, 1992, 2003), Vietnam (Du Pasquier, 1932), China (Zhu et al., 1979; Wang, 1987) and Taiwan (Guan & Yeh, 1977). The adult moths are in general day-flying, visiting nectar sources in the canopy or perching around the host plants, but males are occasionally attracted to light at night, and copulation has been observed at night in at least one of the Japanese subspecies (Owada, pers. comm.). Historically, the systematics and taxonomy of Eterusia moths have been confused due to their very convergent wing patterns (see also Table 1 and Appendix 1 for historical taxonomic problems). Numerous species erroneously placed in Eterusia in the past have been trans-ferred to various other chalcosiine genera, including Cyclosia Hu¨bner, 1820, Pidorus Walker, 1854, Soritia Walker, 1854, Eusphalera Jordan, 1907, Heterusinula Jordan, 1907, Prosopandrophila Hering, 1922, Pseudoscap-tesyle Hering, 1922, and Scotopais Hering, 1922. Others belong in five new genera (Yen, 2004a) (see Tables 1 and 2). The species-level taxonomy of Eterusia has been bed-evilled by problems of delimitation of sympatric sibling species, polytypic and polymorphic species, as well as difficulties in matching the sexes in highly sexually dimorphic species. Although recent studies have attempted to elucidate the taxonomic problems of several species (e.g. Owada, 1989, 1992, 1996, 1998a, b, 2000a, b, 2001, 2002; Yen & Yang, 1998; Kishida & Endo, 1999; Owada & Horie, 1999; Yen, 2003c, 2004b, c), our understanding of their intra- and intergeneric phylogenetic relationships, and knowledge of their morphology and ecology remain very fragmentary. Yen (2003c, 2004a) has attempted to clean up this genus by transferring several species to Eusphalera, and separating the remaining species into nine species groups based on their wing patterns, copulatory structures (genitalia
and pregenital abdominal segments) and scent organs (see Table 2). According to Yen’s (2004a; Yen et al., in press) analyses of the generic phylogeny of Chalcosiinae, none of the previous concepts of Eterusia (e.g. Hering, 1922; Bryk, 1936; Endo & Kishida, 1999) is monophyletic. In this study we follow Yen’s (2004a) cladistic analyses, in which Eterusia is redefined based on synapomorphies shared by the nomi-notypical tricolor species group and its five related species groups (see Table 2). The aim of the present study was to investigate the species-level phylogeny of Eterusia based on all known taxa. The results are then used to examine the evolution of mimetic wing patterns and to interpret the nature of the adaptive resemblances involved.
Compared with butterflies, the diversity and evolution of mimicry among day-flying moths are far less well investi-gated and understood, although the chemical basis of the potential unpalatability has been studied in detail for some taxa, e.g. Geometridae (Nishida, 1995, 2002), Uraniidae (Kite et al., 1997), Arctiidae (Gonza´lez et al., 1999; Weller et al., 1999; Conner et al., 2000; Naumann et al., 2002; Conner & Weller, 2004), Noctuidae (Agaristinae) (Talianchich et al., 2003) and Zygaenidae (Witthohn & Naumann, 1984a, b, 1987a, b, Franzl et al., 1986; Naumann & Witthohn, 1986). In contrast to the related Zygaeninae and Procridinae, chalcosiine moths exhibit a more developed cyanogenic defence system that involves the use of several cyanoglucosides and associated morphological modifications in both adults and larvae (Yen, 2004a), although we still do not know whether the chalcosiine larvae sequester precursors of cyanoglucosides from cyanic host plants, or biosynthesize cyanoglucosides de novo from acyanic host plants. The subfamily as a whole participates in numerous mimicry complexes, which additionally involve members of thirty-two families and four orders of insects (Yen, 2004a). However, their mimicry has never been studied within either a phylogenetic or an ecological framework. Observations of the mimicry complexes involving Eterusia moths were initially presented by Owada (2000b), followed by Owada & Ta’s (2002) interesting observations in northern Vietnam. These authors compared the wing patterns of several Eterusiaspecies and many other insects that have similar wing patterns, and then pointed out that the Eterusia moths in northern Vietnam might demonstrate ‘parallel evolution’ with potentially comimetic agaristine moths, cicadas and lan-tern bugs (Pyrops spp., Fulgoridae). Owada (2000b) also noticed some similarity between the wing patterns of E. taiwana and E. aedea formosana, both of which are endemic to Taiwan but belong to different species groups of the genus. Recently, the discovery of E. vitessa in the Philip-pines and Seram (Yen, 2003a) has extended the mimetic pat-terns of this genus to include members of the Pyraloidea and other agaristine Noctuidae. In the present study, we summar-ize eight potential mimicry complexes involving Eterusia (Table 3) according to their similarity in wing pattern to the human eye. We also examine their sympatric distributions with potential co-mimics. We are mainly interested in two ques-tions: (1) are these suspected mimicry complexes merely arbi-trary assemblages resulting from peculiarities of human visual perception? and (2) do these mimicry complexes fit the
Ta ble 1. The non-Ete rusia chalcosiin e species that were asso ciated with Eterusia (¼ He terusia , Devanica ) and thei r present taxonomi c placem ents. Spec ies/subspecies Original He ring (1922 ) Bryk (1936 ) E n d o & Kishid a (1999 ) Yen (2003c) Yen (2 004a) ; Yen et al ., in press ferre a Walk er, 1854 Eterusia Syno nym of C yclosia papil ionar is ni Swin hoe, 1919 Cyclosia He terusia Eterusia Syn onym of Cyclosia papil ionar is dich roa Jordan, 190 Eterusia He terusinula Heterusinula He terus inula He terusinula Het erusinula circinat a Herrich-Scha ¨ffer, 1854 Heterusia Sorit ia Soritia So ritia Pid orus Pid orus fascia ta Walk er, 1869 Eterusia Syno nym of circ inata culo ti Obert hu ¨r, 1910 Eterusia Ete rusia Eterusia Ete rusia Ete rusia Pid orus duberna rdi Obe rthu ¨r, 1911 Eterusia Ete rusia Eterusia Ete rusia Ete rusia Pid orus bico lor Moore , 1884 Devanica Sorit ia Soritia So ritia Sorit ia Place d in a ne w genu s lata Jo rdan, 1907 Eterusia Sorit ia Soritia So ritia Sorit ia Place d in a ne w genu s circumd ata circ umdat a Walk er, [1865 ] Eterusia Pseudo scap tesyle Pseudosca ptesy le Pse udoscap tesyle Pseudo scap tesyle Pseudo scapt esyle circumd ata purp uralis Jorda n, 1907 Eterusia Pseudo scap tesyle Pseudosca ptesy le Pse udoscap tesyle Pseudo scap tesyle Pseudo scapt esyle moe rens Obert hu ¨r, 1910 Eterusia Ete rusia Soritia So ritia Sorit ia Place d in a ne w genu s mirifica Swin hoe, 1903 Corma Pro sopandro phila Prosopa ndrophila Ete rusia Pro sopandro phila Place d in a ne w genu s repl eta Walker, 1864 Eterusia Ete rusia Eterusia Ete rusia Ete rusia Place d in a ne w genu s alom pra Moore , 1879 Eterusia Syno nym of repl eta uran ia Scha us, 1890 Eterusia Syno nym of repl eta subm argin alis Swinhoe , 1892 Eterusia Syno nym of repl eta pulc hella pulc hella Kollar, 1844 Eterusia Sorit ia Soritia So ritia Sorit ia Sorit ia triliturata Walk er, 1864 Eterusia Syno nym of Sori tia pulc hella pulc hella cicada C . Felder & R. Felder, 1874 Heterusia Syno nym of Sori tia pulc hella pulc hella flavo maculat a Mo ¨sch ler, 1872 Heterusia Syno nym of Sori tia pulc hella pulc hella octop unctat a Mo ¨schler, 1872 Heterusia Syno nym of Sori tia pulc hella pulc hella pulc hella unipu nctata Duf rane, 1936 Eterusia –– – – Sorit ia
lepta linoid es Strand , 1916 Eterusia So ritia Soritia Soritia pulc hella strand i (replacement name ) majo r Jo rdan, 1907 Eterusia So ritia Soritia Soritia So ritia So ritia terioid es Mell, 1922 Het erusia M imascaptesyle Mimascaptesyle Mimascaptesyle S o ritia So ritia eliz abetha W alker, 1854 Eterusia So ritia Soritia Soritia So ritia So ritia micr ocepha la C. Fe lder & R. Fe lder, 1874 Eterusia Syn onym of elizabethae dirup ta W alker, 1864 Eterusia Syn onym of elizabethae prop rimar ginata Prout , 1918 Eterusia He terus ia Eterusia Eterusia So ritia So ritia lacr euzei Obert hu ¨r, 1910 Eterusia Ete rusia Eterusia Eterusia So ritia So ritia shah ama Moore , 1865 Eterusia So ritia Soritia sem iflava Talbo t & Joicy, 1922 Eterusia –– – So ritia So ritia dist incta dist incta Gue ´rin-Me ´ne ville, 1843 Gynau tocera Pro sopa ndrophila Prosopa ndro phila Eterusia Pro sopa ndrophila Pro sopa ndrophila dist incta drat araja M oore, 1859 Eterusia Pro sopa ndrophila Prosopa ndro phila Eterusia Pro sopa ndrophila Pro sopa ndrophila dist incta albin a Jo rdan, 1907 Eterusia Pro sopa ndrophila Prosopa ndro phila Eterusia Pro sopa ndrophila Pro sopa ndrophila yosh imotoi Kish ida & En do, 1999 Eterusia –– Eterusia Pro sopa ndrophila Pro sopa ndrophila xan thina Jorda n, 1907 Eterusia Pro sopa ndrophila Prosopa ndro phila Eterusia Pro sopa ndrophila Pro sopa ndrophila rem ota Walk er, 1862 Eterusia C halcosia Chalc osia Neocha lcosia Neo cha lcosia Neo chalcos ia euc hromioide s Walk er, 1864 Eterusia Pid orus Pidorus Syno nym of Pseudo pidorus fa sciatus phala ena ria pulchella Herrich-S cha ¨ffer, 1854 Het erusia Syn onym of Chalcosia phala ena ria phalaen aria (Gu e´rin-Me ´neville, 1843) subn igra Bet hune-Ba ker, 1911 Eterusia Ete rusia Eterusia Eterusia Eu spha lera Eu spha lera hamp soni hamp soni Holland, 1900 Het erusia Eu corma Eucorm a Eucorm a Eu spha lera Eu spha lera ligat a Rothsc hild, 1903 Het erusia Ete rusia Eterusia Eterusia Eu spha lera Eu spha lera pictu rata pictu rata Ta lbot & Joicy, 1922 Eterusia Ete rusia Eterusia Eterusia Eu spha lera Eu spha lera regi na Rothsc hild, 1903 Het erusia Ete rusia Eterusia Eterusia Eu spha lera Eu spha lera sem iflava Rothsc hild, 1904 Het erusia Ete rusia Eterusia Eterusia Eu spha lera Eu spha lera ve nus Rothsc hild 1915 Het erusia Ete rusia Eterusia Eterusia Eu spha lera Eu spha lera raja M oore, 1859 Eterusia Ete rusia Eterusia Eterusia Ete rusia Pl aced in a new genu s
phylogenetic predictions proposed by Simmons & Weller (2002) for either quasi-Batesian mimicry (Speed, 1993) or Mu¨llerian mimicry (Mu¨ller, 1879)? In Speed’s theory of quasi-Batesian mimicry, mildly unpalatable species may, by mimicking highly unpalatable species, increase the overall attack rate on the model so that this Mu¨llerian mimicry may be ‘parasitic’, or ‘quasi-Batesian’, in species of unequal unpalatability (Mallet, 2001). Simmons & Weller (2002) pre-dict that taxa possessing similar colour patterns within a single lineage under a quasi-Batesian mimicry framework should appear to be convergent. By contrast, classical Mu¨l-lerian mimicry theory suggests that the unpalatable species that are involved in a mimicry complex are mutually bene-ficial. Therefore, the taxa possessing a mimetic pattern within a single lineage are supposed to evolve from a common ancestor and to form a monophyletic clade. We selected three of the potential mimicry types observed by Owada & Ta (2002) (hereafter ‘black’, ‘blue–white’ and ‘yellow’) for phylogenetic tests because these types represent multiple
mimic model systems within a single lineage and because their comimic chalcosiine species are sympatric. We did not include mimicry systems that included only a single Eterusia species, because these do not produce meaningful phyloge-netic signals for the tests employed. In contrast to Simmons & Weller’s (2002) study on wasp moths (Arctiidae, Ctenuchini), considered as having excellent mimicry relationships with polybiine wasps, we did not ‘set’ a model taxon for each potential mimicry complex for Chalcosiinae and their comi-mics because (1) there is no species involved significantly more abundant than another, and (2) if not using their pre-dators (presumably birds) to evaluate the putative mimetic relationships (see Ritland, 1991; Turner & Speed, 1999; Speed et al., 2000; Prudic et al., 2002), there is no available method to compare the unpalatabilities caused by different chemical defence mechanisms, e.g. cyanoglucosides in Eterusia moths vs alkaloids in their comimic agaristine moths. Also, we did not classify species as either ‘perfect’ or ‘imperfect’ mimics (Lindstro¨m et al., 1997) because we do not know the Table 2. The species groups of Eterusia defined for the genus-level phylogenetic analysis in Yen (2003c, 2004a; Yen et al. in press), previous generic placements and present contents.
Species Original
Hering (1922)
Bryk (1936)
Endo & Kishida (1999)
Yen (2003c)
Yen
(2004a) Present study aedeagroup
aedeaLinnaeus, 1763 Papilio (Heliconius) Eterusia Eterusia Eterusia Eterusia Eterusia Eterusia vitessagroup
vitessaYen, 2003a Eterusia – – – – Eterusia Eterusia binotatagroup
binotataMell, 1922 Eterusia Eterusia Eterusia Soritia Eterusia Eterusia Eterusia guanxianaYen, sp.n. – – – – – – Eterusia tricolorgroup
tricolorHope, 1841 Eterusia Eterusia Eterusia Eterusia Eterusia Eterusia Eterusia subluteaWalker, 1852 Eterusia – – – Eterusia Eterusia Eterusia lativittaMoore, 1879a
Eterusia Heterusia Eterusia Eterusia – Eterusia – watanabeiInoue, 1982 Eterusia – – Eterusia Eterusia Eterusia Eterusia taiwanaWileman, 1911 Heterusia Heterusia Eterusia Eterusia Eterusia Eterusia Eterusia nobuoiOwada, 1996 Eterusia – – Eterusia Eterusia Eterusia Eterusia austrochinensisYen, sp.n. – – – – – – Eterusia subcyaneagroupb
subcyaneaWalker, 1854 Eterusia Eterusia Eterusia Eterusia Eterusia Eterusia Eterusia joiceyiTalbot, 1929 Eterusia Eterusia Eterusia Eterusia Eterusia Eterusia Eterusia feminataKishida & Endo, 1999 Eterusia Eterusia Eterusia Eterusia Eterusia Eterusia Eterusia risagroup
risaDoubleday, 1844 Heterusia Soritia Soritia Eterusia Eterusia Eterusia Eterusia angustipennisRo¨ber, 1897 Heterusia Soritia Soritia Eterusia Eterusia Eterusia Eterusia rajagroup
rajaMoore, 1859 Eterusia Eterusia Eterusia Eterusia Eterusia New genus N/A repletagroup
repletaWalker, 1864 Eterusia Eterusia Eterusia Eterusia Eterusia New genus N/A culotigroup
culotiOberthu¨r, 1910 Eterusia Heterusia Eterusia Eterusia Eterusia Pidorus N/A dubernardiOberthu¨r, 1910c
Eterusia Heterusia Eterusia Eterusia Eterusia Pidorus N/A
N/A, not applicable.
a
Eterusia lativittais treated as a synonym of E. sublutea. See Appendix 2 for details.
b
This species group was included in the tricolor group in Yen (2003c).
c
This species was treated as a subspecies of culoti by Bryk (1936) and Endo & Kishida (1999). Owada & Horie (2002) revived its species status and transferred these two species to Pidorus. This generic placement is supported by Yen’s (2004a) phylogenetic analysis.
Ta ble 3. Th e poten tial mimicry comp lexes particip ated by the Ete rusia mo ths and thei r sym patri c comim ic insects. The mim icry types ana lysed in the presen t st udy are shown in bold. Comi metic specie s Mimic ry type Ete rusia species Zygae nidae (Chalc osiinae ) Noctu idae Pyralid ae Cicadae Distribu tion area s ‘blac k’ Ete rusia aedea aedea (edocla form) Eterusia sublu tea sublu tea (lati vitta form) Scob iger a amatri x (blac k form) Exsula vic trix Tosena fasciat a Northeast India to north Indochin a ‘blue –white’ Ete rusia aedea formos ana Eterusia ta iwana Taiwan ‘yellow’ Ete rusia aedea aedea (magnif ica form) Eterusia tr icolor com plex Erasmia obliq uaria Soriti a shahama Scob iger a amatri x (orang e form) Gaeana mac ulata , G. paviei Northeast India to north Indochin a Ete rusia tricolo r Eterusia sublu tea sublu tea Scob iger a amatri x (orang e form), Sorit ia shah ama (orang e form) Gaeana vari egata, G. haina nensis Northeast India to south China Ete rusia nobuoi Eterusia sublu tea yuc hii , E. tricolo r Scob iger a amatri x (orang e form) Gaeana vari egata , G. paviei Southwes t C hina to north Indochin a Ete rusia austro chinens is Scob iger a amatri x (orang e form) Gaeana vari egata South China Ete rusia sublutea sublu tea (orang e form) Eterusia tr icolor, So ritia shah ama Scob iger a amatri x (orang e form) Northeast India Ete rusia sublutea yuc hii Eterusia tr icolor , E. nobu oi Scob iger a amatri x (orang e form) Gaeana mac ulata North Indochin a Ete rusia sublutea yasu norii Erasmic obliq uaris Scob iger a amatri x (orang e form) Gaeana pav iei North Thailan d-Myan mar ‘V itessa ’ Ete rusia vitessa sera mensis Vi tessa ta lboti, V. ze mire V. ternati ca, V . hemiallac tis molu ccana, He ortia sp. Seram (I ndones ia) Ete rusia vitessa vites sa Vi tessa sple ndida , V. philip pina, V. pyralina tr iangulif era, He ortia sp. Mindan ao (Phil ippines) ‘band -1’ Ete rusia risa palawan ica (oran ge form ) (F) Erasmia obliq uaris Palawan (Philip pines) ‘band -2’ Ete rusia risa palawan ica (yello w form ) (F) Euspha lera sp., Psaphis albiv itta semixantha (yello w form) (F) Palawan (Philip pines) ‘band -3’ Ete rusia binota ta (F) Ete rusia guanxia na (F) Soriti a bicolor (F) Indoch ina to sou th China ‘band -4’ Ete rusia femin ata (F) Psaphis albiv itta semixantha (white form ) (F), Psa phis azurea (F), Prosopa ndro phila yoshi motoi (F) Mindan ao (Phil ippines) F, female only.
discriminatory abilities of their predators, which, in any case, are probably very varied.
Materials and methods
Taxa studied
The taxa studied are listed in Table 4, together with their collection localities and the depository institutions and col-lections, and illustrated in Figs 1 and 2 (see also Owada, 1989, 2001; Endo & Kishida, 1999; Owada & Horie, 1999). All species of the redefined Eterusia were included in the analyses. All the type specimens, except for that of Papilio aedeaLinnaeus (¼ E. aedea aedea), which has no surviving
type (Yen, 2004b), were examined. Every subspecies of the polytypic E. aedea was treated as a different terminal taxon because of their extremely variable wing patterns. The three major colour forms of E. aedea aedea were further separated into three terminal taxa because merging them into a single taxon would have involved coding most of the colour pat-tern characters as polymorphic, thus losing phylogenetic signal. Several unrevised subspecies of both E. risa and E. angustipennisare recognized (e.g. Endo & Kishida, 1999; see also Appendix 1), but morphological differentiation is too subtle for analysis. As this species group forms no mimicry complex, we made no separation into different terminal taxa. Two species representing two major lineages of Soritia (Yen, 2003b), the sister genus of Eterusia, were used as outgroup: thirty-two ingroup taxa were represented.
Table 4. Specimens examined for the taxa included in the phylogenetic analyses. Taxon Materials examined Soritia strandi 2?, 2/, Taiwan (NSYSU) Soritia azurea 1?, 1/, Taiwan (NSYSU)
Eterusia aedea aedea(typical form) 10?, 10/, north India, Myanmar (BMNH) Eterusia aedea aedea(edocla form) 10?, 10/, north India, Myanmar (BMNH) Eterusia aedea aedea(magnifica form) 10?, 10/, north India, Thailand, Vietnam (BMNH) Eterusia aedea sinica 10?, 10/, China (BMNH, NSYSU)
Eterusia aedea formosana 4?, 4/, Taiwan (NSYSU) Eterusia aedea sugitanii 4?, 4/, Japan (Honshu) (BMNH) Eterusia aedea micromaculata 1?, 1/, Japan (Tokarashima) (BMNH) Eterusia aedea masatakatoi 1?, 1/, Japan (Nakanoshima) (NSYSU) Eterusia aedea tomokunii 1?, 1/, Japan (Amami o-shima) (NSYSU) Eterusia aedea hamajii 1?, 1/, Japan (Tokunoshima) (NSYSU) Eterusia aedea sakaguchii 2?, 2/, Japan (Okinawa) (NSYSU) Eterusia aedea azumai 1?, 1/, Japan (Kumeijima) (NSYSU) Eterusia aedea okinawana 3?, 4/, Japan (Iriomotejima) (NSYSU) Eterusia aedea virescens 3?, 3/, south India (NSYSU) Eterusia aedea cingala 3?, 3/, Sri Lanka (BMNH, NSYSU) Eterusia vitessa 3?, Philippines (CTC, WMM)
Eterusia risa 10?, 20/, Indonesia, Philippines (BMNH, RMNH, ZMHB) Eterusia angustipennis 1?, 4/, Indonesia (BMNH, MZB, NMNH, RMNH, ZMHB) Eterusia binotata 2?, 1/, China (NSYSU)
Eterusia guanxiana 1/, China (NSYSU)
Eterusia feminata 3?, 2/, Philippines (CTC, SNG, NMNH) Eterusia subcyanea 2?, 2/, Indonesia (BMNH)
Eterusia joiceyi 1/, Sumatra (BMNH)
Eterusia tricolor 3?, 3/, north India, Thailand (BMNH, OXUM, NSYSU) Eterusia watanabei 1?, 2/, Japan (Tsushima Island), China (Jejiang) Eterusia taiwana 2?, 2/, Taiwan (NSYSU)
Eterusia nobuoi 1?, 1/, China (Sichuan) (BMNH) Eterusia austrochinensis 1?, China (Guangdong) (ZMHB) Eterusia sublutea sublutea 10?, north India (BMNH) Eterusia sublutea(fasciata form) 3?, 2/, north India (BMNH) Eterusia sublutea yuchii 3?, 1/, Vietnam (NSYSU) Eterusia sublutea yasunorii 1?, 1/, Thailand (NSYSU)
BMNH, The Natural History Museum, London, U.K.; CTC, Colin Treadaway Collection, Limbach-Wagenschwend, Germany; DEI, Deutsches Entomologisches Institut, Eberswalde, Germany; MZB, Museum Zoologicum Bogoriense, Bogor, Java, Indonesia; NMNH, National Museum of Natural History, Smithsonian Institution, Washington DC, U.S.A.; NSYSU, National Sun Yat-Sen University, Kaohsiung, Taiwan; OXUM, Hope Entomological Collections, University Museum, Oxford, U.K.; RMNH, Rijksmuseum van Natuurlijke Historie, Leiden, the Netherlands; SNG, Forschungsinstitut und Naturmuseum Senckenberg, Frankfurt am Main, Germany; WMM, Thomas Witt Museum, Mu¨nchen, Germany; ZMHB, Museum fu¨r Naturkunde, Humboldt University, Berlin, Germany.
Fig. 1. A potential mimicry complex participated by Eterusia moths and other insects. From left to right, top to bottom: row 1, E. sublutea, E. tricolor(male), E. nobuoi, E. tricolor (female), E. aedea aedea (magnifica form, male); row 2, Gaeana variegata, Gaeana maculata, Gaeana paviei, Soritia shahama; row 3, Scobigera amatrix (male), Scobigera amatrix (female), E. aedea sinica (yellow form, female), E. aedea formosana (yellow form, female). A, Visible light patterns; B, ultraviolet reflectance patterns shown against an ultraviolet reflecting background; C, ultraviolet-induced fluorescence patterns
Fig. 2. A potential mimicry complex participated by Eterusia moths and other insects. Tosena fasciata (middle top); Eterusia aedea aedea (black form, female) (left top, left bottom, right top); Exsula victrix (middle top); Eterusia sublutea (black form, female) (middle bottom); Scobigera vulcania (right bottom). A, Visible light patterns; B, ultraviolet reflectance patterns shown against an ultraviolet reflecting background; C, fluorescence patterns
Morphology
Morphological terminology for Lepidoptera varies with author and taxon: some terms (e.g. transtilla, socii) remain inconsistently and confusingly used. Terms or descriptive phrases used here have some degree of consensus as to meaning and homology. Generally we follow Klots (1970) and Scoble (1992). but for copulatory structures (genitalia and the associated abdominal segments) and wing patterns, we follow Yen (2003c, 2004a).
Dissections were made using standard techniques (Holloway et al., 1987), but the specialized eighth abdominal segment of males was separated from the remaining segments and slide mounted using a supporting plastic ring under the coverslip so as to maintain its shape.
Examining mimetic wing patterns using different lights
The main predators of Eterusia moths and their potential comimics are probably birds. Within the distribution range of Eterusia moths, no lizards (e.g. Agamidae) or frogs (e.g. Rhacophoridae) have been found that might be able to catch these day-flying moths. Orb-web-weaving spiders, such as giant wood spiders (Nephila spp.), are known to prey on various kinds of animals, but the web sites of Nephila are usually much lower than the flight height of Eterusiamoths (Yen, pers. obs.). We therefore excluded the possibility that reptiles, amphibians or spiders are signifi-cant predators of Eterusia. Birds are known to have four-pigment colour vision (Baker & Parker, 1979; Thompson et al., 1992; Cuthill et al., 2000; Church et al., 2001), includ-ing some ability to detect radiation in the near ultraviolet (UV) part of the spectrum (Cuthill et al., 2000). Therefore, both visible and near UV patterns may be relevant to mimi-cry systems. The use of UV reflectance patterns in lepidop-teran research has been widely recognized in clarifying taxonomy of sibling species and investigating inter- and intraspecific recognition (Silberglied & Taylor, 1978; Knu¨ttel & Fiedler, 2001) and sexual selection (Brunton & Majerus, 1995). Most of these studies have concentrated on pierid and papilionid butterflies because these possess highly conspicuous UV patterns (Brunton, 1998) and none of these patterns has been examined in zygaenid moths.
Here we used a simple photographic system to investigate the visual signals presented by the upperside of the moths’ wings and abdomens, revealing whether there are any areas of the wings that reflect UV light, as in many species such areas are reported to have little or no correlation with the human-visible pigmentation patterns (Silberglied & Taylor, 1978; Silberglied, 1984). Additionally, the appearance of the insects may be influenced by fluorescence, especially in dimly lit situations. Cockayne (1924) studied the different appearance of white and yellow areas of different mimetic forms of female Papilio dardanus in the visible spectrum by use of UV-derived fluorescent light, and showed that the different mimetic forms had different degrees of fluores-cence. As yellow and white colours are both important
elements in all mimicry types of Eterusia moth, we also examined the specimens for fluorescence.
Set specimens of Eterusia moths and their potential comimic insects were taken from the collections of the Natural History Museum, London, U.K. (BMNH) and Shen-Horn Yen Collection, Taipei, Taiwan (SHYC) (now deposited in NSYSU), and collection labels were removed to avoid unwanted reflections. A Fuji Finepix 2 digital camera was set up on a camera stand in a dark room. Kodak 18A and 1A filters were used for the UV transmission and fluorescence studies, respectively. The camera was mounted directly above the specimens. The UV lamp was a wide waveband UV source. We used powdered tin chloride (SnCl22H2O) as the
‘white’ reference (Frank Greenaway, pers. comm.). A Kodak colour separation grey scale (Kodak Graphic Imaging Sys-tems) was also used.
Phylogenetic analyses
Selection, organization and coding of characters. There is no doubt that immature stages can be rich sources of char-acters for phylogenetic analyses. Explicit cladistic analyses incorporating data from immature stages have been under-taken on various groups of Lepidoptera. However, in the present study, characters of immature stages were not used because the available information is very incomplete and the known larvae of Eterusia species are either polymorphic within a species or indistinguishable among unrelated spe-cies (Yen, pers. obs.). In general, wing pattern characters are not favoured in character selection because, especially in a group that demonstrates a high sexual dimorphism, poly-morphism and diverse mimicry patterns, the associated variation is expected to increase homoplasy and decrease resolution and reliability of the phylogeny. By being explicit about our character selection criteria we aimed to increase the rigor of our analyses by minimizing subjectivity. Wing pattern characters were included in the present study in reconstructing the initial most-parsimonious phylogeny as suggested by the ‘total evidence’ approach (e.g. Chavarria & Carpenter, 1994; Huelsenbeck et al., 1996).
For this study, eighty morphological characters were identified and partitioned into the following character sub-sets according to their morphological correlations and bio-logical systems: adult head, thorax, wing shape and wing patterns, pregenital abdominal segments 1–7, the eighth abdominal segment, male genitalia and female genitalia. The following references provided information on some characters that were not accessible during the study: Endo & Kishida (1999), Owada & Horie (1999) and Owada (2001). Both binary and multistate characters were employed, all coded as unordered and without a priori weights applied (equal weights analyses: EW). Taxa with more than one state for a given trait were scored as polymorphic (e.g. the forewing pattern of each colour form of E. aedea aedea). When the character state distribution of a character differed between males and females (e.g. wing patterns), the char-acter was coded for males and females separately. The data
matrix was constructed and manipulated usingMACCLADE
4.0 (Maddison & Maddison, 2000), and is presented in Appendix 2. The nexus file is available as an electronic supplement (see Supplementary material). All the charac-ters used in the cladistic analysis are listed in Appendix 3.
Eterusia phylogeny and support measures. Maximum par-simony trees (MPTs) were built for all thirty-four taxa and eighty characters (nineteen binary, sixty-one multistate) usingPAUP* 4.0b10 (Swofford, 1998). Multistate characters
were interpreted as polymorphic rather as uncertain. An initial heuristic search was performed with 10 000 random additions followed by tree bisection reconnection (TBR) branch swapping with no more than one tree held during each search. Multitrees was set to 10 000 and the shortest trees from this initial search were used as starting trees for further TBR swapping. The default setting of character weighting in PAUP* uses the maximum value of the
con-sistency index (CI) as the reweighting factor. However, as Gauthier et al. (2000) and Quicke et al. (2001) have argued that using the rescaled consistency index (RC) or the CI is not suitable, we used the maximum value of the retention index (RI) as the reweighting factor. The resulting phylogenies were compared with those based on EW by generating a strict consensus tree of the MPTs found with EW and successive approximation character weighting (SAW) analyses.
We applied three methods to measure the support and confidence statistics for cladograms. Bremer support (Bremer, 1994) values were calculated usingPAUP* by
con-straining each of the clades recovered by the initial tree, and then following the same protocol described above for searching for MPTs. We also applied both bootstrapping and jackknifing for the whole cladogram and compared their levels of support for each of the clades. For bootstrap and jackknife calculations we used 200 replicates of 10 000 random additions (maxtrees¼ 10 000). A deletion rate of 36.79% for jackknife resampling, suggested by Farris et al. (1996), was adopted. When calculating bootstrap values, uninformative characters were excluded, as suggested by Bryant (1995).
To evaluate the influence of different character subsets on topology, we first implemented the incongruence length difference (ILD) test, as described by Farris et al. (1994) and discussed by Mason-Gamer & Kellogg (1996), DeSalle & Brower (1997) and Dolphin et al. (2000). We partitioned the data matrix into colour patterns (forty-nine characters) vs noncolour patterns (thirty-one characters). We aimed to investigate if the phylogenies generated by different sub-sets tell different evolutionary stories. The ILD test was implemented using heuristic search. For all analyses, 1000 ILD replicates and 1000 random additions were used to estimate the null distribution, and the value of maxtrees was set to 1000. All tests were run with uninformative characters excluded, as recommended by Cunningham (1997).
We conducted another method to compare the trees gen-erated by the data matrix with different character subsets inactivated. We were particularly interested in the charac-ters relevant to mimetic colour patterns and copulatory
structures, so we assigned each of these characters to one of eight categories: (1) all patterns (including colours and shapes); (2) copulatory structures; (3) cephalic colour pat-terns; (4) thoracic colour patpat-terns; (5) wing patpat-terns; (6) abdominal colour patterns; (7) all patterns (colours only); and (8) all patterns (shapes only) (see Table 5). We con-structed MPTs using the same protocol as described above for the whole dataset with each of these groups deactivated. We compared their topological differences with the initial working MPTs using the Templeton test (Templeton, 1983) inPAUP*.
Testing mimicry scenarios. It has been argued that inclusion of the characters that are then being analysed in the tree-building character matrix is circular (e.g. Brooks & McLennan, 1991). However, it has also been argued that inclusion is desirable as it involves analysing all avail-able evidence and the trees obtained using the greatest evidential basis are most likely to be correct (e.g. Vane-Wright et al., 1992; Deleporte, 1993; Luckow & Bruneau, 1997; Zrzavy´, 1997; Zrzavy´ & Nedved, 1999; Kitching, 2002). In the present study, we used both methods, namely including and excluding colour pattern characters. We first examined if excluding the colour pattern characters had a significant influence on the phylogenetic relationships of the taxa (by deactivating those characters), and then used the phylogenetic hypotheses from both treatments for the following tests of the mimicry scenarios.
The three types of mimicry (yellow–orange, black, and blue–white) (Tables 6 and 7) were mapped on to the strict consensus tree from combined EW and SAW results. We examined if the individual mimetic characters evolved as suits or independently by tracing the distribution of each character of interest, and we then examined whether the three mimetic types and the selected characters were phy-logenetically conserved, using the permutation–tail– probability (PTP) test (Faith & Cranston, 1991; Maddison & Slatkin, 1991) as suggested by Simmons & Weller (2002). We used 1000 randomly generated character states for each of the interesting mimetic characters and maintained the observed frequencies. The random character distribution was mapped on to the strict consensus tree (EW and SAW). We compared the mean length (L) and CI of the recon-structed random character vs L and CI of the observed character using a z score (Gravetter & Wallnau, 1992), as suggested by Simmons & Weller (2002). If the observed character is not conserved, then L and CI for the mimetic characters will approach that of random permutations.
To examine if the observed phylogeny was more similar to simulated quasi-Batesian or Mu¨llerian phylogenetic patterns, we generated topologies by moving the minimum number of taxa within, and among, clades to obtain the predicted phylogenetic patterns (Figs 12B, C, 13B, C). L for the nonmimetic characters and overall CIs for these hand-generated topologies were compared with L and CI for the 1000 randomly generated trees using Kishino–Hasegawa tests (Kishino & Hasegawa, 1989). We also generated four different topological constrains for each mimetic
Table 5. List of synapomorphies for the main clades of Eterusia. Synapomorphies
All characters included Colour patterns excluded
Taxon EW SAW EW SAW
7 (1!0) 7 (1!0) x x 11 (1fi0)* 11 (1fi0)* x x 17 (7fi0) 17 (7fi0) x x 22 (0!3) 22 (0!3) x x 23 (0!2) 23 (0!2) x x 29 (2!1) 29 (2!1) x x 33 (3!4) 33 (3!4) x x 36 (0!1) 36 (0!1) x x 37 (0!1) 37 (0!1) x x 44 (0!1) 44 (0!1) x x 48 (1fi0) 48 (1fi0) x x 50 (0!1) 50 (0!1) x x 51 (2fi1) 51 (2fi1) x x 52 (0fi1) 52 (0fi1) 52 (0fi1) 52 (0fi1) 53 (0fi1)* 53 (0fi1)* 53 (0fi1)* 53 (0fi1)* Eterusia 54 (0!1) 54 (0!1) 54 (0!2) 54 (0!2) 57 (0!1) 57 (0!1) 57 (0!2) 57 (0!2) 58 (0fi1)* 58 (0fi1)* 58 (0fi1)* 58 (0fi1)* 60 (0fi1)* 60 (0fi1)* 60 (0fi1)* 60 (0fi1)* 62 (0fi1)* 62 (0fi1)* 62 (0fi1)* 62 (0fi1)* 64 (0!2)* 64 (0!2)* 64 (0!2)* 64 (0!2)* 65 (0fi1)* 65 (0fi1)* x x 67 (0!1) 67 (0!1) 67 (0!1) 67 (0!1) 68 (0fi2) 68 (0fi2) 68 (0fi2)* 68 (0fi2)* 69 (0!1) 69 (0!1) x x 71 (0fi1)* 71 (0fi1)* 71 (0!1) 71 (0!1) 72 (0fi3)* 72 (0fi3)* 72 (0!1) 72 (0!1) 73 (0fi1)* 73 (0fi1)* 73 (0!1) 73 (0!1) 74 (0!1)* 74 (0!1)* 74 (0!1)* 74 (0!1)* 76 (0!2)* 76 (0!2)* 76 (0!2)* 76 (0!2)* 79 (0!1)* 79 (0!1)* 79 (0!1)* 79 (0!1)* (E. binotataþ E. guanxiana) þ ((E. risa þ
E. angustipennis)þ E. aedea) species group
69 (0fi1)* 69 (0fi1)* NRC NRC 76 (2!3)* 76 (2!3)* 77 (1fi2)* 77 (1fi2)* 2 (0!1)* 2 (0!1)* 7 (0fi2) 7 (0fi2) 21 (4!1) 21 (4!1) 23 (2!6) 23 (2!6) x 24 (2fi0) 29 (1fi3) 29 (1fi3) 31 (0fi1) 31 (0fi1) 33 (4!1) x 34 (0!2) 34 (0!2)
E. risaspecies group 36 (1!0) x NRC NRC 37 (1!0) x 40 (0fi2)* 40 (0fi2)* 44 (1!3) x 54 (1!2) 54 (1!2) 57 (1!2) 57 (1!2) 74 (1!2)* 74 (1!2)* 76 (2!5)* 76 (2!5)* 77 (1fi2)* 77 (1fi2)* 79 (1!2)* 79 (1!2)* 1 (0fi2) 1 (0fi2) 1 (0fi2) 1 (0fi2) 2 (1!2)* 2 (1!2)* 2 (1fi2)* 2 (1fi2)* 4 (0fi2)* 4 (0fi2)* 3 (0!2) 3 (0!2) 5 (0fi2)* 5 (0fi2)* x x
Table 5. Continued.
Synapomorphies
All characters included Colour patterns excluded
Taxon EW SAW EW SAW
13 (0fi1)* 13 (0fi1)* x x 16 (0fi2) 16 (0fi2) x x E. risaþ E. angustipennis 17 (0fi2) 17 (0fi2) x x 22 (3!1) 22 (3!1) x x 23 (6!7) 23 (6!7) x x 41 (0fi2) 41 (0fi2) x x 42 (0fi2) 42 (0fi2) x x 46 (0fi1) 46 (0fi1) x x 49 (0fi1) 49 (0fi1) x x 66 (0fi1) 66 (0fi1) x x 68 (2!3) 68 (2!3) x x x x 74 (1fi2)* 74 (1fi2)* x x 76 (3!5)* 76 (3!5)* 79 (2!3)* 79 (2!3)* 79 (2!3)* 79 (2!3)* 8 (0fi2) 8 (0fi2) 2 (0fi1)* 2 (0fi1)* 21 (1!6) 21 (1!6) x x 27 (0fi3) 27 (0fi3) x x 28 (0!4) 28 (0!4) x x 30 (0!4) 30 (0!4) x x 32 (0!4) 32 (0!4) x x 33 (1!2) 33 (1!2) x x 34 (2!3) 34 (2!3) x x E. binotataþ E. guanxiana 43 (0!3) 43 (0!3) x x 50 (1!0) 50 (1!0) x x 51 (1!0) 51 (1!0) 67 (1!2) 67 (1!2) 67 (1!2) 67 (1!2) 72 (3!1)* 72 (3!1)* x x 74 (2!3)* 74 (2!3)* 74 (1!3)* 74 (1!3)* 75 (0!1)* 75 (0!1)* 75 (0!1)* 75 (0!1)* 76 (5!7)* 76 (5!7)* 76 (5!7)* 76 (5!7)* x x 79 (1!2)* 79 (1!2)* 80 (1fi2)* 80 (1fi2)* 80 (1fi2)* 80 (1fi2)* 1 (0!1) 15 (1!0) 19 (0!1) 25 (0!2)
(E. tricolorþ E. subcyanea) þ E. aedea NRC 26 (0!1) NRC NRC
species groups 27 (0!1) 33 (1!4) 35 (0!1) 36 (0fi1) 37 (0fi1) 1 (0fi1) x 1 (0fi1) 1 (0fi1) 18 (0!1) x x x 19 (0!3) x x x 21 (4!3) 21 (4!3) x x 23 (2!3) 23 (2!3) x x 25 (0!1) x x x 26 (0!1) x x x 27 (0fi1) x x x 54 (1!4) 54 (1!4) 54 (1!4)* 54 (1!4)* 57 (1!4) 57 (1!4) 57 (1!4)* 57 (1!4)* E. tricolorþ E. subcyanea species groups 62 (1fi2)* 62 (1fi2)* 62 (1!2)* 62 (1!2)* 63 (1fi2)* 63 (1fi2)* 63 (1fi2)* 63 (1fi2)* 64 (2fi3)* 64 (2fi3)* 64 (2!3)* 64 (2!3)* 65 (1fi2)* 65 (1fi2)* 65 (1!2) 65 (1!2) 67 (1!4) 67 (1!4) 67 (1!4) 67 (1!4)
69 (1!0) 69 (1!0) x x 70 (0fi1)* 70 (0fi1)* 70 (0fi1)* 70 (0fi1)* 71 (1fi2)* 71 (1fi2)* 71 (1!2)* 71 (1!2)* 72 (3fi4)* 72 (3fi4)* 72 (1!4)* 72 (1!4)* 73 (1fi2)* 73 (1fi2)* 73 (1!2)* 73 (1!2)* 74 (1!4)* 74 (1!4)* 74 (1!4)* 74 (1!4)* 79 (1!5)* 79 (1!5)* 79 (1!5)* 79 (1!5)* 15 (1fi2) x x x 21 (3!5) 12 (3!5) x x 22 (3!5) 22 (3!5) x x 24 (0!1) x x x 26 (1!2) x x x 28 (0fi4) 28 (0fi4) x x 29 (1fi4) 29 (1fi4) x x 30 (0fi4) x x x 32 (0fi3) 32 (0fi3) x x E. tricolorspecies group 33 (4!5) 33 (4!5) x x
35 (0fi1) x x x 36 (1!3) 36 (1!3) x x 37 (1!3) 37 (1!3) x x 39 (2fi1) 39 (2fi1) x x 43 (0fi2) x x x 44 (1!0) 44 (1!0) x x 46 (0fi2) 46 (0fi2) x x 47 (0!2) x x x 49 (0fi1) 49 (0fi1) x x 51 (1fi0) x x x 52 (1fi2) x x x 54 (4!5) x x x 67 (4!5) 67 (4!5) 67 (4!5) 67 (4!5) 78 (0fi2)* 78 (0fi2)* 78 (0fi2)* 78 (0fi2)* x 18 (0fi2) x x 24 (1fi2) x x x 25 (1fi2) x x x 26 (2fi1) x x x 27 (1fi2) 27 (1fi2) x x 30 (4fi3) 30 (4fi3) x x 31 (0fi3) 31 (0fi3) x x 34 (0fi6) 34 (0fi6) x x 35 (1fi3) 35 (1fi3) x x E. tricolorspecies group excluding 44 (0!2) x x x
E. sublutea 46 (2!1) x x x 47 (2fi3) 47 (2fi3) x x 51 (0fi2) 51 (0fi2) x x 52 (2fi1) x x x 54 (5fi4) x x x 56 (0fi1)* 56 (0fi1)* 56 (0fi1)* 56 (0fi1)* 57 (4fi5) x x x 76 (2fi6)* 76 (2fi6)* 76 (2fi6) 76 (2fi6) 79 (5fi4)* 79 (5fi4)* 79 (5fi4) 79 (5fi4) 16 (0fi2) 16 (0fi2) x x 28 (4!1) 28 (4!1) x x E. tricolorþ E. taiwana 29 (4!1) x x x 32 (3fi2) 32 (3fi2) x x 33 (5fi3) 33 (5fi3) x x 67 (5fi6) 67 (5fi6) 67 (5!6) 67 (5!6) 23 (3!4) x x 24 (2fi1) x x 25 (2fi1) x x E. sublutea NRC 26 (1fi2) x x 46 (1!2) x x 52 (1fi2)* 52 (1fi2)* 52 (1fi2)* 54 (4fi5)* 54 (4fi5)* 54 (4fi5)* 57 (5fi4)* 57 (5fi4)* 57 (5fi4)*
Table 5. Continued.
Synapomorphies
All characters included Colour patterns excluded
Taxon EW SAW EW SAW
3 (0fi1) 3 (0fi1) 3 (0fi1) 3 (0fi1) x 15 (0!) x x 16 (0!2) 16 (0!2) x x 19 (3!4) 24 (2!0) x x x 35 (1!0) x x x x x x
E. subcyaneaspecies group 40 (0!1) 40 (0!1) x x 41 (0fi1) 41 (0fi1) x x 42 (0fi1) 42 (0fi1) x x 50 (1!0) 50 (1!0) x x 55 (0!1)* 55 (0!1)* 55 (0fi1)* 55 (0fi1)* 57 (4!5) 57 (4!5) x x 68 (2!1) 68 (2!1) 68 (2fi1)* 68 (2fi1)* 17 (0fi2) 17 (0fi2) E. feminataþ E. subcyanea x 25 (2!0) NRC NRC 29 (1!2) 29 (1fi2) 3 (0fi3)* 3 (0fi3)* 3 (0fi3)* 4 (0!1)* 4 (0!1)* x 5 (0!1)* 5 (0!1)* x 14 (1!0) 14 (1!0) x 15 (1!0) 15 (1!0) x 17 (0!5) 17 (0!5) x 19 (0!1) 19 (0!1) x 22 (3!2) 22 (3!2) x 24 (0!2) – x 25 (0!2) – x E. aedeaþ E. vitessa 26 (0!3) 26 (0!3) NRC x 27 (0!2) 27 (0!2) x 28 (0fi1) 28 (0fi1) x 30 (0!1) 30 (0!1) x 31 (0!2) 31 (0!2) x 32 (0fi2) 32 (0fi2) x 34 (0¼ >4) 34 (0¼ >4) x 35 (0!1) 35 (0!1) x 43 (0fi1) 43 (0fi1) x 46 (0!3) 46 (0!3) x x x 54 (2!1)* x x 57 (2!1)* 61 (0!1) 61 (0!1) x 66 (0fi1) 66 (0fi1) x 68 (2!3) 68 (2!3) x 76 (2!3)* 76 (2!3)* x 77 (1!3)* 77 (1!3)* 77 (2!3)* 1 (0fi1) – 1 (0fi1) 1 (0fi1) 11 (0fi2)* 11 (0fi2)* x x 16 (0fi5) 16 (0fi5) x x 21 (4fi2) 21 (4fi2) x x 30 (1!2) 30 (1!2) x x 33 (4!3) 33 (4!3) x x 35 (1!2) 35 (1!2) x x 36 (1!2) 36 (1!2) x x E. aedea 37 (1!2) 37 (1!2) x x 39 (2fi0) 39 (2fi0) x x 45 (0fi1)* 45 (0fi1)* x x 49 (0fi1) 49 (0fi1) x x x x 54 (2fi1)* – x x 57 (2fi1)* –
Table 6. Results of Templeton tests that compared the trees based on the whole dataset and the trees with colour pattern and wing shape characters inactivated. The test for wing patterns shows that the null hypothesis must be rejected, whereas the test for wing shapes shows that the null hypothesis is accepted.
Treatment No. of characters excluded No. of characters included Tree length (steps)a No. of
treesb CI RI Test results Pc
All characters included 0 80 356 907 0.725 0.858 – All pattern characters deactivated 49 31 83 17 0.964 0.988 <0.0001* All copulatory structure characters deactivated 28 52 270 48 0.685 0.818 0.1025–0.3537 All cephalic pattern characters deactivated 3 77 348 868 0.718 0.853 1.000 All thoracic pattern characters deactivated 6 74 345 906 0.722 0.857 0.7630–1.000 All wing pattern characters deactivated 32 48 132 92 0.886 0.958 0.0348* All abdomen pattern characters deactivated 8 72 324 788 0.735 0.863 0.7055–0.7963 All colour characters deactivated 17 63 258 1092 0.721 0.870 0.5361–0.9837 All shape characters deactivated 31 49 180 3885 0.850 0.931 0.0588–0.1088
a
Under equal weighting. The numbers indicate the difference in length between the maximum parsimony and constrained trees.
b
Only the best tree from each constraint was included in this test.
c
All values are significant at P < 0.05. CI, consistency index; RI, retention index.
Table 7. Comparison of selected observed mimicry characters with randomly generated character distributions based on the phylogeny with all characters included (see Appendix 3 for detailed descriptions of characters).
EW SAW Observed characters 1000 random characters Observed characters 1000 random characters
Character L CI Mean L SD Mean CI SD L CI Mean L SD Mean CI SD Mimetic types: (0) nonmimetic; (1) blue–white;
(2) yellow–orange; (3) black 7* 0.43* 9.22 1.28 0.33 0.05 7* 0.43* 10.48 0.74 0.29 0.02 Character 16: male forewing ground colour 10* 0.50* 14.77 1.22 0.34 0.03 10* 0.50* 16.17 1.18 0.31 0.02 Character 17: female forewing ground colour 8* 0.63* 15.68 1.28 0.31 0.03 8* 0.63* 17.11 1.28 0.29 0.02 Character 21: male forewing medial zone 7* 1.00* 14.4 1.17 0.49 0.04 8* 0.88* 15.40 1.09 0.46 0.03 Character 23: female forewing medial zone 7* 1.00* 13.91 1.09 0.51 0.04 8* 0.88* 15.38 1.05 0.48 0.03 Character 24: male discoidal spot 6* 0.33* 10.74 1.38 0.19 0.03 6* 0.33* 12.65 1.16 0.16 0.02 Character 25: female discoidal spot 6* 0.33* 9.82 1.34 0.21 0.03 6* 0.33* 11.89 1.05 0.17 0.02 Character 26: male submarginal spots 4* 1.00* 14.71 1.34 0.28 0.02 5* 0.80* 16.62 1.37 0.24 0.02 Character 27: female submarginal spots 4* 0.75* 10.66 0.99 0.28 0.03 5* 0.60* 11.36 1.27 0.52 0.03 Character 28: male hindwing colour zone c 10* 0.50* 13.02 0.92 0.39 0.03 11* 0.45* 14.31 0.99 0.35 0.03 Character 29: female hindwing colour zone c 10* 0.50* 13.95 0.88 0.39 0.03 11* 0.45* 14.46 0.88 0.35 0.03 Character 32: male hindwing zone d–f 6* 0.67* 11.85 1.33 0.34 0.03 7* 0.57* 13.22 1.33 0.31 0.03 Character 34: female hindwing zone d–f 10* 0.60* 13.60 1.03 0.44 0.03 11* 0.55* 14.37 0.73 0.42 0.02 Character 39: male hindwing submarginal spots 3* 0.67* 11.51 1.42 0.17 0.02 4* 0.50* 13.31 1.23 0.15 0.01 Character 41: female hindwing submarginal spots 4* 0.50* 5.54 0.61 0.37 0.05 4* 0.50* 5.69 0.47 0.35 0.03 Character 47: male tergal patterns 6* 0.67* 7.56 0.56 0.53 0.04 4* 1.00* 7.71 0.46 0.52 0.03 Character 49: male pleural dots 3* 0.33* 5.06 0.87 0.21 0.04 3* 0.33* 5.59 0.56 0.18 0.02 Character 50: female pleural dots 3* 0.33* 5.57 0.85 0.18 0.03 3* 0.33* 5.64 0.54 0.17 0.02 Character 51: sternal patterns 5* 0.40* 8.97 1.72 0.22 0.03 5* 0.40* 10.54 1.09 0.19 0.02
*Significantly different at P < 0.05.
L, length; CI, consistency index; SD, standard deviation; EW, equal weighting; SAW, successive approximation character weighting.
68 (3!4) 68 (3!4) 68 (3fi4)* 68 (3fi4)* 72 (3fi2)* 72 (3fi2)* 72 (3fi2)* 72 (3fi2)* 76 (3!4)* 76 (3!4)* 76 (3!4)* 76 (3!4)* 77 (3!4)* 77 (3!4)* 77 (2fi4)* 77 (3fi4)*
‘‘–‘‘: same as that in the EW analysis based on the whole dataset; ‘‘x’’: synapomorphy not recovered in the analysis;! indicates change which does not occur in all reconstructions;fi indicates unambiguous change; * unique character; NRC: clade not recovered in the analysis.
type (Table 8) and performed a heuristic search according to the same protocol of searching for MPTs. An additional mimetic assemblage, ‘yellow-1’, was added by deleting the magnifica form of E. aedea aedea. We hoped to examine whether deletion of this form from the yellow type, which is dominated by the E. tricolor species group, would have any influence. In total, six alternative phy-logenetic null hypotheses (H0) (all under EW) were to be compared with the consensus of the observed MPTs (under EW).
Results and discussion
Examining mimetic wing patterns under different lighting conditions
Yellow–orange type (Fig. 1). Under visible light con-ditions, Eterusia species differed mainly from one another in the darkness of the yellow colour and the metallic blue (zone e) of the hindwings. The wing pattern of E. aedea has markedly larger white patches in the forewings. The ‘normal form’ of the agaristine Scobigera amatrix is simi-lar to the various Eterusia species in having cellusimi-lar, cubital, anal and radial white spots, but the submarginal white patches and discoidal spots of Eterusia are lacking in this species. The hindwing of Scobigera amatrix has a
white stripe along the humeral margin, but this pattern is completely absent in Eterusia. Soritia shahama was included in the comparison because it might have mimetic relationships with E. tricolor in northeast India or with E. nobuoi in southwest China. Its hindwing has the basal CuA1 patch separated from zone c and this feature is shared by the male of E. tricolor and E. sublutea sublutea. However, the submarginal white patches and metallic blue sheen in Eterusia are not present in this species. Owada & Ta (2002) considered that Gaeana cicadas belong to the same mimetic type with Eterusia and Scobigera amatrix because they have medial and sub-basal yellow patches and transverse white stripes on the forewings and a light yellow ground colour on the hindwing.
Under UV reflectance conditions, E. sublutea sublutea, E. tricolor, E. nobuoi, E. aedea aedea (magnifica form) and Scobigera amatrix appeared more similar, having much darker forewing and hindwing ground colours, whereas the forewing ground colours of E. aedea formosana and E. aedea sinicaseemed to be much lighter (Fig. 1B). Except for the female of E. tricolor, the metallic blue sheen on the hindwing does not differ markedly among the species. None of the three cicada species showed any correspondence with their potential moth comimics in this regard because they all lack UV-reflective forewing patches.
The fluorescence images (Fig. 1C) showed a different pattern. In E. sublutea sublutea, E. nobuoi and E. tricolor,
Table 8. Comparison of selected observed mimicry characters with randomly generated character distributions based on the phylogeny with all colour pattern characters excluded (see Appendix 3 for detailed descriptions of characters).
EW SAW Observed characters 1000 random characters Observed characters 1000 random characters
Character L CI Mean L SD Mean CI SD L CI Mean L SD Mean CI SD Mimetic types: (0) nonmimetic; (1) blue–white;
(2) yellow–orange; (3) black 7 0.44* 7.75 1.13 0.39 0.06 7* 0.43* 7.60 1.28 0.41 0.08 Character 16: male forewing ground colour 8* 0.63* 12.05 1.08 0.42 0.04 8* 0.63* 11.77 1.31 0.43 0.05 Character 17: female forewing ground colour 8* 0.63* 12.92 1.75 0.38 0.04 8* 0.63* 12.68 1.79 0.41 0.05 Character 21: male forewing medial zone 8* 0.88* 12.77 1.25 0.55 0.06 8* 0.88* 12.80 1.10 0.55 0.05 Character 23: female forewing medial zone 7* 1.00* 12.54 1.18 0.56 0.05 7* 1.00* 12.58 1.03 0.56 0.05 Character 24: male discoidal spot 7* 0.29* 8.50 1.25 0.24 0.04 7* 0.29* 8.42 1.06 0.24 0.03 Character 25: female discoidal spot 7* 0.29* 8.50 1.19 0.24 0.12 7* 0.29* 8.41 1.03 0.24 0.03 Character 26: male submarginal spots 5* 0.80* 12.19 1.23 0.33 0.03 5* 0.80* 12.22 1.08 0.32 0.03 Character 27: female submarginal spots 5* 0.60* 8.77 1.17 0.35 0.05 5* 0.60* 8.82 1.09 0.34 0.05 Character 28: male hindwing colour zone c 11 0.45 11.38 0.99 0.44 0.04 11* 0.45* 11.41 0.87 0.43 0.04 Character 29: female hindwing colour zone c 10 0.50* 11.85 1.04 0.42 0.04 10* 0.50* 11.79 1.01 0.41 0.03 Character 32: male hindwing zone d–f 7* 0.57* 10 1.05 0.40 0.05 7* 0.57* 10.21 1.09 0.41 0.05 Character 34: female hindwing zone d–f 11 0.55 12.05 1.17 0.51 0.05 11* 0.55* 12.03 1.29 0.50 0.05 Character 39: male hindwing submarginal spots 4* 0.50* 9.58 0.97 0.21 0.02 4* 0.50* 9.63 0.95 0.22 0.01 Character 41: female hindwing submarginal spots 4* 0.50* 4.92 0.91 0.42 0.11 4* 0.50* 4.89 0.87 0.41 0.01 Character 47: male tergal patterns 4* 1.00* 6.79 0.69 0.59 0.06 4* 1.00* 6.74 0.72 0.55 0.06 Character 49: male pleural dots 2* 0.50* 3.83 1.07 0.26 0.09 3* 0.33* 4.17 0.98 0.26 0.08 Character 50: female pleural dots 2* 0.50* 4.30 1.02 0.25 0.07 2* 0.50* 4.20 1.01 0.26 0.06 Character 51: sternal patterns 5* 0.40* 7.46 1.09 0.27 0.44 5* 0.40* 7.86 0.98 0.26 0.65
*Significantly different at P < 0.05.
the celluar, cubital and discoidal patches of the forewing and the discoidal white spots of the hindwing were much lighter than those of the other species. The light-coloured patches in Gaeana variegata were markedly brighter than those of other insects. The dorsal part of the prothorax of Gaeana variegataand G. maculata also showed very bright and symmetrical dots, which were completely absent in the moths. Therefore, we are not convinced that cicadas belong to the same mimetic group.
Black type (Fig. 2). We selected all the species speculated by Owada & Ta (2002) as belonging to the black mimicry complex, including three specimens of the edocla form of E. aedea aedeabecause this form has a wide variation in the metallic blue sheen in zone e of the hindwing. Under visible lighting conditions, each of these species appeared slightly different in the development of the metallic blue sheen, although every species had prominent white patches on the forewings and reduced white patches on the hindwings (compared with the yellow–orange mimetic type). The UV reflectance of the forewing white patches of E. aedea aedea and E. sublutea sublutea (fasciata form) was much stronger than in the two agaristine species and Tosena fasciata. The metallic blue sheen did not appear very differently in these lighting conditions. The UV-induced fluorescent forewing white patches of Eterusia and Tosena fasciata were stronger than those of the Agaristinae moths, whereas specimens with a metallic blue sheen had very similar appearance under this lighting. Including Tosena in this mimicry complex (Owada & Ta, 2002) is questionable because this species does not fly actively in the day time, unlike Eterusia species (Yen,
pers. obs.). Furthermore, its reflectance of eyes, thoracic band and abdomen under the UV-derived fluorescence is not seen from either Eterusia or Scobigera moths.
Blue–white type (Fig. 3). Under the visible light, the main differences between E. taiwana and E. aedea formosana were the congregation of the white patches in zone c of the forewing and the development of submarginal spots in the hindwing. UV reflectance and UV-induced fluorescence showed no marked differences in the hindwing ground colour between E. taiwana and the two colour forms of E. aedea formosana. Because the presence of the yellow hindwings of E. aedea formosanais not sex limited, we do not expect that this unusual colour pattern is correlated with sexual selection within the subspecies. In Taiwan, E. taiwana and E. aedea formosana are sympatric, but the adults of the latter are generally more common (Yen, pers. obs.). Owada (2001) indicated that E. aedea formosana is nocturnal, whereas the first author’s observations contradict this because only the male of this subspecies can be attracted by light traps, but both sexes are very active in the forest canopy during the day.
Phylogenetic relationships of Eterusia
Relationships of Eterusia species based on the full dataset. The initial analysis of the dataset (Appendix 2) generated 907 MPTs (tree length¼ 356, CI ¼ 0.752, RI¼ 0.858). The strict consensus cladogram of these (Fig. 11A) recovered Eterusia sensu Yen (2004a) as a
Fig. 3. A potential mimicry complex participated by Eterusia taiwana (top) and E. aedea formosana (yellow form, middle; white form, bottom). A, Visible light patterns; B, ultraviolet reflectance patterns shown against an ultraviolet reflecting background; C, fluorescence patterns
Fig. 4. Antennal shapes of Eterusia species (characters 1–2). A, Eterusia risa risa, female; B, E. risa risa, male; C, E. nobuoi, male; D, E. tricolor, female; E, E. guanxiana, female; F, E. binotata, male; G, E. taiwana, female
Fig. 6. Wing patterns of the Eterusia and Soritia (outgroup) moths. A, Definitions of wing patterns; B, definitions of wing areas; C, E. aedea aedea (edocla form), female, upperside; D, E. aedea aedea (edocla form), female, underside; E, E. aedea aedea (magnifica form), male; F, E. aedea sinica, male; G, E. aedea sugitanii, male; H, E. aedea sakaguchii, female; I, E. aedea azumai, male; J, E. aedea azumai, female; K, E. aedea okinawana, male; L, E. aedea virescens, male; M, E. aedea virescens, female; N, E. aedea cingala, female; O, E. vitessa, upperside; P, E. vitessa, underside; Q, Soritia strandi, male; R, Soritia azurea, male; S, Soritia azurea, female (I–J, photograph by M. Owada)
Fig. 7. Wing patterns of the Eterusia moths. A, Eterusia risa risa, male, upperside; B, E. angustipennis angustipennis, male, upperside; C, E. angustipennis angustipennis, underside; D, E. risa palawanica, female; E, E. angustipennis angustipennis, female, upperside; F, E. angustipennis angustipennis, female, underside; G, E. binotata, male, upperside; H, E. binotata, male, underside; I, E. guanxiana, female, upperside; J, E. guanxiana, female, underside; K, E. binotata, female; L, E. joiceyi, female; M, E. feminata, male, upperside; N, E. feminata, male, underside; O, E. feminata, female, upperside; P, E. feminata, female, underside; Q, E. subcyanea, male, upperside; R, E. subcyanea, male, underside; S, E. subcyanea, female, upperside; T, E. subcyanea, female, underside (K, photograph by M. Owada)
Fig. 8. Wing patterns of the Eterusia moths. A, Eterusia taiwana, male; B, E. taiwana, female; C, E. watanabei, upperside; D, E. watanabei, underside; E, E. tricolor, male; F, E. tricolor, female; G, E. sublutea sublutea, male, upperside; H, E. sublutea sublutea, male, underside; I, E. sublutea sublutea, male, upperside; J, E. austrochinensis, male, upperside; K, E. austrochinensis, male, underside; L, E. nobuoi, male; M, E. sublutea yuchii, male; N, E. sublutea yuchii, female; O, E. sublutea yasunorii, male; P, E. sublutea yasunorii, female; Q, E. sublutea sublutea (fasciata form), male, upperside; R, E. sublutea sublutea (fasciata form), male, underside; S, E. sublutea sublutea (fasciata form), female, upperside; T, E. sublutea sublutea (fasciata form), female, underside (M–P, photographs by M. Owada)
Fig. 10. Stylized drawings of the male eighth abdominal segment (A–H, O), male genitalia (I–N, P–V) and female genitalia (W–Z). A–H, Lateral view of the eighth tergite and sternite; I–N, lateral view of the uncus–tegumen complex and vinculum; O, dorsal view of the eighth tergite; P, ventral view of the vinculum; Q–T, tegumenal apodemes and associated sclerites; U, aedeagus; V, valvae; W, ostium and the surrounding structures; X–Z, locations of appendix bursae
Fig. 11. The phylogenetic relationships of Eterusia species using different character sets and weighting strategies. A, All characters included under equal weighting (EW); B, all characters included under successive approximation character weighting (SAW); C, all colour pattern characters excluded under EW; D, all colour pattern characters excluded under SAW. Branch support is given in the form of bootstrap values above the branches and jackknife values below the branches
monophyletic group, supported by thirty-one synapomor-phies, of which thirteen are unique to Eterusia. The clade (E. risaþ E. angustipennis) is supported by sixteen synapo-morphies and the clade (E. binotataþ E. guanxiana) by seventeen synapomorphies. These two clades are sister groups and are referred to here collectively as the risa species group, which is supported by eighteen synapo-morphic characters, including six unique ones.
The E. tricolor species group sensu Owada was recovered as a monophyletic group supported by twenty-four synapomorphies, but except for the sister-group relation-ship of E. tricolor and E. taiwana (supported by six synapomorphies), the relationships of the remaining taxa were unresolved. The E. tricolor species group was recovered as the sister group of the subcyanea species group (supported by ten synapomorphies), which has an internal relationship of ((E. feminataþ E. subcyanea) þ E. joiceyi). All of the subspecies of E. aedea were recovered as a monophyletic group, which was placed as the sister group of E. vitessa by the support of twenty-five synapomorphies. Internal relationships within E. aedea were poorly resolved, except for the sister-group relationship of the Sri Lankan E. aedea cingala and the south Indian E. aedea virescens. The relationships of the (E. tricolorþ E. subcyanea), E. risa and E. aedea groups were not resolved in this analysis.
SAW resulted in the stabilization of trees after three iterations. The strict consensus cladogram of the final fifteen MPTs (Fig. 11B) additionally recovered the sub-species of E. sublutea as a monophyletic group supported by eight synapomorphies. It also displayed greater reso-lution of the terminal nodes, providing a more detailed hypothesis of the relationships within the species groups. The synapomorphies that appear to support the E. risa species group are slightly different from those recovered by the EW analysis. The absence of a male discoidal spot (24:0) was additionally recovered to support this group, but four characters, male hindwing with black margin confined to zone f (33:1), male hindwing discoidal spot absent (36:0), female hindwing discoidal spot absent (37:0) and female hindwing with basal CuA1 spot con-nected with basal M3 spot (44:3), were not recovered in the SAW analysis. The clade comprising the E. tricolorþ E. subcyanea species groups was recovered as a monophyletic group as well, but six of the twenty-two synapomorphies recovered in the EW analysis were not found by SAW. In general, the apomorphies supporting the clades recovered by both EW and SAW are congruent except for some characters missing in the SAW analysis. For detailed character distributions and descriptions, see Appendices 2 and 3.
Fig. 12. A comparison of the consensus based on different phylogenetic scenarios of mimicry based on the whole dataset. A, The selected three types of mimicry (see Table 3) mapped on the observed phylogeny of Eterusia (see also Fig. 11A). Bremer support values are given above the branches; B, quasi-Batesian mimicry; C, Mu¨llerian mimicry; three types of mimicry constrained on one single phylogeny; D–G, Mu¨llerian mimicry with three types of mimicry constrained; D, all ‘yellow’ taxa constrained; E, ‘yellow’ taxa constrained by excluding E. aedea aedea (magnifica form); F, ‘blue–white’ taxa constrained; G, ‘black’ taxa constrained. All of the hypotheses of Mu¨llerian mimicry are rejected (see Table 9)
