The Journal OF Research ON THE LEPIDOPTERA Volume 42 2003 (2010) ISSN 0022 4324 THE LEPIDOPTERA RESEARCH EOUNDATION The Journal of Research on the Lepidoptera ISSN 0022 4324 PuBi.isiiKD Bv: The Lepidoptera Research Foundation, Inc. 9620 Heather Road Beverly Hills, California 90210-1757 TEL (310) 274 1052 E-mail: Editorial: niattoni@ucla.edn Business: mattoni@ucla.edu Technical: jrl_lepidoptera@yahoo.com Eounder: William Hovanitz (1915-1977) Editorial Staff: Rudolf H. T. Mattoni, acting editor E-mail: mattoni@ucla.edu Nancy R. Vannucci, information manager E-mail: jrl_lepicloptera@yahoo.com Associate Editors: The Lepidoptera Research Eoundation is in process of reorganization. A board of associate editors will be established. Manuscripts and notices material may at present be sent to the acting editor, Rudolf H. T. Mattoni, e-mail: mattoni@ucla.edu. Please note the instructions to authors on the back inside cover of this Journal. Past journals and other publications, and all matters relative to subscriptions, mailing and billing are being handled by BioQuip Products, Inc. Phone: (310) 667-88()(). E-mail: bqinfo@bioquip.com. Information manager at tfif. address: Nancy R. Vannucci, Maza 3340, Moron B1708GOP, Buenos Aires, Argentina. Fhe Journal is sent to all members of the Foundation. Glasses of membersiiir: Regidar (Individual) (Contributing Subscription Rate/ Institutions Life $ 25.00 year $ 30.00 or more year $ 35.00 year $ 300.00 STATEMENT OE OWNERSHIP AND MANAGEMENT n IE JOfTRNALOF RESEARCH ON THE LEPID(4PTERAwill be published two timesayear by the LEPIDOPTERA RF,SEARCH FOUNDAITON, INC. Publication and business offices are located at the Beverly Hills, California address given above. The Foundation is a non-profit organization incorporated in the State of California in 1965. The president is Rudolf H. T. Mattoni, the vice-president is Jeremiah George, the secretary-treasurer is Leona Mattoni. The board of directors (2005-) is comprised of Konrad Fiedler, Dan Rubinoff, Jeremiah George, and Rudolf 1 1. T. Mattoni. Pa.sf isstiES ARE A\'AiLABLF. AF OLR I'F.Mi’ORARV’ WEBSFFE : www.lepidopteraresearchfoundation.org journal of Research on the Lepiilojjlera 42: 1-4, 200.4 (2010) Differences in thermal responses in a fragmented landscape; temperature affects the sampling of diurnal, but not nocturnal fruit-feeding Lepidoptera Danilo B. Ribfjro' - and Andre V. L. Freitas'** 'De|)artamento de Zoologia, Iiistituto cle Biologia, Univer.sidade E.stadiial de (4ini|jinas. (4’ 6109, 14083-970 Oampinas, SP. Brazil, tel. -t-+.6.5 19-452 16.3 10 -Prog'iama de Po.s-Giadiiagao ein Ecologia, hi.sliluto de Biologia, Uiiiversidade Estadual de (’.anipinas, Brazil. hlo(lhr@yahoo. com.br -’DepartanieiUo de Zoologia, In.stiiuto de Biologia, Univer.sidade Estadual de Cam|tinas. UP 6109, 1,3084-970 Uainpinas, SP, Brazil, tel. +-r55 19-452 16.3 10 Ab.stract. Weather is the primary determinant in hiittei lly activity, with tem|3eralnre one of the key factors affecting the hiolog)’ and behavior of most Lepidoptera. Despite et idence about the inllttence of temperature in Lejtidoptera capture jtatterns, few stttdies have analyzed microclimatic characteristics. Most studies focused on broad geographic scales and historical climatic jtatterns. The present stitdr contrasts the effect of local temperature on the capttire rates ol two groups of ft iiit-feeding Lepidoptera, btttterllies (dittrnal) and noctuid tnoths (noctttrnal), iti a fragmented landscape. Ten fragments with live traps each were sampled in sotttheast Brazilian Atlantic Forest dttring eight days iti |tily-2005. We fotmd a jtositive relation between mean tem|teratiire and both richness and abundance of captttred butterflies, but not the noctnid moths. These difleretices are jtrobably a result ol the miothermic natttre of moths, making the moth assemblage less dependent of solar radiation than butterflies. 4'he differences between moths and htittei flies coitld affect the distribittion of these insects across fragmented latidsca|3es and sttggest that adult moths are |)robably less sensitive to changes in the amottnt of solar radiation than adtilt btitlerflies, Key word.s: frtiit-feeding btitterllies, Nym|3halidae, Noctitidae, temperatitre. Introduction Ehrlich (1984) stated that weather is the [triinary determinant in btitterfly activity. Temperature is a key factor affecting most Lepidoptera because it has direct effects on larval behavior and performance, flight activity and associated behaviors such as foraging and territoriality (Broersma et al., 1976; Scriber & Slansky, 1981; Hrdy el al, 1996; Shirai el al., 1998; Kiilirt et al., 2005; Nabeta et al., 2005; Merckx et al., 2006). Many btitterfly species are able to maintain their body temperattire above ambient through basking and/or shivering behaviors (Kemp, 2002), such that individuals with higher body temperatnres can be active for longer periods than their cooler bodied counterparts (Dudley, 1991). On sunny days some butterflies have been reported to fly for longer periods and maintain higher body temperatnres than on cloudy days (Shelly & Ludwig, 1985). Temperattire *■ Corresponding author. Departamento dc Zoologia, In.stiluto cle Biologia, L'niversidacle E.stadual de Ciampina.s. C.P 6109, 14083- 970 Campina.s, SP, Brazil, tel. -1+5519-35216310 baku@unicamp. br Rereived: 5 March 2009 Accepted: 6 March 2009 is also known to affect the seasonal dislribtition of some species (Turner et al, 1987), probability of capttire in temperate regions (Yela &: Holyoak, 1997; Holyoak et al., 1997), and the abundance and diversity patterns of both butterflies (Turner et al., 1987; White & Kerr, 2007) and moths (Brehin et al, 2007; Choi, 2008). Despite the evidence that ambient temperature influences activity patterns in Lepidoptera, few studies have analyzed the effect on a microclimatic scale (e.g. Yela & Holyoak, 1997) . Most have focused on broad geograjihic scales and historic climatic patterns (e.g. Brown Sc Freitas, 2000; Menendez et al, 2007). However, anthropogenic activity is known to prodtice major changes in microclimatic conditions (Saunders, 1991; Vitonsek et al, 1997) which likely in turn could affect the insects that occtipy a given habitat. Accordingly the present study attempts to test the effect of local temperature in the capttire rates of fruit-feeding butterflies and moths in a fragmented landscape. Here we ask whether there are differences in the responses of diurnal versus nocturnal Lepidoptera to average temperature, linking the possible differences with the effects of forest fragmentation among these two sets of insects. /. Res.Lepid. 0 D m 0 Richness Abundance Figure 1. Linear regressions between temperature and species richness and between temperature and abundance for both butterflies and moths sampled with baited traps in a fragmented landscape in Brazilian Atlantic Forest, a) Linear regression between mean temperature and richness of fruit-feeding butterflies; b) linear regression between mean temperature and abundance of fruit-feeding butterflies; c) linear regression between mean temperature and richness of bait attracted moths; d) linear regression between mean temperature and abundance of bait attracted moths. Materials and methods Field work was carried out in the Sao Liiiz do Paraitinga municipality (Fig. 1), Sao Paulo State, south-eastern Brazil (centered in 23°20’ S, 45°20’ W). The original vegetation of the area was mainly a dense humid forest (MME 1983); however, the process of forest fragmentation drastically changed the plant community cover across the region (Schmidt, 1949; Petrone, 1959; Dean, 1997). Today, large parts of the native vegetation has been removed with the resulting landscape now being composed of small patches of disturbed forest scattered in a matrix of farm fields and abandoned pastures and including some patches of Eucalyptus plantations. Ten forest fragments were chosen at random for sampling. In each fragment, five traps were arranged along a linear transect for a total of 50 traps. Detailed information of sampling methods and the study site are given in Ribeiro ct al. (2008). Butterflies (diurnal) andiioctuid moths (nocturnal) were sampled during |une 2005. The traps remained in the field for eight days and were visited at 48-hoiir intervals. During each visit bait was replaced and each captured individual was identified, marked and released. Species not determinable in the field were taken for later identification (following Ribeiro et at, 2008). All butterflies were identified to species level and the moths (all Noctuidae) were discriminated as morphospecies. Ambient temperatures were measured with a niax- min thermometer at each visit. The arithmetic mean of the maximal and minimal temperature measured during the period between visits (48 hours) was used as mean temperature. We use simple linear regression to test if there was a relationship between abundance and richness of butterflies and moths with mean temperature. Results A total of 233 individuals comprising 27 species of fruit-feeding Nymphalidae and 475 individuals of nine morphospecies of noctuid moths were captured. 42: 1-4, 2003 (2010) 3 The mean temperatures recorded during tlie study period ranged between 1 1.5 and 21.()°(1. We found a significant relation between mean temperature and species richness (p < O.OOOl; R- = 0.362; N = 40; y = 0.2481x+ 14.84) (Fig. la) and individual abundance (p < 0.0001; R" = 0.362; N = 40; y = 0,2481x + 14.84) (Fig. lb) in our butterfly samples. In contrast, we did not find a significant relation between mean temperature in either richness (p = 0. 581 1; R- = 0.0084; N = 40; y = -0.1608x+ 16.599) (Fig. Ic) or abundance (p = 0.8936; R2= 0.0004; N = 40; y = -0.0079x + 16.094) (Fig. Id) of the moth samples. Discussion We found a strong relation between microhabitat temperature and richness and abundance in samples of diurnal, but not nocturnal noctuid Lepidoptera. The relation between butterfly capture and temperature is likely the result of differences in daily activity of the diurnal set of species compared with the nocturnal noctuids. Despite the great variety of behaviors related to thermoregulation in butterflies (Clench, 1966), air temperature remains important in determining butterfly activity (Douwes, 1976). Tbus we would expect greater butterfly activity and capture rate on hot rather than in cold days. Although weather, especially temperature, is usually considered the most important factor determining butterfly and moth diversity (Holyoak et al, 1997; Yela & Molyoak, 1997; Hawkins & Porter, 2003; Brehm et al., 2007; Menendez et al., 2007; Choi, 2008), few studies demonstrate significant relations between microhabitat conditions with species richness and abundance of Lepidoptera (e.g. Yela & Holyoak, 1997; Doha et al, 2008). The differences found in diurnal versus nocturnal fruit-feeding Lepidoptera are likely the result of the latter being miothermic and therefore independent of solar radiation to enable their activities (Daily & Ehrlich, 1996). If the conclusion is correct, we would expect that butterflies in the subfamily Brassolinae should respond similarly to nocturnal moths, since brassolines are not as dependent upon solar radiation as other butterflies are (Siygley, 1994). The degree of fragmentation has significant influence on microclimatic conditions, as temperature, humidity and amount of solar radiation (Saunders, 1991), with many studies of diurnal fruit-feeding Lepidoptera reporting changes in the community correlated with forest fragmentation (Ki'auss et al, 2003; Shahabuddin & Ponte, 2005; Uehara-Prado et al., 2007). The change in microclimatic conditions also affects the distribution of butterflies in forest fragments, probably because fragmentation impacts the activity of these insects by being beneficial to heliophylous species that are more likely to displace shade-loving species. Another important alteration caused by microclimatic changes in temperature is the effects upon life-history traits of butterflies by changing daily fecundity and lifetime number of eggs of females (Ktirlsson & Van Dyck, 2005) . However, the same effect may be less important in moths because they are not directly influenced by solar radiation (Daily & Ehrlich, 1996) and suffer only the indirect effects of the changes (e.g. alteration in leaf quality, resource offer and so on). Both .sample richness and abundance are positively correlated with mean temperature in the present study. Since trap capture is considered as a reliable measure of activity and density in fruit-feeding Lepidoptera, we can conclude that temperature had an undeniable effect in butterfly richness and abundance. I lowever, in view of the great importance of temperature in determining the behavior and distribution of diurnal Lepidoptera, other studies are necessary to verify if this relation could be found in other habitats, with different climates and land covers. Acknowledgements We es|)ecially thank ('.aria Pen/, and Niklas Walilberg lor critically reading and connnenting tt]>on oitr mamtscript. W'e also tliank tlie severtil land owners who permitted held work on their |jroperties. DBR thanks FAPKSI’ tor ;i tellowship (grants #03/ 1 1697-0, 07/.5089(>0). AVLF acknowledges the FAPESP (grants #00/01484-1 and #01/05269-9), the Bra/ihan CNPq (lellowshi]) #300315/200.5-8), and the National Science Foundation (DEB grant #0527441 ). This project is part ol BIOTA-FAPFSP program (grants #98/05101-8 and #02/08558-6). Literature cited Bruim, G., R. K. Cot.WF.t.i. & J. Ki.uca;. 2007. The role otenvironment and mid-domain ef fect on moth species richness along a tropical elevational gradient. Globtil Ecolog)' and Biogeographv 16: 20.5-219. Broersma, D. B., J. R. Barreei', & J. O. Sit.i.iNcs. 1976. Activity and blacklight induced flight of black cutworm Le])idoptera- Noctnidae and European corn-borer L.epidojttera-Pyralidae as related to temperature and relative httmidily. Environmental Entomolog)’ 5: 1 191-1 194. Brown, K. S. Jr & A. V. L. Freit.vs. 2000, Atlantic forest hntterllies: iiulicators for landscape consen'ation. Biotropica 32: 934-956. Ciioi, S. W. 2008. Diversity and com|)osition of larger moths in three different forest types of Sottthern Korea. Ecological Research 23: 503-509. Ci.ENCH, H. K. 1966. Behavioral thei inoregnlation in btitterflies. Ecolog)' 17: 1021-10;34. Daii.v, G. C. & P. R. EiiRt icii. 1996. Noctiimality and species sttn'ival. Proceedings of The National Academy ol Science 93: 11709-11712. Dean, W. 1997. A ferro e logo. A histdria e a devasttigao da Matti Atlantica brasileira. Companhia das Letras, Sao Patilo. Dot.tA,J., M. S. Dew, N. A. Ar,\vind& A. KfMAR. 2008. Adult butterfly 4 J. Res.Lepid. coniinunities in ct)free plantations around a protected area in the Western C'.hats, India. Animal Clonservation 1 1: 2t>34. Douwes, I’. 1976. Activity in Heodes virgaurear (Lep., Lycaenidae) in relation to air temperature, solar radiation, and time of day. Oecoloo-ia 22: 287-298. Duni.iiV, R. 1991. Thermoregulation in unpalatable Danaine butterllies. Functional Ecology .'5: .503-506. EiiRMOii, P. R. 1984. The structure and dynamics of butterlly populations. Pp. 16(>2()I. hr. Vane-Wright, R. I. & Ackeiy, P. R. 1984 (Eds.). The biology of butterflies. London, Academic Press Inc., 430 p. Hawkins, B. A. & E. E. Porter. 2003. Does herbivore diversity depend on plant diversity? The case of Clalifornia butterflies. The American Naturalist 161: 40-49. Hoi.voak, M., V. |arosik& I. Novak. 1997. Weather-induced changes in moth activity bias measurement of long-term population dynamics from light trap samples. Entomologia Experimentalis et Applicata 83: 329-335. Hrdy, I., F. Koc:ot'REK, J. Berankova, & j. Kuldova. 1996. Temperature models for ]>redicting the flight activity of local populations of Cydia fiinebmna (Lepidoptera: Tortricidae) in Central Europe. Euro]>ean Journal of Entomology 93: 569-578. K,\ri.s,son, B. & 11. Van Dyck. 2005. Does habitat fragmentation affect temperature-related life-history traits? A laboratory test with a woodland butterlly. Proceedings of The Royal Society B. 272: 12.57-1263. Kemp, D. L. 2002. Visual mate-searching behaviour in the evening brown butterfly, Melanilisleda (L.) (Lepidoptera: Nymphalidae). Australian Journal of Entomolog)' 41: 300-305. Kraii.ss, j., I. Stefan-Dewenter & T. T.sc:harntke. 200.3. How does landscape context contribute to effects of habitat fragmentation on diversity and po]>ulation density of butterllies? Journal of Biogeography .30: 889-900. KOhre, Lh, J. Samietz&S. Dorn. 2005. Thermoregulation behaviour in codling moth larvae. Physiological Entomology 30: 54-61. Menf.ndez, R., a. GoN7,.Ai.E7,-ME(a.A.s, Y. Coi.i.iNOM.AM, R. Fox, D. B. Roy, R. Ohef.mui.ler & C. D. Thomas. 2007. Direct and indirect effects of climate and habitat factors on butterfly diversity. Ecology 88: 60.5-61 1 . Merc.kx, T, B. KARi.ssftN & 11. Van Dyc:k. 2006. Sex- and landscape-related differences in flight ability under suboptimal temperatures in a woodland butterfly. Functional Ecology. 20: 4.36-441. MME-Minksterio 1),\.s min.as e enercta. 198.3. Projeto Radam Brasil. Programa de Integrayao Nacional. Levantamenlo de Recursos Naturals, Vol. 32. Publication Division, Rio de Janeiro. N.abeta, F. H.. M. Naiv\i & Y. Kunimi. 2005. Effects of temperature and photoperiod on the development and reproduction of Adoxophyes liotimni (Lepidoptera: Tortricidae). Applied Entomology and Zoology 40: 231-2.38. Petrone, P. 1959. A regiao de Sao Luiz do Paraitinga: estudo de geografia luimana. Revista Brasileira de Geografia 21: 239X336. Ribeiro, D. B., R I. Prado, K. S. Brown Jr., & A. V. L. Freitas. 2008. Additive partitioning of butterfly diversity in a fragmented landscape: importance of scale and implications for consen,'ation. Diversity and Distribution 14: 961-968. Saunders, D. A., R. J. Hobbs & .R. Margui.es. 1991. Biological consequences of ecosystem fragmentation: a review. Gonservation Biology 5: 18-32. Schmidt, C. B. 1949. A vida rural no Brasil: a area do Paraitinga- uma amostra representativa. Boletim de Agricultura. Numero unico. Secretaria da Agricultura do Estado de Sao Paulo, Sao Paulo. ScRiBER, J. M. & F. SianskyJr. 1981. The nutritional ecology of immature insects. Annual Review of Entomology 26: 183-211. Shahabuddin, G. Sc C. A. Ponte. 2005. Frugivorous butterfly species in tropical forest fragments: correlates of vulnerability to extinction. Biodiversity and Conservation 14: 1137-1152. SiiELi.Y, T. E. & D. Ludwig. 1985. Thermoregulatory behavior of the butterfly Cafisto nubila (Satyridae) in a Puerto Rican Forest. Oikos 44: 229-233. SiiiR.\i, Y., Y. Kosugi, & I. Noguchi. 1998. Effects of sex, mating status and temperature on flight activity of the Oriental tea tortrix, Homona magnanima (l.epidoptera: Tortricidae). Applied Entomology and Zoology 33: 413-418. Sryc.ley, R. B. 1994. Shivering and its cost during reproductive behavior in owl butterllies Caligoiind Opsiphanes (Ntyuphalidae: Brassolinae). Animal Behaviour 47: 2.3-32. Turner, }. R. G., C. M. Gatehoiese & C. A. Corey. 1987. Does solar energy control organic diversity? Butterflies, moths and the British climate. Oikos 48: 195-205. Uehar.a-Pr.4do, M., Brown Jr., K. S., & A. V. L. Freita.s. 2007. Species richness, composition and abundance of fruit-feeding butterflies in the Brazilian Atlantic Forest: comparison beftveen a fragmented and a continuous landscape. Global Ecolog)' and Biogeography 16: 43-54. ViTousEK, P. M., H. A. Mooney, J. Lubchenco &J. M. Melillo. 1997. Human domination of earth’s ecosystems. Science 277: 494- 499. White, P. |. T. &J. T. Kerr. 2007. Human impacts on environment- diversity relationships: evidence for biotic homogenization from butterfly species richness patterns. Global Ecolog}' and Biogeography 16: 290-299. Yela, J. L. & M. Hoi.yoak. 1997. Effects of moonlight and meteorological factors on light and bait trap catches of noctuid moths (Lepidoptera : Noctuidae). Environmental Entomology Vol. 26, lY 6, pp, 1283-1290. journal oj Ri'searrh on Uw IjjndojHm 42; 5-12, 2003 (2010) Use of sound and aerial chases in sexual recognition in Neotropical Hamadryas butterflies (Nymphalidae) Onildo }. Marini-Fii.ho' -^ and Woodruff W. Benson'^ ‘Dep. Biologia Animal, Universidade Estadual de Oampinas, C. R. 6109, 13083-970 Oam|)inas, SP, Brazil -Instituto Ohico Mendes de Oonsei vacjao da Biodiversidade, EQSW 103/104, Bloco D, 1- aiidar. Setor Sudoeste. CEP: 70670-350. Bra.silia, DF, Brazil 0. m a ri ni @gm a il. com '14ei3. Biologia Animal, Universidade Estadual de Campinas, C. P. 6109, 13083-970 Camjtinas, SP, Brazil Abstract. Neotropical Hamadryas butterllies are notorious for the clicking noise they produce in flight. Tests of the sotind production cajtacitv of nine species of Hamadryas and observations on aerial interactions of six .s|tecies showed that, among the three species grotips, onh males of the jeronia grotip produced sounds, whereas females of all species and males of the jcbrna and laodamia gron|ts never did so. .Vlost of the aerial interactions occurred during mid-day, generally initiated by males in exploratory flights. Chasers were always males. Male.s engaging in chases with other males continnallv nroduced clickine sotmds dnrinsi: the interactions whereas males chasing females usnallv restricted their auditory displays to the initial jt of both sexes of six species of Hamadryas, m; observed, with one single male participating it females and male.s of //. jeronia. There was no i successfully defended a feeding jterch and the n is probably involved in finding mates, while at dominance hierarchy among male Hamadrya.s. Key words: agonistic interactions, behavior, connntmication, sexual recognition, sound. Introduction Butterflies couununicate by means of a variety of stereotyped acoustic, visual, chemical, and tactile signals (Swihart, 1967; Wickman & Wikluud, 1983; Boppre, 1984; Silberglied, 1984; Bernard & Remington, 1991; Lees, 1992). Among these, acoustic signaling is relatively infrequently cited in the literature (Swihart, 1967). The Neotropical Hamadryas (Hiihner, 1806) butterflies, a tight-knit group containing 20 species (Jenkins, 1983), are famous for their production of loud clicking sounds during aerial chases (Bates, 1865; Darwin, 1871; Swihart, 1967; Otero, 1990). Adult butterflies feed on tree sap and rotting fruits, and individuals of both sexes may display and fight to defend feeding sites from congeners (Marini-Filho, 1996). Ross (1963) classified Hamadryas ‘pugnacious,’ although he did not obseiwe site fidelity or other evidence for territorial behavior. //mMadryzo butterflies are capable of hearing sounds produced by other conspecific butterflies and consider that once these sounds are produced during social interactions, these are probably involved in conspecific communication (Yack et at, 2()()0) . Yack et Received: 27 Aj>ril 2009 AccejHed: 1 May 2009 lia.se of the pursuit. In a cage contaiuiug 24 iiulividuals ‘s were involved in 100% of the aerial interactions 1 57% of them. The most chased iudividttals were elatioti between the ntimber of times an individual utiiber of aerial chases initiated. Sound prodttctiou ■rial chases may be used in the establishment of a btttterflies, comiminicatioti, Hamadryas, acoustic al, 2000 also discuss that the po.ssible origin of sound production in the Papilionoidea btUterllies could be derived from bat-detecliou in the basal clade of Hedyloidea moths, thus being a degeneration ofthe.se former structures. Although Seitz (1913) reported that almost all Hamadryas [members of the genus Agmrnia and jWidromia (genus Hamadryas Stcnsn Jenkins, 1983)] make sound on llight, species typical ofden.se tropical forests (e.g. H. chloe, H. alicia, H. rosandra and H. Jenkins, 1983) apparently do not ttiake sounds. Otero (1986), based on the observation that males of //. jeronia intensely produce sound when pursuing other Hamadryasm flight, whereas H. j'ebrua were ‘mute’ and performed a spiral flight, argues that the behaviors represent alternative means of sexual recognition in the genus. Hamadryas were placed by Jenkins (1983) in three species groups (subgenera) based mainly in wing venation differences (note that species names preceded by * are possible exceptions): (1) jeronia species group (subgenus Hamadryas) , including H. jeronia, H. gualemalena, H. il)litliim.e, H. ej)inome, H. fornax, *H. alicia, *H. rosandra, H. amphinome, H. belladonna, //. arinome, (2) yf/zrR// species group (subgenus Ageronia), including H. februa, H. amphichloe, H. glaiiconome, H. honorina, H. allanlis, II. chloe, H. albicornis; and (3) laodamia species group 6 /. Res.Lepid. (subgeniis Peridromia) , including H. laodamia, H. arete, //. veiutina. The mechanism of sound production by Hamadryas is still subject of controversy. At least seven different places on the thorax, abdomen and fore wings have been proposed as sound-producing organs (Jenkins, 1983; DeVries, 1987; Monge-Najera & Hernandez, 1991; Swihart, 1967). Otero (1990) presents good evidence that Hamadryas feronia produce percussive sound through the striking of the wings. He points out that the structures responsible for the loud snapping sound are the swollen veins at the distal end of the fore wing discal cell. These modihcations also occur in the Australian Hecatesia moths and were called castanets by Bailey ( 1 978) . Some Satyrinae species in the genus Pharneuptycliia and Euptychoides also seem to have identical structures a.ssociated with sound production (Kane, 1982; Murillo-Hiller, 2006). According to Otero (1990), castanets were only observed in the males of the sound-producing pnvnia. species, while males and females of H. fehrua in Venezuela do not show these structures. The swellings of the snb-costal venation of the forewings have also been supposed to be associated with sound production by Monge- Najera and Hernandez (1991), however, these authors pointed that swollen snb-costal veins occur in all Hamadryas species they dissected {amphinome, feronia, gualemalena, glauconome, and fehrua) independent of sex. Further studies made by Yack et al. (2000) contradicts the percussion mechanism arguing that sound can be produced by a single forewing, thus pro])osing a wing deformation mechanism for the production of sound. This ‘flip-flop’ changes the side of the wing concavity as shown in the photographs ])resented by Monge-Najera et al. 1998. In the present study we evaluate the sound production capacity of nine species of the genus Hamadryas and analyze the hypotheses related to the possible functions of sound production and aerial chases in sexual recognition. After determining which species and sexes are capable of producing sounds, we use behavioral and morphological data to conjecture if ( 1 ) aerial interactions with sound production are used for sexual recognition between Hamadryas, and (2) aerial interactions are related to the establishment of a dominance hierarchy between the males present in the feeding arena. Study sites and methods But ter flies were captured using standard Van Someren-Rydon butterfly traps baited with a mixture of fermented sugarcane juice and banana (DeVries, 1 987) . Traps were set several times from 1992 to 2090 along the edges and interiors of forests and woodlands at eight different localities: Linhares Forest Reserve (tropical semi-deciduous forest), Linhares, ES, Brazil (19°04’ S; 40°08’ W); Santa Genebra Forest Reserve (subtropical semi-deciduous forest), Campinas, SP, Brazil (47°04’ S, 22°50’ W); Brasilia Botanical Garden and Agna Limpa Reserve (tropical cerrado savanna and galleiy forest) , Brasilia, DF, Brazil (15°57’ S; 47°56’ W); Pipeline Road (tropical evergreen rain forest), Gamboa, Panama (09° 10’ N, 79°51’ W); the restinga dry-forest of Praia das Neves (semi-deciduous tropical dune forest) ES, Brazil (21°18’ S, 41°02’ W) ; Ecological Station of the University of Minas Gerais (upland tropical semi-deciduous forest), Belo Horizonte, MG, Brazil (19°43’ S, 43°57’ W) ; Serra da Canastra National Park (cerrado savanna and gallery forest) , MG, Brazil (20° 14’ S, 46°33’ W); and the cerrado woodlands of Fazenda Jatoba, Correntina, BA, Brazil. We tested freshly captured butterflies for sound production capacity by holding the hind wings dosed over their backs with forceps and passing the fore wings below the hind wings, so that the fore wings were free to beat and produce their typical clicking sound (Otero, 1990). Data on flight interactions were obtained from butterflies kept in an outdoor cage (4x6 m at the base and 4 m in height containing the trunks of two trees) set in a plantation of native trees in the Linhares Forest Reserve. This area was regularly used by at least five Hamadryas species. Caged butterflies permitted observations on butterfly social behavior during a period of Hamadryas scarcity between June 1993 and March 1994. Observations of flight interactions were all from October 1993 on 24 caged butterflies, comprising 13 males (m) and 5 females (f) H. feronia, 2 m atid 1 f H. amphinome, 1 m H. iphthime, 1 f //. arinome, and 1 m H. laodamia (Table 1). Butterflies were fed with fermented banana and sugarcane juice every morning. Each butterfly was individually numbered on the hind wing underside with India ink or marked with highly visible colored spots on the upper side of each fore wing to indicate its sex and species. For each interaction, the species and sex of both individuals and which of them was the chaser and the chased were recorded. Sounds produced and, when possible, the individual responsible for it were also recorded. Nomenclature is based on Jenkins (1983). The non-parametric statistic Chi squared was used to test the heterogeneity of the interactions among the sexes while the binomial statistic (Z) was used to compare the distributions of probabilities of interactions among individuals of the same sex or species and those of different sex or species. 42: 5-12, 2003 (2010) Table 1 . Observed and expected [ ] number of aerial interactions from 24 caged Hamadryas butterflies kept in an outdoor cage (A/= 129 interactions). Expected values of the main diagonal were calculated as [n(n-^)/N] x 129, while the other elements were calculated as {if/N) x 129. FR = H. feronia, AM = H. amphinome, ARI = H. arinome, IP = H. iphthime, LA = H. laodamia. FR m (»?=13) FR/(fi=5) AM m (n=2) AM/(>j=l) ARI/(j!=1) IP III (»!=1) LA III (;/=l ) FR m 91 [55]*** 28 [23] 1 [9|** 0 [5]* 2 [5] 6 [ 5] 0 [ 5]* FR/ 0 [7]*** 0 [-/]* 0 [2] 0[2] 1 [2] 0 [2] AM m 0 [ /] 0 [ /] 0 [i] 0 [1] 0 [i] AM/ 0 [0] 0 [0] (» [9] ARI / 0 [0] « [0] IP m 0 [0] Hypothe.sis te.st.s Lor .sample proportion vs. hypothesized value: * = P < 0.05, ** = P < 0.01, *** = P < 0.001. Spearman rank correlation was used to assess the relationship between two discrete variables describing the number of events observed. Results Sound production. The sound production tests carried out with recently captured butterflies showed that only males of the five species of the yrrortm group (sub-genus Hamadryas'. H. feronia, H. iphthime, H. epinorne, H. amphinome, H. arinome) produced sounds under the test conditions (Table 2). These were also the only species in which one can perceive .swollen veins in the middle of the proximal border of the males’ forewings. The feronia group females and individuals of both sexes of the laodamia (H. laodamia and H. arete) and februa {H. februa and H. chloe) groups never produced sounds in the hand tests (Table 2) nor in the flight cage. Males of//, feronia, H. epinorne and H. iphthime (/eroruVegroup species) produced most .sounds, whether in flight interactions, when perched over food, or while walking on tree trunks. Data a.ssembled from the literature (Table 3) suggest that there may be variation in sound production capacity of some species; however, since many Hamadryas species have veiy similar color patterns, these results need confirmation. Flying chases. Of the 104 observed flight interactions in which the departing points of the participating individuals were observed, 74 occurred between two H. feronia males and 23 between a male and a female H. feronia. The other 7 interactions involved a male H. feronia and another species. The most frequent interactions between two males were initiated by a flying chaser (64%), while the most frequent interactions between a male and a female were between perched individuals (39%) (y- = 58.7, d. f. - ?>, P < O.OOOl). Aerial chases occurred almost exclusively during periods of intense sunshine, mainly from 1 130 h to 1330 h on hot windless days. Hamadryas males commonly made exploratory flights ca. 1.5 to 2.5 m high, surrounding the two trees inside the cage. Such behavior normally stimulated other individuals in the cage to fly. When encountering a perched individual, flying Hamadryas of the feronia-growp frequently made a pendidar flight display. This display comprised a .semicircular flight with constant .sound production for 5 to 10 s, 10 cm below the perched individual, which could be of either sex. Some perched individuals did not respond to the display, but others took flight after or ahead of the displaying individual. When this happened with two males, a chase was normally initiated with one or both butterflies vigorously producing sounds during the first 10-20 s of the interaction. During a chase both individuals flew rapidly and performed complex aerobatic maneuvers which included downward spiral fliglits, fast dives- and-rises, zigzags, and sudden, momentary (<1 sec) perches. Twenty-five percent of the aerial chases (n = 27) began when an individual walked on the trunk and found and touched another individual. Apparently individuals walking on trunks after feeding were deliberately looking for other Hamadryas w\d\ which to interact (and not for food). Another 33% of the aerial chases (n = 35) began with a perched butterfly darting after a hovering individual, and 18% of the cha.ses (n = 19) began with a flying butterfly provoking a perched individual to fly using the pendular display. The other 24% of the aerial interactions (n = 26) were initiated when two individuals met in flight; these occurred mainly during the hottest hours when many butterflies were flying. Interactions between more than two butterflies were not considered. s /. Res.Lepid. Table 2. Number of individuals producing sound in the hand tests carried out with males and females of nine Hamadryas species in Brazil and Panama. Species group (Siibgenus) Species Sex Number producing sound Number soundless Percent individuals producing sound febma {Agmmi(i) m 0 21 0 I 0 18 0 chlod' m 0 3 0 I 0 2 0 laod a m id ( Peri d ro in id) Idoilnmid- m 0 8 0 I 0 4 0 ari’tr m 0 1 0 I 0 1 0 feronia ( Haniadryds) feronid'- m 35 7 83 f 0 26 0 iphihimr- ^ m 39 9 95 f 0 20 0 cpinomr- m 17 1 94 I 0 15 0 amphinomr' '■ “ m 18 3 86 I 0 19 0 drinomr m 5 0 100 I 0 5 0 1 ) .SV«,v(( Jenkins (1983). Butterflies front: (2) Linhares, ES, Brazil; (3) Campinas, SP, Brazil; (4) Balboa, Panama; (5) Brasilia, DF, Brazil; (6) Praia das Neves, ES, Brazil; (1) Belo Horizonte, MG, Brazil; (8) Serra da Canastra National Park, MG, Brazil; (9) Faz, Jatttba, Correntina, Bahia, Brazil. Table 3. Summary of published reports of sound production in Hamadryas butterflies based on field observations (FO) and manual tests using Otero’s (1990) hand test (OHT). References; (1) Jenkins (1983); (2) Ross (1963); (3) Monge-Najera and Hernandez (1991); (4) Otero (1990). Group/Species Sex Locality Method (FO/OHT*) Sound production Ref. felmid species Group II. amphichlop P Dominican Republic FO Yes 1 //. chloe p ? (dense rainforest) FO No 1 II. fcbrun P Mexico FO Ye.s 9 II. pdmia ??? ??? FO Yes 3 II. februd m and / Venezuela FO/OHT No 4 II. guatemdlend P Mexico FO Yes 9 pronid species Group H. feronia m Venezuela FO/OHT Yes 4 II. feronia / Venezuela FO/OHT No 4 42: 5-12, 2003 (2010) 9 Of the 129 aerial chases with interacting individuals of known sex, among 17 ni and 7 f, 76% (n = 98) were between two males, which was more than expected (Z = 3.097, P < 0.001 ) , 24% (n = 3 1 ) with a male chasing a female (Z = -0.600, P > 0.05), and none with a female as the chaser, which was less than expected (Z = -3.802, P< 0.001) (Table 1). Hamadryas feronia \yAi'\\c\p'Aied in all interactions, generally chasing a conspecific (90.1%, n = 120). Only 8 aerial interactions (6.2%) involved other species, and 7 of these were between H. feronin'cmd a male H. iphthime (Table 1). Chases between two male H. feronia occurred in a higher frequency than was expected (Z = 6.409, P < 0.001), and those involving a male and a female H. feronia occurred in a smaller frequency than was expected (Z = -2.721, P < 0.005) (Table 1). Chases involved fast flights and much sound production. These interactions nsnally ended with one of the individuals perching while tlie other continued flying. Interactions between a male and a female nsnally began with some sound production by the chasing male in a short period of fast flight, after which the male followed the female in a slow flight with little or no sound production until she perched. Frequently, the male would alight behind and court the female (described below). Individuals that were chased a lot seemed to avoid interactions and apparently flew less during the periods of greater exploratoi^ flight activity. The majority of the cha.ses were carried out by only a few individuals (Fig. la) , with one H. feronia (no. 89) performing no less than 57% of all chases (33 of the 58 chases by identified individuals). The next most active males initiated only four chases each. Three of the ten most chased individuals were females, but the frequency distribution of clia,sed butterflies was much more uniform than the frequency distribution of the chasers (Fig. lb). Sometimes more than two /-/rt/wr/rfr'yrw engaged in aerial interactions. These interactions tended to be intense, long lasting, and difficult to keep track of the individuals. Most involved three or four individuals, but a few had up to seven. These happened when individuals met during exploratory flights in the hottest hours of the day and perched individuals joined the flying party. Some of these interactions in which all individuals cotild be identified consisted of two or more males chasing a female. All Hamadryas species of both sexes may defend feeding sites through displays and physical interactions. However, there was no relation between the success of an individual in defending a feeding site and the number of times it chased after other individuals. Only one individual seemed to be efficient in both Figure 1. Frequencies of aerial chases of identified Hamadryas individuals in a flight cage (n = 58). A) Chases effected by a chasing individual. B) Chases suffered by a chased individual. Individuals are represented by a species/sex/number code: AM = Hamadryas amphinome, API = H. arinome, FR = H. feronia, IP = H. iphthime, m = males, f = females, and the capture number of the individual. tasks (f/. feronia no. 89). Of the species engaging in aerial chases, H. amphinome did much better at defending feeding sites than at chasing other Hamadryads (Table 1; Fig. 2, Spearman correlation, N = 14, R = 0.25, t, = 0.90, T= 0.39). Individuals of other species participated in le.ss chases than was expected (Table 1). Courtship. In the cage, six observations were made on male H. feronia cotirting females perched on a tree trunk. All of them around noon. Usually a male perched 5 to 10 cm behind a female and facing her and remained 1 to 5 min slowly opening and closing 10 J. Res.Lepid. Figure 2. Relation between the number of times individuals successfully defended food and the number of aerial chases initiated by the same individuals. AM = amphinome, FR = feronia, IP = iphthime. his wings while slowly approaching the female by taking a few steps at a time. The female responded to this approach with similar wing movement. When the male got close (< 1 cm), she usually walked 15 to 100 cm down the trunk. If the male touched her with his wings or legs, she usually flew to another tree trunk or to the cage walk The male either followed her to the new perch or quit after some unsuccessful attempts. After this he would fly around the trees apparently looking for another individual with which to interact. The first author once observed a much longer courtship in H. epinom.em a reforested area containing Joannesia princeps (Euphorbiaceae) and Senna nmltijuga (Leguminosae). The butterflies behaved essentially like the H. feronia described above, differing principally in that when they changed perches they usually flew over a much larger area (about 20 x 20 in) from 1 to 7 Ill high (canopy height), and usually perched higher up (4-6 m) than did the caged H. feronia. Discussion Our results, added to the observations of Jenkins (1983), Otero (1988) and Yack et al. (2000), indicate that sound production capacity in Hamadryas is restricted to the males of tiie feronia species group, with the possible exception of H. amphichloe, which belongs to the februa species group. Monge-Najera and Hernandez (1991) allude to one population of sound producing H. februa which disagrees with our data and those of other workers. Species of the feronia group are unique in their possession of swollen wing veins associated with sound production (Otero, 1988 and our observations), and the rare reports of sound production in the februa group (Monge- Najera & Hernandez, 1991; Ross, 1963) may result from identification errors. The five populations of H. februa tested in the present study did not produce sound, as is the case of the Venezuelan population studied by Otero (1988, 1990), which contrasts with the results of Monge-Najera and Hernandez (1991) and Ross (1963). Hamadryas feroniais monomorphic, suggesting that sound production or other non-visual cues may be related to the discovery of females for mating. The absence of sound production and the presence of sexual color diinoiq)hism in the laodamia species group suggest that sexual recognition between individuals of this group may have a visual component, since butterflies that possess sexual color dimorphism are supposed to recognize each other visually (Silberglied, 1984), although there is little evidence for this. The absence of both sound production and color dimorphism in the species of the februa species group suggests that chemical signals may be used for determination of gender. All the species mentioned by Jenkins (1983) and Otero (1988) in which sound production has not been noted belong either to the laodamia or februa species groups. Species of the februa group have been reported as capable [H. amphichloe, and H. februa) or incapable of producing sounds {H. chloe (Jenkins, 1983) and H. februa (Otero, 1988)]. This suggests that there may exist different mechanisms of sexual recognition between species of the same group or that the dogma that has developed in relation to the sound production capacity of the whole group Hamadryas has been due to misidentifications of flying individuals because of their great similarity. In the past, this genus had a very confusing taxonomy, comprising about 100 species. Jenkins (1983) synoiiymized more than two thirds of these, leaving only 20 species and 21 subspecies in the genus Hamadryas. Probably many reports prior to the revision of Jenkins (1983) contain identification errors, which may have added to the confusion about the sound production capacity of each species. Jenkins (1983) recorded two copulations for Hamadryas, one at 1130 h {H. amphinome) and the other at 1400 h (H. velutina). Most of the aerial interactions and courtship behavior in the present 42: 5-12, 2003 (2010) Study occurred around noon, simultaneously with the majority of the flight interactions. The sounds were produced at will during flight interactions by male butterflies. Haviadryas feronia has been proven to hear the sounds in the same acoustic range they produce (Yack et ai, 2000) . Therefore, these situations may be used to promote sexual recognition which apparently takes place quickly at the very beginning of chases carried out by males after females. Thus, males initiating chases shift to courtship behavior almost immediately upon approaching a female or proceed in aerial pursuit when finding a male. Some moths {Hecatesia thyridion) and Satyrinae butterflies are among the few Lepidoptera known to produce high frequency sounds, similar to those produced by Haviadryas males, which are voluntarily used in intraspecific communication while in courtship flight (Bailey, 1978; Kiine, 1982; Murillo-Hiller, 2006). Thus, it seems difficult to ascertain what role the clicking sounds play in Haviadryas' behavioral repertoire. We conjecture that the clicks are used as an early recognition of sexual partners and that sound intensity may also be used by females as a means to assess the male’s htness, as bigger healthier males may produce louder clicks. Behavioral experiments are needed to provide evidence for either hypotheses. Although Haviadryas butterflies of both sexes defend feeding sites, the aerial chases observed here do not seem to be associated \vith resource defense or to territoriality (Ross, 1963). Food was provided ad lihituvi in two localized spots in the flight cage and the majority of the individuals had ceased feeding when aerial chases reached a climax (Marini-Filho, 1996). With the exception of one very successful male, males that had the greatest success in defending feeding sites were different from those that initiated aerial chases. W'hile H. feronia was the species that performed the majority of the aerial chases, H. aviphinotne was the one that defended the feeding resource more fiercely. This is consistent with their size difference, as body size determines to a great extent the winner of those interactions (Marini-Filho, 1996). Aerial chases apparently play two intimately related functions: the discovery of receptive females and the establishment of dominance hierarchies for mating priority, although we did not consider this during the experiment. Dominance hierarchies may be the result of natural selection over intraspecihc differences in flight capacity or other fitness-related character, promoting the individual spacing in natural populations and enhancing the mating chances of hierarchically superior individuals (see Rutowski etal, 1989), the priority order being generally established through previous agonistic encounters (Archer, 1988: ,1 14). The establishment of dominance hierarchies may come about by the recognition of the individual aggre.ssiveness by the butterflies present in the .same feeding area (usually one or a few tree trunks oozing fermenting sap). It is unlikely that monomorphic butterflies as other monomorphic insects are able to visually recognize others of the same or similar species (Ewing, 1984). Males and females of most species of the groups feronia and februa have a cryptic marbled color pattern, making them difficult to be recognized in flight. It is more likely that after a series of aerial chases the individuals present in the area can recognize behaviorally that there are other individuals more aggre.ssive a,ssuming then a submissive attitude, either evading chases or refusing to fly during the period of most aerial interactions. Possibly the most chased Haviadryas were receptive, while non- receptive females would not stay close to the food source during the period of aerial chases, and would otherwise be looking for suitable host-plants for their larvae if they could have left the cage. These results may also help the definition of the Haviadryas phylogeny once it seems that there is a high agreement between the ability to produce .sounds observed on the males of the feronia species group and the inability to produce sound in the other two species groups {laodaviia And februa). The ability of //. aviphkhloeio produce sound must be checked by hand test and further considered with the morphologic factors to find if it is an exception from this pattern or in fact more associated to the feronia species group than to the februa species group. Acknowledgements We thank the direction and staf'f'oftlie Linharcs Forest Preserve for essential logistic stijtport dtiring this study. We are grateltil to Andre V. 1.. Fieitas, Carla M. Penz. Rogerio P. Martins, Dottglas Yanega, Robert Srygley and .Stanley Rand Ibr discussions and/or critical cotnnients on earlier drafts of the inaiuiscript. Sttidy grants to 0|M-F were |)rovided by Coordenadoria de Aperfeicoainento de Pessoal de Nivel Sttjterior (CAPES) and tbe Smithsonian Tropical Research Institute. Financial support for this sttidy was |>rovided by FAPESP to WWB. Literature Cited Archkr, J. 1988. The behavioural biology of aggre.s.sion. Cambridge University Press, Cambridge. B.vii.i'.v, W. J. 1978. Resonant wing systems in the Atistralian whistling tnoth Hecatesia (Agarasidae, Lepidoptera). Nature 272: 444-446. B.vtes. H. W. 1865. Ciontrihtitions to an insect fauna of the Amazon Valley. Lepidoptera-Nymphalinae. Journal of Entomology 2: 311-315. Bernard, G. D. & C. L. Reminoton. 1991. Color vision in Lycaena htitterflies: spectral timing of receptor arrays in relation to 12 J. Res.Lepid. beliavioral ecolog)'. Proceedings of the National Academy of Sciences (USA) 88; 2783-278?/ Boppre. M. 1984. Chemically mediated interactions between butterflies. In: Vane-Wright, R.I. & P.R. Ackery (eds.), The biology of butterflies, Symposium of the Royal Entomological Society of London, no. 11, pp. 259-275. Academic Press, L.ondon. Darwin, C. 1871. The descent of man and selection in relation to sex, Vo!. 1. John Murray, London. Jenkins, D. W. 1983. Neotropical Nyniphaiidae 1. Revision of Hamadryns. Bulletin of the Allyn Museum 81: 1-146. K.,\ne, S. 1 982. Notes on the acoustic signals of a Neotropical satyrid butterfly. Journal of the Lepidoptertsts’ Society 36: 200-206. Lees, D. C. 1992. Foreleg stridulation in male Urania moths (Lepidoptera: Uraniidae). Zoological Journal of the Linnean Society 106: 163-170. M.\rini-Fii,h(), O. J. 1 996. Defesa de reciirsos alimentares e interagoes aereas entre borboletas simpatricas do genero Hamadryas. Master’s tliesis, Lhiiversidade Estadual de Campinas, Brazil, ix -h 73 pp. Monc;e-NA|Era, J. & F. Hern.andez. 1991 . A morphological search for the sound mechanism of Hamadryas bulterflies (Lepidoptera: Nymphalidae). Journal of Research on the Lepidoptera 30: 196-208. M(>n('.e-N.4|er.-\, J., F. Hernandez, M. 1. Gonz.-vlez, J. Soley, J. Araya & S. Zoi.i.A. 1998. Spatial distribution, territoriality and sound production by tropical cryptic butterflies (Hamadryas, Lepidoptera: Nymphalidae): implications for the ‘industrial melanism' debate. Revista de Biologia Tropical 48: 297-329. Murii.i.o-Hii.i.er L. R. 2006. A noise producing butterfly, Yphthimoides castrensis (Nymphalidae, Satyrinae) from south Brazil. Journal of the Lepidopterists’ Society 60: 61-64. Otero, L. D. 1988. Contribucidii a la historia natural del genero Hamadryas (Lepidoptera: Nymphalidae). Tesis Doctoral, Universidad Central de Venezuela, Macaray. viii + 108 pp. Otero, L. D. 1990. The stridulatory organ in Hamadryas (Nymphalidae): preliminary observations. Journal of the Lepidopterists’ Society 44: 285-288. Ross, G. N. 1963. Evidence for lack of territoriality in two species of Hamadryas, Nymphalidae. Journal of Research on the Lepidoptera 3: 207-229. Rutowski, R. L., j. Ai.cock & M. C.arey. 1989. Hilltopping in the pipevine swallowtail butterfly (Battus philenor). Ethology 82: 244-254. Seitz, A. 1913. The macrolepidoptera of the world. Alfred Kermen, Stuttgart. Sii.BERca.iED, R. E. 1984. Visual communication in sexual selection among butterllies. In: Vane-Wright, R. I. & Ackery, R R. (eds.). The biology of butterflies, Symposium of the Royal Entomological Society of London, no. 11, pp. 207-223. Academic Press, London. SwiiLART, S. L. 1967. Hearing in butterflies (Nymphalidae: Heliconius, Ageronia). Journal of Insect Physiology 13: 469-476. Wickman, R O. & C. WiCKi.UND. 1983. Territorial defense and its seasonal decline in the speckled wood butterfly (Pararge argeria) . Animal Behaviour 31: 1206-1216. Yack, J. E., L. D. Otero, J. W. Daw.son, A. Suri.ykke & J. H. Fuliard. 2000. Sound production and hearing in the blue cracker butterfly Hamadryas feronia (Lepidoptera: Nymphalidae) from Venezuela. The Journal of Experimental Biology 203: 3689-3702. Journal of Research on the Lepidoftera 42: 13-20, 2003 (2010) Ecobiology of the common castor butterfly Ariarfwc merione merione (Cramer) (Lepidoptera: Rhopalocera: Nymphalidae) Janaki Bai Atluri'*, Samatha Bodapati', Bhupathi Rwalu Mataia', Sandhya Deepira, Devara' and SUBBA Reddi Chiiakaia- ‘Department of Botany, Andhra Lhiiversity, Visakhapatnam - 530003, Andhra Pradesh, India -Department of Environmental Sciences, Andhra University, Visakhapatnam - 530003, Andhra Pradesh, India jhalluri_adit'(a007@\ahoo.com Abstract. We describe the life histoiy of the comtnon castor butterfly, Ariadne merione merione, monthly occurrence and seasonality' of eai ly stages and larval performance in terms of food consnmjition and ittilization, and the length of life cycle. Onr study was conducted during 2002 itt the Andhra University campus at Visakhapattiam (17'’42' N, 82°18' E), South India. Field study itidicated that A. merione merione wna in continitoiis flight and reproduction, with highest densities of earlv and achtlt stages occurring during June - Se|)tember, the time of the entire South-West monsoon. Occurrence of the early stages was positively, but non-significantly correlated with rainfall, relative humidity, tetnperatnre and day-length. Multiple regression analysis showed that the effect of any combination of weather parameters on the rejtroductive activity was less than 40%. The Sotith-West monsoon period jtrobably infhienced the reprodttetive activity by promoting fresh growth of the larval host plant, Ririnus ronimunis, which in turn sup|rorted development of early stages. Ariadne merione meiionew’-As. exemplified by a life cycle of 27.4 ± 3.57 days (eggs 3-4, larvae 1 3-18, and pupa 6-9 days) permitting a maximum of 8-9 overlapping generations per year. The values of the tuitritioniil indices across the instars were A.l). 87.02-95.50%; E.U.l. 3.80-20.90%; E.U.D. 4.00-24.08%, measured at 28°U in the laboratoi-)'. These relatively high values, at least partially explain the ecological success of A. merione merione 'm the urban environment. Key words: castor butterfly, Ariadne merione, life history, po|)itlatioti index, mttritional indices. Introduction Of the estimated 20, ()()() - 30, GOO sj^ecies of butterflies occurring globally, at least 1,500 species occur in India. Several field guides for the identification of the Indian btitterflies are available (Wynter- Blyth, 1957; Haribal, 1992; Gay et al, 1992; Gunathilagaraj et ai, 1998; Kunte, 2000 and the references therein). A list of the works giving the descriptions of the life histories was given by Pant and Ghatterjee (1950), of which those of Bell (1909 - 1927) are important. However, review of these early works indicated that for many species data, particularly on the duration and phenology of early lifestages, are either absent or incomplete. Haribal (1992) noted that the life histories of nearly 70% of the Indian species require description. We began sttidies to address the sittiation. Here we describe the life history of Ariadne merione merione (Gramer), the common castor butterfly, of the Oriental region. It is a sj^ecific pest of the castor ^Corresponding author Received: 24 July 2008 Accepted: 20 August 2008 seed jjlant Jiicinus communis (Nayar el ai, 1976) and the larvae also feed on the stinging nettles Tragia involucrata And T. Jduhenetti (Euphorbiaceae) (Ktmte, 2000). Because reprodtictive efficiency dej^ends on life style and feeding pattern (Boggs, 1981; Slansky & Scriber, 1985; Muthnkrishnan Sc Pandian, 1987), we also studied larval performance with resj^ect of food tuilization by feeding them on a daily stij3|)ly of pieces of fresh leaf of the castor j^lant. Materials and methods The sttidy was conducted dtiring the year 2002 in the Andhra Lhiiversity campus (168 ha) at Visakhapatnam (17°42’ N, 82°18’ E) situated in the east coast of India. The natural plant commtmity of the camjius was searched for the distribution and reproductive activity of the common castor btitterfly Ariadne merione merione (Cramer). Adult btitterflies were seen mostly near the larval host plant Ricinus communis Linnaetis. Once located detailed observations were made at 10 sites in order to observe the flight activity and abundance of adtilLs, the period of cojiulation and ovijiosition, following which we collected fresh eggs to study the life history and the duration of early stages. After oviposition, the leaf 14 J. Res.Lepid. with egg(s) was collected in Petri dishes (15 cm x 2.5 cm depth) and brought to the laboratory. The piece of the leaf with the egg was then placed in a smaller Petri dish (10 cm x 1.5 cm depth) the inside of which lined with moistened blotter to prevent the leaf from drying. Five such samples were placed in a cage covered with wire mesh. The laboratory temperature was 28 ± 2° C and relative humidity 80 ± 10% with normal indirect sunlight conditions that varied in duration between 12h during November/Jamiary and 14h dnring june/july. The eggs were then examined at 6h intervals daily for recording time to eclosion. The larvae were subsequently reared on a weighed quantity of fresh leaves supplied daily. The time of each moult was noted. The morphological characters, body measurements, body weight of each instar and the faeces egested were taken daily. The prepiipal behavior of the final iiistar, pupal particulars and the time of adult eclosion were also recorded. Larval performance in terms of food utilization indices were calculated as described by Waldbauer (1968) as: arbitrary scale of rare, less common, and common. The relation between the monthly distribution of early stages and prevailing rainfall, relative humidity, temperature, and day-length was assessed by statistical correlation and multiple regression analysis using Minitab Statistical Software 14, 2003. Results Adult stage (Fig. la, b) Both male and female adults were nearly identical, characterized by their reddish brown colored wings bearing black colored wavy lines. Copulations occurred during mid-day, mostly between 1100 - 1500 h and lasting for more than one hour. Adults were found feeding on spoiled flowers of Lantana camam, overripe, fallen and damaged fruits of Annona squamosa, Syzygiuni cumini and Artocarpus heterophyllus, and the sap oozing from wounds in the tree trunks of Citrus aurantifolia. Food c{>nsuinpti()ii index (C'.l.) Relative gnjwth rate (Ci. R.) Approximate digestibility (A. D.) F.tliciency ol conversion of digested food (E. C. D.) Egg stage (Fig. Ic) Wl. of foofl consumed Wt. of instar X No. of feeding days Wt. gained by the instar Mean wt. of instar x No. of feeding days Wt. of food ingested - Wt. of faeces - X 100 Wt. of food ingested Wt. gained bv the instar - X 100 W't. of food consumed * Wt. of faeces Gravid females lay eggs singly on the under surface of the leaves of the castor plant mostly before mid-day, between 0900 -1200h. Females spread their wings during egg laying, depositing 1 to clutches of 15. There was no bias for the age of the leaf. During one survey old leaves had 1-7 and young tender leaves 1 - 3 eggs. The eggs were round, 0.80 - 0.90 (0.83 ± 0.04) mm in diameter. At oviposition they were white, the color changing to light brown before hatching. Wlien first laid eggs appeared soft in texture, but within 6 - 10 seconds they became hairy. They hatched in 3 - 4 days. Soon after hatching, larvae ate their egg-shells. Each larva passed through five distinct instars over a period of 13 - 18 days. Ff ficicncy ol conversion of ingested food (E. C. 1.) Wt. gained bv the inslar - X 100 Wt. of food ingested To determine the developmental success of each of the early stages, a number of eggs were placed in Petri dishes in each month and the number of larvae hatched, pupae formed and the adults eclosed were recorded. To record the different early stages on the natural host plant, one plant at each of the 10 study sites was thoroughly searched at 10 day intervals eacli month and the early stages found were enumerated and pooled for each month. During the same visits, the (light frequency of adults was also noted using the Larval stage (Fig. Id-h) Instar I lasted for 2-3 days. Larvae were 1 .8 - 2.0 ( 1 .9 ± 0.08) mm on Dl, growing to a length of 2.50 - 3.00 (2.80 ±0.21) mm and width ofO.30- 0.50 (0.43 ±0.09) mm before moult. Body was somewhat rectangular in siiape, but slightly narrowing posteriorly. Its color was pale brown immediately after hatching, later turning brownish green with three brown colored horizontal bands on dorsal side. Head was very minute, and brown. Instar II also lasted for 2-3 days and attained a length of 3.30 - 4.00 (3.73 ± 0.30) mm and width of 0.60 - 0.90 (0.73 ± 0.12) mm. Whitish green spines with branched ends appeared over the entire body. Head was brown with a pair of brown horns. There 42; 13-20, 2003 (2010) 15 were no changes in other characters seen in instar I. Instar III lasted for 3-4 days. Developing to a length of 6.00 - 8.00 (7.00 ± 0.81) mm and width of 1.10 - 1.50 (1.36 ±0.18) mm. Dorsally they had a yellowish green broad stripe with brown edge longitudinal to the body. The body spines present on the three brown horizontal bands were also brown. Head was 1 mm in size, blackish brown in color with white markings. The head horns were 0.80 - 1.00 (0.90 ± 0.08) mm long and branched. Legs were clearly visible. The larva did not move much, but moved its head continuously when disturbed. There were no changes in other characters from previous instar. Instar IV also lasted for 3-4 days, growing to a length of 1 0.00 - 1 5.00 ( 1 2.00 ± 0.21) mm and a width of 1.50 - 2.00 (1.73 ± 0.20) mm. Body became green in color. The dorsal stripe turned iDiown with yellowish cream edges. The three black horizontal bands began to disappear. Head was blackish brown in color, square shaped and measured 1 .00 - 2.00 ( 1 .53 ± 0.41 ) mm in diameter. There were three triangular white markings on the head. The head horns were reddish brown in color and measured 2 mm in length. Segmentation was clear. Body spines were green in color, arranged in four lines on each side of the body on all the segments. The legs were green. Instar V also lasted for 3-4 days. When full grown the larva was 23.0 - 30.0 (25.6 ± 0.32) mm long and 2.20 - 3.00 (2.73 ± 0.37) mm wide. Body was dark green. The dorsal stripe changed to orange with black edges showing numerous small white to cream colored spots. The dorsal three horizontal bands disappeared completely. Head was 2.00 - 3.00 (2.56 ± 0.41 ) mm in diameter. It had prominent white triangular markings with black border two present above and one below. The horns became orange in color, with black tips, and measured 3.00 - 4.00 (3.60 ± 0.43) mm in length. Light and dark green crossed lines developed on both lateral sides of the body. The color of spines changed to brown with black tips and with yellow to orange colored spots at their base. Pupal stage (Fig.li) During the prepupal period of 1 - 2 days the full- grown larva stopped feeding, turned brown and its lateral crossed lines changed to brown and white. The body contracted and the larva attached itself to the substratum with its posterior end hanging downwards. It measured 20.00 - 25.00 (22.60 ± 0.20) mm in length and 3 mm in width. The pupal stage lasted for 5-7 days. The brown color changed to black with pupal maturation until adult eclosion. It measured 15.00 - 17.00 (16.00 ± 0.08) mm in length and 6.00 - 7.00 (6.46 ± 0.41) mm in width at the broadest end. The anterior end was narrow. At the broadest point both lateral sides were curved inwards, between which two pointed projections appeared on dorsal side. Average pupal weight was 202.3 mg. Development success and population index Hatching success varied between 40 and 100%, being highest duringjune to September. Both larval and pupal development success varied between 50 and 100%, (Table 1). The numerical frequency of eggs, larvae, pupae recorded on the host plants and adult abundance, along with the prevailing weather data are given in Table 2. The three early stages and adults could be found under natural conditions throughout the year. However, the period ofjune and September provided the highest frequency of all stages, with peak numbers in July. Correlation between the counts of early stages and monthly average temjterature, average relative humidity, total rainfall, and average day-length was positive, but non-significant, the coefficient values being 0.566, 0.333, 0.468, and 0.521 respectively. The four weather variables jointly influenced the distribution of early stages to the extent of about 40%, as indicated by multiple regression coefficients, R- 0.216-0.396 (Table 3). Other combinations including temperat lire /rain fall-/ day-length, temperature/ Table 1 . Hatching, larval and pupal development success of Ariadne merione merione in the laboratory. Life cycle stage Calendar nioiith.s J F M A M J J A S O N D # eggs incubaled 4 4 5 4 5 10 17 10 6 7 5 0 # larvae hatched 2 3 2 3 3 10 17 8 5 5 5 4 # pupae formed 1 2 1 9 3 9 12 8 5 4 4 3 # adults emerged 1 1 1 2 9 9 11 8 4 3 3 3 16 /. Res.Lepid. Table 2. Distribution of early stages of Ariadne merione merione on Ricinus communis and the associated weather conditions. Life cycle stage, Calendar months J F M A M J J A S O N D Eaiiy stages 7 7 9 8 24 42 117 61 33 21 14 16 Adults ❖ 4: ** Temperature (°C) 24.15 25.45 27.85 29.15 30.7 29.4 30.75 28.2 29.25 28.3 26.2 24.55 Relative humidity (%) 74 68 74 74.5 , 71.25 77 73 80.5 76.5 74 62.25 69 Rainfall (mm) 014.1 000.0 000.0 085.2 015.1 143.2 075.4 143.5 023.5 118.4 007.8 000.0 Daylength (h) 1112 1148 1215 1312 1337 1316 1304 1322 1232 1224 1132 1105 * Rare, ** Common, *** Very common. Table 3. Multiple regression of the counts of the early stages in relation to the prevailing weather parameters. Constant (A) X. X., X3 X, R2 -229.6 7.342 0.758 0.331 -153.7 6.336 0.139 0.369 -195.5 7.707 0.009 0.321 -7.1 0.338 0.239 0.216 -208.4 0.188 0.183 0.272 -146.2 0.118 0.138 0.297 -137 6.419 -0.270 0.152 0.370 -229.6 9.48 1 .094 -0.068 0.335 -87.7 11.88 0.220 -0.183 0.395 -121.1 -0.497 0.137 0.146 0.300 -101.3 12.21 0.301 0.211 -0.196 0.396 X| - Monthly average temperature; X„ - Monthly average relative humidity; - Monthly total rainfall; X^ - Monthly average daylength. Table 4. Food consumption, growth and food utilization efficiencies of Ariadne merione merione larva fed with Ricinus communis leaves. Instar Wt. of food Wt. of faeces (mg) Wt. gained by GR Cl AD ECD ECI number ingested (mg) larva (mg) (mg/day/mg) (mg/day/mg) (%) (%) (%) 1 11 45.0 ± 10.03 2.0 ± 0.35 1.72 ±0.1 6 0.42 11.02 95.50 04.00 03.8 III 150.0 + 16.39 13.0 ±2. 16 10.85 ±0..59 0..34 04.80 91.30 07.90 07.2 IV 2,50.0 ± 05.65 25.0 ± 5.09 31.00 ± 1.65 0.36 02.90 90.00 13.70 12.4 V 925 ± 22.22 120.0 ±5.88 193.87 ±2.61 0.45 02.16 87.02 24.08 20.9 - ln(licate.s no data due to very small size of first instar. 42: 13-20, 2003 (2010) 17 Figure 1 . Photographs of the sequential stages in the life history of Ariadne merione merione. a) Adult pairing, b) Adults feeding on the damaged fruits of Annona squamosa, c) Egg. d) Instar I. e) Instar II. f) Instar III. g) Instar IV. h) Instar V. i) Pupa. 18 J. Res.Lepid. relative humidity/ rainfall, and temperature-rainfall also liad similar to lower values. Food consumption, growth and utilization The data for the weight of food consumed and weight gained by the larvae are given in Table 4. The same data could not be collected for instar I due to its small size with consequent danger in handling. The amount of food consumed increased from instar to instar, the proportion of total food consumed in instars from II to V being 3.28, 10.94, 18.24, and 67.51%. Thus, there was greatest consumption in instar V. The weight gain corresponded to the food consumption trend of the respective iiistars. The weight gain in instar V was 81 .65% of total larval weight. The weight of successive instars plotted against the food consumed indicated a clear relationship between these two parameters (y = 0.227 X and 18.383; r = 0.9963). The values of growth rate (G. R.) decreased from instar II to III and then increased to instar V, the values varying between 0.34 and 0.45 mg/day/mg. Consumption index (C. I.) progressively decreased from iiistar to instar, the values ranging between 2.16 and 1 1.02 mg/ day/mg. Table 4 also includes the indices of food utilization efficiencies A. D., E. C. L, and E. C. D. The range of A. D. values was 87.02 to 95.5%, that of E. C. I. 3.8 to 20.9% and E. C. D. 4.0 to 24.08%. While E. C. I. and E. C. D. decreased, A. D. increased as the larvae progressed. Discussion The year round occurrence of early stages on the host plant Ridnus communis showed that Ariadne merione merione breeds continuously, corresponding with the usual pattern noted for most tropical butterflies (Owen, 1971; Owen et ai, 1972). The period of highest frequency from June to September correlates with the South - West monsoon. Thus rainfall appears to be the most important factor promoting higher reproduction rates in A. merione merione as is the case for both Catopsilia crocale (Chirstopher & Mathavan, 1986) and Catopsilia pyranthe (Atluri et ai, 2004a). However, statistical correlation of the distribution and abundance of early stages with the rainfall, though positive, was non-significant. Precipitation during the South-West monsoon likely had its influence on reproduction via the host plant. During this season, the host plant had its greatest fresh growtii, a resource needed by the larvae for better performance due to the likely higher levels of nitrogen and water content (Slaosky & Feeny, 1977; Scriber, 1977; Mattson, 1986). Although Figure 2. Relationship between food consumption and growth in Ariadne merione merione on Ridnus communis. the host plant was available throughout the year, leaf quality in terms of nitrogen and water content might have varied through the year, hence the observed trend in the pattern of reproduction of A. merione merione. The work of Pullin (1987) on the growth of larvae of Aglais urticae fed with foliage with different water contents also suggested the likely variations in the breeding success as being due to variations in rainfall. Pollard et al. (1997) also examined a similar relationship. The low incidence of early stages during periods other than the South - West monsoon could have been due to a decrease in mature egg number as reported by Braby (1995) in the Satyriiie butterflies, which also breed continuously. Few other species noted at the study biotope also reproduced all year, but at a higher rate during different periods: Pachliopta arisolochiae April to May, and October to November (Atluri et al, 2004b) , Papilio polytes August to Februaiy (Atluri et al, 2002) , Graphiurn agamemnon August to December (Veiikataramana et al, 2003a), Eurema hecabe September to November (Veiikataramana et al, 2003b), Euploea core November tojaniiaiy (Veiikataramana et al, 2001). For most of India, Wynter - Blyth (1957) rated spring as the most favorable period, followed by post monsoon and South - West monsoon. In the northern western Ghats, Kimte (1997) observed highest flight activity during late monsoon (August to September) and early winter (October to November). These differences in the phenology of butterflies suggest that different species respond differently to the prevailing environmental seasonality and exhibit different life history patterns. Even different species of a genus may behave differently as observed by Jones and Rieiiks (1987) in tlie three species of the tropical Eurema they studied. 42: 13-20, 2003 (2010) 19 The overall effect of weather on population trends is complex and difficult to predict, as also expressed by Pollard (1988). The characters of full grown larva observed in this stndy substantiate those given in Bell (1910) and Sevastopulo (1939) as well as pupal duration. The total development time from egg laying to adult eclosion was determined as 27.4 ± 3.57 days at about 28°C, thus permitting a maximum of 8 to 9 overlapping broods per year. This behavior is in line with the expectation of tropical butterflies to have a short life cycle, and multiple broods over the year (Owen, 1971). Since temperature influences instar duration and the overall development time (Mathavan & Pandian, 1975; Palanichamy et ah, 1982; Pathak & Pizvi, 2003; Braby, 2003) , tlie brood ntun!)er in other parts of A. nierione merione distribution may vary from our records depending on the prevailing temperattires. As no temperature extremes occur at Visakhapatnam, especially at the Andhra University site, the duration of life cycle did not vary much over the overlapping seasons. Adult feeding on the damaged and ripened fruit helps them obtain proteins and carbon sources (Tevey & del Rio, 2001 ), with such nutrient uptake improving egg productivity (Fischer et al, 2004) . The larval food also appears to be highly nutritional as indicated by the observed values of assimilation efficiency (A. D.), the efficiency of conversion of ingested food (E. C. 1.) , and the efficiency of conversion of digested food (E. C. D.) into the body substance. The chemistr)’ of the leaf, particularly its nitrogen and water content, influences the assimilation efficiency (Pandian & Marian, 1986). The castor leaves contain 2.54% nitrogen and 75.20% water (Senthamizhselvan & Mnrngan, 1988). Hence (he observed high A. D. value, mean 90.97%. Stich high values are characteristic of the foliage feeders (Slansky & Scriber, 1985) and indicative of their high growth efficiency (Singhal, 1980). The values of E. C7 D. and E. C. 1., particularly those of the last two instars, are also relatively high (12.4%, 20.9%; 13.7%, 24.1%), thus respectively indicating tissue growth efficiency and ecological growth efficiency, which enabled A. merione nierione io thrive successfully in the urban environment. Acknowledgements riie autliors thank Profe.ssnr Frances S. Cliew of tlie De|3artinent of Biology of Tufts University. Medford, USA for appraising ottr work. We also thank Mr. S. Naresh of the Andhra University Statistics Department for help in the statistical analysis of the data, and the Librarian, BNHS, Mtnnbai, for the supply of literature. Two anonymous reviewers improved otir original manuscript for which we thank them. Literature cited An.L Ri, f . B., S. B. VF.N'ivVtA Rvmana & C. Sriiiw Rmni. 2002. Autecology of the common mormon htitterfly Papilio polyles {Le]jido])tera: Rhopalocera: Pajtilionidae). Journal of Environmental Biology 23(2): 199-204. Art.i'Ki, |. B., S. P. Vknk.\ta Ramana, & C. Suisiu Ri nni. 2004a. Ecohiology of C.alopsilia pyranlhe, a trojtical pierid hutterlly. (airrent Science 86(3): 4.57-461. ATtx'Ri, |. B., S. P. Venk,\ia Ramana, D. Krishna Ri nni & SuititA Ri nni, C. 2004b. Ecohiology of the common rose htitterfly /-’rtc/t/t'o/t/r/ arislolochiae (Lepidoptera: Rhopalocera: Papilionidae) . Proceedings AP Akademi of Sciences 8(2): 147-154. Brabv, M. F. 1995. Larval and adult food plants for some tropical satyrine butterflies in northern Queensland. Australian F.ntomologist 22( I ): .5-13. Br.\bv, M. F. 2003. Effect of temperature on develo|)ment and survival in Deltas iiigriria (Fabricius) (Lepidoptera: Pieridae). Australian Journal of Entomology 42(2): 138-143. Bttt.i,, 'L R. 1909-1927. The common hutterilies of the |)lains of India. Journal of Bombay Nattiral History .Society vol. 19(1) -vol. 31(4). Bf.i.i,, T. R. 1910. The common butterflies of the plains of India. 72 Krgolis merione. Journal of Bombay Natural History Society 20: 319. Btrc.os, C. L. 1981. Nutritional and life histoiy determinants of resource allocation in holometaholous itisects. Atnerican Naturalist I 17: 692-701. Chris roi’tiFR M. S. M. & S. M vfhav.an. 1986. Population dynamics ofa tr( ipical le|tido|3tcran CntopsiUa rroeale ( Pieridae) . Proceedings of Indian Academy of Sciences (Animtil Sciences) 95(3): 303-324. Fi.schfr, K., D. .M. O'Brifn & C. L. Boons. 2004. Allocation of larval and adult resources to rejtroduction iti a fruit feeding butterfh'. Functional Ecology 18: 656-663. G.\v, T, 1. D. Ki Hi.MtLVR R J. C. Pt'NiniiA. 1992. Common butterflies of India. Oxford University Press, Bombay. 67 pp. Gunahiii.aoarai, K., T. N. A. PERfMAt. K. Javaicam &: M. Ganesh Kumar. 1998. Field guide. Some .South Indian butterflies. Nilgiri Wildlife &: Environment Association, Udhagamandalam, India. 271 pp. Haribai..\L 1992. The hutterilies of Sikkim Himalayas and their natural history. Sikkim Nature (ionservation Foundation (.SN(;F), Gangtok. 217 pj). JttNF.s, R. E &• J. Ri INKS. 1987. Reproductive seasonality in the tro|rical genus Eurema (Lepidoptera: Pieridae). Biotropica 19(1): 7-16. Kuntf, K. 2000. Butterflies of Peninsular India. Utiiversities Press (India) Limited, Hyderabad. 254|)p. Kuntf. K. J. 1997. Seasonal ])atterns in hutterlly abundance and s|)ecies diversity in four tropical habitats in northern Western Ghiits. Journal of Bioscience 22: 593-603. Leaky, D. J. & C. .M. Dei, Rio. 2001. It takes guts (and more) to eat fruit: lessons from avian nutritional ecologv'. Auk 118: 819-831. M.aih.aa'an, S. & T. J. Pandian. 1975. Effect of temperature on food utilization in the monarch butterfh' Danaus chrysippus. Oikos 26: 60-64. M.vnsoN, W. J. 1980. Herbivoty in relation to plant nitrogen content. Annual Reriew of Ecololgv' and .Systematics 11: 119- 161. Miniiab. 2003. Statistical SoftAvare 14. Mu fhukrishnan.J. & T. j. Pandi an. 1987. Insect energetics, |5]3. 373- 511. In: Animal Energetics, Pandian, T. J. and F. J. Vernherg (eds), Academic Press, Netv York. Naaar, K. K., T. N. Ananth.akrishnan & B, V. D.aaid. 1976. General and applied entomology. Tata McGraw Hill Company Lttl., NeAv Delhi. 590 pp. 20 /. Res.Lepid. Owen, D. F. 1971. Tropical biiUoi flies. Clarendon Pre.ss, Oxford. 20,5 p. Owen, I). K, |. Ov\t,n & 1). O. Cei.vnter. 1972. Sea.sonal changes in relative abundance and estimates of species diversity in a family of tropical butterflies. Oikos 2.S: 200-205. F.M.\Ni(ai..\MV, S.. R. I’oNNUCii.wiY & T. Th.vnu.IiRXI. 1982. Effect of temperattire on food intake, growth and conversion efiiciency Eupterote mollip’ra (Insect: Lepidoptera) . Proceedings of Indian Academy of Sciences (Animal Sciences) 91: 417-422. P.vnuian T. |. & M. If M.\ri.\n. 1980. Prediction of assimilation efiiciency in lepidopterans. Proceedings of Indian Academy ofSciences (Animal Sciences) 95(6): 641-665. Pan i , C. D. & N. C. Cii.vrrER|EE. 1950. A list of described immatitre stages of Indian Le[)idoptera. Part 1. Rhopalocera. Indian Forest Records (New' Series) 7: 213-225. P,\iii.\K, .VI. & P. Q. Pi/A'i. 2003. Age specific stirvival and fertility table ot Papilio demokus At different set of temperatures and host plants. Indian Journal of Entomology 65(1 ): 123-126. P()i,i.ARi), E. 1988. Teinjaerature, rainfall and bittterfly numbers, joitrnal of Applied Ecology' 25: 819-828. P01.1.ARI), E., J. N. Ce..vi()RE\-1)..\vie.s & |. A. Tiiom.as. 1997. Drought reduces breeding success of the butterfly Aglais urticae. Ecological Entomology 22: 31.5-318. Pui.i.iN, A. S. 1987. C.bauges in leaf t]uality following clijjping and I egrowtb in Urtira tiioicti, and consequences for a specialist insect herbivore Aglais iirlirar. Oikos 49: 39-45. ScRiiiER, J. .VI. 1977. Limiting effects of low’ leaf water content on nitrogen utilization, energy budget, and lanal growth oi Hxlapliora arropia (Lepidoptera: Saturniidae). Oecologia 28: 269-287. SEN'riiAMi/,ii.sEi.v.4N, M. & K. .Vlt'RL'UAN. 1988. Bioenergetics and rejrroductive ef ficiency of /l/rr/c/owe(/;/(rt cmitilalaF. (Orthoptera: Insecta) in relation to food quality. Proceedings of Indian Academy of Sciences (Animal Sciences) 97(6): 505-517. Sev.vstopui.o, D. G. 1939. The early stages of Indian Lepidoptera. Journal of Bomhay Natural Histoiy Society 41(1): 72-81 . SiNGH.M., R. N. 1980. Relationships betw’een ecological efficiencies of a herbivore and a carnivore insect. Indian Jotirnal of Ecology 7(1): 71-76. Si.vnskvF. & j. M. Scriiser. 1985. Food consumjition and ittilization, ]3p. 8.5-163. In: Comprehensive Insect Physiology, Biochemistiy and Pharmacology, Kerkuit, G. A. & L. 1. Gilbert (eds), Pergamon, Oxford. StAN.SKY, F. & P. Feeny. 1977. Stabilization of the rate of nitrogen accitmulation b\ larvae of the cabbage butterfly on w'ild and cultivated food jilants. Ecological Monographs 47: 269-228. Venryia Rymana, S. P, |. B. Ati.uri & C. Si bba Reddi. 200 1 . Autecology' of the common crow' butterfly. Ecology Environment & Conservation 7(1): 47-52. \'enk/\ta R.xmana, S. P., J. B. Atluri & C. Subba Reddi. 2003a. Autecology of tailed jay butterflv Graphium agamemnon (Lepidoptera: Rhopalocera: Pajiilionidae) . Journal of Environmental Biology' 24(3): 29.5-303. Venkata R,\mana, S. P., J. B. Ati.uri & C. Subba Reddi. 2003b. Biology and food ittilisation efficiency oi Eumna hecabc (Lepidoptera: Rhopalocera: Pieridae). Proceedings of the National Academy ofSciences, India 73,B(1): 17-27. W.M.DBAUER. G. P. 1968. The consutnption and utilization of food bv insects, pp. 229-288. In: Beament, Treherne & Wigglesworth (Eds.) , Advatices in insect physiology', Acadetnic Press, London and New' York. Wynter-Bi.vtii, M. a. 1957. Butterflies of the Indian Region. Bombay Natural flistory Society, Bombay. 523 pp. fournal of Research on the Lepidoftera 42: 21-33, 2003 (2010) Larval feeding behaviour and myrmecophily of the Brenton Blue, Orachrysops niobe (Trimen) (Lepidoptera: Lycaenidae) David A. Edcif* and Hliib van Hambl'rc; Scliool of Environmental Sciences and Development, North-West University, I’rivate Bag XOOOl. Potchefstrooni, 2.520, Sotith Alrica. daveedge@xnets. co. za Abstract. The larval feeding hehaviotir and myrinecophih' of the Brenton Bltie Orachrysops niohe. an endangered polyommatine htitterfly from Knysna in Sotith Africa, were investigated by field observations and captive larval rearing. The aerial and stihterranean parts of the Indigofera erecta legtmie host plants were searched for O. niohe eggs, larvae and potential host ants. I’hird and fonrtli instar larvae and ptipae were fotind in association witfi Camponotns baynei dwts on tfie host plant rootstock. Ant colonies in viewable artificial C. baynei nests were sited near host jtlants bearing nudtiple O. niobe eggs, bm no larvae were taken into the nests. Uannibalism was observed between larvae raised in captivity on cm host plant. A third instar captive larva was enclosed with a potted host plant connected to a similar artificial ant nest. The larva disappeared and was later fotind feeding on the depleted plant rootstock, tended Itv the ants, and tfiis behaviour was confirmed by field observations. O. niobe’s ant association is inferred to be obligate. Leguminous Indigoferahosi plants have amino acid enriched rootstocks, wliich inav have pre-adapted the larval digestive system to a cannibalistic or carnivorous lifestvle. I.arval growth characteristics are used to compare African polvommatine genera and Orachrysops is intermediate fietween the facultative myrmeco|thilous genera and the predaceous/parasitic Lepidochrysops species. A cladistic analysis based on host jilants, ant associations and feeding behaviour leads to a hypothetical phylogenv of the Af rican im i inecopliilotis jiolyoimnatines. Keywords: myrmecophily, Orachrysops niobe. polyommatine, phytophagv', rootstock feeding. Introduction South Africa lias a wealth of niyrniecophiloiis lycaeiiicls (Terblaiiche & van Hamburg, 2003), many of which exhibit restricted ranges (endemism) and are Red Listed species (Henning Sc Henning, 1989; Henning et al, 2009). The phenomena of endemism and rarity are believed to restdf from the narrow environmental niches available to species that reqtiire the overlapping jiresence of host plants and tending ants (Pierce et al., 2002). Nearly all of the ohligately myrmecophilous South African lycaenid butterflies are in the tribes Aphnaeini and Polyouuuatiui (sensu Pringle et al, 1994). Orachrysops is a recently erected polyommatine genus (Vari & Kroon, 1986), for which the life history and myrmecophily of its 1 1 species and one subspecies are little known. Clark and Dickson ( 1971) were otily able to rear larvae of Orachrysops lacrimosa (Bethnne- Bakei', 1923) to the end of the second instar, after which the larvae died. Recent work on the two most ^Corresponding author Received: 27 October 20()H Accepted: 5 November 2008 endangered species in the genus has extended this knowledge to all stages of their life history. Edge and Pringle (1996) reported that the larvae oi Orachrysops niobe (Trimen, 1862) were phytophagous in all instars during captive rearing to the adult stage, and whilst a dorsal nectary organ (DNO) was present no ant association appeared to be necessary. Ln and Sainways (2001; 2002a; 2002b) made field observations of all larval stages and jDupae for Orachrysops ariadne (Butler, 1898) and detected an apparently obligate ant association with Camponotus natalensis (F. Smith). Polyommatine larvae display a range of ant associations, including predacious parasitism, facultative mutualisms and myrmecoxeny (no ant association). Larval diets vary from phytophagy to entomophagy, or combinations thereof (Cottrell, 1984; Fiedler, 1991b; Fiedler, 1998; Pierce et al, 2002). Variation is evident within genera (e.g. Maculinea) as well as between genera, with significant implications for the ecology and population dynamics of each species (Thomas et al, 1998). Consequently extrapolations between even closely related species can be misleading, and detailed field observations as well as laboratoiy experiments are the only way to establish with any certainty the laiwal diet and the exact nature of the myrmecophily for each species (Thomas et (d., 1989; Ehnes & Thomas, 1992). 99 /. Res.Lepid. Materials and methods Study site riie study site was the Breiiton Blue Butterfly Reserve (BBBR) at Breutoii-ou-Sea nearKuysnain the Western (lape Province of South Africa. It has a total area of 14.670 m-, is centred at co-ordinates 34®()4’2()” S, 23-()2’()()” E, and lies at 90-115 metres above mean sea level on a well-drained south-facing slope with an average inclination of 1 in 3 (18°), vaiying between 10° and 26°. The climatic, topographical and geological featttres ol' the site and its vegetation communities have been fully described elsewhere (Edge, 2005; Edge et al, 2008a). Field observations All O. nioh('hos\ plants {ludigofera erertaThunhevg, Eabaceae) were systematically searched for eggs between November 2001 and April 2003 and all plants with >5 eggs were searched repeatedly to detect the j)resence of any larvae, pupae or ants, at various times of day including the evening. The size, stage and behaviour of any larvae discovered were recorded, and samples were taken of ants for identification. Sizes were meastired with a vernier scale using a hand lens. Erom April 2002 not only were the leaves and stems of the plants down to ground level searched, but also some of tbe rootstocks were carefully excavated to a depth of 2-4 cm. Captive rearing on cut host plant I lost plant sprigs bearing eggs were cut off and placed in clear air-tight plastic containers 25mm diameter x 55 mm high, with a drop of water maintained in the bottom of the container to prevent desiccation of the plant. The oviposition date (if known), hatching date and all subsequent dates and measurements were written on labels attached to the container lids. If there were two or more ova on a sprig, the larvae were separated into individual coutainers on emergence. Every few days the contents were carefully removed, the larvae examined and the overall length (from the tip of the mandibles to the end of the final segment for the first instar larvae, and of the dorsal carapace for the second, third and fourth instars) measured to an accuracy of 0.5 mm with a vernier calliper. The containers were cleaned otit with water and fresh cut sprigs of host plant were inserted after careftilly transferring the larvae to the new leaves. Erom the third instar onwards, the larvae were transferred to larger flatter plastic containers 90 mm diameter x 50 mm high that would accommodate larger pieces of host plant. Upon pupation, the pupae were removed and placed on cotton wool under a netting eclosion cage. Any adults emerging were preserved as voucher specimens. Artificial ant nests Artificial ant nests similar to those used by Britton (1997) 300 mm xl50 mm x 20 mm deep, with labyrinthine passages, were made from wood with sealed transparent tops, and covered by a detachable hardboard lid to exclude light. Three queen right colonies of the host ant Camponotus baynei Arnold were collected on 5 October 2002, at a location away from the BBBR, by breaking open decayed logs lying on the ground under dense bushes. Each ant colony was kept in a large plastic box 320 x 220 x 60 mm deep with flnon (active ingredient: polytetrafluoroethylene) coated walls to prevent escape, and the ants were offered access into one of the artificial nests through a translucent plastic tube. The ants quickly took up residence in the artificial nests and feeding stations were set up in the large plastic box where a 50% v/v solution of sugar, plain water and chopped up dead insects was provided. On several occasions a third instar larva of O. riiobe on its sprig of host plant was placed in the plastic box to observe any ant interactions. Two of these ant nests were slightly bitried (covered with 10-20 mm soil) on 16 October 2002, close to host plants on which a large number of O. niobeeggs had been laid, to see whether butterfly larvae would be taken into the artificial ant nest. Transhtcent plastic tubing provided access from the ant nest to the base of the host plant. The nests were inspected every week until 27 January 2003, when one of the nests was removed to the laboratory to prepare for a captive rearing experiment (see below). The other nest remained in the field until January 2004, when it too was removed to the laboratory for another captive rearing experiment. Captive rearing with live host plant and ants Two /. erecta plants were tratisplanted from the field with their sitrrounding soil into pots I75mm diatneter x 100 mm deep in December 2002, and w'atered regidarly. An experiment was set up in February 2003 with the two potted and caged /. erecta plants, an artificial ant nest containing an ant colony with brood of all stages, and one of the large plastic boxes with atit feeding stations, all connected by clear 6 mtn diameter plastic tubes (Fig. 1 ). A 3“’ instar larva 42: 21-33, 2003 (2010) 23 Figure 1. Experimental set up for captive rearing with live host plant and ants (7 mm long) was placed on each of the I. erecta plants on 8 March 2003. The plants and the ant nest were examined regularly to detect any larval activity and any ant-larva interactions. Morphology of the immature stages of O. niobe Larvae were examined with a Wild M5 stereomicroscope at magnifications of tip to 50x. The various stages were photographed under magnification with a Nikon (loolpix E4600 digital camera. Particular attention was given to the dorsal nectary organs (DNOs), perforated cupola organs (PCOs), tentacular organs (TOs) and the mandibles of the 4"' instar larva. Growth characteristics of O. niohe larvae and comparison with other polyommatines Data were obtained from Clark and Dickson (1971) and Elmes et al. (2001) to enable a comparison to be made between the growth patterns observed in the early stages of O. niobe and other polyommatine species. Host plants recorded for other Orachrysops species Data were obtained from variotis published sources and from fellow lepidopterists of the host plants recorded for the genus Orachrysops. Localities for other Orrtc/try,5o/is species were \asited, the females were observed ovipositing, and specimens were tiiken of the host plants and sent to an expert for identification. Ova of the Orac/nyvo/« species were collected and it was confirmed that the larvae stirvived and fed on the host plant on which they were laid. High magnification photographs were taken of the eggs and the larvae that hatched from them. Results Field observations - larvae and pupae The 1*‘ and 2"'' instar larvae of O. niobe make 0.5 mm - 1.0 mm grooves in the epidermis and pali.sade parenchyma of the glabrous uppersides of the leaflets of I. erecta. When not feeding, the larvae descend to the lowest part of the plant and rest on tlie stem in a head-down position, making them very difficult to find in the field. The first 4‘'’ instar larva was discovered on 27 March 2002 at 15.50 pm on a cool, cloudy day. A vertical hole abotit 8 mm diameter was alongside the rootstock of this plant, from which several ants emerged. Down the hole about 20 mm deep was a 4th instar larva, which was careftilly removed for meastirement and found to be 18 mm long x 4 mm wide, and then replaced in the hole. A sample was taken of the ants and H. G. Robertson of the South African Iziko Museum identified them as Camponotus baynei Arnold. The next day the larva had pupated (dimensions 12 mm X 4 mm). A few days later the ptipa could not be found, so possibly the attendant ants must have taken it deeper underground. Most of the subsequent observations were also made later in the day and early evening, when the O. ^4 /. Res.Lcpid. Figures 2-6. 2. O. niobe larval feeding marks can be seen on the 18mm diameter rootstock of I. erecta host plant, where 3 larvae were found at different times. 3. Rootstock of /. erecta showing feeding damage inflicted by larva of O. niobe (original diameter of 6mm reduced to 2mm) (x20). 4. 2"'' instar O. niobe larva (2.5 mm long). 5. 3'"’ instar O. niobe larva (7mm long) showing the head shield (Photos by D. A. Edge). 6. 4"' instar O. niobe larva (18mm long) showing an everted tentacular organ (TO) on abdominal segment A8 (top right) (Photo by L. du Preez). 42: 21-33, 2003 (2010) 25 attendant ants appeared to be more active. On two occasions two fully-grown 4‘'’ instar larvae were found on the same I. erecta rootstock. In one instance the C. originally in attendance were supplanted by a Pheidole species (in large numbers), and the two larvae could no longer be found. The ant attendants were C. baynei in thirteen out of hfteen observations made (seven 4'*' instar larvae; hve pupae and three pupa cases) , with Camponohis berichti in attendance in the other two cases. Mature rootstocks of I. erecta are up to 18 mm diameter (Fig. 2). The holes alongside the rootstocks of I. erecta appear to be excavated by the C. ba^jnei-c\nis, but these holes do not lead to ant nests. C. baynei was only found to be nesting above ground in decayed dead wood with holes bored out by a beetle larva, and this was usually some distance away from the 1. erecta plants. Captive rearing with cut host plant The size and duration of the early stages of O. niobe during captive rearing on cut host plant are summarised in Table 1. The few adults that were reared were dwarfs, notwithstanding their rarity in nature (Edge, 2008). During the 2004 and 2005 captive rearing experiments a number of new observations were made. It was conhrmed that the first and second instars (and presumably the third) normally consume their shed cuticle, including the head capsule, after ecdysis. In April 2005 experiments were conducted whereby pairs of well-fed fourth instar larvae were placed in the same container with fresh host plant. Within 24 hours in each case one of the larvae disappeared and the survivor grew in size. In one Table 1. Summary of the size and duration of the early stages of O. niobe, reared on cut host plant. Stage Size Duration Ovum 0.6 dia X O.Snnn high 6 -7 days P' instar 0.8 - 1 .5inm* 5 - 6 days 2'"' instar 1 .5 - 3. Omni* 8-12 days 3"‘ instar 3.0 - 7.5mm* 35 - 57 days 4"' instar 7.5 - 12.0mm* 26 - 61 days Pupa 7.5 - 8.0mm 13 -23 days Adult 10 - 13mm L'p to 15 days * For the larval instars the sizes are at the start and Hnish of the instar. instance the act of cannibalism was observed. Whilst the prey larva was feeding on a host plant leaf, the predator larva attacked it from behind and below, through the soft ventral parts of abdominal segments A7 and A8. After penetrating the integument with its jaws, the predator larva sucked out the prey’s body contents, eventually reducing it to just a skin and head shield, which was also subsequently consumed by the predator larva. The prey larva thrashed about during the attack trying to free itself, with the tentacular organs (TOs) being very active. The predator lar%'a increased in size from 9 mm to 1 1 mm before and after this attack. The attacking behaviour appears to be calculated and instinctive. Artificial ant nests Regular inspections of the artificial C. baynei ant nests showed that the ant colonies remained healthy, with stable abundances of adult ants and brood. During the 1 03 days that the first nest was in place, no larvae of O. niobe were observed inside the nest, nor were any (). niobe larvae seen at all. The other nest was in the field for 15 months and although the nest remained active, no larvae were observed to come into the nest. Both host plants had large numbers of O. niobeeggs laid on them, and there is a high probability that (). niobe larvae were in the immediate vicinity of the nests. Captive rearing with live host plant and ants The two third instar larvae that were placed on the two host plants could not be found after the first 24 hours. When the ants were given access to the plant, they were observed crawling on the plant and on the soil under the plant in the evenings, but no larvae were detected. The larvae were also not seen in the ant nest. Three months later, the plants were isolated from the rest of the experiment and the soil around the rootstock of the plants was carefully excavated. Nothing was found around the rootstock of the more healthy plant. The stems of the other plant were badly withered and many were dead. A large hole (±20 mm) was discovered alongside the rootstock. The plant became loose and it was lifted out of the hole. A final instar O. nta/zclarva ( 1 5 mm long x 4 mm wide) w'as clinging to the rootstock in a hunched position. The rootstock was badly damaged (reduced to 2 mm diameter from 6 mm). This larva was observed for the next few months with ants remaining in attendance at all times. Since both the larva and the ants were photophobic, the 26 /. Res.Lepid. larva was relocated to a vertical wooden box with a red perspex translucent viewing window. The box was filled with soil and a rootstock from a live plant above was placed in a groove visible through the window. Ants from the artificial nest were given access, and were observed tending the larva and imbibing secretions from the dorsal nectary organ (DNO) . The larva mostly stayed on the rootstock until it became clear that the larva was eating the rootstock, cutting out deep grooves (Fig. 3). The length of the larva was now about 18 mm. The larva eventually lost interest in the rootstock, and accompanied by ants made its way to the bottom of the box where the ants had excavated a hole. The C. baynei ants continued to imbibe secretions from the DNO, cover the larva with soil when exposed to light, and occasionally bodily carried the larva into deeper holes that they had excavated. At no stage was there an attempt to carry the larva into the ant nest, nor were the ants observed to feed the larva by trophallaxis or any other method. When the soil was excavated again on 5 August the larva had pupated (pupa 15 mm long), still tended by the ants. The pupa was removed in mid-October and placed under a hatching cage. A full size male butterfly (forewing length 17 mm) eclosed on 3 November 2003. Morphology of the immature stages The following features of the morphology of the immature stages of O. niobe were not reported by Edge and Pringle (1996). The 2'”' instar has much shorter dorsal and ventral setae than the P' instar, and already has active TOs (Fig. 4) . Tlie head shield of the 3'^'* instar larva completely covers the head (Fig. 5). The 4'’' instar larva frequently everts its TOs (Fig. 6). Comparative growth characteristics of polyommatine larvae The ratios between the lengths at the finish to the lengths at the start of each larval instar are depicted in Fig. 7. Note particularly the contrast between the normally phytophagous taxa, Latnpides boeticus (Linnaeus, 1767) and Euchrysops barkeri (Trimen, 1893), and the myrmecophagous species Lepidochrysops palricia (Trimen & Bowker, 1887), Lepidochrysops varinhilis (Cottrell, 1965) and Maculinea avion (Linnaeus, 1767). The difference between O. niobe reared purely on leaves cut from the host plant and the same species reared on live host plant, including rootstock and with ants in attendance, is also notable, particularly in the 4’'' instar. Host plants and ant associates recorded for other Orachrysops species The data currently available for the host plants and known ant attendants for Orachrysops species are summarised in Table 2. All host plants recorded so far are in the genus Indigofera or the closely affiliated genus Indigas trum.'And all ant associates yet known are in the genus Camponotus. Discussion Phytophagy The life histories of several Lepidochrysops species have already been described: L. patricia - by Clark and Dickson (1957). L. variabilis - by Cottrell (1965). L. methymna (Trimen, 1862) - by Clark and Dickson (1971). L. trimeni (Bethune-Baker, 1923) - by Claassens (1972; 1974 and 1976). L. ign.ota (Trimen & Bowker, 1887) - by Henning (1983b). L. plebeia (Butler, 1898) - by Williams (1990). It was generally accepted from these observations that the larvae of Lepidochrysops are mostly myrmecophagous. Clark and Dickson (1971) reared Orachrysops lacriniosa (Bethune-Baker, 1923) as far as the start of the 3"' instar, when the larvae died. When Vari (1986) separated the genus Orachrysops from genus Lepidochrysops Hedicke on adult morphological Figure 7. Ratios between lengths at the finish and lengths at start of larval instars for six polyommatine butterflies: O. niobe 1 = reared on cut host plant; O. niobe 2 = reared on live host piant with ants; E barken and L. boeticus = Clark & Dickson 1 971 ; L. patricia - Clark & Dickson 1 957; L variabilis = Cottreil 1 965; M. arion^ Elmes et al. 2001 . SPECIES 42: 21-33, 2003 (2010) 27 Table 2. Host plants and ant associates recorded for Orachrysops species. Orachrysops species Indigofera host plant Locality region Camponotus ant associate Sources 0. niobe (Trimen) I. erecta Thnml>. Brenton S. Cape C. baynei Arnold Williams, 19‘)(i; Lnbke et at., 1997 (). firiftdne (Buder) I. 7ww/;var. laxaW. Bolii.s Ivirlool KZN C. nalatensis (F. Smith) Ell K: Samways, 2001 0. tacrimosa (Betlume- Baker) I. nbscura N. F,. Br. Cireylingstad Gauteng Not known Edge ]iersonal observations 2004 (). tacrimosa (Betlmne- Baker) Indigastrum fastigium (E, Mey.) Verloren Valei Mptimalanga Not known Edge persomil observations 2004 (). brinkmarii Heath I. dectinala E. Mey. Kammanassie .S. Cape Not know'll Heath 1997 O. subraxnis G. A. & S. F. Henning I. woodiwAY. woodiiH. Bolus I. trislis E. Mey. Wahroonga KZN Camponotus sp. Not known Samways & I.tt, 2007 Lit, 2003 0. nasnlus nasulus G. A. & S. F. Henning 1. elandsbergensis P. B. Phillipson I logsback E. Cajte Not known Edge personal observations 2004 (). mijburghi G. A. & S. F. I lenning /. exiansiana Bnrtl Davy Heilbron OFS Not known Pringe et at., 1994 (). near myburghi I. dimidiata Woge\ ex Walp. sensu stricto Snikerbosrand Gauteng Not know'll Terblaticlie & FAlge, 2007 0. regalis G. A. & S. F. Henning I. accepta N. E. Br. Wolkberg Eimjtopo Not ktiown Edge ]iersonal observtuions 2004 (). warreni G. A. & S. F', Henning /. dimidiala \'ogel ex Walp. sensn stricto X'erloren \'alei Mjttimalanga Not ktiown Edge personal observations 2004 0. montanus G. A. & S. F. Henning I. di midiala Vogel ex Wal|3. sensti lato Golden Gate GFS Not known Edge jiersotial observations 2004 N. B. All Indigofeni And Indigasirumphwu names deteriiiined bySchrire (200r)a) grounds, the larvae of both genera were still assumed to be myrmecophagous. Edge and Pringle (1996) reared O. niobe larvae in captivity on host plant cuttings without ants, and this discovery added a clear biological jtistification for the separation of Orachrysops fro m Lepid och rysops. The O. niobe larvae reared on host j)lant cuttings restdted in dwarf adtilts. In held observations P' and 2"^' instar larvae were nearly always fotind feeding on the leaves of the host plant (with a single observation of a P‘ instar larva feeding on the rootstock). 3”^ and 4''' instar larvae were always found tindergrotind feeding on the rootstock attended by ants, and have never been found feeding on the leaves. Rootstock feeding in the Lycaenidae has only been recorded once before. Jackson (1937) observed that the larvae oi Eiiclirysops crawshayi crawshayi (Butler) fed “on the fleshy otiter cortex of Cynoglossum coeruleum Hochst. et D.C., Boraginaceae, always below the grotind; and they are attended by many species of ants.” Rootstock feeding in O. appears to be essential to produce full size adults, since there is no evidence that the diet of Orachrysops larvae incltides any ant provided food. Rootstock feeding has also sidjsequently been obseiA’ed in another Orar/ny.^e/As species (Terblanche & Edge, 2007). Rootstocks as a dietary source Pierce (1985) noted that many myrmecophilotts lycaenid larvae feed on nitrogen rich plants and nitrogen rich parts thereof (such as flowers and legtmie pods) . More recent work has questioned this association (Fiedler, 1995; 1996), but the rootstock feeding behaviour of the 3"' and 4"’ instar larvae of O. accords with Pierce (1985), if the rootstock is indeed protein rich. The rootstock of the legtime I. erecta is a storage organ from which the plant can resprout, storing carbohydrates and bearing nitrogen fixing root nodules probably rich iu amino acids and protein. Analysis of /. erecta leaves and rootstocks showed that although their overall amino acid content is quite similar (11600 nmol per g dry weight), the rootstocks had more than double the content of essential amino acids, particularly threonine, histidene and allo-isoleucine. These amino acids could play 28 /. Res.Lepid. an important role in the development of 4''' instar larvae and pupae (e.g. threonine is important for the synthesis of collagen, a basic constituent of the more complex connective tissites generated in the lepidopteran ptipal stage). Cannibalism Cannibalism by O. niobe larvae is another potent source of protein. Freqtiently mtiltiple eggs are laid on a single host plant (tip to 31 eggs were recorded), so encounters between T', 2”'' and 3"' instar larvae on the host plant (where attacks are most likely to sncceed) must be quite freqtient. A larva grows rapidly after it has successfully attacked and constmied one of its siblings, and its sttrvival chances are considerably enhanced. The habit that the larvae have of resting on the lower thicker part of the stem no doubt protects their vtilnerable ventral jjarts from such attacks. Laiwae that stirvive long enotigh to reach and start feeding on the rootstock are much better protected from attack, since the thick dorsal integument (carapace) extends down to the rootstock on both sides. Morphological and behavioural adaptations of the larvae of O. niobe Cottrell ( 1984) described a ntimber of adaptations of lycaenid larvae that enable them to ptirstie a myrmecophilous life style. These adaptations have great vahte in not only creating “enemyTVee space” (At.satt, 1981 ), but also by providing access to a more reliable nutritious diet restthing in more rapid growth and larger adtilts (Pierce et al., 2002). A number of these adaptations have been observed in the larx ae of (). niobe. They have a thick, tough integument, which defends the larvae from attack by the ants or from conspecilic larvae. They have an onisciform body shape with an expansion of the dorsal and dorsolateral areas, which slope down to well-develo]>ed setadVinged marginal ridges that can be brought in close contact with the snbstrate, and seal off the ventral areas inchiding the retracted head and legs. The ants (and the larva’s siblings) therefore cannot gain access to the softer ventral parts and vital organs when the larva crawls or rests on a hard stibstrate. (). niobe larvae have a dorsal nectary organ (DNO) on the seventh abdominal segment in the 2"'', 3"'atKl 4''’ instars, producing a secretion that aids myrmecojthily (Pierce, 1989; Fiedler & Maschwitz, 1989). 4’hey have tentacular organs (TOs) on the eighth abdominal segment that appear to excite the ants in attendance and draw them to the DNO (C.laassens & Dickson, 1977; Fiedler & Maschwitz, 1987). The larvae of O. niobe '^ho have a number of other adaptations not yet observed in other lycaenid larvae, including powerful mandibles, which adapt them for rootstock feeding and cannibalism, and an ability to burrow in the soil - although the attendant ants may assist this burrowing. Larval shelter and overwintering The sitbterranean, myrmecophilous lifestyle of the larvae of O. niobe insulates them from extremes of temperature and humidity changes; provides protection from bres; and gives them access to a food source (the rootstock) and a safe place to shelter during the winter when the growth of the host plants pauses. Larval growth patterns Dyar (1890) postulated that insect dimensions increased by the same factor at each moult. Phytocarniverous lycaenid larvae such as Maculinea that parasitise ant colonies are exceptions to this rule (Elmes et ai, 2001). Maculineei larvae show a normal growth pattern during the first three instars, but after adoption by their M-yrrairaant hosts, in the final instar they increase in length by a factor of five, and by >50 times in mass (Elmes et al, 1991; Thomas & Wardlaw, 1992). Elmes et al. (2001) hypothesi.sed that these growth patterns may have evolved to make the newly moulted 3"* instar butterfly larvae approximately the same size as the ant larvae at adoption and better able to mimic the larvae of their ant hosts. Once in the ant nest and accepted by the ants they could grow to a larger size than other lycaenids becaitse of the readily available, high quality food source. In Eig. 6 it is clear that the phytocarnivorous larvae of Maculinea and Lepidochrysops show similar growth patterns. Whilst O. niobe larvae raised under artificial (1) or natural conditions (2) have a higher rate of growth in the third instar than any of the other examples, in the final instar the growth rate of O. niobe under natural conditions (2) is intermediate between the phytophagous larvae {E. barkeri and L. boeticus) and the phytocarnivorous larvae. Rootstock feeding may be the key to the higher growth achieved in the final instar relative to the normally phytophagous taxa. Specialisation of Orachrysops associations All the known host plants of the Orachrysops genu?, are in the genus Indigofera or the very closely affiliated genus Indigastrum (Table 2). Monophagy is common in the localised species and allows such species to 42: 21-33, 2003 (2010) 29 Table 3. Ant associations of southern African polyommatine butterflies. Key to third column [coding adopted from Fiedler (1991a)] 0 = No ant relationship (myrmecoxenous) 1 = Very few ant associations reported (weakly myrmecophilous) 2 = A varying proportion of larvae attended by ants (moderately myrmecophilous) 3 = Most if not all mature larvae ant-associated (steadily myrmecophilous) 4 = Larvae dependent on ants as commensals or parasites (obligately myrmecophilous) ** = DNO -r TOs * = DNO only = PCOs only ( ) = hypothetical ? = likely but not confirmed Genus Known ant a.s.sociates Degree of myermecophily Sources U ra n oth a u ma Butler ? a Ps/’udon/KYuiub/i Stem|)tfer (0/1) t'-g Carireus Butler (O/D* a, c Harpmdymts I leron (0/1)* c Brephid'tum Scudcler (1)** c Oraidium Bethune-Baker (1)? g Tuxenlius Larsen (1)** c ZiiUlia Eliot (1)** g Ziziila Ghapman (1)** c Actizera Ghapman c Lrploles Scudcler Indeterminate \ 1 Zizina Ghajnnan (2)^* c Cupidopsis Etarsch a, c Eirochrysops Bethuue-Baker 9** c Lampides Hubuer ('.ampoiiotii.s Plagioli’pis 9^1:* c, f Zizeeria Ghapman Papinoma c, f Azanus Moore Pheidole a, c. f Chilades Moore Pheidote 3** c, f Ihniru.s Moore Plairiolepis Monomorium c, f Euchrysops Butler Monomorium 3:tc:tc a, c Oral hrysops Vari Canipoiiiitiis 4** i Erpidochrysops I ledicke Camponotus 4** b, c, d. e Sources: a = Jackson. 1937; b = Gottrell, 1965; c = 1991a; g = Pringle el at.. 1994; h = Williams, 1999; Glark & Dickson. 1971; d = Glaaseus, i = Lu & Samways, 2901. 1974 & 1976; e = Henning, 1983a; r= Fiedler, avoid direct competition and co-exist at a locality (e.g. O. lacrimosa and (). warreni at Verloren Valei and O. subravus and O. ariadne at Wahroonga - Samways & Lu, 2007). Although only two ant associations are known as yet for Orachrysops species, the ants concerned, C. baynei and C. natalensis have some ecological similarities (both are primarily nocturnal ants). C. baynei is only found in fynbos or thicket and nests in dead wood above ground level (Edge et at., 2008b), whereas C. natalensis can be found in fynbos as well as grasslanrl and nests in the ground (Lu & Samways, 20()2a). Ant associates of South African polyommatines The records of known or hypothesised ant associations within the South African polyommatines [members of the tribe Polyonnnatini as dehned by 30 /. Rfs.Lepid. Eliot ( 1973) J are listed in Table 3. (Sotirces: Claassens, 1974, 1976; Clark & Dickson, 1971; Cottrell, 1965; Fiedler, 1991a; Henning, 1983a; Jackson, 1937; Lu & Samvvays, 2001; Pringle et al, 1994; Williams, 1999). The coding system used in the final column has been adopted from Fiedler (1991a). Idle close affinities between Orachrysops and Lepidochrysops wotild have always made a species of Camponotus the most likely host ant for Orachrysops, and this has now been confirmed for two of its species. However, with only 1 1 out of 126 Lepidochrysops species ant associates known (Pierce et al, 2002) and 2 otit of 1 1 Orachrysops ant associates known, there is insufficient evidence to conclude that Camponotus species are the only ant associates for both genera. Nature of the ant association of O. niohe 3’'* and 4''' instar O. larvae were almost always tended by the same ant, C. bayncL O. niohe s congener O. ariadnehds only been fotind in association with one ant (C. natalensis) (Lu & Samways, 2001). Cottrell (1984) considered that relationships specific to one ant species tended to be obligate. Even when more than one ant species interacts with the larvae, in obligate relationships one ant species is the most sticcessful host (Thomas et al, 1989). Facultative relationships by contrast tend to be formed with several species of ant, even from different genera (Fiedler, 2001). The larvae of O. niobe did not enter into ant nests either in the field (artificial nests) or in the laboratoiy The nests of the attendant ants found to date are relatively remote from the plants on which O. niohe feeds, but the plants need to be within the foraging range of worker attendant ants so they can find the larvae. No trophallaxis or other feeding of the larvae by ants has been observed. The larvae are rootstock feeders in the later instars, and appear to need the assistance of the ants to access the rootstock. After pupation, clear access to the surface through a hole or tunnel is necessary for the eclosed adult to escape and expand its wings. Ants have been observed both in captivity and in the field repositioning pupae and their assistance may be essential to place the pupa in a safe position to eclose. The balance of evidence is therefore that O. niobe is obligately dependent on an ant association with C. baynei. Phylogenetic origins of Orachrysops and Lepidochrysops Various atithors have proposed phylogenetic hypotheses regarding the origins and evolution of lycaenid ant associations (Hinton, 1951; Eliot, 1973; Fiedler, 1991b; Pierce et al, 2002). Hinton (1951) believed that the possession of a DNO was a primitive feature in the Lycaenidae, and that its absence was a Figure 8. Hypothetical cladogram of the myrmecophilous polyommatine genera of South Africa based on genus of ant associates, larval feeding behaviour and host plant families and genera. Polyommatine genus Azanus Chilades Tarucus Euchrysops Orachrysops Lepidochrysops AFRICAN POLYOMMATiNI 42: 21-33, 2003 (2010) 31 derived condition. Fiedler (1991b) pointed out that the lycaenid subfamilies that were apparently more primitive on other morphological considerations (Poritiinae, Miletinae and Curetinae) were not generally ant associated; whereas the more advanced subfamily Lycaeninae [sensu Eliot (1973) = Theclini + Aphnaeini -i- Lycaenini + Polyommatini] contained most of the myrmecophilous lineages. Within Lycaeninae sensu Eliot Pierce ei al. (2002) predicted that the Aphnaeini and certain subtribes of Theclini would be shown to be basal and that Lycaenini and Polyommatini were derived groups. In Table 3, southern African polyommatine genera are listed in ascending order of their degree of myrmecophily. A tentative cladogram of the steadily myrmecophilous polyommatine genera (denoted as 3 or 4 in Table 3) , based on genus of ant associates, lanal feeding behaviour and host plants is presented in Fig. 8. Azanus and Chilades are associated with Pheidole ants, which are the dominant ants in some habitats. Tarucus is associated with a few ant genera, including Camponotus. The genera Oreic/irysops, Euchrysops and Lepidochrysops are predominantly Camponotus associated. Ornchrysops has evolved rootstock feeding, cannibalism, an obligate ant relationship and specialisation on Indigofera. Euchrysops has a looser ant association, and normal phytophagy, with the exception of A. crawshayi. This interesting taxon has genitalia similar to Euchrysops (Stempffer, 1967), facies similar to Harpend.yreus, and larval behaviour with features found in Orachrysops. It is of note that this species has been placed in different genera by various authors (by Butler in ScoUtantides\ by Aurivillius in Harpendyreus Awd later Cupido\ by Jackson in Cyclirius; and finally by Stempffer in Euchrysops) . The larvae of Eepidochrysops are phytophagous in the first two instars, and myrmecophagous in the last two instars (they feed on the ant brood). This is considered to be the closest ant relationship (Fiedler, 1998), with the larvae being treated as if they were ant brood within the ants’ nests. It is here hypothesised that the larvae of the common ancestor of the three genera Euchrysops, Orachrysops, and Lepidochrysops could have had rootstock feeding habits, which physiologically adapted them to a higher protein diet. Furthermore, the cannibalistic behaviour of O. niobe larvae may have evolved in the common ancestors. This wovdcl have enabled them to develop a taste and a need for insect protein, a trait that could have evolved in a sister lineage into the myrmecophagous behaviour of Lepidochrysops larvae. A higher protein diet may have made closer ant associations possible because of the enhanced ability to produce nutritious (high protein) secretions from the DNO (Pierce, 1985). This would have enabled the larvae to attract more ant attendants and given them the opportunity to evolve more advanced chemical camouHage and signaling skills (Fiedler, 1991b; 1998) . Such adaptations would have generated further selective advantages in these butterfly lineages, residting in more complex ant associations and greater interdependence of the butterllies and their ant associates. Conclusions O. niobe is a specialised organism, being monophagons and having an obligate relationship with a single ant species. Its habit of rootstock feeding, which may be shared with other Orachrysops species, distinguishes it from nearly all other afrotropical polyommatines. A close phylogenetic relationship between the genera Orachrysops, Euchrysops and Lepidochrysops is inferred from a cladistic analysis. Acknowledgements (^apeNature is thanked ft)i' allowing access to the BBBR to conditct research. The Green IVitst’s financial sn|t])ort is gratefully acknowledged, and North West Lhiiversity kindly provided a stereoinicroscope oti loan. Dr. Brian Schrire of the Royal Botanic Gardens, Kew identified the Indigofera and Indigaslnim specimens. Literature cited At.saii, B. R. 1981. Lycaenid butterllies and ants: selection for enemy-free space. American Nattiralist 1 18: 638-6.'i4. Bruton, D. R. 1997. Atit trajt nests enable detection of a rare and localised butterfly, Arrodipsas myrmccophila (Waterhouse anti Lyell) (Lepidt)]ttera: Lycaenidae) in the field. Menutirs of the Miisetim of X'ictoria 56: 38.3-387. Ct.vvssitNS, A. j. M. 1972. Lepidochrysops trimeni (Bethnne-Baker) (Lepido|)tera: Lycaenidae) reared from larvae foitnd in the nest of Camponotus maeulatusYAh. (Hytiienoptera: Formicidae). Journal of the Entomological Society of South Africa 35: 359- 360. Ci.vvssENS, A. |. M. 1974. Methods etnployed in the snccessfnl retiring, tinder artificial conditions, of two Gape species of Lepidochrysops (Lepidoptera: Lycaenidae). Journal of the Entomological Society of South Af rica 37: 387-.392. CiA-tssKNS, A. |. M. 1976. Observations on the myrmecophilous relationships and the parasites of Lepidochrysops methymna nuihymna (Tiimen ) and L. trimeni (Bethnne-Baker) (Lepidoptera: Lycaenidae). Jottrnal of the Entomological Society of South Africa 39: 279-289. Claassens, a. j. M. & C. G. C. Dicicson. 1977. A study of the myrmecophilous behaviour of the immattire stages of Aloeides thyra (L.) (Lepidoptera: Lycaenidae) with special reference to the function of the retractile tubercles and with additional notes on the biology of tbe species. Entomologist’s Record and Journal of Variation 89: 225-231. Ci.ARR, G. G. & C. G. G. Dickson. 1957. The life-history of Lepidochrysops patricia (Trim.) (Lepido|Dtera: Lycaenidae). 32 /. Res.Lepid. lounial of the Entoinoloi>ical Society of South Africa 20(1): I 11-117. (ii .VRK, G. G. & (1. G. G. Dickson. 1971. Life histories of tlie Sotith African lycaeiiid Ittitterflie.s. I’tirnell, Cape Town. Co ri RKt.i., G. B. 1965. A study of the A/ct/ty?«t;('/-grotip of the genus I .I'pidorhrysops l iedicke (Lejtidoptera: Lycaenidae). .Memoir of the Entomological Society of Sotithern Africa 9: 1-1 10. GoriRi i.i,, C. B. 1984. A|)hyto]tliag}' in hutterllies: its relationshijj to myrmecophily. Zoological Jotirnal of the Linnaean Society 79: 1-57. Dt.\R, 1 1. G. 1890. The ntimber of motilts of lepidopterotis larvae. Psyche 5: 420-422. Encr., D. A. 2005. Ecological factors infhiencing the stirvival of the Brenton Bine butterfly, Oracltiysops niobe (Trimen) (Le]>idoptera: Lycaenidae). North-West University, Fotchefstroom. (Thesis - D. Phil.) Enor., D. A. 2008. Adult behaviour of Orach ijsops itiohe (Trimen) (Le|)ido|ttera: Lycaenidae). Metamorphosis 19(3): 116-126. E.ih;e, D. a., S. S. Gii.i.iers. &: R. F. Tr.RittANCHE. 2008a. Vegetation associated with the occtirrence of the Brenton Bine bntterfly. South Af rican Journal of Science 104: 505-510. Edoe, D. a., M. G. Roberison & H. V.an H.VMiiURt.. 2008b. Ant assemblages at jrotential breeding sites for the Brenton Blue btilterllv Orarhrysryps (Trimen) (Lepidopteni : Lycaenidae). African Entomolog)' 16(2): 259-262. Eix.e, 1). A. & E. 1.. Prinoi.e. 1996. Notes on the natnrtil history of the Brenton Bine Orachry.sops niobe (Trimen) (Le]tidoptera: Lycaenidae). Metamoiphosis 7(3): 109-20. El lot, J. N. 1973. The higher classification of the Lycaenidae (Le]rido|5tera): a tentative tirrangement. Btilletin of the British .Vlnseiim (Nattiral Llistory) Entomolog}' 28: 37.3-516. Femes, G. W. & |. A. d'noM.vs. 1992. Comjrlexity of species conservation in managed habitats: interaction between MacuHnea butterflies and their ant hosts. Biodiversity Conservation 1(3): 155-169. Femes, G. W., j. A. Tiiom.vs, M. I,. Mt'N(;uiR..\ & K. Fiedeer. 2001. Larvae of lycaenid btitterflies that |tarasitise ant colonies |trovide exceptions to normal insect growth rules. Biological |oin nal of the Linnaean Society 73: 259-278. Femes, Ci. W., J. G. W.vrde.wv & |. A. Tiiom.vs. 1991. L.arvae of Maculinea rebeli, a large-bhie btitterfly and their Myrmira host ants: patterns of cateipillar growth and stirvival. Journal of Zoology 224(1): 79-92. Eieoeer, K. 1991a. European and northwest African Lycaenidae and their associations with ants, [otirnal of Research on the Lepidoptera 28(4): 239-257. Fiedeer, K. 1991b. Systematic, evohttionary, and ecological implications of mvrmecophilv within the Lycaenidae (Insecta: l.epidojjtera: Papilionoidea). Bonner Zoologisches Monographien 31: 1-210. Fiedeer, K, 1995. Lycaenid butterflies and jjlants: is myrmecophily associated with partictilar host plant preferences? Ethology, Ecology and Evolution 7: 107-132. Fiedeer, R, 1996. llo.st-|)lan( relationships of lycaenid btitterflies: large-scale patterns, interactions with jtlaiit chemistry, and nuittialism with ants. Entomologia Experimentalis et Applicata 80: 259-267. Fiedeer, K. 1998. Lycaenid-ant interactions of the Maculinea lype: tracing their historical roots in a comparative framework. Jotirnal of Insect Con.servation 2: 1-1 1. Fiedeer, K. & IVfAsoHwrr/, U. 1989. Etinctional analysis of the niyrnieco])hilotis relationshijrs between ants (1 lymenoptera: Eormicidae) and lycaenids (Lepidoptera: Lycaenidae). 1. Release of food recrtiitment in ants by lycaenid larvae and pupae. Ethology 80: 71-80. 1 Ieaiii, a. 1997a. Description of a new species of ( )rachrysops\' Ai'i from the western (ia|)e pror ince of Sotith Africa (Lepidoptera- Lycaenidae). Metamorphosis 8( 1 ): 26-32. Hennino, (L a., R. F. Terbianche & J. B. B.m.e (eds.). 2009. South African Red Data Book - Btitterflies. South African National Biodiversity Instittite, Pretoria, (in press). Hennino, S. F. 1983a. Biological groups within the Lycaenidae (Lepido|ttera) . Jotirnal of the Entomological Society of Sotithern Africa 46: 65-85. Hennino, S. E. 1983b. Chemical commtmication between lycaenid larvae (Lepidoptera: Lycaenidae) and ants (Hymenoptera: Eormicidae) [Aloeides dentath, Acantholepis capensis, Lepidochry.wps ignola, C.amponolus niveosetosus]. Journal of the Entomological Society of Southern Africa 46(2): 341-366. Hennino, S. F. 1 984. The effect of ant a.ssociation on lycaenid larval duration (Lycaenidae: Lepidoptera). Entomologist’s Record and Jotirnal of Variation 96: 99-102. Hennino., S. E & G. A. Henning. 1989. Sotith Af rican red data book: butterllies. South African National Scientific Programmes Report 158. Council for Scientific and Inditstrial Research, Pretoria. Hinton, H. E. 1951. Myrmecophilons Lycaenidae and other Lepidoptera: a summaiy. Proceedings of the South l.ondon Natural History Society 1949-19,50: 1 11-175. JAOK.SON, TIL E. 1937. The early stages of some African Lycaenidae (l.epidoptera), with an accottnt of larval habiLs. Transactions of the Royal Entomological Society of London 86(12): 201-238. Lu, S-S. 2003. Biology atid conservation of the threatened Karkloof blue btitterfly Orcichrysops ariadne (Btitler) (L.epidoptera: Lycaenidae). University of Natal, Pietermaritzburg. (Thesis - D. Phil. ) 112 jt. l.u, S-S. & M. J. Samways. 2001. Life history of the threatened Karkloof bhie btitterfly, Orachrysops ariadne (Lepidoptera: L.ycaenidae). African Entomology 9(2): 137-151. Lu, S-S. & M. J. S.AMWAVS. 2002a. BehaMounil ecology of the Karkloof blue butterfly, Orachrysops ariadne (l.epidoptera: Lycaenidae) relevant to its conservation. African Entomology 10(1): 137- 147. Lu, S-S. & M. J. Samways. 2002b. Conservation management recommendations for the Ktirkloof blue butterfly, Orachrysops rtt7rtt/Hc(Lepido|)tera: Lycaenidae). African Entomolog)' 10(1): 149-159. Pierce, N. E. 1985. Lycaenid butterflies and ants: selection for nitrogen fixing and other protein rich food plants. The American Nattiralist 125: 888-895. Pierce, N.E. 1989. Bntterfly-ant mtuiialisms, pp. 299-324. /«: Grubb P. J. & Whittaker, J. B. (eds.). Towards a more exact ecolog)’. The 30th .sym|jositim of the British Ecological .Society. Blackwell Scientific Publications, I.ondon. Pierce, N. E., M. F. Braby, A. Hiatii, D. J. 1.ohman,J. M.vnTiEW, D. B. Rand & M. A. Tr.\v.\ssos. 2002. The ecology' and evohition of ant association in the Lycaenidae (Lepidoptera). Annual review of entomology 47: 733-771. PRiNot.E, E. 1,., G. A. Henninc, & j. B. B.m.e. 1994. Pennington’s Btttterflies of Sotithern Africa. 2'"' edition. .Struik Winchester, Cape Town. 800 pp. Stemi’efer, H. 1967. The genera of the African Lycaenidae (Lepidoptera: Rhopalocera). Bulletin of the British Miisetim (Natttral History) Entomology' 10. S.VMWAYS, M.J. K- .S-S Lu. 2007. Key traits in a threatened butterfly and its common sibling: implications for conservation. Biological Conservation 16: 409.5-4107. Sw.ANE.i’OEE, D. A. 1953. Btitterflies of South Africa. Maskew Miller, Cape Town. 320 pp. Terbi.anciie, R. F. & H. Van Hamburg. 2003. The taxonomy, biogeogra])hy and conservation of the myrmecophilotis CJirysorilis btttterilies (Le|>ido|ttera: Lycaenidae) in Sotith Africa. Koedoe 46(2): 65-81. Terbianche, R. F. & D. /V. Edc.e. 2007. The first record of an 42: 21-33, 2003 (2010) 33 0?w/ir)’io/M I'roni Gauteng. Metamorphosis 18(4): 131-141. Thomas, J. A., R. T. Ci.arkk, G. W. Ei.mes & M. E. 1 Iocmberu. 1998. Population dynamics in the genus Marulhira (Lepido]jtera: lA'caenidae), pp 261-290. In: Dempster, J.P. & McLean, L.F.G. (eds.). Insect Populations. Kluwer Academic Publishers, Dordrecht. Tuom.vs, J. a., G. W. Ei.mes, J. C. W.aruiaw & M. Wowciechowski. 1989. Host specificity among Maculinea butterflies in Myrmka ant nesLs. Oecologia 79(4): 4.52-457. Tuom.vs, J. A. & J. C. W.vrdiaw. 1992. The cajjacity of a Myrmira ant nest to support a predacious species of Maculinea btitterfly. Oecologia91(l): 101-109. V.vri, L. 1986. Introduction, pp. vii-x. /w: Vari, L. & Kroon, 1). 1986. Southern African l,e])idoptera - a series of cross- referenced indices. The Lepido])terists' Society of Southern Africa & The Transvaal Museum, Pretoria. Wii.t.iAMS, M. G. 1990. Gbservations of the biolog)' oi Lepidorhrysops plehela plebeia (Butler) (Lycaenidae: Polyomattinae) . Metamorphosis 1(27): 10-1 1. Wii.i.iAMS, M. G. 1996. Report on research findings concerning the life history and ecology of the Brenton Blue ( Orarhrysops niohe) . Metamorphosis 7(1): .3-7. Wii.i.iAMS, M. C. 1999. Taxonomical changes affecting the afrotropical butterflies and skippers (Lepidoptera) since the ptiblication of Garcasson's Butterflies of Africa - Part 1: Lycaenidae. Metamorjrhosis 10(2): 52-69. journal oj Ri'searrh on the Lepidoptera 42: 34-49, 2003 (2010) Life history of the Imperial Moth Fades imperialis (Drury) (Saturniidae: Ceratocampinae) in New England^ U.S.A.I distribution^ decline, and nutritional ecology of a relictual islandic population Paul Z. Goldstein 1 1 1 Gay Head Ave., Vineyard Haven, MA 02568. drpzgoldslein @gmail. com Abstract. The decline and current status oi Earles impeiialis (Drury) (Saturniidae: Ceratocampinae) in New England are re\4ewed, and primary data surrounding the life history and nutritional ecology presented. Though common throughout much of its historical North American range, this species declined precipitously in New England during the 20''' century. Suggested explanations for this region-wide decline include the deployment of pesticides and metal halide street lamps and the introduction of parasitoid flies. The existence of a remnant population of E. mperialis, thought to have been extirpated from New England as early as the 195()s, is reported from Martha’s Vineyard Island, Dukes County, Massachusetts, U.S.A., representing the last relict of a phenotypically, phenologically, and possibly ecologically infrasubspecific entity. Based on comparisons with museum specimens, adults from this population appear to be indistinguishable from the few historical specimens known from nearby mainland Massachusetts, smaller than those from now extirpated populations in Connecticut, New York, and New Jersey, and slightly larger than members of, the northernmost E. /m/tcnVf/A populations (£. i. /tinlMichener) found in the Great Lakes region. The Martha’s Vineyard population is univoltine, peaking in late July and exhibiting a more contracted flight season than other extant North American populations. Both parentage and food plant significantly affect larval growth and development, and although laiwae on Martha’s Vineyard feed extensively if not exclusively on pitch pine (Pinus rigida) in the wild, they grow significantly faster, attain greater pupal weights, and more efficiently convert ingested and digested tissues of post oak (Qiierrus stellata) to biomass in the laboratory. Performance as measured by relative growth rate and the efficiency of conversion of ingested and digested food to biomass (ECI and ECD) are correlated with foliar nitrogen and water content: post oak foliage during the larval growth season contains more nitrogen and water than corresponding amounts of pitch pine foliage growing in the same soils. It is suggested that there exists a non-ntitritional explanation for the association of E. imperialis with pitch pine and for its pattern of decline and persistence. The taxonomic and possible biogeographic affinities of this population are discussed from within the context of a growing understanding of New England’s diverse yet threatened lepidopteran fauna, and the potential for reintroducing E. imperialis to mainland New England is discussed. Keywords: Earles imperialis, invertebrate conservation, Martha’s Vineyard, pitch pine. Introduction The imperial moth Eacles imperialis (Drury) (Saturniidae: Ceratocampinae), is one of North America’s largest and familiar saturniids. Throughout its extensive range, this species exhibits considerable regional variation in coloration and size — on the basis of which three North American subspecific epithets are currently attributed — as well as variation in phenology and liost plant utilization (Ferguson, 1971; Coveil, 1984; Tiiskes et ai, 1996) . Considered a common moth in tlie southern United States, E. imperialis, like many large bombycoid moths, underwent a precipitous decline in much of northeastern North America during the mid-2()''' century (Hessel, 1976; Ferguson, liereived: 24 July 2006 Accepted: 24 August 2006 1971; Schweitzer, 1988). Ferguson (1971: 25) writes “[the imperial moth]... like some of the other large saturniids, is said to have largely disappeared from heavily populated areas such as those in Connecticut and the vicinity of New York City, where it was formerly common.” Various hypotheses for such declines, ranging from the expanded use of metal halide bulbs in street lamps (Hessel, 1976) to the widespread deployment of pesticides aimed at gypsy moths (Goldstein, 1991) and to the introduction of parasitoids, specifically the tachiiiid fly Compsilura concinnata, for the same purpose (Boettner et al, 2000) have been invoked to explain these declines. Believed extirpated from New England, a relict population of E. imperialis was verified by the author in 1982 on Martha’s Vineyard Island (Dukes Co.), off Cape Cod, Massachusetts, where it quickly became of interest to conservationists. In this paper, following a 42: 34-49, 2003 (2010) 35 review of the taxonomy and regional variation in this species complex, I present life histoiy obseiwations and larval growth and performance data on this population between 1986 and 1989, discuss this species’ decline by summarizing historical records and information on pesticide deployment and parasitoid impacts, and discuss the nutritional and life history requirements of this species in light of putative reestablishment on mainland New England. Distribution and variation in the Fades imperialis complex in North America Extending from (lanada to Ai gentina, the imperial moth E. imperialis is both the widest-ranging and the northernmost occurring species in its genus and one of the most widely distributed saturniids that occurs in North America (Lemaire, 1988), representing a complex of ecologically and possibly phylogenetically distinct intra-nominal entities. Authors have differed in their treatment of subspecific epithets, variously recognizing the western E. oslari as a fidl species versus a subspecies of imperialis and the southern E.i. nobilis Neumoegen as a subspecies versus a synonym of nominate imperialis (reviewed in Tuskes el al, 1996) . Not including these, two recognized subspecies of /f. imperialis occur in North America (Eemaire, 1988; Tuskes ei al, 1996): the nominate subspecies and E. i. pini, which is distributed in the Great Lakes region and the Adirondacks. According to Eerguson (1971: 24), E. imperialis “occupies nearly all of the United States east of the Great Plains, with the exception of northern New England and northern portions of Michigan and Wisconsin.” Eerguson (1971: 25) further reports records “from southern New Hampshire... through New York State, the Niagara Peninsula of Ontario, and southern Michigan, westward to the eastern edge of the Great Plains, and southward to the Gulf Goast and southern Florida (but not the Florida Keys). It has not been taken in Maine.” The New England states are represented by specimens in collections at the American Museum of Natural History (AMNH), the Museum of Gomparative Zoology, (MCZ), the Peabody Museum of Natural History (PMNH), which houses the bulk of the F. M. Jones collection from Martha’s Vineyard, and the collection of the now defunct Boston Society of Natural History (BSNH), which curreittly resides at Boston University (Fig. 1). A single 1900 specimen from Kittery Point (southern Maine’s York Gounty) is housed at the MGZ, and Patch (1908; cited in Tuskes et al. [1996: 65]) reported it from Girmberland Gounty, ME. Beyond these r ecor ds, Farquhar’s (1934) thesis enumerated various other Figure 1. Historical distribution of E. imperialis in New England by county. Arrow indicates presence of lone extant population on Martha’s Vineyard Island, Dukes Co., MA. New England r ecords, also included in Fig. 1. The species’ crrrrent occurrence iir the rrortheastern portion of its range sorrth of New Errgland, iirchtdiirg Long Island, N. Y. and southern New Jer sey, is associated at least in part with habitats char acter ized by sandy soils such as pitch piire-scr uh oak barreirs, where its larval host plant, pitch j)ine {Pinas rigida), aboutrds. Srrch habitats have been heavily impacted, in large part due to the ease with which sandy soils are maniprrlated for building and constrirction pitrposes. Sandy, well-drained soils nray be a requirement for Eacles imperialis in that, like all ceratocarrrpine satrrnriids, Eacles lar vae do not constrirct cocoons but burrow into and pupate within the soil, from which the pupae themselves emer'ge so that the adults may edose above-gr'ound. The cotrspicrroits dearth of historical recor ds fr om Gape God, where sandy soils predominate, may he attributable to a combination of small |)rivate collections’ having been orphaned. Each'S imperialis exhihxis apparently clinal variation in wing maculation and shading as well as in size and pherrology, with southern popirlations (formerly attribirted to E. i. nobilis) often exhibiting rrror e intense bi'owtr postmedial shading than northern poprrlations. It has also been observed that adrrlt individuals of northern popirlations attributed to E. i. pini in 36 J. Res.Lepid. nortliern Michigan , Ontario, Quebec, and New York, are markedly smaller than their southern and eastern coimteiparts (Ferguson, 197l;Tuskes etaL, 1996), and bear more intensive peppering with blackish spots; the larvae exhibit pronounced dorsolateral scoli relative to nominate imperialis. The maciilation on specimens taken from Martha’s Vineyard is consistent both with museum specimens from the island and with those taken from mainland New England prior to the species’ decline there. Based on an examination of these and other specimens from the northeast, the mean forewing length for male E. imperialis from Martha’s Vineyard was 49.64 mm (N = 60, se = .401), significantly smaller than that of 78 specimens examined from New York, Connecticut, and New Jersey (one-way AN OVA, p < .0001, DF = 1, F-ratio = 88.241; Fig. 2), and at the lower end of the range of 47.59 mm reported by Tuskes et al. (1996). This number is higher than the average male wing length of 47 mm reported for A. i pinihy Michener (1950) in the original description and out of the 42 mm - 48 mm range reported for male pini by Tuskes et ai (1996). Michener (1950) also reported an average wing length of 54 mm for male nominate imperialisfrom the vicinity of New York City; specimens I measured from this area averaged 54.92 mm (N = 36, se = .92), by way of comparison. Although female wing lengths for mainland New England specimens were not analyzed due to their scarcity in collections, Martha’s Vineyard female specimens averaged 57 mm in wingspan, just below tlie range of 58-68 mm reported by Tuskes et al, greater than tlie 51 mm average pini female wingspan reported by Michener and out of the 47 mm - 54 mm range reported for pini by Tuskes et al The familiar green/brown larval polymorphism of E. imperialis is apparent both among lab-reared caterpillars and those obseiwed in the wild on Martha’s Vineyard. Both maternity and food plant affect the expression of this polymorphism: among the larvae reared for this study and in situ (P. Goldstein, unpubl.) , the lime green color form was less prevalent than the dark brown, with the majority of reared green larvae developing on pine versus oak. Ferguson (1971: 26) writes that “[a] brood from Massachusetts, described by Eliot and Soule [1902], contained only one green larva.” Flarris (1890: 404), on the other hand, writes “[the caterpillars are] for the most part, of a green color, slightly tinged with red on the back; but many of them become more or less tanned or swarthy, and are sometimes found entirely brown.” During the course of my rearing, I observed seven cases in which larvae switched from brown to green at one molt only to shift back at a subsequent molt; a common color morph of oak-fed larvae. The most common color morpli Figure 2. Mean forewing length (base to apex, in mm) of 1 38 male Eacles imperialis from Martha’s Vineyard, Connecticut, New York, and New Jersey. The mean forewing length for male E. imperiaiis from Martha’s Vineyard was 49.64mm (N=60, se=.401), significantly smaller than that of specimens examined from New York, Connecticut, and New Jersey (one-way ANOVA, p<.0001 , DF=1 , F-ratio=88.241). See text. * of Individuals Observed Figure 3. Flight season of E. imperialis on Martha’s Vineyard, 1984-1989 inclusive, based on observations of 107 individuals, 86 male and 21 female (see text). of oak-fed larvae is orange to reddish brown, often with the spiracles and dorsolateral scoli surrounded by paler patches. With respect to life history timing, Harris (1890: 404) writes: “The moth appears here [Massachusetts] from the 12''' of June to the beginning of July, and then lays its eggs on the buttonwood [sycamore, Liquidambar sty rad folia] tree. The caterpillars maybe found upon this tree, grown to their full size, between the 2()th of August and the end of September, during which time tliey descend from the trees to go into the ground.” Phenologically, based on a sample of 107 individuals (86 male, 21 female) collected or observed 42: 34-49, 2003 (2010) 37 on Martha’s Vineyard between 1982 and 1989, the population is protandrous, and both males and females peak at the end of July: 65% (56) of the males were observed between 21 July and 1 August, and 71% (15) of the females were observed between 25 July and 1 August (Fig. 3). This flight season is consistent with the collection dates of museum specimens from the nearest known mainland historical records (e.g. Cohasset, MA; Bristol, RI), has remained predictable in the years since this study was conducted, and may represent a more contracted flight season than that reported in Tuskes et al. (1996) for northern E. i. imperialis. Most indiMduals appeared at lights after 2200h, and individuals were observed coming to light until 040()h. Taken collectively, these observations hint at the possibility that northern E. imperialis represents at least one and possibly two biological entities distinct from southern imperinlis. It is noteworthy that Lemaire (1988: 31) portrays the range o^ E. i. pmii\s crossing the Appallachian divide to northern New England. Although there exists a phenetic similarity of adult New England specimens to individuals typical of T. i. piniwith respect to size and macnlation, the diagnostic characters of adult and larval pini (Tuskes et al, 1996: 67, 68; Pis. 1 , 7, & 8) do not appear in specimens from New England. Historically, the superficially clinal nature of variation in this complex has presented obstacles to a clear understanding of what may be taxonomically and biologically distinct natural entities. Notwithstanding the perhaps controversial criteria on which the retention of saturniid snbspecific epithets rely (Goldstein, 1997; but see Tuskes et al, 1996 for an alternative viewpoint), the similarities between nominate E. impeiialis cmd E. i. pini in wing macnlation, size, phenology, and especially host plant use are relevant to the holobiology of this complex. It is conceivable that the small size of E. i. pini and the New England T. i. imperialism s related to the combination of a short growing season and an apparently exclusive association with pines at northern latitudes. I would recognize E. pini as a full species and anticipate that the northeastern, pine-feeding popidations of E. imperialis and those comprising pini will ultimately be recognized as separate sister species. Since the type locality of Earles imperialis is in New York, I would further anticipate that nohilis be resurrected. Eacles imperialis on Martha’s Vineyard: past and present Jones and Kimball (1943), in their extensive treatment of the Lepidoptera of Martha’s Vineyard and Nantucket Islands, described E. imperialis as Figure 4. Current distribution of E. imperialis on Martha’s Vineyard as of 28 August, 2009. Courtesy Massachusetts Natural Heritage & Endangered Species Program. Westboro, MA. occurring regularly on Martha’s Vineyard, where it persists and is widely distributed on both moraine and ontwash plain soils (Fig. 4). It is not known from Nantucket, and virtually all mainland New England specimens deposited in museum collections I examined were taken during the first half of the 20th century, with only a few taken as late as the early 1950s. By all accounts, this species had essentially declined dramatically in New England more than a decade before Ferguson’s (1971) publication, and indeed was considered extirpated from Massachusetts during the early drafting stages of the Massaclursetts Endangered Species Act (MESA; M.G.L. c. 131Aand regulations 321 GMR 10.00). It is currently listed as “threatened” in Massachusetts. The decline of A. imperialism northeastern North America is not tiniqne, but rather consistent with a well-known pattern of decline among saturniids (Hes.sel, 1976; Fergitson, 1971; Frank, 1988; Schweitzer, 1988). In fact, the three most dramatic declines of saturniids in New England have been ceratocampines: E. imperialis, the royal walnut moth Citheronia regalis (Fabricius) , another of North America’s most massive saturniids, and the pine devil moth C. sepulcralis Grots & Robinson, another barrens species that feeds on pine and the type locality of which is Andover, MA. Martha’s Vineyard appears to have served as a regional refugium for several species (e.g. Actias Inna) that remained extremely abundant on the island while undergoing anecdotal declines — even temporaiy ones — on the mainland. It is not at j)resent possible to single out any one of the various possible causes for these declines; none are mutually exclusive. Moreover the weak coincidence of the deployment 38 /. Res.Lepid. of metal halide street lights and pesticides several decades following the introduction of the tachinid fly Compsilura concinnata (Diptera: Tachinidae) in 1906 to combat gypsy moths and other pests (Howard & Fiske, 1911) makes parsing the relative importance of these factors difficult. In hindsight, it is not surprising that the sole New England population of E. imperialis to persist did so on an island that was not as heavily subjected to as heavy aerial deployment of pesticides or high-wattage metal halide streetlights, or to the intentional release of parasitoids, as the mainland. Available data on the use of DDT and other pesticides, recorded on a per-county basis between 1948 and 1965, indicate that Dukes Co. (including Martha’s Vineyard) was one of four counties sprayed only once during this period, and was subjected to less intensive treatment (as measured by total spray- acres) than any other county in the Commonwealth excepting Middlesex (Bewick, 1979, reproduced in Table 1). Martha’s Vineyard is approximately 100 square miles, or 64,000 acres; its single documented pesticide treatment of 17,000 acres in 1956 is less intensive, for example, than the 31,071 acres sprayed on Nantucket, whose land mass is roughly half that of the Vineyard; and far less than the hundreds of thousands of acres of maritime barrens habitats in Plymouth and Barnstable comities routinely subjected to spraying before the use of DDT was discontinued. Prior to the widespread deployment of DDT, the parasitoid fly C. concinnatah'Ad become well-established in New England, shortly after its 1906 introduction (Culver, 1919) . Introduced to combat the gypsy moth Table 1. Massachusetts DDT spray history 1948-1965 (Reproduced from Bewick, 1979). County Total spray-acres # Years treated Avg. acres sprayed per treatment year PKniouth 525,962 5 105,193 1 taiii])shire, 1 taiiKlfii, Franklin 505,952 7 72,279 Barnstalrlf 490,130 10 49,013 Worcester, Norfolk 374,905 2 187,453 Berksliire 324,765 7 46,395 Nantucket 31,071 1 NA Essex 29,685 2 14,823 Bristol 18,496 1 NA Dukes 17,000 1 NA Middlesex 16,934 1 NA Lymantria dispar (L.), the browntail moth Euproctis chrysorrhoea (L.) and other pests, it quickly became apparent that this animal attacks a large and diverse assemblage of macrolepidopteran larvae (Webber & Schaffner, 1926; Ai nauld, 1978). Recent observations by Boettner et al. (2000 and unpnbl.) confirm the extraordinaiy impact of C. concinnata on several native saturniids on mainland New England and that the flies will attack E. dw/icria/A caterpillars when presented with the opportunity. Compsilura concinnata is considered a stong candidate for playing a role in the demise of E. imperialis and other large moths on mainland New England (Boettner et al, 2000); Compsilura concinnata occurrence on Martha’s Vineyard has not been verified. Regional variation and host plant use Eacles imperialis (and can certainly be reared) on a wide range of tree hosts (Ferguson, 1971; Stone, 1991; Tuskes et al, 1996), to which it may be locally specialized, and the moth’s range exceeds that of any recorded host plant species. Ferguson (1971) lists an impressive array of recorded hosts: “oak, hickory, walnut, sycamore, basswood, maple, honey locust, chokecheriy, sumac, sweet gum, sassafras, elm, beech, hornbeam, birch, alder, pine, spruce, hemlock, cedar, cypress, and juniper.” However, the degree to which a given population of E. imperialis is ever genuinely polyphagous in nature is not well studied. On Martha’s Vineyard, E. imperialis appears to feed exclusively on pitch pine {Pinus ri.gida), a common tree associated with sandy, well-drained soils and barrens habitats. Pine-feeding is not nnusnal for Eacles impeiialis elsewhere, nor is it restricted to the northernmost tier of the moth’s range: Abbot and Smith (1797) observed pine-feeding in Georgia over two centuries ago, and pitch pine is a common host in the New Jersey pine barrens (D. F. Schweitzer, pers. comm.). All the recorded hosts of E. i. pini are conifers (Tuskes et al, 1996), including jack pine {Pinus hanksiana), which replaces P. rigida to the north and west of southern New England. Erom the perspective of consumption and digestion, conifer foliage represents a diet high in fiber, relatively low in water and nitrogen content, and possessed of allelochemical defenses such as monoterpenes stored in resin ducts (Raffa, 1991). The observation of localized conifer specialization on the part of A. imperialis i?, of interest both from the perspective of pine-herbivore biology and from that of the species’ northeastern decline. Host plant specificity among herbivorous insects in general and Lepidoptera in particular has been of broad interest to entomologists at least since the 42: 34-49, 2003 (2010) 39 works of Brues’ (1920, 1924), which were followed by seminal ecological studies and reviews highlighting- general patterns in the ecology and evolution of host use breadth (e.g. Ehrlich & Raven, 1964; Futuyma, 1976; Gilbert, 1979; Cates, 1980; Jenny, 1984). Adult oviposition cues play a critical role in host plant specialization (Wiklund, 1975), and variables such as host plant chemistry and architecture (Cates & Rhoades, 1977; Scriber & Feeny, 1979; Berenbaum, 1981; Bowers, 1983, 1984; Bernays, 1989), foliar water, nitrogen and fiber content (Scriber, 1977, 1979b; Mattson, 1980; Mattson & Scriber, 1987), and maternity (e.g. Mousseati 8c Dingle, 1991) effect lepidopteraii digestive efficiency. Poptilation genetic and phylogenetic data have been brought to bear on the evolution of host specialization and the role of host shifts in speciation (Berlocher, 1998), and the co-cladogenesis of insects and plants (Farrell 8c Mitter, 1990, 1998; Futuyma 8c McCafferty, 1990; Funk et ciL, 1995; Becerra, 1997; Farrell, 1998). Studies of the ecophysiological roles of host plant growth form and seasonality (e.g. Tilton, 1977;James 8c Smith, 1978; Scriber, 1978) in mediating insect-plant associations are directly relevant to the evolution of pine feeding. From a nutritional perspective, folivores of trees and other woody plants represent a guild that must often contend with low nitrogen and water contents relative to graminivores or forb feeders, for example. Nitrogen is an important growth-limiting factor for many herbivores which may covary with water content (Mattson, 1980); low foliar water content can impede the growth and development of herbivores (Scriber, 1977) as well as a given herbivore’s ability to utilize available nitrogen (Scriber, 1979a, b). These effects can be particularly severe with regard to the performance of tree feeders (Scriber, 1979b); since low nitrogen and water contents are typically associated witli woody plants as opposed to forbs and grasses, folivores of trees often display low growth and utilization efficiencies (Scriber & Slansky, 1981). I pursued a line of investigation tow'ards better understanding host plant use in the relictual population of this moth, specifically the question of whether the apparent confinement to pines is nutritionally imposed or the result of some unknown non-nutrionally based or abiotic factor. This endeavor comprised laboratory-based experiments focusing on the role of food plant in influencing larval growth and development. Given the considerable range¬ wide variation in host use among geographically delimited populations of £. imperialis, the notion that such local specialization may have been accompanied by physiological adaptation is of interest from the perspective of understanding host shifts, regardless of whether they accompany speciation per se. Materials and methods Five treatments from fotir species of host plants were chosen on the basis of recorded use by E. imperialis, availability and, in the case of pitch pine, comparability between material from wild populations known to be used by £. imperialis arboretum-raised material. Pitch pine is the only known host of E. imperialis on Martha’s Vineyard (personal obseiwations of both larvae and wild oviposition behavior of adult females) , and was reported as the favored lan al host of other Massacliusetts E. imperiaUs (Eliot 8c Soule, 1902). The jack pine P. beraksiana was chosen because it is associated with E. imperialis pini (Michener, 1950; M. C. Nielsen, pers. comm.; B. Scholtens, pers. comm.), comprising the northernmost named subspecies of E. imperialis and the only other regional populations associated exclusively with conifers. Jack pine may be considered an ecological analog of pitch pine, in that it replaces the latter in habitats to the north and west characterized by granitic soils that, like maritime pitch pine-scrub oak barrens, are sandy, well-drained, and acidic soils (Little, 1979; Schweitzer & Rawinsky, 1988). M. C. Nielsen (pers. comm.) reports that E. i. pini appears most commonly in association with jack pine on Grayling soils. With the exception of tiny remnant stands such as that at West Rock, New Flaven, CT, the post oak (X stellala is represented in New England primarily as a di.sjunct population on Martha’s Vineyard. The nearest large stands of Q. stellala occur on Long Island, N. Y. and in the New Jersey pine barrens, which also hosts the largest northeastern population of E. imperialis. The black walntU J. nigra was chosen because it is a recorded host of E. imperialis, yet does not occur regularly on Martha’s Vineyard. Food plant material was harvested every five days from the Felix Neck Wildlife Sanctuai'y, Edgartown, MA (pitch pine and post oak) and the Arnold Arboretum, Jamaica Plain, MA (pitch pine, jack pine, and black walnut). LIpon cutting, plant material was placed in water pics and maintained at 4()°F until used. Foliage fitted with water pics was placed in rearing containers, the supply of fresh food maintained, and the container cleaned of frass every five days. Since foliar water and nitrogen concentrations have been demonstrated to decrease with leaf age (Axelsson & Agren, 1979; Slansky 8c Scriber, 1985) , it was therefore important to be as consistent as possible when selecting plant material so that foliage treatments of wildly disparate ages were not lumped under the same heading. 40 /. Res.Lepid. Gravid females from Martha’s Vineyard were collected in 1988 and 1989 at light and placed in paper bags for oviposition. Ova were harvested and segregated according to parent, then maintained in 4” diameter petri dishes tinder constant conditions of light and temperature (per day, 14 hours light at 25°G and 10 hours darkness at 20°C). The 1988 livestock came from four females taken between 31 July and 5 August at three different sites oit Martha’s Vineyard: Gedar Tree Neck, West Tisbnry (4 Atigust, 1 10 ova); Makonikey, West Tisbnry (3 and 5 August, 155 and 78 ova); and Pennywise Path, Edgartown (31 Jtily, 107 ova). During the hrst (1988) season’s experiments, ova were selectively cooled at 50°F for up to four days to synchronize hatching. Larvae from each clutch were weighed tipon hatching, placed on each of live of the food plant cohorts, and weighed every five days (116 hours) through day 15 in all cases and day 20 when ]4ossible. Larvae were reared individually, first in 4.0” petri dishes for their hrst hve days, and then in plastic containers 4.5” in diameter and 3.5” in height. Humidity was maintained iisitig 1” x 2” cut sections of paper towels secured to the container lid and moistened twice every live days. All larvae were maintained under constant conditions of temperature and lighting as described above. Since leaves that remain attached to the stem are generally less apt to lose water than if they are cut (Schroeder, 1984), food plant freshness was maintained by placing plant sprigs (or petioles, in the case of/, nigia) in water pics sititated in larval rearing containers. Weights of surviving larvae from different clutches and on different host plants were natural log- transformed and compared (2-way ANOVA) for each weighing interval through Day 15. The influence of food plant treatment was further analyzed (1-way ANOVA) by lumpittg all suiwiving larvae of a giveti food plant treatment and age regardless of parentage. In 1989 these experiments were repeated in part and with several modihcations, using ova from seven lemales collected between 26 and 28 Jitly from three sites on Martha’s Vineyard; Lobsterville, Aquinnah (26 July, 178 ova); Gedar Tree Neck, West Tisbnry (three females, all 27 July; 55, 125, and 169 ova); and Makonikey, M'est Tisbtiry (three females, 26, 27, and 28 Jtily; 178, 50, and 104 ova). The 1989 work focused exclusively on two food plants taken from ctirrent E. habitat (71 rigida, the native host, and Q. strllala, with all plant material taken from Felix Neck M'ildlife Sanctuary, Edgartown, MA. Eggs were not cooled to synchronize hatching, and in order to minimize handling of yotmg larvae, all hatchling caterpillars were started not in petri dishes but in the larger plastic containers. Given observations that hatchling larvae lost weight to desiccation quickly, care was taken to ensure that every larva was weighed and placed on the appropriate food plant immediately upon hatching. To maintain humidity, I used 1” x 1” X 2” sponges, washed at each weighing instead of paper towel sections. The sponge holds more water and for a longer period of time and does not require additional moistening between weightings. All sponge blocks were washed repeatedly prior to use to clear them of chemical additives, and thoroughly cleaned of larval frass at each weighing to avoid mold. Food plants were maintained as in 1988 at 40°F, misted with distilled water and given refreshed water pics every 24 hours. Feedings were made as close to identical as possible; all the feedings for a given weighing interval and food plant species consisted of material cut from a single tree. Sample sizes were increased to fifteen larvae per female per food plant treatment initially, for 104 larvae in total on each food plant. Color morph and instar were noted at each weighing and larvae were reared to pupation, sexed and weighed again. Weights were analyzed as described through day 20 for all larvae and pupae. Since E. imperialis is a sexually dimorphic species, male and female pupae were compared separately (1-way ANOVA) for each food plant cohort. Weights upon hatching were compared independent of a food plant effect (one-way ANOVA) to evaluate bias in assigning larvae to food plants. There was not a large enough sample size per sex per clutch to retrieve any significant data regarding a maternal effect on pupal weight. However, all pupal weights representing each sex were lumped for each food plant in order to test for a food plant effect via one-way ANOVA. The gravimetric assessment of digestive and growth indices involved forty offspring of a single female on four of the 1988 food plant treatments: both wild and arboretnm-grown P. rigida, arboretum-grown P. hanksiana, and wild Q. stellata. Low survivorship on /. nigra precluded there being enough data to analyze. These indices were: relative growth rate (RGR); relative consumption rate (RGR); approximate digestibility (AD); efficiency of conversion of ingested food to biomass (ECI) ; and efficiency of conversion of digested food to biomass (ECD). The experimental regime employed here involved monitoring the food uptake and utilization during the course of a single instar, from the beginning of the third to the beginning of the fourth instar. All vegetative material used in these experiments consisted of pre-weighed individual leaves or sprigs; unconsumed food and frass was dried at 95°F and re-weighed. Control samples of plant material for each feeding were also weighed. 42: 34-49, 2003 (2010) 41 dried, and re-weighed for tlie purpose of calculating conversion factors. Larvae were weighed at the beginning and end of the experiment, then frozen, dried, and re-weighed to determine the dry weight gained by each larva. Each larva thus provided its own conversion factor, its final dry weight treated as a percentage of its final fresh weight and used to estimate initial dry weight. Based on these weights of plant and larval material, the hve food plant utilization measures were determined following Waldbatier (1968). For each larva used in the calculation of utilization indices, the dried control food plant material was ground and analyzed for percent-nitrogen using a Kjeltec nitrogen analysis system. Mean digestive indices were compared via one-way AN OVA. The gravimetric assessment of performance has been reviewed and critiqued numerotis times since its introchiction (Scriber & Slansky, 1981; Schroeder, 1984; Agren Sc AxeLsson, 1979; Bowers et al, 1991), and a substantial body of literature has been devoted to evaluating the various sources of error associated with this method. A significant methodological source of error in the calculation of such nutritional indices derives from the indirect calculation of the dry weight of plant material and of the initial dry weight of the larva (Bowers et «/.,199 1 ) . Since the host plants used differ in megaphyll morphology, achieving similarity between treatment and controls is difficult to standardize. This necessitated frequent feeding of fresh plant material in order to minimize the differential effects of desiccation on digestibility. For each feeding of stellata, a single leaf was bisected, half used to determine the dry weight conversion factor. Conversion factors for P. ngidz/ involved the use of individtial needles excised from sheaths: for each needle-bearing sheath, a single needle was removed for immediate drying wliile the others were tised for feeding. This method has the dual advantages of providing as accurate a control as possible, since all needles within a given sheath are the same age and size, and of minimizing damage to both treatment and control, since the sheath can be removed from its enclosed needles without tearing or severing significant mesophyll ti.sstie. Preliminary work showed that the short needles of P. banksiana, which occtir in pairs, were more vulnerable to desiccation than those of P. rigida. To offset desiccation, single needles were therefore retained intact within bisected sheaths, the dry weight of the inedible sheath later stibtracted from the estimated dry weight of the initial feeding prior to calculation of fresh weight/ dry weight conversion factors. Results Food plant and maternal effects on growth Larval growth varied considerably across the 1988 food plant treatments. The (1988) data indicated that E. imperkdis larvae grew more slowly at first on the decidtiotis plants than conifers, with growth most markedly retarded among the larvae fed /. nigra (Table 2, Fig. 5). Based on weight, larvae responded best to the diet of arboretum-grown P banksiana, followed by that of arboretum-grown P. rigida, wild (X stellata and finally wild P. rigida. However, for a period between the second and third weightings (days 5 and 10), the growth of larvae fed Q. stellata stirpassed that of the other food plant cohorts. Most larvae fed arborettim- grown jack pine and pitch pine weighed more at Day 20 than those given other food plants. ANOVA indicated signiheant food plant as well as maternal effects on larval growth (Table 3), the latter far less marked than the former. Larvae in the 1989 follow-iqD growth ex])eriments, conducted under modified conditions where foliar water was more rigorously ensured, consistently gained biomass and molted earlier when fed post oak than when fed pitch pine, nnambignotisly demonstrating enhanced performance of E. iniperialis on wild Q. stellata relative to wild P rigida based on larval growth and pupal weight (Fig. 6; Tables 4, 5), again with significant differences attribtitable both to food plant and to maternity in eacli of the host plant treatments (Table 6; Figs. 7, 8). The variation in larval growth attributable to maternal effects was, as in 1988 (4able 3), small relative Figure 5. Fresh weight (mg) of E. imperialison five food plant treatments taken at 5-day intervals, 1988. Cf. Table 2. 42 /. lies.Lepid. Table 2. Weights (mg) of larvae reared on five food-plant treatments, 1988. x= mean; se = standard error; N= sample size; AA = foliage used from Arnold Arboretum, Jamaica Plain, MA; FN = foliage used from Felix Neck Wildlife Sanctuary, Edgartown, MA. Cf. Fig. 5. Age (clay.s) Pimis rigida (FN) x(se) Pinus rigida (AA) x(se) Qiiercus stellata (FN) x(se) Pinus banksiana (AA) x(se) Juglans nigra (AA) x(se) 0 9.384 (0.083) Ar=19 8.9 (0.081) 7V=19 9.47 (0.0653) AA23 7.632 (0.069) N=19 8.411 (0.083) N=19 5 43.947 (0.213) N=19 45.968 (0.21) fV=19 40.335 (0.179) W=23 59.058 (0.311) W=19 34.211 (0.191) N=19 10 157.789 (.495) W=19 203.158 (0.45) Afcl9 246 (0.562) 7V=23 243.342 (0.556) N=19 68.618 (0.409) N=17 15 652.829 (1.279) 7V=14 988.174 (0.865) N=\9 973.099 (1.068) N=19 1219.421 (1.185) iV=19 160.769 (0.851) N=13 20 1964.592 (2.479) N=14 2471.57 (2.54) N=\0 1867.557 (3.629) N=7 3320.9 (1.869) N=14 296.788 (1.783) N=8 Table 3. Analysis of variation due to food-plant and maternal effects on larval growth, 1 988. NA not applicable. Variation due to foodplant Variation due to maternal effects Larval age (days) DF F-ratio P-valiie DF F-ratio P-valiie 0 4 NA NA 3 1.08 0.362 5 4 3.988 0.005 3 6.297 0.001 10 4 22.329 <.001 3 6.259 0.001 15 4 55.356 <.001 3 4.393 0.007 Table 4. Weights (mg) of E. imperialis reared on P. rigida and O. stellata at ages 0 through 20 days, 1989, followed by pupal weights for males and females. Range refers to mean values for offspring of a given female (not applicable for pupal weights). Cf. Fig. 6. Pinus rigida Quercus stellata L.arval age (days) x(se) Range x(se) Range 0 10.221 (0.013) N=104 8.334-11.727 10.315 (0.014) N=104 8.531-12.358 5 47.803 (0.035) N=104 41.508-58.789 59.227 (0.044) N=100 53.389-66.627 10 315.042 (0.118) N=102 229.862-468.191 555.959 (0.213) N=92 383.782-647.46 15 1516.205 (0.268) N=96 1 161.357-2013.193 2195.288 (0.328) N=86 1516.931-2711.417 20 3561.943 (0.349) N=93 3007.108-4226.031 4484.786 (0.387) N=87 3517.45-5599.417 .Vlale pupal weight (mg) 3719.304 (0.936) N=26 NA 4751.297 (0.786) N=29 NA Female pupal weight (mg) 4473.514 (1.199) N=22 NA 5705.4 (1.277) N=22 NA 42; 34-49, 2003 (2010) 43 Table 5. Summary and analysis of development of E. imperialis in therms of instar on R rigida and Q. stellata. Day 5 Day 10 Day 15 Day 20 Instar 1 2 2 3 3 4 4 5 # Piiic-fcd larvae 24 80 13 89 8 88 28 65 # Oak-fed larvae fi 96 2 90 1 91 12 73 G-value 13 9 6 7 P-value <.001 <.005 <.025 <.025 Table 6. Analysis of food-plant and maternal effects on larval growth, 1989. NA = : not applicable; NS = insufficient data. Variation due to foodplant Variation due to maternal effects Larval age (day.s) DF F-ratio P-\aluc DF F-ratio P-value 0 NA NA NA 6 I.5..547 <.001 5 1 26.194 <.001 6 3.643 0.002 10 1 55.112 <.001 6 6. 1 76 <.001 15 1 25.382 <.001 6 6.507 <.001 20 1 24.34 <.001 6 5.593 <.001 Male pupa 1 40.049 <.001 NS NS NS Female pupa 1 28.115 <.001 NS NS NS Table 7. Summary and analysis of nutritional indices obtained on four food-plant treatments. AA = foliage used from Arnold Arboretum, Jamaica Plain, MA; FN = foliage used from Felix Neck Wildlife Sanctuary, Edgartown, MA. Nutritional Pinus rigida (FN) index Pinus hanksiana Pinus rigida (AA) Qiiercus stellata One-way AN OVA F-ratio P-value x(se) x(se) .v(,se) x(se) AD 31.09(0.6) N=4 31.64 (0.431) N=7 28.78 (0.2.59) A''=9 26.5 (0.283) N=7 0.898 0.457 FXID 32.16(0.714) N=4 39.38 (0.647) N=7 41.96 (0.331) iV=9 46.22 (0.43) N=7 1 .056 0.387 ECl 15.01 (0..304) N=4 16..35 (0.197) N=7 18.08 (0.126) yV=9 18.48 (0.208) N=7 4.778 0.01 RGR 0.205 (0.04) N=4 0.244 (0.021) N=7 0.251 (0.022) A’=9 0.264 (0.032) N=7 2.157 0.121 RGR 0.473 (0.072) iV=4 0.49 (0.04) N=7 0.414 (0.02) N=9 0.425 (0.04) N=7 2.178 0.1 18 %N 5.83 (0.052) 7V=8 6.79 (0.086) 1V=8 6.84 (0.048) N=S 8.07 (0.091) A=8 48.973 <.001 to that attributable to host plant. One trend is tliat of the consistent relative growth of offspring of several females; mean larval weights for each of four females on post oak were more massive than on pitch pine by Day 5, remaining so at every consecutive weighting through Day 20. The offspring of each of the four females grew, relative to one another, in exactly the same order on oak and pine, with the offspring of female 5 outdistancing tho.se of females 6, 7, and 2, respectively (Figs. 7, 8). This trend implies that there 44 /. Res.Lepid. Mean Individual l.arval wgt (mg) Larval Age (Days) Figure 6. Fresh weight (mg) of E. impehalis on pitch pine P. rigida and post oak Q. stellata, 1989. Cf. Table 4. Figure 8. Fresh weight (mg) of pine-fed offspring of four females of E. imperialis, 1989. 6000 5500 5000 4500 I 4000 t 3500 .? 3000 > 2500 I 2000 j 1500 1000 500 0 0 5 10 15 20 25 1 arval Age (Days) Figure 7. Fresh weight (mg) of oak-fed offspring from four female E. imperialis, 1989. exists little trade-off in relative efficiency from oak to pine. That is, a larva vvell-eqnipped to utilize pine relative to another larva may be just as relatively well equipped to utilize oak. This explanation is consistent with the fact that no statistically significant interaction exists between the parental effect and the food plant effect. Significant differences appeared among mean weights of larvae from different clutches at each weighing and among mean weights of hatchling laiwae, but these relative differences did not persist as such at snb.sequent weighings. Since there were no significant differences between mean weights of hatchling larvae given different food-plant treatments, these results can not be attributed to bias in the initial separation of larvae into treatment cohorts. A comparison of the numbers of larvae that had achieved a given instar by a given age revealed the following; significantly more of the oak-fed larvae than the pine-fed larvae molted to second instar by Day 5 (DF=1, G=12.996), third instar by Day 10 (G=8.5()2), fourth instar by Day 15 (G=:6.162), and fifth instar by Day 20 (G=6.688). Additionally, mean male and female pupal weights were significantly higher for oak- fed larvae (4751.297 mg and 5705.4 mg, respectively) than for pine-fed larvae (3119.304 mg and 4473.514 mg, respectively). In both years the number of larvae exhibiting the green color morph was higher for the pine-fed cohort than tiie oak-fed cohort. Exactly 50% (52 of 104) of all the 1989 larvae reared on pine exhibited green coloration; most of these turned green at the .second molt (beginning of the third instar), and seven reverted to brown at subsequent molts, five at the third and two at the fourth. Of the 52 green larvae from both 1989 food plant treatments, only four had been reared on oak. As many as 13 of 15 and as few as 5 of 15 offspring of a given female fed pitch pine were green. Almost all larvae reared on oak exhibited a lighter brown or reddish body color. Nutritional indices Nutritional indices, calculated exclusively during the third instar, corroborated enhanced growth rate and efficiency on oak relative to pine, as did the higher pupal weights among oak-fed versus pine-fed larvae (Table 7). The efficiency of conversion of ingested food (ECD), the efficiency of conversion of digested food (EGI), and the relative growth rate (RGR) were positively correlated with foliar nitrogen and 42: 34-49, 2003 (2010) 45 water content across food plants; the approximate digestibility (AD) and the relative consnmption rate (RCR) were not. The only significant differences, however, were among the ECI and percent nitrogen values. The foliage of wild grown post oak and arboretnm grown pitch pine contained significantly more nitrogen and water than corres]>onding amounts of wild pitch pine and arboretum-grown jack pine, and these numbers were paralleled by larval ECIs. (Table 7). Discussion The life history constraints of host specialization have been an important focus within the study of herbivore evolution. It has been hypothesized that the limits imposed by a contraction in the spectrum of potential host species are offset by an enhanced efficiency with which an herbivore utilizes that narrower host range (Brties, 1924; House, 1962; Emlen, 1973; Gilbert, 1979). The validity of this “feeding specialization hypotliesis” is central to our understanding of evolved herbivoi'y (Slansky & Scriber, 1985). Not only have consistent patterns of higher utilization efficiencies among monophagous versus polyphagotis or oligophagous herbivores failed to be demonstrated, but Scriber and Feeny (1979) have contended that host plant chemistry is responsible for most of the variation seen in larval performance. That is to say the “costs” of specialization have not been well-dehned or demonstrated in a broad sense because the axes along which organisms specialize may or may not intersect. It has become increasingly clear that in order to effect proper experimental and analytical procedures to test the feeding specialization hypothesis as a general paradigm, one must recognize a range of organism-specific variables, from elements in plant foliage that affect herbivore development to life history manifestations of constraints imposed by the host plant. These variables defy simple patterns, being too numerous and interdependent for their roles to be parsed except very broadly (Gaston 8c Reavey, 1989). Scriber (1983) suggested that one reason for our relative lack of understanding derives from the paucity of studies narrowly focused on groups of taxonomically and ecologically similar organisms. To these I would add studies of ecologically similar but phylogenetically independent and phylogenetically well understood groups. Notwithstanding the predictable effects of nutrient rich foliage among arboretum-grown plants relative to wild foliage, the growth rates and efficiencies of consumption and digestion of oak versus the wild pine host suggest that, ntitritionally. New England imperial moth caterpillars do not require pitch pine alone in order to survive. At the same time, E. imperialis may be adapted or pre-adapted physiologically to conifers. Conifer feeding on the part of herbivorous insects represents a nutritional dynamic different from deciduous leaf feeding. Pitch pine, in particidar, is a complicated fire-adapted plant, and perhaps a more relevant comparison than that between larval performance on arboretum-grown, well-fertilized plants and performance on native hosts of disjtinct moth populations might be undertaken between geographically disparate populations on the wild northern conifer hosts and among different age cohorts of pitch pine foliage growth in viable habitats of E. imperialis. That said, differential patterns in nutritional content between evergreen and deciduous trees have been demonstrated (Miller & Stoner, 1979), with evergreen foliage having generally lower nutrient contents. Pines and other conifers typically contain less foliar nitrogen than deciduous angiosperms under similar conditions of growth and development (Bidwell & Dnrzon, 1975; Slansky & Scriber, 1985), and wild pitch pine from barrens and typically nutrient- poor (Forman, 1979; Schweitzer & Rawinski, 1988). Eolivores of nutrient-poor, woody plants tend to show greater breadth of dietary tolerance (i.e. be more polyphagotis) than those on highly nutritions foliage (Matt-son & Scriber, 1987). Mattson and Scriber (1987) cite Holloway and Hebert (1979) who found that conifer-feeding Lepidoptera “are less specific in host plant choice than species feeding on angiosperms.” The data presented in this study are consistent with this claim in that the larvae of E. imperialis ure capable of sustained development on different hosts. Both the published host records of E. imperialis and the restilts of this study support the contention that this species, including populations functionally restricted to pine, can metabolize a broad range of potential if not realized host plants. Large body size is also considered an advantage when feeding on low-nutrient diets (Wasserman & Mitter, 1978; Peters, 1983; Mattson & Scriber, 1987), such as pines or late-season tree foliage. In fact, tree¬ feeding lepidopteran species active late in the growing season tend to be large (Mattson, 1980; Niemela et ai, 1981), as would be expected especially for those in which the adults do not feed (Slansky & Scriber, 1985) such as E. impenalis. There may also exist such a trend for folivores of evergreen versus deciduous plants. Opler (1978) noted that leafminers feeding on evergreen oak species tended to be larger than those on deciduous species. Although phylogenetic data were not yet available to evaluate the evolution of size in a cladistic framework, ntimerous authors /. Res.Lepid. 4(') have observed phylogenetically biased patterns in lepidopteran size associated with host plant use and life histoiy (Mattson, 1977; Wasserman & Mitter, 1978; Nieinela et at., 1981; Hayes, 1983; Gaston & Reavey, 1989), and although adaptive speculation is frivolous, it is woi'th noting that E. imperialis, the most massive saturniid extant in New England, is also the latest feeding saturniid in the region, active as larvae as late as October. By this time many of the host plants utilized further south, where the flight season of E. imperialis is more protracted, are senescent or nearly so northward, potentially accounting in part for the more strict association with conifers northward. In contrast to pitch pine, wild grown post oak contained significantly higher amounts of nitrogen than even arboretum grown j)itch pine, even tbough ECl’s were not significantly different for larvae fed the two food plants. This implies either more efficient nitrogen utilization of pitch pine versus post oak on the part of Earles or simply that oak contains more nitrogen than EV/cfes larvae can effectively metabolize. Despite the fact that the only two plant cohorts for which nitrogen content was not significantly different were the two arboretum-grown pines, the ntean EGI was significantly higher for larvae on arboretum- grown pitch pine than for those on Jack pine; both were significantly higher than for wild pitch pine. The mean EGI for arboretum-grown pitch pine was comparable to (i.e. not significantly different from) that of Martha’s Vineyard post oak, on which larvae had the liighest ECIs, and which supported a significantly higher nitrogen content (in fact the highest of all food plants measured). Taken collectively, these results suggest a potential physiological adaptation to pine¬ feeding in general, and pitch feeding specifically on the part of northern E. imperialis. The results presented here go to show that simply because a particular food plant species meets an berbivorous organism’s nutritional requirements and the organism is ]ihysiologically capable of growth and development on that food is not an indication that it is an actual, realized host in nature. There might be any of a number of possible explanations for the fact that Massachusetts E. imperialis larvae appear to grow faster and more efficiently on a non-utilized host (Q. stellala) than on the native host {P. rigida). Ghemical oviposition cues such as terpenes specific to conifers, selective predation of larvae on one host versus another, abiotic habitat requirements of soil pupation, and even simple availability may all play a role in the restriction of northern E. imperialis to conifers. Eor example, I observed late instars of wild- reared larvae placed on (7. stellala undergo heavy predation by vespid wasps {Vespa vulgaris-, pers. obs.) relative to those reared in situ on P. rigida. The frass of oak-feeding larvae is less dry and more prone to mold than that of pine-feeding larvae, and may serve to attract predators. The restriction of Earles imperialis to the common pitch pine on Martha’s Vineyard is of interest from the standpoint of conservation as well as evolutionary ecology. Earles imperirdis is one of 24 regionally threatened moth species occurring on Martha’s Vineyard protected under the Massachusetts Endangered Species Act (MESA; M.G.L. c. ISlAand regulations 32 1 GMR 1 0.00) , not including at least one additional species, Dntana rontrarta (Notodontidae) that appears to have been impacted severely on mainland southern New England and may be locally extirpated. As thorough an understanding as possible of why species such as these have declined — and what they require to persist — is a mission-critical prerequisite to any reintroduction and restoration effort. As conservationists consider potential sites at which to reintroduce and restore this species, we must weigh a variety of considerations, among them suitability of habitat, availability of host plant, probability of success, verifiability of historical occurrence, and legal logistics. Biologically, the most obvious candidate sites, those showing the greatest promise for success, are barrens habitats on Cape God and in Plymouth County and on Nantucket Island. Ironically, historical records of E. imperirdis from Cape Cod are lacking, and the conspicuous absence of this moth and its near relatives from Nantucket has long been noted: Jones and Kimball (1943) made the observation that although four species of ceratocampine saturniids occur on Martha’s Vineyard, none were known at the time of that writing from Nantucket. Jones and Kimball speculated that such heavy bodied moths found it difficult to distribute across water barriers. As was the case during Jones and Kimball’s time, four species of Ceratocampinae {E. imperialis, Anisota senatoria, A. stigma, and A. virginiensis) persist in numbers on Martha’s Vineyard. Anisota stigtua, at one point listed under the Massachusetts Endangered Species Act, occurs less ubiquitously on mainland New England than on Martha’s Vineyard (Mello et rd., 1999). However, this species has apparently colonized Nantucket (K. Coombs-Beattie, pers. comm.; Goldstein, 1997), where it now occurs commonly, possibly obviating the argument that all ceratocampines have difficulty crossing water barriers. Although Jones and Kimball did not discuss the historical ecology or land use history per se of either island, the possible role of habitat destruction and fragmentation of barrens habitats must be considered. 42: 34-49, 2003 (2010) 47 There can be little debate that viable habitat persists at mainland barrens sites, including the extensive maritime barrens in Plymouth County at Myles Standish State Forest (approx. 16, ()()() acres) and at the Massachusetts Military Reservation (roughly 15,000 acres), as well the 2,000 acre inland barrens at Montague Plain, Franklin Co., MA and on Nantucket Island. Jones and KimbalFs (1943) observation that ceratocampines were absent from Nantucket during the20th century of course begs the question of whether they were ever there. The land use history of Nantucket, like that of Martha’s Vineyard, involved significant alteration and conversion of forested and shrubland habitats for the purposes of agriculture (Dunwiddie, 1992). Although both Martha’s Vineyard and Nantucket were part of an extensive coastal plain as recently as 10, GOO years ago, Nantucket was almost completely denuded of forest during the Revolutionaiy War, which no doubt had an impact on the lepidopteran fauna. It may be observed that, in addition to the ceratocampines, other groups of forest tree Lepidoptera are depauperate on that island relative to Martha’s Vineyard. There is a marked contrast, for example, between the islandic faunas of Limacodidae: whereas nine species of limacodids {Euclea delphinii, Isa textiila, Phobetron pitchecium, Prolhnacodes hadia. Apod a bigut tat a, Lithacodes fasciola, Packardia elegans, P geminata, Torticidia flexuosa) occur regularly oil Martha’s Vineyard, only two {E. delphinii and L. fasciola) were recorded on Nantucket byjones and Kimball (1943: 123-125) . Jones and Kimball’s data also suggest a comparative dearth of leaf litter feeding deltoid noctuids on Nantucket relative to Martha’s Vineyard, which would be expected following systemic deforestation. Pitch pine, however, is now an extremely common plant on Nantucket and, ironically, it is conceivable that the introduced tachinid C. concinnata, to be verified from the island of Martha’s Vineyard or Nantucket, may prevent the re-establishment of E. imperialis on mainland New England. It has yet to be determined whether or not the introduced parasitoid C. concinnata poses a barrier to recolouizing the mainland, whether or not the fly’s absence on Martha’s Vineyard is, if not an artifact of under-sampling, a reason for Eacles' persistence there. Acknowledgements Much of this \vt)rk was conducted as an undergraduate tltesis under the mentorship of M. Deane Bowers and James M. Carpenter. All errors, omissions and otlicr such gaffes remain the author’s sole responsibility. Many then at the .Vluseum of Comparative Zoolog}' and the Biological Laboratories at Harvard University are to be thanked for their help, guidance, and friendship: Eric Fajer, Ed Armstrong, Charlie Vogt, Scott Shaw, David Furth, Mark Skinner, Katht' Brown-Wing, and Peter Frumlioff. E(|nally iinjtortant were my fellow naturalists in southeastern Massachusetts who kindly contributed field work and observations and, through the various conservation organizations for which they worked, granted |3ermission to conduct that work on a variety of protected areas. Of particular help were Tim Simmons (Sheriff’s Meadow Foundation), Gus Ben David (Massachusetts Audubon Society), anti 'Fttm Chase (The Trustees of Reservations). Mary Lesniak kindly enabled acce.ss to the collection of the Boston Society of Natural History, currently housed at Boston University. Jeff Boettner (University of Massachusetts) and Tim Simmons (Massachusetts Natural Heritage & Findangered Species Program) |)rovided tnany valuable insights and clarifications. I thank Mo Nielsen (Michigan State University), Paul Opler (Colorado State University), Rav Pupedis (Peabody Museum of Natural History, Yale Lhiivesity), and Brian Scholtens (College of Charleston) for relevant observations and collection records from their respective institutions. Rodger Gwiazdowski (University of Ma.ssachusetts, .Amherst) provided real-time checking of EV/efevdata labels at UMass. I thank Sean Bober for assisting with the assemblage of figures and Mike Nelson (Massachusetts Natural Heritage & Endangered Species Program) for invaluable eleventh hour help with maps. Literature cited Abbot, J. & J. E. Smu h. 1797. The natural history of the rarer lepido]5terous insects of Georgia. 2 vols. London: printed hv T. Bensley lot J. Edwards. ARNAtut), P. H.,Jr. 1978. A host-parasite catalog of North American Tachinidae (Diptera). USD.A. Science and Education Administration, Washington. D. C. publication 1319. Acren & Axei.sson. 1979. Energ-y budgets do balance. Oecologia 42: 375-370. Be(:err.\, J. X. 1997. Insects on plants: macroevolutionaiy chemical trends in host use. Science 270: 253-256. Berenbaum, M. 1981. Patterns of furanocoumarin distribution and insect herbivory in the Umbelliferae: plant chetnistry and commnnity structure. Ecology 62: 125T1266. Bereociier, S. 11. 1998. Can sympatric speciation be proven from biogeographic and phylogenetic evidence? I’p. 3-15. hv. 1). J. 1 loward and S. H, Bet locher (eds). Endless forms: species and s|)eciation. New York. NY: Oxford University Press. Bernavs, E. a. 1989. Host range in jthvtophagous insects: the potential role of generalist itredators. Evolutionary Ecolog)' 3:299-311. Bewick, J. A. 1979. Gypsy moth control |5rojcct report. Ma.ssachusetts Executive Office of Environmental Affairs. Boston, AIA. BinwEi E, R. G. S. & D. J. Dur/.on. 1975. Some recent aspects of nitrogen metabolism. Pp. 152-155. hv. P. [. Davies (ed.). Historical and current aspects of jtlant phy.siolog}': a svmposium honoring F. G. Stewart. State College of Agricultural and Life Sciences, New York. Boeitn'er, C. J.. J. S. Ei.kinton, & C. G. Boeitner. 2009. Effects of a biological control introduction on three nontarget native species of saturniid moths. Conservation Biology 14(6): 1798- 1806. Bowers, M. D. 1983. Iridoid glycosides and larval hostplant specificity in checkerspot butterflies (Euphydrya.s, Nymphalidae). J. Chem. Ecol. 9: 47.5-493. Bowers, M. D. 1984. Iridoid glycosides and hostplant specificity in larvae of the buckeye butterfly, yitttothrt coenia (Nymphalidae). J. Chem. Ficol. 10: 1567-1577. Bowers, M. D., N. E. Stamp, & E. D. Fa|er. 1991. Factors affecting calculation of nutritional indices for foliage feeding insects: an experimental approach. Entomologia Exjterimentiilis & 48 /. Rf.'t.Lepid. A])plicacla 61: 101-116. Brues, C. T. 1920. The selection of food-plants by insects, with special reference to lepidopterons larvae. Ain. Nat. 54: 313- 332. Brues, C. T. 1924. The s]>ecificity of foodplants in the evolntion of irhytophagons insects. Am. Nat. 58: 127-144. C.vnts, R. G. 1980. Feeding patterns of nionophagous,oligophagous, and polyphagous insect herbivores: the effect of resource abundance and plant chemistiy. Oecologia (Berl.) 46: 22-31. (Ivies, R. G. & D. F. Rhoades. 1977. Patterns in the production of anti-herbivore chemical defenses in plant communities. Biochem. Syst. Ecol. 5: 185-193. GovEi.i,, G. V. 1984. A field guide to the moths of Eastern North America. Boston: Houghton Mifflin, Go. Gui,ver, J. J. 1919. A study of Comfmlura concinnala, an imported tachinid parasite of the gipsy moth and the brown-tail moth. United States Department of Agriculture, Bulletin 766. DuNwinniE, P. W. 1992. Changing landscapes: a pictorial field guide to a century of change on Nantucket. Privately published. Eiirlk:!!, P. R. & P. H. Rwen. 1964. Butterflies and plants: a study in coevolution. Evolution 18: 586-608. Ei.io'it I. M. & C. G. SoiiEE. 1902. Caterpillars and their moths. New York: The Centuiy Co. Emi.en, J. M. 1973. Feeding ecology. Pp. 157-185. /n: Ecology: an evolutionary approach. Reading, MA: Addison-Wesley. F.vrrei.l. B. D. 1998. “Inordinate fondness” explained: why are there so many beetles? Science 281: 555-559. Farrell, B. D. & C. Mrm'.R. 1990. Phylogenetics of insect/plant interactions: have Phyllohrolka leaf beetles (Chrysomelidae) and the Lamiales diversified in parallel? Evolution 44: 1389- 1403. F..\rrell, B. D. & C. MirruR. 1998. The timing of insect/plant diversification: might Tetmopes (Coleoptera: Cerambycidae) and Asdepias (Asclepiadaceae) have co-evolved? Biol. J. Linn. Soc. 63: 553-577. F.vrquh.vr, D. W. 1934. The Lepidoptera of New England. PhD dissertation. Harvard University. Fergii.son, D. C. 1971 . The moths of North America, Fascicle 20. 2A Bombycoidea: Saturniidae (Part). London: Curwen Press. Forman, R.T.T (ed.). 1979. Pine barrens: eco.system and landscape. New York: Academic Press. Fr..\nk, K. D. 1988. Impact of outdoor lighting on moths: an asessment. Journal of the Lepidopterists’ Society 42: 63-93. Funk, , D. j., D. J. Fu'euvma, G. Or'i i, & A. Meyi-:r. 1995. A histoiy of host associations and evolutionary diversification for Oplimdla (Coleo|)tera: Chrysomelidae): new evidence from mitochondrial DNA. Evolution 49: 1008-1017. Futuyma, D. J. 1976. Foodplant specialization and environmental predictability in Lepidoptera. Am. Nat. 1 10: 285-292. Futuyma, D. J. Sc S. S. MoC.-vekeriy. 1990. Phylogeny and the evolution of host plant associations in the leaf beetle genus Ophmella (Coleoptera: Chrysomelidae). Evolution 44: 1885- 1913. Ga.si()n, K. j. & D. Re,\vi-:y. 1989. Patterns in the life histories and feeding strategies of British macrolepidoptera. Biol. J. Linn. Soc. 37: 367-381. Gilbert, L. E. 1979. Development of theory in the analysis of iiisect- jjlant interactions, pp. 1 17-154. Im D. J. Horn, G. S. Stairs, and R. 1). Mitchell (eds.). Analysis of Ecological Systems. Columbus: (Dhio State University Press. Goldstein, P. Z. 1991. Investigation of the natural history and nutritional ecology' of a remnant jropiilation of Eades imperia/is Druiy on Martha’s Vineyard Island, MassachusetLs. Lhidergraduate thesis. Harvard University, Cambridge, MA. Goi.iwtein, P. Z. 1997. Review of: Tuskes, Tuttle, and Collins. The wild silk moths of North America. J. New York Ent. Soc. 105(1-2): 121-125. Harris, T. W. (C. L. Flint, ed.) 1890. Treatise on some of the insects injurious to vegetation. New York: Orange Judd Co. 640 pp. [2'“' edition of: Harris, T. W. 1841. A report on the insects of Massachusetts, injurious to vegetation. Cambridge, MA: Folsom, Wells and Tluirston.] H.4YES,J. L. 1983. Acomparison of the life history and moiphological character patterns in temperate butterflies. J. Kans. Ent. Soc. 56: 547-551. Hessel, S. A. 1976. A preliminaiy scan of rare and endangered Nearctic moths. Atala 4: 19-21. Hollow.ay.J. D. & P. D. N. Herbert. 1979. Ecological and taxonomic trends in macrolepidopteran host plant selection. Biol.J. Linn. Soc. 11: 229-251. House, H. L. 1962. Insect nutrition. Ann. Rev. Biochem. 31: 653-672. Howard, L. O. & W. F. Fiske. 1911. The importation into the United States of the parasites of the gipsy-moth and the browntail moth. USDA Bur. Entomol. Bull. 1911, 1. J.AME.s, T. D. 'W. 8c D. W. Smith. 1978. Seasonal changes in the major ash constituents of leaves and some woody components of trembling aspect and red osier dogwood. Can. J. BoL 56: 1798-1803. " Jermy, T. 1984. Evolution of insect/host plant relationships. Am. Nat. 124: 609-6.30. JoNE.s, F. M. & C. P. Kimball. 1943. The Lepidoptera of Nantucket and Martha’s Vineyard Islands, Massachusetts. The Nantucket Maria Mitchell Association. Vol. IV. Nantucket, MA. Lem.vire, C. 1988. The Saturniidae of America. Ceratocampinae. Museo Nacional de Costa Rica, San Jose. Little. S. 1979. Fire and plant succession in the New Jersey Pine Barrens. Pp. 297-.314. /«: Forman, R. T. T. (ed.). Pine Barrens: ecosistem and landscape. Orlando, FL: Academic Press. M.\tt.son, W. j. 1977. Size and abundance of forest Lepidoptera in relation to host plant resources. Colloques Inteniatioiiaux du C. N.R.S. 265: 429-441. M.-vit.son, W. j. 1980. Herbivory in relation to plant nitrogen content. Ann. Rev. Ecol. Syst. 11: 119-161. M/Vrr.soN, W. J. & J. M. Scriber. 1987. Nutritional ecology' of insect folivores of woody plants: nitrogen, water, fiber, and mineral considerations. Pp. 114-139. /w: F. Slansky and J. G. Rodriguez (eds.) , Nutritional ecology of insects, mites, spiders, and related invertebrates. John Wiley and Sons, New York. Meli.o, M., M. Ai.iberti, S. Galllsi-la, K. Gerber, F. Hohn, A. Lawrence, D. Luers, R.J. Nagel, T. Ruehi.i & B. Stephen.son. 1999. Lloyd Center for Environmental Studies. Inventory of state-listed Lepidoptera and other insects at Massachusetts Military Reservation 1996-1998. Report to Massachusetts Natural Heritage & Endangered Species Program. Michener, C. D. 1950. A northern subspecies of Backs imperialis (Lepidoptera: Saturniidae). Jour. Kaiis. Ent. Soc. 23: 17-21. Miller, P, C. & W. A. Stoner. 1979. Canopy structure and environmental interactions. Pp. 428-458. Im O. T. Solbrig, S. fain, G. B. Johnson and P. H. Raven (eds.), Topics in plant population biology. New York: Columbia University Press. Moli.s.seau, T. A. & H. Dingle. 1991. Maternal effects on insect life histories. Ann. Rev. Entomol. 36: 511-534. Niemeia, P., S. Hanhimaki, & R. M.anniia. 1981. The relationship of adult size in noctuid moths (Lepidoptera: Noctuidae) to breadth of diet and growth form of host plants. Annales Entomologici Fennici 47: 17-20. Opi.er, P. a. 1978. Interaction of plant life histoiy components as related to arboreal herbivoiy. Pp. 23-32. Im G. G. Montgomeiy (ed.). Ecology of Arboreal Folivores. Washington, D.C.: Smithsonian Institute. Patch, E. M. 1908. Insect notes for 1908. Maine Agric. Exp. Stn. Bull. 162: 351-.387. PF.TER.S, R. H. 1983. The ecological implications of body size. 42; 34-49, 2003 (2010) 49 (Cambridge: (lambt idge University Press. R\iia, K. F. 1991. Induced del'ensive reactions in conifer-bark beetle .systems. Pp. 245-276. In: D. W. Tallainyand M.J. Raupp (eds.) Phytochemical induction by berbivores. New York: Academic Press. ScHWF.n zKR, D. F. 1983. Rare Lepidojitera of the Kiitama Plain on Martha’s Vineyard, Massachtisetts. Report to Massachusetts Natural Heritage and Engaiigered Species Program, Boston. Sc.HWEirzKR, D. F. 1988. Stattis of Satnrniidae in the northeastern USA: a quick review. News of the Lepido|>terists’ Society [1988] 1:2-3. Sc.HWEiT/.FR, D. F. &T. j. Rawin.ski. 1988. TNU Element stewardship abstract: northeastern itch jDine/scrtib oak barrens. The Nature Conservancy, Eastern Heritage Task Force, Boston, MA. SciiROEDER, L. A. 1984. Comparison of gravimetry and planimetry in determining dry matter budgets for three species of phytophagous lepidopteran larvae. Entomol. Appl. 35: 25.5- 261. St.RiBER, ). M. 1977. L.imiting effect of low-leaf-water content on the nitrogen utilization, energy budget, anrl larval growth of Hyaloplwm cecropia (Lepidoptera: Satnrniidae). Oecologia 28: 269-287. ScRiBER, J. IVl. 1978. Fhe effects of larval feeding specialization and plant growth form on the consumption and utilization of plant biomass and nitrogen: an ecological consideration. Ent. Exp. App. 24: 694-710. ' ScRiBER, |. M. 1 979a. The effects of sequentially switching foodplants upon biomass and nitrogen utilization by polyphagous and stenophagous larvae. Fait. Exp. Appl. 25: 203-215. ScRiBER, J. M. 1979b. Fdfects of leafwaler .su]i]>lementation u|ion postingestive nutritional indices of foil), slmib, vine, and treefeeding l.e|)idoptera. Ent. Exj). Appl. 25: 240-252. ScRiBER, J. .M. 1983. fhe evohuion of feeding s|)ecialization. physiological efficiency, and host races in selected Papilionidae and Satnrniidae. Pp. 373-412. In: Denno, R. F’. and M. S. McClure (eds.). Variable plants and herbivores in natural and managed .systems. NewYoi k: Academic Press. ScRiBER, J. M. & P. Feeny. 1979. Growth ol herbivorous cateipillars in relation to feeding s|)ecialization and to the growth form of their food |)lant.s. FTology 60; 829-850. ScRiBER, J. M. & F. Seansky. 1981. The nutritional ecology of immature insects. Ann. Rev. Entomol. 26: 183-21 I. Si.ANSRY, F. & J. M. ScRiBER. 1985. Food consumption and utilization. Pp. 87-163. In: G. A. Kerkut and T. 1. Gilbert (eds.). Comprehensive Insect Physiology, Biochemistry, and Pharmacolog)' Vol. 4. Pergamon Press. Sro.NE, S. E. 1991. Foodplants of world Sattirniidae. Tepid. Soc. Memoir 4. L.epidopterists’ Society, Lawrence, Kans. Tii.ton, D. L. 1977. Seasonal growth and foliar nutrients oi Lai ix laricina in three wetland ecosystems. Can. j. Bot. 55: 1291- 1 298. Tuskes, P. M., j. P. Tuiti e, & .VI. M. Cott.iNs. 1996. The wild silk moths of North America. A natural history of the saturuiidae of the United Stales and Canada. Ithaca; Cornell University Press. W.M DBAUER, G. P. 1968. The consumption and utilization of food bv insects. Advances in Insect Physiology' 5: 229-288. W.\s,SERMAN, S. S. & C. Miri'ER. 1978. The relationship of body size to breadth of diet in some Le|)idopiera. Flcol. Fait. 3: 15.5-161). Webber, R. T, & j. Sch.aeener Jr. 1926. 1 lost relations of Cowpsilnm concinnutn Meigen, an important tachinid parasite of the gipsy moth and the lirown-tail moth. Washingoii 1). C., USDA bulletin 1363. WiKi.L'Ni), C. 1975. The evolutionary relationships between adult oviposition preference and larval host plant range in VnpiUo machaonl,. Oecologia 18: 185-197. journal oj Rrsearrh on the Lepuloptera 42: 50-55. 2003 (2010) Association of three species of Strymon Hiibner (Lycaenidae: Theclinae; Eumaeini) with bromeliads in southern Brazil Simone Schmid*'--, Volki:r S. Schmid'-^, Rafael Kamke'^, Josefina Steiner^ and Anne Zillikens'-'^ 'Mecl.-Naturwissenschaftliches Forschiingszentrum, Universitat Tubingen, 72074 Tubingen, Germany -De])artnieiit of Cell Biology, Embiyology and Genetics (BEG), CCB, Federal Lbiiversity of Santa Catarina, Campus Universitario Trin- dade, 88.040-900 Florianopolis, SC, Brazil simigrohme@holmaiLcom Abstract. As part of a project studying the species richness of bromeliad flower visitors and the diversity and nature of their animal-plant interactions, three species of the lycaenid butterfly genus Strymon were recorded as pests of bromeliad inflorescences. Strymon ziba fed on the fruits of Aechmea n udicaulis, S. oreaki on those of Ae. lindenii and Ae. caudata and S. serapio on the dry capsules of Vriesea friburgensis. The caterpillars of 5. zAfland S. oreala, pests t)f cultivated pineapple, were facultatively associated with ants. One S. ziba pupa was parasitized by a chalcidid wasp. One S. oreala pupa was parasitized by an ichneumonid wasp. BehaGoiir and life history data of the caterpillars are described and aspects of the host specificity of the lycaenids and potential pest control by parasitoid wasps are discussed. Key words: Aechmea, animal-plant interactions, Atlantic rain forest, Bromeliaceae, Chalcididae, herbivory, Ichneiimonidae, parasitism, restinga, Vriesea. Introduction Bromeliaceae, a neotropical plant family, can be considered keystone species by providing microcosms for the richness of tropical rain forests due to the high diversity of animal taxa, especially arthropods, associated -with them (Frank & Lounibos, 2008). Among the latter, Lepidoptera are major lierbivores with many caterpillars feeding upon bromeliad foliage: Napaea eiicharilla Bates (Riodinidae) on Werauhia sanguinolenta (Cogniaux & Marchal) R. Grant (syn. Vriesea sanguinolenta, Schmidt & Zotz 2000), Aechmea bracteata Grisebach, Ae. nudicaulis (L.) Grisebach (Beiitelspacher, 1972) And Ananas comosus {V.) Merrill (Schmidt & Zotz, 2000), Cana domitianus moFabridiis (1793) (Riodinidae) on TUlandsia caput-medusae E. Morren (Beuteispacher, 1972; Frank & Lounibos, 2008), Dynastor darius darius Stichel and D. niacrosiris Westwood (Nymphalidae, Urich & Emmel, 1991a, b) on Ae. nudicaulis and Castnia boisdiivalii Walker (Castniidae, Biezaiiko, 1961; Frank & Lounibos, 2008) on T. aeranthos (Loiseleur) L. B. Smith. Beuteispacher ( 1972) also mentioned Theda te|;m7«ButIer & Druce 1872 (Lycaenidae), which is a mistaken record of Ziegleria hesperilis, feeding on TUlandsia caput-medusae, but voucher specimens of “Theda hesperitis" in UNAM *Correspon d i ng a ulhor ReceixH'd: 25 Sepletnber 2009 Accepted: 15 October 2009 (Universidad Nacional Autonoma de Mexico) examined by Robert K. Robbins are in fact S. serapio Godman & Salvin (1887) (R. K. Robbins, pers. comm.). Not only the plants’ vegetative parts, but also their inflorescences contribute significantly to the local fauna! biodiversity by providing resources for a great variety of flower visitors that act as pollinators or pollen and nectar robbers (Sazima & Sazima, 1999; Machado & Semir, 2006; Caneia & Sazima, 2003; Schmid et al, b, submitted). In addition to causing leaf damage, some herbivorous arthropods associated with bromeliad inflorescences directly interfere with plant reproduction by feeding on reproductive tissues of flowers and fruits, like beetles, butterflies and moths, grasshoppers and even crabs (Fischer et al., 1997; Caneia & Sazima, 2003; Frank & Lounibos, 2008). An example of inconspicuous herbivory affecting plant reproductive success was observed in the bromeliads Vriesea friburgensis Mez and Werauhia gladioliflora (H. Wendland) J. R. Grant whose buds are parasitized by Eurytoma wasps (Hymenoptera, Eurytomidae) so no fruits are formed (Gates & Cascaiite-Marm, 2004; Grohme et at, 2007). Beyond that, cases of seed predation by C/jo/ms and Metamasius weevils (Coleoptera, Ciircolioiiidae) (Frank 1999) and Epimoritis testaceellusVAigonoi 1887 (Pyralidae) larvae that develop in flower pods of TUlandsia fasciculata Swartz (1788) (Biigbee, 1975; Heppiier, 1992) have been I'eported. New World hairstreaks (genus Strymon, Lycaenidae: Theclinae: Eumaeini) use ornamental bromeliads (genera Aechmea, TUlandsia) and the commercial pineapple {Ananas comosus) as host plants 42; 50-55, 2003 (2010) 51 (Robbins & Nicolay, 2002). Strymon megarus Godart 1824 (syn. Theda basilides, also misspelled as T. basalides) larvae feed on Ananas and odier bromeliads like Ae. bracteata (Bentelspacher, 1972; Frank & Lonnibos, 2008) and can be considered pest species. Strymon ziba, S. serapio (Robbins & Nicolay, 2002) and .S', oreala (Zikan, 1956) , the species examined in our study, were reported to eat bromeliads, S. ziba Hewitson 1868 and S. oreala Hewitson 1868 are known pests of A. comosus (Harris, 1927; Zikan, 1956). Since caterpillars and other immature stages of the Eumaeini are small and ci'yptically coloured, food plants have been recorded for only 25% of the species (Duarte et al., 2005). Studying the species richness of bromeliad flower visitors and the diversity and nature of their animal-plant interactions in tlie Atlantic rain forest of southern Brazil, we found lycaenid caterpillars attacking developing fruits. In order to assess the specificity of these associations we examined inflorescences of four common sympatric bromeliad species, Aechmea nudicauUs, Ae. Undenii (E. Morren) Baker, caudato Lindman 1891 (Bromelioideae) and Vriesea friburgensis (Tillandsioideae), for the presence of larvae and reared them for identification. also recorded basic data on development, behaviour and natural enemies of the caterpillars. Materials and methods Bromeliads with inflorescences were searched for eggs and larvae between November 2006 and June 2008 at four study sites (frost-free subtropical habitats) : Santo Antonio de Lisboa and the Environmental Conservation Unit Desterro UGAD (both secondary forest; 27°.80’26” S, 48°30’28” ’V\^; 27°.3r 50” S, 48°30’50” W) (Zillikens et al, 2001; Zillikens & Steiner, 2004) as well as Joaquina Beach and Campeche Beach (dune vegetation, Sampaio et al, 2002; 27°40’.38” S, 48°28’48” W; 27°37’37” S, 48°26’59” W), on Santa Catarina Island, southern Brazil. All bromeliads examined were growing terrestrially although Aechmea nudicaidis Aho occurs on trees. In total, 20 infested bromeliads of four species {Aechmea nudicauUs, n = 11; Ae. lindenii, n = 2; Ae. caudata, n = 5; and Vriesea friburgensis, n = 2), growing terrestrially on rocks, in sand or shallow soil, were taken to the laboratory. Presence, size and colour of eggs and caterpillars on the inflorescences were observed regularly eveiy 1-2 days. Ants a.s.sociated with lycaenid larvae were also collected. When larvae had finished feeding and retreated for pupation, tlie bromeliad plants were enclosed with fine gauze to capture the emerging adult butterflies. Voucher specimens of the recorded butterfly, ant and parasitoid species were deposited in the entomological collection of J. Steiner at the Native Bee Laboratory (LANUESC), BEG, Federal University of Santa Catarina, Florianopolis, Santa Catarina, Brazil. Results Three lycaenid species of the genus Strymon were reared from the caterpillars found on four bromeliad species. All constitute new records of parasite/ host association. Up to four caterpillars were found simultaneously on one inflorescence. Strymon ziba (Hewitson 1868) We found 24 laiwae of S. ziba on 1 1 inflorescences of the bromeliad Aechmea nudicauUs, yielding an average of 2.2 caterpillars per inflorescence (range 1-4) from November 2006 to January 2007. Seventeen adults emerged in the laboratory, overall sex ratio was 0.7 (M/F) . On all inflorescences we detected small white spherical bodies, the eggs from which the larvae had hatched (Fig. lA, B). The colour of the larvae was cryptic and changed during their growth from wliitish-yellow to reddisli- pink (Fig. 1C, D). The former matched well to the fruits whereas the latter matched well to the inflorescence stem. The larvae appeared shortly after the end of the flowering period and stayed close to the ripening fruits. Larval feeding behaviour consisted of gnawing a hole into the fruit base large enough for the smaller larval stages to enter the fruit completely and for the larger stages to insert the head anrl anterior part into the cavity. Through this hole they fed on the soft nutritive tissues of the ovaiy and ovules, leaving the rigid cortical outer wall of the developing fruit mostly intact (Fig. lA, B). On an inflorescence of Ae. nudicauUs with two larvae feeding, 30 fruits were damaged, resulting in a mean of 15 fruits damaged per larva. Mean fruit loss per inflorescence was 84.5% (n = 2). Occasionally, the larvae drew back from the fruits to hide under the bracts for about lialf a day, probably for moulting. Development in the egg took five days (n = 2). The larval phase lasted 1.3-15 days (ii = 1 ). The fully grown larvae (-12-15 mm length) moved into the bromeliad rosette where they pupated on the upper side of the leaves half way between tip and base; one male pupated on a bract of the inflorescence. Pupation took 8-11 days (n = 3). Imagines (Fig. IE) emerged between mid November until end of January, synchronized to the flowering/fruiting period of Ae. nudicauUs. In the laboratory, caterpillars were occasionally 52 /. Res.Lepid. Figure 1. Strymon larvae on bromeliads on Santa Catarina Island, Southern Brazil. A-P; Strymon larvae and imagines and associated ants A-B: Infested fruits of Aechmea nudicauiis, Santa Catarina Island, Brazil. A: Fruit with feeding hole and larval faeces of Strymon caterpillar and a hatched egg of Strymon ziba at the base of the fruit. B: Fruit with feeding hole and a closed egg. C-D: Colour change in Strymon ziba larvae. C: Small, whitish-yellow. D: Larger larva after colour change to reddish-pink. E: Strymon ziba female, collected on Aechmea nudicauiis, Santa Catarina Island, Brazil. Right-hand side ventral view, left-hand side dorsal view. F-l: Ants associated with Strymon ziba caterpillars. F; Crematogaster limata. G: Linepithema iniquum. H: Monomorium sp. (floricola). I: Paratrechina sp. J: Strymon oreala, female, right-hand side ventral view, left-hand side dorsal view. K; Strymon oreala larva feeding on fruits of Aechmea caudata. L: Larva of S. oreala feeding on withered flower petals. M: Tapinoma melanocephalum ant on the back of a S. oreala larva (v/hite arrow). N: Strymon serapio, female, right-hand side ventral view, left-hand side dorsal view. O: Larva of S. serapio on dry fruit of Vriesea friburgensis with feeding hole. P: Dry fruit of Vriesea friburgensis with exuvia of S. serapio inside. Length of exuvia: = 1 2 mm. Q-W: Strymon larvae at A. lindenii and parasitoids. Q-R: Strymon larvae. Q: Strymon larva on A. lindenii feeding on fruit. R: Strymon pupa on infructescence. S-T: Anisobas, a parasitoid of Strymon sp. S: Lateral view of the Anisobas imago that hatched from the Strymon exuvia. T; Opened Strymon exuvia besides hatched Anisobas imago. U-W: Conura, a parasitoid of Strymon sp. U: Opened pupal case of Strymon sp. V: Imago of Conura sp., dorsal view. W: Lateral view. 42: 50-55, 2003 (2010) 53 tended by ants of four species: Crematogaster limatn^mith 1858, Linepithema iniquum Mayr 1870, Monomorium //onro/fl Jerdon 1851 and Paratrechina (Fig. lF-1). Tending worker ants walked over the bodies of the caterpillars and took up small droplets secreted posterodorsally (Fig. IF). Additionally, we observed S. z//;a imagines sucking extrafloral and floral nectar from inflorescences of Ae. nudicaulis and one female laying one single egg on each of two recently withered flowers, respectively. The female flew around the inflorescence and sucked nectar of several flowers before ovipositing. The initially greenish eggs turned white after a few minutes. Thereafter, the female left the inflorescence. Strymon oreala (Hewitson 1868) We discovered seven larvae of ,S’. oreala (Fig. IJ) on five inflorescences of the bromeliad Ae. caudata (April 2008) and two larvae on two inflorescences of Ae. lindenii (August - September 2007) . As described foi¬ ls'. ziha on Ae. nudicaulis larvae fed on the developing fruits of A^'. lindenii And Ae. caudata (Fig. IK-M) and retreated into the rosette for pupation. Additionally, a larva was seen feeding on withered flower leaves, probably eating old reproductive structures inside the petals (Fig. IL). On an inflorescence of Ac. caudata with one larva feeding, 15 fruits were damaged. The pupal stage took 15-16 days (n = 4). The larvae were reddish-pink like the inflorescence stem (Fig. 1 K-M). Occasionally, single ants of the species Tapinoma melanocephalum Fabricius 1793 were observed on .S', orm/rt larvae (Fig. IM). Strymon serapio (Godman 8c Salvin 1887) We detected four browuish-yellow larvae of S. serapio (Fig. IN) on two inflorescences of the bromeliad Vriesea friburgensis in secondary forest in December 2007. The larvae chewed a hole into the hard capsule of the developing fruit (Fig. lO) and fed on the seeds within. Pupation took place inside the empty fruit capsule (Fig. IP), the imago emerged after 1 1 days (n = 1). Parasitoids On two occasions we found Strymon brood infested with a parasitoid. The first case (14 November 2005, Oampeche Beach) was a larva on an inflorescence of Ac. lindenii (Fig. IQ). Since the only identified Strymon infestation of this bromeliad was by .S’, orecda (see above) we assume that the parasitized larva belonged to the same species. The larva pupated on the infriictescence on 18’'’ November 2005 (Fig. IR). This might, however, not be the usual location for pupating because Aurum® insect glue had been applied to the infriictescence stem, thus preventing the caterpillar from moving down to the rosette. After 18 days an ichneumonid wasp of the genus Anisohas (subfamily Ichneumoninae) emerged from the pupa (Fig. IS, T). In the second case, discovered 6 January 2007 at Santo Antonio, a pupa (Fig. lU) was located at the upper margin of a leaf of Ac. nudicaulis (plant with infriictescence). So far, we found this bromeliad species only to be infested with .S', ziba (see above); hence we assume that the pupa belonged to this species. On 23 January 2007, a chalcidid wasp of the genus Conura (subfamily Chalcidinae), “most probably of the //aw; group” (Gerard Delvare, pers. comm.), emerged from the pupa (Fig. IV, W). Discussion Larval behaviour and host plants Our findings constitute new host records for the associated Strymon species. The only lycaeuids so far recorded on Ac. lindenii Are larvae of an uuidentified species of Theda on Santa Catarina Island (Lenzi et al. 2006) with a similar feeding behaviour and life history data as described here for .S', oreala. It is therefore possible that they did in fact observe larvae of a species of Strymon. Our observations also confirm some life history traits rejiorted by Duarte et al. (2005) such as the cryptic coloration of the larvae w4uch is well adapted to parts of the plants on which they move or feed. Besides the evident association of Strymon larvae with infructescences, they were even more selective in the sense that they fed only on the internal parts of the developing fruits, i.e. the ovaries, but not on leaves, sepals or other plant tissues. By hollowing out the fruits they create their own shelter for feeding or even pupal chambers (in case of .S', serapio) . We further report here the first data ou the life cycle of the three species, all of which develop within about one month and without diapause. It is therefore likely that the Strymon species studied by us are multivoltine having several generations per year. This is in agreement with the pattern reported for other Strymon species in the tropics (Opler et al. 2009). The choice of hosts by ovipositing females depends on seasonal availability of fruiting plants. The bromeliads studied by us all have relatively short and seasonal fruiting periods, so that only one or two generations can develop on a given ])lant population. 54 J. Res.Lep 'td. Adults emerging at the end of a flowering period have to seek for alternative hosts for egg laying. In our study, Strymon orealawAS the only species recorded on two hosts both in the genus Aechmea. Interestingly, the inflorescences and flowers of Ac. lindenii and Ae. caudata are very similar in floral morphology and coloration (Kiimke, pers. obs.), but Aechmea lindenii flowers from August to November (Dorneles et at, ms) whereas Ae. caudata flowers from March to June without overlap, though single plants of both Aechmea species can be found flowering outside the main flowering period, for example Ae. lindenii on more open areas in restinga sites throughout the year (Lenzi et al. 2006) and Ae. caudata in secondary forest in September (Ktimke, pers. obs.). Nevertheless, there is a gap of several months for which we do not yet know the host plants. We know, however, that S. oreala does not attack the infructescences of Ae. nudicaulis or V. frihurgensis, which flower between. Therefore, a switch to another host, whether bromeliad or not, must occur in .S’, orecdn, S. serapio and S. ziba. As .S', ziba and S. oreala are pests of cultivated pineapple it would be interesting to further identify alternative host plant species in order to better understand under which circumstances the larvae reach pest status and to assess their damage to the crop. In this context it is also worth emphasizing our record of a possible natural enemy of S. ziba, a parasitoid wasp of the family Chalcididae. Its potential as natural biological control agent should be assessed by elucidating its life history, abundance and host specihcity. Association with ants As has been reported from other lycaenid caterpillars, the larvae of .S', ziba possess a dorsal secretory organ, the Newcomer’s gland (Malicky 1970), that might exude honey-like droplets to appease ants. Of the species recorded, Monomorium Jloricola dnd Tapinoma melanocephalumme invasive ants (Delabie et al. 1995; Campos-Farinha 2005) and only occurred in the laboratory. We consider the same to be true for Pciratrechina sp. for this species was only observed in the laboratory. Cremcitognster limata and Linepithema inicjuum, however, were frequently found nesting in the bromeliads or visiting their inflorescences (Rostimek et al. 2008, Schmid et al a, ms) and were thus brought to the laboratory together with the plants taken in the field. These two, at least, may be considered associated with the Strymon cater|)illars under natural conditions, albeit only facidtatively since larvae observed in the held were mostly not tended by ants. Diversity of the Strymon - bromeliad association Our hnding that four species of bromeliads were parasitized in very similar ways by Strymoyi larvae is remarkable in yet another aspect. It conhrms a relatively high diversity and abundance of sympatric, even syntopic, Strymon species in bromeliad-rich Atlantic forest and restinga habitats of southern Brazil. Thus, the fact that three co-occurring species were recorded in studies on only four bromeliad species sitggests that a thorough examination of further bromeliad inflorescences might result in the hnding of more Strymon species and underlines the importance of these plants for sustaining a high diversity of the lepidopteran fauna in the Mata Atlantica. Acknowledgements We tliank Robert K. Robbins for lycaenid identibcation and valuable help with Strymon taxonomy and literature, as well as David Wahl and Gerard Delvare for parasitoid identification. This study is part of the Project “Internal dynamics of rain forest: specificity of animal-plant interaction” within the Brazilian-German program “Mata Atlantica”, and we acknowledge the financial sup]>ort by BMBF (01LB02()5A1) and CNPq (59004()/2()0(>5). Literature cited Bi:uiEt..st>ACHKR, C. R. 1972. Some observations on the Lepidoptera of bromeliads. Jonrtial of the Lepidopterists’ Society 26: 133- 137. Bikz.vnko, C. M. 1961. Castniidae, Zygaenidae, Dalceridae, Encleidae, Megalopygidae, Cossidae et Hepialidae da Zona Missioneira do Rio Grande do Sul. Arqiiivos de Entomologia Serie B, Escola de Agronomia “Eliseu Maciel” 14 : 1-12. Buoisee, R. E. 1975. A new species of the getius Eurytoma (Hymenoptera: Emytomidae) from a |tyralid occurring in the flower pods of Titlnndsia famrulata. Journal of the Georgia Entomological Society 10: 91-93. Gami’os-Farinha, A. E. 2005. Urban pest ant.s of Brazil (Hymenoptera: Formicidae). In: Proceedings of the Fifth International Conference on Urbati Pests. Chow-Yang Lee and William II. Robinsoti (editors), 81-84. C.ANEt.A, M. B. F. & M. .SAZtMA. 2003. Florivoiy by the crab Arntnsn angustipes influetices hntmningbird visits to Aechmea pectinata (Bromeliaceae). Biotropica 35: 289-294. Delabie, |. H. C., I. C. Do Nascimenio, P. Pacheco & A. B. Casimiro. 1995. Community structure of house-infesting ants (Hymenoptera: Formicidae) in southern Bahia, Brazil. Florida Entomologist 78: 264-270. Dorneles, L. L., A. Zillikens, B. Harier-Marques & J. Steiner. Submitted. Mixed pollination by hummingbirds and bees in Aechmea lindenii (Bromeliaceae) in Atlantic Rain Forest at Santa Catarina Island, southern Brazil. Sid>mitted to Entomologia Generalis. Duarte, M., R. K. Robbins &• O. H. H. Miei.ke. 2005. Immature stages of Cnlycopis rantonia (Hewitson, 1877) (Lepidoptera, Lycaenidae, Theclinae, E.umaeini), with notes on rearing detritivorous hairstreaks on artificial diet. Zootaxa 1063: 1-31. Fischer, E. A., L. F. I.. Du.vri e & A. C. Ar-VUIo. 1997. Consumption of 42: 50-55, 2003 (2010) Bromeliad flowers by the crab Melasesarma ndmpes in a Brazilian coastal forest. Crustaceana 70: 118-123. Fr..\nk, ]. H. 1999. Bronieliad-eating weevils. Selbyana 20: 40-48. Fr.-\nk, J. H. & L. P. Lolinibos. 2008. Insects and allies associated with bronieliads: a review. Terrestrial Arthropod Reviews 1: 125-153. G. vtes, M. W. & A. Cascan’I'e-M.-vrIn. 2004. A new phytophagotis species of Euryloma (Hymenoptera: Eurytoniidae) attacking Wmiuhia gladioUflora (Bronieliales: Bronieliaceae). Zootaxa 512: 1-10. Grohme, S., J. Steiner & A. Zii.i.ikens. 2007. Destruction of floral buds in the bi'oineliad Vriesea friburgensis by the phytophagous lan'ae of the wasp Ewjlomasp in southern Brazil (Hymenoptera: Eun'tomidae). Entomologia Generalis 30: 167-172. H. arris, W. V. 1927. On a lycaenid butterfly attacking pineapples in Trinidad, B.W.l. Bull. Entomological Research 18: 183-188. Heppner, J. 1992. Bromeliad pod borer, Epimorms leslaceellus (Lcpidoptera: Pyralidae: Gallerinae). Florida Department of Agriculture and Consumer Services, Division of Plant Industiy, Entomology Circular 351: 1-2. Lenzi, M., }. Z. De M.\tos & A. 1. Ortu. 2006. Variacao morfologica e reprodutiva de Aechmea Undenii (E. Morren) Baker var. lindenii (Bronieliaceae). Acta Botanica Brasilica 20: 487-500. M.ACH.41X), C. G. &J. Semir. 2006. Fenologia da lloracao e hiologia floral de bromeliaceas ornitofilas de uma area da Mata Atlantica do Sndeste brasileiro. ReMsta Brasileira de Botanica 29: 163- 174. M.ai.k'.ky, H. 1927. New asjiecLs on the association between Ivcaenid larvae (Lycaenidae) and ants (Formicidae, Hymenoptera). Journal of the Lepidopterists’ Society 24: 190-202. Opi.er, P. a., K. Lorrs & T. N.vberh.aus. 2009. Butterflies and moths of North America. Bozeman, MT: Big Sky Institute. Available from: http:/ /www.butterfliesandmoths.org/ (Accessed 23/09/2009). Robbins, R. K. & S. S. Nic:oi..av. 2002. An overview of Stnmoti Hiibner (Lycaenidae, Theclinae, Eumaeini). Journal of the Lepidopterists’ Society 55: 85-100. Rosumek, F. B., M. a. Ui.yssea, B. C. Lopes, J. Sieiner & A. Zii.i.ikens. 2008. Formigas de solo e de bromelias em uma area de .Vlata Atlantica, llha de Santa Catarina, sul do Brasil: Levantamento de especies e novos registros. Biotemas 21: 81-89. S.\.\ipaio, .M. C., L. E. Perisse, G. A. De Oliveira & R. 1. Rios. 2002. The contrasting clonal architecture of two bronieliads from sandy coastal plains in Brazil. Flora 197: 44.3-451. S.azima, M. & I. S.AZiMA. 1999. The perching bird Com’ba flaxml a as a co-pollinator of bromeliad flowers in southeastern Brazil. Canadian Journal of Zoology 77: 47-51. ScHMin, V. S., S. Schmid, J. Steiner & A. Zii.i.ikens. a. Accepted. High diversity of ants foraging on extrafloral nectar of bronieliads in the Atlantic rainforest of southern Brazil. Accepted for publication in Studies on Neotropical Fauna and Environment. Schmid, S., V. S. Sch.mid, B. Harier-Marques, J. Sieiner & A. Zii.i.ikens. b. Submu'Ied. Biniodal pollination system of the bromeliad Aechmea nudicaulis involving hummingbirds and bees. Submitted to Plant Biology. Schmidt, G. & G. Zotz. 2000. Herbivoiy in the epiphyte, Vriesea sanguinolen/a Cogn. & Marchal (Bronieliaceae). Journal of Tropical Ecology 16: 829-839. Urich, F. C. & T. C. Emmel. 1991a. Life history of neotropical butterflies from Trinidad 4. Dynastor macrosiris (Lepidoptera: Nymjihalidae: Bra.ssolinae). Tropical Lepidoptera 2: 141-144. Urich, F. C. & T. C. Emmet. 1991b. Life history of neotropical butterflies from Trinidad 5. Dynastor darius darius (Lepidoptera: Nymphalidae: Brassolinae). Tropical Lepidoptera 2: 14.5-149. ZiK.\N, J. F. 1956. Beitrag zur Biologie von 12 Theclinen-Arten. Dusenia 7: 139-148. Zii.LiKENS, A., J. .Steiner & Z. Mihalko. 2001. Nests of A ugochlora (A.) raoxin bronieliads, a previously unknown site for sweat bees (1 Ivmenoptera: Halictidae). Studies on Netitropical Fauna and Environment 36: 137-142. Zii.i.ikens, A. & J. Steiner. 2004. Nest architecture, life cycle and cleptoparasite of the neotropical leaf-cutting bee Megachile {(ihrysosanis) pseudanthidioides .Moure ( Hymenojitera: Megachilidae). Journal of the Kansas Entomological Society 77: 19.3-202. journal of lirsearch on the I .epidoptera 42; 56-63, 2003 (2010) Temporal and spatial segregation of Battus devilliers and B. polydamm cubemis (Papilionidae) in La Habana, Cuba Ormaily Madrliga Rios* and Alejandro Barro Canamero Departamento de Biologia Aiiinial y Humana, Faciiltad de Biologi'a, Universidad de La Habana, calle 25 # 455 e/ I yj, Vedado, CP: 10 400, Ciudad de La Habana, Cuba. al)airo@fhio. uh. ru Abstract. The spatial and temporal distribution of two syntopic species of the genus Battus (Papilionidae) that inhabit two areas in the north coast of La Habana, Cuba, is analyzed. The study was carried out from April 2006 to March 2007. Samples were taken using transects 100 m long, separated from each other by 80 m gaps. Populations of E. polydamas rubensis and B. devillim of (he evergreen forest at Boca de Canasi, the most natural habitat, had similar abundance, while in the secondary scrub at Boca dejaruco, an extremely degraded area, B. polydamas cubmsis wd.s much more abundant than B. devilliers. Analysis of structural subniche usage by season showed significant differences in daily actirity (from 9:00 am to 12:00 m and from 12:{)() m to 3:00 pm) and habifcit type for each species. Flight stratum had lesser influence than sunlight intensity on both butterflies. Populations of the latter seemed to be influenced by seasonality, while those of if. polydamas cubensis seemed more infhienced by habitat. Keywords: Battus, Cuba, Papilionidae, .segregation, syntopic species. Introduction Ecological roles of butterflies are important for ecosystems functioning, with their study needed to better understand their ecological interactions and functions. In Cuba, only Fontenla (1989) and Genaro et al. (1994) have conducted studies focusing on butterflies communities and their resource partitioning in a particular habitat. Due to their rapid reproduction and close association with specific physical factors and plant resources butterflies are highly sensitive to environmental changes, so they are good indicators of ecosystem health (Brown, 1991; Kremen, 1992; New et al, 1995; New, 1997; Brown & Freitas, 2000). Battus presents a Neotropical distribution (Sims & Shapiro, 1983) with 12 or 14 species (Tyler et at, 1994; Racheli & Pariset, 1992, respectively). The coevolutionary association of Battus with their host plant, Arist(}lochias\i\). (Aristolochiaceae), as well as the mimetic interactions among members of this butterfly genus may be the subject of many studies on butterfly community structure (Young, 1972, 1973). Battus devilliers (Godart, 1823) and B. polydamas ^Current addres.r. Vicedirecdoii Curatofial, Miuseo Nacioiial de Histoi'ia Natural de Guba. Obispo 61, esq. Oficios, Habana Vieja, CP: 10 100, Ciudad de La Habaua. ormaily@ninhtic.inf.cii Received: 21 May 2009 Accepted: 30 July 2009 cubensis (Dufraiie, 1946) are the only members of the genus that inhabit Cuba. The former occurs in Bahamas (Knowles & Smith, 1995) and the latter was reported at Cayman Islands in 1938 (Carpenter & Lewis, 1943), but there has been no recent mention of further records for these islands (Askew, 1980, 1988; Schwartz et al, 1987). On several occasions (Alayoii & Solana, 1989; Racheli & Pariset, 1992; Nunez & Barro, 2003) both species were recorded in Cuba as syntopic {sensu Rivas, 1964) . The interaction becomes more interesting considering the different habitat requirements of B. polydamas and B. devilliers mentioned by Tyler et al (1994), who established that Battus polydamas is very common in disturbed forest, while B. devilliers prefer more natural seasonal forest. Since resource use for both species in Cuba is completely unknown, we undertook a study of temporal and spatial patterns of two populations on the northwestern coast of Cuba. Materials and methods Study areas The study was conducted at two localities of tlie north coast of La Habana. Both areas are dose to human populations and show different levels of degradation. Boca de Jariico is 45 km east of La Habana city, at 23° 1 r N, 82°0r W. The site is typified by secondaiy vegetation with many herbs and bushes and is the most impacted area, which is why the habitat is categorized 42; r>(>63, 2003 (2010) 57 Figure 1 . Study Sites. A: Secondary scrub of Boca de Jaruco. B: Sea grape forest of Boca de Canasi. C: Evergreen forest of Boca de Canasi. D: Summit of the hill where the evergreen forest canopy of Boca de Canasi becomes open. as secondaiy scrub (Fig. la). The unique eniei'gent large trees are isolated Ficus sp. (Moraceae). There are many invasive plants, the most abundant being Acacia famesiana (L.) Willd., (Mimosaceae), Comocladia dentaia ]-AC(\. (Anacardiaceac) and members of the family Poaceae. Common species also includes Lantaua Camara L. (Verbenaceae) and Viguiera dcntata (Cav.) Spreng., (Asteraceae), and vines as Merrcmia dissecta (|acq.) Hallierf., (Convolvulaceae), Cucumis dipsacnis Ehrenb., (Curcubitaceae) and Aristolochia hilahiata L.. During the dry season there are natural burns. Boca de Canasi is 20 km east from Boca de jaruco, at 23°09’ N and 81°47’ W. Two different habitat were analyzed there: a sea grape [Coccoloba uvifera (L.) L. (Polygonaceae)] forest (Fig. lb), and an evergreen forest (Fig. Ic, d). The latter is crossed by several paths, but is less frequented by persons than the former. It also has higher relative humidity due to the cover of arboreal species such as Burscra simaruha (L.) Sarg. (Burseraceae) and Coccoloba divcrsifolia Jacq. that reduce incident sunlight. Endemic plants like Coccolriuax borhidiaua O. Mnhiz (Arecaceae), Lcptocercus lurightii Leon (Cactaceae) and Eugenia molUfoUaXJrh. (Myrtaceae) are also common (Borhidi, 1996). On the summit of the hill (ca. 100 m above sea level) the evergreen forest canopy opens and bushes like Acacia faruesiana and Croton lucidus L. (Euphorbiaceae) are abundant. Ecological counts Boca de Jaruco was sampled from April 2006 to March 2007. At Boca de Canasi sampling started July 2006 in the evergreen forest, while in the sea grape habitat counts began on April 2006, ending on March 2007 and January 2007, respectively. The transect 58 /. Res.Lepid. iiietliod was followed. Individuals were counted only when seen from the sides or the front of observer, never from behind, within a range of approximately 5 m. At least once a month, eight transects were sampled each hour, from 9:00 a.m. to 3:00 p.m., at each studied site. Transect length was 100 m, the transects se|)arated from each other by 80 m. At the end of the study a total of 992 transects in the secondary scrub of Boca dejarnco were covered, 352 in the sea grape and 336 in the evergreen forest. Simlight intensity and flight stratum were recorded for each specimen. These dimensions of structural snbniche were used to analyze part of the spatial segregation. Border values of each category in a particidar dimension were determined following previous Cid^an ecological studies on butterflies (Fontenlak, 1989; Genaro et al, 1994). For sunlight intensity three categories were dehned: sunny (when individual flies directly in sunlight, without any vegetation cover), filtered sun (when individual flies under some vegetation cover, but still in sunlight) and shading (when individual flies where vegetation cover is so dense that it is difficult for sunlight to enter). Three categories of flight stratum from ground to above 3 m were defined: first level from the ground (0 - 1.5 m), second 1.5 - 3.0 m, and third > 3.0 m. Wdien the same individual moved from one level to another only the first stratum seen was recorded. Due to variation of climate throughout the day, the ttvo variables were analyzed separately over a two hours range, 9:00 am -12:00 m, and 12:00 m -3:00 pm. Relative abundance was correlated with mean precipitation of the previous month. This correlation was due to both the influence of rain on vegetation and the time it taikes for plant growth (rains increase flower abundance, for example). We used a correlation between these variables based on published residts showing that nectar source availability is an important condition for a well developed butterfly community (('Jansen et al, 2001). 'Values of mean precipitation per month were supplied by the Instituto Nacional de Recursos Hidraulicos. Statistical analysis (iraphPad InStat, version 3.01 (1998) software was used for data analysis. The Kohnogorov-Smirnov test was the first step in every case to evaluate data normality (p < 0.10). Median and 25 and 75 percentils were calculated. A Mann- Whitney U-test was used to compare the values of the same variable due to the nonparametric natitre of data. Comparison of three or more values was performed with a Kruskal-Wallis test. W'hen the later was significant (p < 0.05), a Dunn’s Multiple Comparisons Post-Test was applied. Correlation between two variables was analyzed with a Nonparametric Correlation (Spearman r). Results and discussion Spatial segregation Habitat use. The greatest difference between populations of both species relates to their abundance in each habitat type. In the three habitats B. polydamas cubensis is common while B. dexnlliers is rare (Fig. 2). Human disturbance of both studied areas may be one of the factors affecting rarity, considering the habitat requirements mentioned by Tyler et al. (1994) . Boca de Jarnco is the most disturbed site and has the highest proportional abundance of B. polydamas cubensis across all the sites sampled. On the other hand, the evergreen forest of Boca de Canasf is the best preserved habitat of the triad and possesses the highest proportion of B. devilliers individuals, even during the dry season (Fig. 3). Concerning habitat requirements, Alayon and Solatia (1989) reported that both species coexist in Cuchillas del Toa, specifically in the ecotone between forest and cleared areas. No statistical significance was found correlating precipitation mean of the previous month and species relative abundance, except in the sea grape forest (Table 1). This may be due this site being mostly a feeding area, since no host plants were found. Thus the presence of butterflies is strongly related with flower availability. Additionally, the presence, in this habitat, of non native plants probably might increase tlie flower availability annually (Nunez & Barro, 2003) , and could be the reason why the rain acts like the primary factor influencing flowering. Although the secondary scrub presents almost the same non native plants as the sea grape forest, there is no correlation of butterfly abtmdance with rain. It is likely that, in this habitat, flower availability is not the primaiy factor Table 1. Values of the Nonparametric Correlation (Spearman r) between relative abundance of the populations (measured as the mean of the major number of individuals in a single count) and precipitation mean of the previous month. Specie.s/I labitat Secondary scrub Evergreen forest Sea grape forest Battus pots'damns 0.3082 0.3531 0.6786 cubensis Battus dndttiers 0..3935 -0.09258 No 42; 56-63, 2003 (2010) 59 2006 2007 I I Battus polydamas cubensis ■■■ Battus devilliers — Precipitations B 45 240 200 160 120 80 40 0 2006 2007 c 2006 2007 I I Battus polydamas cubensis I Battus devilliers —h— Precipitations I I Battus polydamas cubensis Battus devilliers — Precipitations Figure 2. Population densities of Battus devilliers and B. polydamas cubensis, from April, 2006 to March, 2007. A: Secondary scrub of Boca de Jaruco. B: Evergreen forest of Boca de Canasi. C: Sea grape forest of Boca de Canasi. Gray line represents precipitations per month. Black squares represents relative abundance of Battus devilliers and the white ones those of B. polydamas cubensis. The blank spaces were not sampled. Relative abundance was the highest number of individuals observed in one hour. B 100% 80% 60% 40% 20% 0% ■o o S 62, 100% 80% 60% 40% 20% 0% Rain Season Dry Season Rain Season Dry Season Figure 3. Proportional abundance of Battus devilliers and B. polydamas cubensis in the habitats in which they coexist. A: Secondary scrub of Boca de Jaruco. B: Evergreen forest of Boca de Canasi. The black colour represents Battus devilliers and white shows B. polydamas cubensis. influencing the abundance of the.se butterflies. It may be host plant availability, for example. At Boca de Canasi both species coexist only in the evergreen forest (Fig. 2), thus, no statistical analysis was made for the sea grape liabitat. However, nearness of the two sites and the floristic diversity at the latter (Nunez & Barro, 2003) seem to support the idea of delimited feeding and breeding areas for a single population of B. polydamm cubensis. In this manner, the sea grape forest may be the foraging area and the evergreen forest the reproductive site (liost plants observed only in the latter). This behavior was previously described for a Costa Rican population of this species by Young (1972). It may also explain /. Res.Lepid. (■)() the differential abtindance of /j. polydamas cubensis Al each habitat, secondary scrub and evergreen forest. Another consec|nence of the behavior may be the similarity between relative abtindance of both species in the evergreen forest dtiring the rainy season (U = 3.5, p = 0.2469) due to an underestimation in counts of the actual ntimber of individttals in the population of B. polydai/ias cuhensis (Fig. 4). Data for the dry season were instifficient for analysis. By contrast, populations of both species in Boca de Jarnco differed significantly (U = 1.5, p = 0.01 04) during the rainy season and more so dtiring the dry period (U = 0, p = 0.0022). Battus dnrilliers was always rare compared with B. polydamas cuhensis (Fig. 4). However, OMR saw both species in equal proportion during the study period at another site approximately 1 km west of transects and nearest to Jarnco River, where native vegetation was best preserved. Another reason concerning life cycle and reprochictive strategy may contribtite to the success of B. polydamas cuhensis. Gregarionsness of eggs and larval stages may be the most important becatise of the benefits described for the behavior as stated by Stamp (1980), Matsttmoto (1989) and Reader and Hochtili (2003). Gregariotis early stages are absent in B. dmilliers. Structural subniche Sunlight intensity. Large sunny patches are common in Boca de [ariico due to the presence of abundant herbs and bushes and only few isolated large trees. Accordingly, sunny patches were the category most tised in that habitat. The Kruskall-Wallis Test shows differences among the different levels for both s})ecies. The Dunn’s Mtiltiple (ionijjarisons post test demonstrates that in the morning those differences occur between sunny patches and the other levels (p < 0.001 for both species). During the afternoon differences in the tise of three levels were also present. The results of post test were p < 0.001 comparing stmny patches w'ith the other two, although sun filtered and shading patches usage by B. polydamas cuhensis increases (Fig. 5a, b). At Boca de (ianasf, vegetation cover is more highly developed, .so snn filtered and shading patches are more freqtient than at Boca de Jarnco. Neverthele.ss, vahies of Krnskall-Wallis Test show differences among the mentioned levels. Only B. devilliers increased significantly the using of sun filtered sites in the same |)roportion as stmny patches. Statistical differences were between stmny and shading levels (p < 0.05) and between filtered and shading ones (p < 0.01) in the first hours of day, while dtiring the afternoon the same relation occurs, btit without individuals flying in shading patches (Fig. 5c, d). On the other hand, B. polydamas cuhensis always preferred sunny patches. Differences in the morning were between first level (stmny) and the filtered and shading patches (p < 0.001 in both cases). During the afternoon no individuals were observed flying in shading patches (Fig. 5c, cl). Our results show that the tise of this structural dimension depends not only c:m habitat, but on the ecological requirements of species. At Boca de Jarnco, sunny sites ccwer most of the sttidy area and coirsequently both species made major use of them. Nevertheless, in Boca de Canasf, even when vegetation cover increased in the forest, t^nly B. devilliers used filtered patches as well as sunny ones. This tendency matches witli the primary habitat of each species, mentioned by Tyler et al. (1994) , that explains why B. polydamas \s a widespread and flexible species, whereas B. devilliers is sensitive and vtilnerable, dependent on forested areas. Flight stratum In the morning, at Boca de Jarnco both species mostly fly in the two lower strata (Fig. 6a, b). Battus dn)illi(ms c\id nc:)t change this strategy during the entire all day, while B. polydamas cuhensis started to use any strattim indiscriminately after midday. Krtiskall-Wallis valties were significant in the morning for both species. The post hoc test demonstrated differences between the highest strattim and the lowest one for both species (p < 0.001), and between the intermediate and the highest stratum with p < 0.01 for B. devilliers and p < 0.001 for B. polydamas cuhensis. After midday, KW vahies did not show any difference among strata used by either species, although graphically we observed that B. devilliers keeps similar proportions to those of the morning (Fig. 6a, b). At Boca de Canasf, there was no difference in flight stratum use during the day. Bcdtus devilliers mostly flies in the lowest strattim, with patches of filtered sun. On the other hand, B. polydamas cuhensis iWes in all of the three strata. For B. devilliers the post test display differences between the lowest and the intermediate strata with p < 0.01 at both time intervals. Between the lowest and the highest strata we found the same p value for the morning, and p < 0.001 for the afternoon. By contrast, there was no statistical difference in the tise of flight stratum by B. polydamas cuhensis in this habitat (Fig. 6c, d). Flight strata were more indiscriminately used by both species in the two habitats. The restdts demonstrate almost no selection for this structtiral dimension, perhaps because strata are not limiting 42: 5r)-63, 2003 (2010) 61 A 45 T 40- 35- 30- 25- 20- 15- 10- JC. Rain Season Dry Season B Figure 4. Relative abundance (individuals/hour) of Battus devilliers (gray) and B. polydamas cubensis (black) In the two seasons defined in Cuba as Rain (May-October) and Dry (November-April). A: Secondary scrub of Boca de Jaruco. B: Evergreen forest of Boca de Canasi. The graphic represents the median and the 25% and 75% percentils. A B ■ shading □ filtered sun □sunny B. devilliers B. polydamas cubensis B. devilliers B. polydamas cubensis c D 100% 75% 50% 25% ■ 0% ■ shading □ filtered sun □sunny B. devilliers B. polydamas cubensis B. devilliers B. polydamas cubensis Figure 5. Spacial segregation of Battus devilliers and B. polydamas cubensis in the structural subniche, specifically in the light intensity used by each one. A and B: Secondary scrub of Boca de Jaruco (morning and afternoon, respectively). C and D: Evergreen forest of Boca de Canasi (morning and afternoon, respectively). Analysis only for the rainy season. Morning 9;00 am -11 :00 am and afternoon 12:00 pm - 2:00 pm. resources like flowers and host plant availability or because of' low ecological relevance for both species. Similar results were obtained by Fontenla (1989) in a larger butterfly community. Temporal segregation Seasonality. Another relevant divergence in population ecology is the seasonal pattern of B. devilliers (U = 3.5, p = 0.0425) by contrast to the continuous dynamics of the other species (U = 10.5, p = 0.2620) (Fig. 2a, b; Fig. 4). This strategy and its presence in all sampled habitats show the greater ecological plasticity of B. polydamas cubensis. It is relevant that absence of B. devilliers in the sea grape habitat is recent, as five years ago it was observed by Nunez and Barro (2003) and again in lower frequency than B. polydamas cubensis. Annual fluctuations of popnlations, like those /. Res.Lepid. (i2 A 100% ■ 3,0m 75% □ l,5-3,0m 50% □ 0-1, 5m 25% 0% B. devilliers B. polydamas cubensis B. devilliers B. polydamas cubensis B. devilliers B. polydamas cubensis Figure 6. Spacial segregation of Battus devilliers and B. polydamas cubensis in the structural subniche, specifically in the flight height or stratum used by each. A and B: Secondary scrub of Boca de Jaruco (morning and afternoon, respectively). C: Evergreen forest of Boca de Canasi (there were not found significant differences in both day-hours). It was analized only rain season. Morning 9;00 am - 11:00 am and afternoon 12:00 pm -2:00 pm. Battus polydamas cubensis Battus devilliers Figure 7. Daily activity of Battus devilliers (gray line) and B. polydamas cubensis (black line) for the rain season in the two studied habitats. A: Evergreen forest of Boca de Canasi. B: Secondary scrub of Boca de Jaruco. obsen'ed in pa.st years in Boca clejarnco by the second author or changes in habitat structure may influence population dynamics. Daily activity. We analyzed the daily activity for each species and habitat during the rainy season (Fig. 7). Although no statistical analyses were made with these data, two different trends were observed in each population. While the abundance of B. polydamas cubensis decreases during the day, with a peak of activity at 10:00 am, abundance of B. devilliers tends to increase until midday with the maximum number of individuals at noon. Thus at 12:00 the means of both pojtulations were similar in the evergreen forest. At the same time, in the secondary scrub of Boca de Jaruco, although the mean values of abundance were so different, the standard tleviations were too overlapped. This behavior is not surprising since B. polydamas cubensis flies in sunny patches more than 42: 5f>63, 2003 (2010) 03 B. devilliers that prefers sliade and is less influenced by high temperature of noon. This mechanism may affect thermoregulation permitting the species a longer diurnal flight period. Battus deiulliers and B. polydamas cubensis present more selectivity for the dimension sunlight intensity of tlie structural subniche than for the flight stratum since the late was more indiscriminate used. Temporal segregation seems to occur between these species, B. dniilUersx^ influenced by seasonality wliile B. polydamas cubensis is present all tlie year. Acknowledgements We thank especially the persons who helped with the field work: Rayner Nunez, Alejandro Garcia, Joel Lastra, Elier Fonseca and Rene Marzuk. We also are gratel'ul to Ariam jinienez. who was ol great help w’ith the statistical analysis. Our gratitude to Gilberto Silva, Rayner Nunez and Oraily Madrtiga tor their helpful conunents about the manuscript and also our thanks tor the botanical assistance to Iralis Ventosa and Antonio l.opez. Three anonymous reviewers contributed to clarifying the submitted manuscri|5t. Literature cited Ai.ayon, G. & E. Sor.vNA. 1989. Lista de mariposas diurnas (Lepidoptera: Rhopalocera) colectadas en la Reserva de la IMostera Guchillas del Toa (Holgtiin- Guantanamo), Cuba. Garciana 7: 2-4. Askkw. R. R. 1980. The butterfly (Lejtidoptera, Rhopalocera) fauna otthe Cayman Islands. Atoll Research Bulletin 241: 121-138. Askf.w, R. R. 1988. Butterflies ot Grand Cayman, a dynamic island fauna. Journal ot Natural Histoiy 22: 87,'i-888. Borhidi, a. 1996. Phytogeogra]thy and vegetation ecology' of Cuba. Akademiai Kiado. Budajtest. 923 jtp. Brow'n, K. S. Jr. 1991. Conservation of Neotrojtical environments: insects as indicators, |rp. 349-404. In: N. .\1. Collins and J. A. Thomas (eds.), The conservation of insects and their habitats. London, Royal Entomological Society Symposium XV; Academic Press. Brown, K. S. Jr. & A. V. L. Frkit.as. 2000. Atlantic forest butterflies: indicators for landscape conservation. Biotropica 32 (4b): 934-956. Carpenter, G. D. H. & Lewis, C. B. 1943. A collection of Lepidoptera (Rhopalocera) from the Cayman Islands. Annals of the Carnegie Museum 9: 371-396. Ci.ausen, 11. D., H. B. Hoi.beck & J. Rei)der.sen. 2001. Factors influencing abundance of butterflies and moths in the uncultivated habitats of an organic farm in Denmark. Biological Conservation 98: 167-178. Ft)NTENLA, J. L. 1989. Particion de recursos en una comunidad de mariposas (Lepidoptera: Rhopalocera). Poeyana 385: 1-26. Genaro, J. A., A. C. Sanchez. & V. Berovides. 1994. Solapamiento del nicho ecologico de dos especies del genero Anarlia (Le|>idoptera: Nymphalidae). Revista Biologi'a 8: 141-144. Gr.ai’IiPad InStat. 1998. InStat Prism, version 3.01 from GraphPad Prism (http://w'ww.graphpad.com). (Active in June, 2007) Knowles, D. O. & D, S. Smi ih. 1995. First records oi Parachomnthus magdalia (l lesperiidae) from the Bahamas, and extension of the baliamian range oi Battus dnnlliers (Papilionidae). Journal of the Lepido|jterist.s Society 49 ( 1 ): 91-94. Kremen, C. 1992. Assessing the indicator properties of species assemblages for natural areas monitoring. Ecological Applications 2: 203-217. Matsumoto, K. 1989. Effects of aggregation on the survival and development on different host plants in a papilionid butterdy, Luehdorfia japonica Leech. Japan Journal of Entomology 57 (4): 853-860. New, r. R. 1997. Are Lepidoptera an effective “umbrella group” for biodiversity conservation? Journal of Insect Conservation I: .5-12. New, T. R., R. .VI. Pvi.e.J. A. Thom.as, C. D. Thomas, & P. G. Ha.mmond. 1995. Butterfly conservation and management. Annual Review of Ecology and Systematics 40: 56-83. Nunez, R. & A. Barro. 2003. Composicion y estructura de dos comunidadesdemariposiLs (Lepidoptera: Papilionoidea) en Boca de Canasi, La Ilabana, Cuba. Retista Biologia 17 (1): 8-17. Racheli, T. & L. P.ARisET. 1992. II genere Battus. Tassonomia e Storia Naturale. Fragmenta Entomologica 23 (Su|rlemento). 163 pp. Reader, T. & D. F. Hochuli. 2003. Understanding gregariousness in a larval Lejridopterau: the role of host |rlant. jrredation and microclimate. Ecological Entomology' 28: 729-737. Rivas, L. R. 1964. A reinterpretation ol the concepts “simpalric” and “allo|jatric”yvith propo,sal of the additional terms “syntopic” and “allotopic.” Systematic Zoology 13 (1): 42-43. Sims, S. R. & A. .M. Shapiro. 1983. Pupal dia|5ause in Battus phiknor (Lejtidoptera: Pa|)ilionidae). Annals of the fintomological Society' of America 76 (3): 407-412. Sciiyv.yRTZ, A., F. L. Gonzalez, & R. M. Hender.son. 1987. New records of butterflies from the West Indies. )ournal of the Le]3ido|3terist.s' Society 41 (3): 14.5-450. Stamp, N. E. 1980. Egg deposition patterns in butterflies: yvhy do some species cluster their eggs rather than de|)osit them singly? American Naturalist 115 (3): 367-380. Tvt.ER, H. A., K. S. Brown Jr. & K. H. Wtt.soN. 1994. Syvallowtail butterflies of the Americas. A study iti biological dynamics, ecological diversity, biosystematics and consen'ation. Scientific Publishers. 376 pp. Youno, a. M. 1972. Mimetic associations in populations of tropical butterflies. II .Vlimetic interactions of Battus polydamas atid Battus b(dlus. Biotropica 4(1): 1 7-27. Youno, A. M. 1973. Notes on the life cycle atid natural history of Parides areas mylotes (Papilionidae) in a Gosta Rican premontane wet forest. P.syche 80: 1-21. /oitrnal of Research on the Lepidoptera 42: 64-73, 2()()3 (2010) A tale of two species: detritivory, parapatry, and sexual dimorphism in Lamprospilus collucia and L, orcidia (Lycaenidae: Theclinae: Eumaeini) Roberi K. Robbins', Annexe Aiello-, Julie Feinstein'\ Amy Berkov'\ Astrid Caldas*, Robert C. Busby"’ AND MaROELO DliARTE" 'DejDartmciit of Entomolog)'. I’O Box 37012, NllB SU)p 105, Smithsonian Institution, Washington, DC 20013-7012, USA robhinsr@si.edu -Smithsonian 'IVo]>ical Research Institute, Smithsonian Institution, A|3artado 0843-03092 Balboa, Ancon, Panama aieUoa@si.edu ■’Department of Biology, City College of New York, City University of New York, Convent Avenue at 138th Street, New York, NY, 10031, USA jfstein@aninh.org, herhov@.sri. rrny.euny.edu 'De|5ariment of Entomology, PO Box 37012, NHB Stop 105, Smithsonian Institution, Washington, DC 20013-7012 USA ast rid ra ! das @yah oo.ro m ■’’7 Countryside Way, Andover, MA 01810-6041 USA rr. busby@romrast. net '’Coleyao de Lepidoptera, Museu de Zoologia, Lhiiversidade de Silo Paulo, Av. Nazare 481, 04263-000 Sao Paulo, SP Brasil mduartes@usp.br Abstract. Lamprosj)ilus rolluria (I lewitson) and L. orridia (Hewitson) are facultatively detritivorous hairstreaks. Eeinales in nature lay eggs on dead twigs and leaves that are on or near the ground. In the lab, females oviposit readily on dead leaves. Caterpillars of both species eat dead plant material in nature and can be reared in the lab to the adult stage on artificial diet to which no plant material has been added. Lamprospilus rolluria and L. orridia have parapatric distributions; the former species is endemic to the Transandean Region and the latter to the Amazonian and Atlantic Regions. Both species have similar male behavior, which is consistent with the hypothesis that their parapatric distributions are maintained by mating interference. The sexes of L. rolluria and L. orridia have been incorrectly associated in compendia of Neotropical butterllies and are associated in this paper by geographic distribution, wing pattern similarity, and rearing data. Although L. rolluria and L. orridia have been considered to be conspecific, an analysis of geogra|rhical variation supports the hypothesis that they are distinct biological species. Key words: Amazonian Region, biogeography, hairstreak systematics, Lecythidaceae, Transandean Region. Introduction Lantpw.spilits collucia (Hewitson) and L. orcidia (Hewitson) are common and widespread lowland Neotropical lycaenids (Theclinae: Eumaeini) that are biologically significant for a ntimber of reasons. First, L. collucia and T. orcidia are ecologically tmusual. LaiYal detritivory occurs rarely in the “Macrolepidoptera” (Powell e! al., 1998; Holm & W’agner, 2002), but has been reported in Lamprospilus Geyer (Duarte & Robbins, in press), specifically in L. collucia and L. orcidia. Second, L. collucia and L. orcidia are biogeographically significant because they have been cited as a representative parapatric species |)air with a Gentral/South American distribution (Robbins, 2004a). This biogeographic pattern, while well-known in forest-dwelling aposematic butterfly laxa (Brown, 1982), has not been doctimented Rerewed: 15 September 2009 Accepted: 12 October 2009 previously in the Eumaeini. Third, L. collucia and L. orcidia are of taxonomic interest because they are widely misidentihed in publications on Neotropical btitterflies (e.g., Godman & Salvin, 1887; Weeks 1911; Draudt, 1919-1920; Kitye, 1921; Barcant, 1970; Robbins & Small, 1981; D’Abrera, 1995). Both species are sexually dimorphic. The males have similar wing patterns (Figs. 1-4, 13-16, 21-22), forwhich rea.son they have been considered to be conspecihe (Godman & Salvin, 1887-1901; Kitye, 1921). Alternately, the female wing patterns (Figs. 5-12, 17-20, 23) are different from each other and from those of the males, with which they have rarely been associated (e.g., Dratidt, 1919-1920). The purpose of this paper is to address the ecology, biogeography, and taxonomy of L. collucia and L. orcidia by answering basic questions abotit them. Where do females oviposit? What do their caterpillars eat? When and where do males set tip mating territories? What are the distributions of L. collucia and L. orcidia? In which habitats do they occur? How are T. collucia and L. orcidia distinguished? How 42: 64-73, 2003 (2010) 65 do they vary seasonally and geographically? On what basis are the sexes associated? What is the available evidence that they are different biological species? We add brief notes on nomenclature to confirm that we are using the correct names. The placement of/.. coUuciam\d L. orcidiam /.rtw/;ros/>//«5Geyer is dealt with elsewhere (Duarte & Robbins, in press). Materials and methods Eggs were obtained in the lab following the methods detailed in Duarte et al. (2005) . Larvae from these eggs were reared on artificial diet to which no vascular plant material was added other than wheat germ and linseed oil (Duarte et al., 2005). Rearing methods for immatures collected in nature generally follow Feinstein et al. (2007). Depositories for vouchers are noted. Biogeographic and taxonomic results for L. collucia are based on 96 males (6 genitalic dissections from Mexico, Panama, western Ecuador, and eastern Colombia) and 70 females (6 genitalic dissections from Mexico, Costa Rica, Panama, and Trinidad). Analogous results for L. orridia are based on 50 males (6 genitalic dissections from Ecuador, Peru, and 3 states in Brazil) and 57 females (6 genitalic dissections from Peru and Brazil). We map the distributions of each species by sex because these distributions are evidence for associating the sexes. Although mitochondrial “barcodes” are reported for L. collucia and L. orridia (BOLD website, http://w\\'w.barcodinglife.org/\’iews/ login. php, accessed 26 Aug 2009), the barcodes are not publically available and the “barcoded specimen” of L. orcidia is misidentified. Genitalic terms follow Klots (1970), as modified for theEumaeini (Robbins, 1991). Wing venation follows Comstock (1918), and other morphological terms follow Snodgrass (1935). Geographic distributions are mapped by gender. Months are abbreviated by their first three letters in English. Vouchers for the distribution maps and other results are deposited in the following collections: (AA) Annette Aiello Collection, Ancon, Panama; (BMNH) Natural History Museum, London, UK; (DZUP) Universidade Federal do Parana, Curitiba, Parana, Brazil; (MCZ) Museum of Comparative Zoology, Harvard University, Cambridge MA, USA; (MECN) Museo Ecuatoriano de Ciencias Naturales, Quito, Ecuador; (MUSM) Museo de Historia Natural, Universidad Nacional Mayor de San Marcos, Lima, Peru; (MZUSP) Museu de Zoologia, Universidade de Sao Paulo, Brazil; (RCB) Robert C. Busby Collection, Andover, MA, USA; (USNM) National Museum of Natural History, Smithsonian Institution, Washington, DC, USA. Results Ecology and biogeography Oviposition and food “plants.” As part of a study of plant fungal diseases (Davidson et al., 2000), a “mostly dead” seedling of Anacardium excelsurn (wild cashew, Anacardiaceae) was collected by Davidson about 9 Jun 1996 near the Rio Frijoles, Pipeline Road, Canal Area, Panama (see Ridgely, 1976 for information on the Pipeline Road locality). A dark reddish brown larva of L. collucia was found three days later eating the cotyledon of the dead seedling. Aiello fed the caterpillar the peduncle of Anacardium occidentale to complete its development. On 18 Jun 1996, a dark brown pupa with erect setae on the sides of the abdomen was formed. The pupa turned black on 1 Jul 1996, and a male of L. collucia emerged later that day. The reared adult male is deposited in AA (Aiello lot: 1996-10). A female of L. collucia was collected by Robbins and Caldas on 30 Mar 2000 in Ancon, Canal Area, Panama. She laid 22 eggs over 6 days in the lab on dead leaves and on the side of a \'ial. Aiello reared the hatched larvae on artificial diet without any added plant material. A female emerged on 13 May 2000. The reared female and her mother are de[)osited in USNM. Robbins and Caldas observed a female of L. collucia ovipositing on a twig on the ground in Ancon on 31 Mar 2000. After caj)ture, the female butterfly laid another 44 eggs over the next 5 days on dead leaves in the lab. Aiello reared the resulting caterpillars on artificial diet without any added plant material, and three males and one female emerged 14-16 May 2000 (Figs. 1,5). The mother and her reared offspring are deposited in USNM. Robbins and Caldas observed a female of L. collucia ovipositing on a green leaf about 10 cm from the ground on 2 Apr 2000 (Fig. 9) . She was not captured and the egg was not collected. A male of L. orcidia was reared from the fallen androecia of Esclnoeilera coriacea (Lecythidaceae, plant vouchers deposited in New York Botanical Garden) from lowland moist forest 7 km north of Saiil, French Guiana (3°37’ N, 53° 12’ W). The androecia were collected by Berkov 21 Oct 1995 in the diy season, and the adult male of L. orcidia emerged 8 Nov 1995 (voucher deposited in USNM, Fig. 1 6) . Another eight Lycaenidae that belong to another genus were also reared from these androecia (Feinstein, Robbins, & Berkov, in prep.). /. Res.Lepid. (i6 Figures 1-12. Lamprospilus coilucia adults. 1. c? ventral, Panama, reared, male sibling of 5. 2. ventral, Panama, form typically seen in the wet season. 3. (J ventral, Panama, form typically seen in the dry season. 4. dorsal of 1 . 5. $ ventral, Panama, reared, female sibling of 1 . 6. $ ventral, Panama, form typically seen in the wet season. 7. $ ventral, Nicaragua, form typically seen in the dry season. 8. $ ventral, no locality, reproduction of figure from the original description. 9. $ ventral, Panama, female walking on a twig near the ground before laying an egg. 10. $ dorsal, Venezuela. 11. 5 dorsal, Venezuela. 12. $ dorsal, no locality, reproduction of figure from the original description. A male and two females (Fig. 20) of L. orcidia (identified as Lycaenidae #2 in Feinstein el at 2007) were reared from larvae found in fallen androecia of L('cylhis corrugata (I.ecythidaceae) during the first three months of 2003 in the wet season at Les Noiiragues Re.search Station in French Guiana (4°05’ N, 52°4rW; IK) km south of Gayenne). Although 17 other Lycaenidae were reared, none belong to Lamprospilus. Reared adult vouchers are deposited in USNM. Male behavior. Lamprospilus males display “territorial” behavior that is similar to that reported in other eiimaeines (e.g., Alcock & O’Neill, 1986; 1987); males wait for receptive females to fly through the territory and “defend” these areas by flying at other males that enter tlie territory. Males of L. colluda and L. orcidia set up mating territories in the morning on hilltops (vouchers below are deposited in USNM, observations are by Robbins; times are standard time 42: 64-73, 2003 (2010) 67 1 cm Figures 13-23. Lamprospilus orcidia aduWs. 13. c? ventral, Brazil (Para), presumed holotype. 14. ventral, Peru. 15. (S ventral, Peru. 16. c? ventral, French Guiana, reared from fallen androecia of Lecythidaceae. 17. $ ventral, Peru, arrow points to brown scales basal of the postmedian line. 18. $ ventral, Peru. 19. $ ventral, Peru. 20. $ ventral, French Guiana, reared from fallen androecia of Lecythidaceae. 21. (5* dorsal, Brazil (Para), presumed holotype. 22. (5' dorsal, Peru. 23. $ dorsal, Peru. at that locality). Lamprospihis collucia in Panama, 0730-1045 hours 4 S observed (2 vouchers), 5 Oct 1978, 0730-0745 hours, Canal Area, Paraiso, Cerro Paraiso. 5 S (5 vouchers), 1 Jan 1979, 1000-1030 hours. Canal Area, Paraiso, Cerro Paraiso. 1(5' (1 voucher), 5 Mar 1979, 1045 hours. Canal Area, Paraiso, Cerro Paraiso. >25(5' observed (2 vouchers), 17 May 1979, 0830- 1030 hours, Canal Area, Cerro Calera. \S (1 voucher), 28 Jiil 1979, 1000 hours, Canal Ai ea, Paraiso, Cerro Paraiso. Lamprospihis orcidia in Brazil, 0904-0920 hours 68 /. Res.Lepid. 1(5' (1 v(3ucher), 18 Mar 1991, 0904 hours, Sao Paulo, 17 km west of Teodoro Sampaio. 1(5 (1 voucher), 24 May 1998, 0920 hours, Rio de Janeiro, Iguaba Grande. Habitat. Lmnprospilus collucia and L. orcidia occur in wet and dry lowland forest, ranging from “relatively virgin” forest (e.g., Parque Manu, Peru) to mature secondary forest (e.g., Gamboa, Canal Ai'ea, Panama) to patchy disturbed forest in urban areas (e.g.. Ancon, Canal Area, Panama) . We have seen no specimens of L. coZ/Mcm collected above 1,000 m elevation in Central America, but in western Ecuador they have been found in wet forest at 1,500 m and on a ridge with dry forest at 2,100 m where there is often a strong westerly wind. Most individuals of L. orcidia m'e recorded from lowland forest, but some have been recorded from 1,000 m elevation in southern Brazil. Adults of L. collucia and L. orcidia are most abundant at the end of the dry season and beginning of the wet season in Panama and southeastern Peru, a pattern typical of the I Mniprospilus ^eciion (Duarte & Robbins, in press). Distribution. Males of L. collucia are recorded from northeastern Mexico to Ecuador west of the Andes and to Trinidad, northern Venezuela, and central Colombia east of the Andes (circles in Fig. 24) while males of L. orcidia are known east of the Andes from central Venezuela to southern Brazil and Bolivia (squares in Fig. 24). Males of L. colliiciamd L. orcidia are not sympatric. Females of L. collucia'dre recorded from northeastern Mexico to the northwestern tip of Peru west of the Andes and to Trinidad, and central Venezuela east of the Andes (circles in Fig. 25) while females of L. orcidia are known from east of the Andes from the Guianas and southern Venezuela and southern Colombia to southern Brazil (squares in Fig. 25). Females of L. collucia 'And L. orcidia Ate not sympatric. There is one ntale of L. orcidia and one female of L. collucia from the Rio Suapure, Venezuela (MCZ), a tributary of the Rio Orinoco in central Venezuela (Bolivar state) that flows through llanos (savannah) and Amazonian forest habitats (arrows in Figs. 24-25) . These specimens lack collection date or more specific locality data. Weeks (1911) noted only that they were collected in “the neighborhood of the Suapure River in Venezuela.” It is unknown if both were collected at the same locality along the Rio Suapure, but if so, it is the only locality where both species have been found. The female from Rio Suapure was listed and illustrated as Theda madie Weeks, but the male was apparently misidentilied as Theda xenata (a misspelling of Theda xeneta Hewitson, see taxonomy section below) (Weeks, 1911). There are no males of Calycopis xeneta from the Rio Suapure in the Weeks Collection (MCZ). Taxonomy Distinguishing male characters. Location of the charcoal-black patch on the ventral forewing is the most consistent and easy way to distinguish males of L. collucia and L. orcidia (Figs. 1-3, 13-16). In L. collucia, this patch is distal of the postmedian line whereas in L. orcidia, it is distal and basal with the basal part darker in some individuals. We have not seen a male with an intermediate wing pattern. Godman and Salvin (1887-1901) and Kaye (1921) apparently considered this difference to be intraspecific variation, but Comstock and Huntington (1962) noted that the two wing patterns were distinct. The ventral wing patterns of these males are similar to those of other species with charcoal-black patches. Males of some other Lamprospilus species, such as L. coelicolor (Butler & Druce) and L. aunus (Cramer), are easily distinguished by the better defined and more triangular shape of the dark brown patch on the ventral forewing (Fig. 42 in Duarte 8c Robbins, in press). Males of Calycopis xeneta (Hewitson) have a brown spot in ventral hindwing cell Cip-2Ajust distal of the postmedian line (Fig. 58 in Duarte & Robbins, in press) that is lacking in Lamprospilus. Variation of male wing pattern. Wing pattern variation in male L. collucia is most evident on the ventral wings (Figs. 1-3). The width and exact shape of the postmedian line on both wings is perhaps the most variable element. The darkness and extent of the charcoal-black patches is also variable. Those individuals with a ventral wing pattern which is a bit lighter than average (Fig. 3) are more prevalent in the dry season, but we find no evidence for geographical variation. Wing pattern variation in male L. orcidia is also most evident on the ventral wings (Figs. 13-16). Again, the shape of the postmediaii line and the extent and darkness of the charcoal-black patches on both wings are the most variable elements. We do not have sufficient data to assess seasonal wing pattern variation, but find no evidence of geographical variation. Distinguishing female characters. Shape and color of the ventral forewing postmedian line is the most consistent way to distinguish females of L. collucia and L. orcidia. This line is relatively thick and reddisli to dark maroon in L. collucia (Figs. 5-9) and is a relatively thin black and white line with diffuse liglit brown scaling basally in L. orcidia (Figs. 17-20, arrow points to brown scaling). The ventral wing pattern of female L. collucia could be confused with that of female L. lanckena (Schaus), 42: 64-73, 2()()3 (2010) 69 Figures 24-25. Distribution of L. coHucia (circles) and L. orcidia (squares). Arrows point to possible sympatry on the Rio Suapure (Venezuela). The shaded area is an extremely close approximation to the Transandean Region of Brown (1982: 456); this area of endemism was proposed without exact borders. 24. Males. 25. Females. but the later has the ventral forevving postmedian line of L. collucia in cell Car,-2A and more rounded hinchvings. The black and white forewing postmedian line with brown basal scaling is the best way to distinguish L. orcidia from other hairstreak species, hut this character is sometimes inconspicuous (Fig. 19). Even with genitalic dissection, some females of L. orcidia may he difficult to identify definitively. Variation of female wing pattern. Wing pattern in female L. collucia is quite variable. Dorsal ground color has variable amounts of blue scaling (Figs. 10- 11), which varies in hue from shining blue to chalky gray. The ventral brownish-black patch of scales distal of the postmeclian line varies from absent (Fig. 7) to conspicuous (Figs. 5-6). The color of the ventral postmedian line varies from reddish to dark maroon, but the thick forewing line from the costa to vein Cu,, is a constant feature. As in the male, intlividuals with a lighter ventral wing pattern (Fig. 7) tend to be most frequent in the dry season. Wing pattern variation in female L. orcidia is similarly variable. Dorsal ground color varies from blue to chalky gray. The ventral wing pattern is rather “non-descript”, but the black and white postmedian line with basal brownish scaling appears to be constant, even if its expression is variable (Figs. 17-20). Male genitalia and their variation. The male genitalia of L. collucia and L. orcidia are typical of Lamprospilus (Duarte & Robbins, in press) with a single medium sized tooth on each gnathos arm (arrow in Fig. 26). The only evident genitalic difference between the two species is that the penis of L. orcidia consistently has a small second cornutus (arrow in Fig. 27) while that of L. collucia may or may not (Fig. 26) have the second cornutus. Otherwise, the illustrated differences in the saccus, penis, and valvae (Figs. 26- 27) fall within the range of intraspecific variation. Female genitalia and their variation. The female genitalia of/., collucia me typical of Lamprospilus Wnh “fan-shaped” signa (Figs. 28-29) and an inwardly curved sclerotized ridge on the distal end of the 8th abdominal tergum (illustrated in Duarte & Robbins, in press). The shape of the ductus bursae, especially the posterior end, varies intras(}ecifically, but does not distinguish the species. The signa of the two species differ in the sample (Figs. 28-29), but are structurally quite variable, for which reason we suspect that this difference might not be confirmed by a larger sample size. Nomenclature. Theda collucia Hewitson was 70 /. lies.Lepid. Figures 26-27. Male genitalia, ventral aspect (left), lateral aspect (right), penis in lateral aspect (bottom). 26. L collucia, Panama (Canal Area), arrow points to single “tooth” on the gnathos. 27. L. orcidia, Brazil (Minas Gerais), arrow points to small second terminal cornutus. Scale 1 mm. described from at least one pair in the Hewitson Collection (now in BMNH), but only the female was illustrated (Figs. 8, 12). No type locality was given. Johnson (1993: 22) designated a female lectotype (B.M. Type Rh 1010) from Esp. Santo (presumably Espirito Santo, Brazil) that fits the original illustration very well even though it is missing most of its forewings. However, there was no type locality in the original description, and a photograph of the lectotype and its labels from the 1970s shows that this specimen lacked a locality label at that time. Johnson (1993: 22) did not list Brazil as part of the South American distribution of collucia, so his “Esp. Santo” citation is difficult to interpret. Primary types of Theda collucia's Junior synonyms have been examined: madie Weeks ($, MCZ), am/Wradc Schaus ($, USNM, the original description erroneously listed the BMNH), iodiniis Kaye {S, BMNH), Dyar ($, USNM), and shueyi Johnson ((5',AMNH). lliecla orcidia Hewitson was described from at least one male in the H. W. Bates collection (now in BMNH) from the Amazon. There is one male in the BMNH that fits this description (Figs. 13, 21, B.M. Type Rh 872) and is presumed to be a liolotype. Illustrations of the holotypes (by original designation) of junior .synonyms to/i'cm'w Johnson (AMNH), Johnson & Kroenlein (BMNH), .vi/uw Austin & K. Johnson (I)ZUP), ron doni a Xmiin & K. Johnson (DZUP), ofaarra Austin & K. Johnson (DZUP), j&cr/ifexa Austin Sc K. Johnson (DZUP) , and purpura Kiisiin Sc K. Johnson (DZUP) can be found in the original descriptions (Johnson, 1993; Johnson & Ki'oenleiii, 1993; Austin & Johnson, 1997). The wing pattern of female L. orcidia is non¬ descript, as noted. Perhaps for that reason, a female of L. orcidia was included in the type series of the unrelated Theda ceromiaHeMntson. However, Johnson and Kroenlein (1993:4) designated another specimen as the lectotype, which is the reason that Theda ceromia is now placed in Ziegleria (Robbins, 2004b; Duarte Sc Robbins, in press). Discussion Detritivory, Females of L. collucia have been recorded in nature ovipositing on dead twigs on the ground and on a leaf near the ground. In the lab, females oviposit readily on dead leaves. Caterpillars of L. collucia and L. orcidia in nature have been found eating a “nearly” dead seedling and the androecia of Lecythidaceae flowers on the ground. In the lab, larvae complete development on live and dead organic matter. Although many butterflies, including Lycaenidae, can be reared on an artificial diet to which dried, ground leaves of the food plant are added (Morton, 1981; Mark, 1993; 1995), larvae of 42: 64-73, 2003 (2010) 71 Figures 28"29. Female genitalia, dorsal (left) and lateral aspects. 28. L. collucia, Panama (Canal Area). 29. L. orcidia, Peru (Madre de Dios). Scale 1 mm. L. collucia and L. orcicUa readily ate and completed development on an agar-ba.sed artificial diet without the addition of'leaves. These results are very similar to those reported for Calycopis (S. Johnson, 1985; Rohhins et al., 199(i; Duarte et ah, 2005), and are consistent with the hypothesis that L. collucia and L. orcidia are facultative detritivores. Different kinds of detritus provide different kinds of nutrition for a caterpillar. A preliminaiy analysis of some Lecythidaceae androecia showed that they have higher sugar and phosphorus content than “leaf litter” (nitrogen levels were variable), but a lower content of other minerals and fiber (A. Wliigham pers. comm.). Detritivores may also eat micro-organisms living on detritus (Findlay & Tenore, 1982; Hohn 8c Wagner, 2002) , but to date, the ntitrition that lycaenid detritivorous caterpillars derive from different food objects is an unexplored subject. Maximal adult abundance of L. collucia and L. orcidia at the end of the dry season and beginning of the wet season suggests that larvae find more suitable food or suffer lower mortality during the dry season. Many trees are deciduous during the dry season, hut whether fungi and other caterpillar pathogens and predators are less abundant at that time is an open question. Parapatry. Brown (1982) partitioned the distribution of Neotropical forest butterflies into four slightly overlajjping “fuzzy-edged” biogeographic regions of endemism, three of which (Transandcan, Amazonian, and Atlantic) consist primarily of areas under 1,500 m elevation. The biogeographic distribution of L. collucia is a “textbook” example of Brown’s Transandean Region; this species occupies virtually the entire Transandean Region (shaded part of Figs. 24-25). The distribution of L. orcidia is a combination of Brown’s Amazonian and Atlantic Regions. So far as we are aware, this is the first time that a clear-cut Transandean/Amazonian parapatric distribution has been documented in the Eumaeini. In most other potential cases, such as Lamasina drauclti (Lathy) and L. ganimedes (Cramer) (Robbins & Lamas, 72 J. Res.Lepid. 2008), species are not sufficiently well-represented in ninsenin collections to determine whether distiibntions are allopatric or parapatric. The parapatric distributions of L. collucia and L. orcidia (Figs. 24-25) are unlikely to be maintained by competition for larval food; it is difficult to visualize the dead organic matter that the caterpillars eat as a limiting resource. However, males of both species set up mating territories in the morning on hilltops and occur in similar habitats. These findings suggest the testable hypothesis that mating interference is responsible for maintaining parapatry between the two species. Associating males and females. The evidence that males and females of L. coZ/wr/a are correctly associated is that the distribution of each sex is almost identical (Figs. 24-25), both sexes have a dark brown patch on the ventral forewing distal of the postniedian line (Figs. 1-3, 5-9) , and both sexes have been reared from eggs laid by the same mother (Figs. 1,5). The evidence that the male and female of L. orcidia are the same species is that the distribution of males and females is almost identical (Figs. 24-25), both sexes have darker scales (albeit, much reduced in the female) basal of the ventral forewing postniedian line (Figs. 13-20), and both have been reared from fallen flowers of Lecythidaceae (Figs. 16, 20; no other Lamprospilus species were reared from these flowers). Finally, no other “unassociated” Lamprospilus male or female has the same distribution as either species. Biological species. With the possible exception of the old Rio Siiapure specimens mentioned above from Weeks ( 1 9 1 1 ) , the distributions of L. collucia and L. orcidia 'Ave parapatric (Fig. 24-25). Distinguishing characters are consistent throughout the range of each species and do not vary in the areas where the distributions meet. This evidence is consistent with the hypothesis that the two taxa do not interbreed. Lamprospilus collucia and L. orcidia are likely to be phylogenetic sisters. In a phylogenetic analysis intended to determine relations among the genera of the '‘'Lamprospilus Section” (Duarte & Robbins, in press), the morphological character coding for L. collucia and L. orcidia was identical. However, the coding was also very similar to that for L. coelicolor And L. aivnus. For this reason, an analysis of phylogenetically informative characters among Lamprospilus species is needed to test whether L. collucia and L. orcidia are indeed sister species. Acknowledgements We are grateful to Karie Darrow for preparing the plates of adults, to Vichai Malikul for preparing the genitalic illustrations, to Brian Harris for help with technical aspects of this project, to Gerardo Lamas for providing data on L. collucia in Peru, and to Amie Wliigham for allowing us to cite lier preliminary iinpublislied information. For supporting travel by both Duarte and Robbins and for substantial funds, we thank Fundagao de Amparo a Pe.squisa do Estado de Sao Paulo/FAPESP (as part of the project “Systematics, Bionomy, and Evolution of Neotropical Lepidoptera”; process numbers 02/ 1 3898-0 and 03/ 1 3985-3) and Pro-Reitoria de Pesqitisa da Universidade de Sao Paulo/USP/Projeto 1. For excellent suggestions on the manuscript, we thank Paul Goldstein, Andre Freitas, and Rudi Mattoni. Literature cited AlcockJ. & K. M. O’Neu.l. 1986. Density-dependent mating tactics in the gray hairstreak, Strymon melinus (Lepidoptera, Lycaenidae). Journal of Zoology 209: 105-113. Aixock, J. & K. M. O’Neill. 1987. Territory preferences and intensity of competition in the grey hairstreak Strymon melinus (Lepidoptera: Lycaenidae) and the tarantula hawk wasp Hemipepsis usiulata (Hymenoptera: Pompilidae). American Midland Naturalist 11: 120-138. Au.stin, G. T. & K.Jt)HN,soN. 1997. Theclinae of Rondonia, Brazil: Gigantorubra and Anguloph, with descriptions of new species (Lepidoptera: Lycaenidae). Insecta Mundi 1 1(3/4): 255-272. Barcant, M. 1970. Butterflies of Trinidad and Tobago. London, Collins. 314 pp., 28 pis. Brown, K. S. 1982. Historical and ecological factors in the biogeography of aposematic neotropical butterflies. American Zoologist 22(2): 453-471. Comstoc:k, J. H. 1918. The wings of insects. The Comstock Publishing Company, Ithaca. 430 pp. CoM.STOCK, W. P. & E. 1. Huntinciton. 1962. An annotated list of the Lycaenidae (Lepidoptera: Rhopalocera) of the Western Hemisphere. Journal of the New York Entomological Society 70: 39 46. D’Abrer,\, B. L. 1995. Butterflies of the Neotropical region. Part VII. Lycaenidae. Black Rock, Hill House, i-xi + 1098-1270 pp. D.AVID.SON, J. M., S. A. Rf.hner, M. Santana & E. Lasso. 2000. First report of Phytnphthora heveae and Pythiuni spp. on tropical tree seedlings in Panama. Plant Disease 84(6): 706. Divuidt, M. 1919-1920. Thedini F., pp. 744-812, In: A. Seitz (Ed.). Die Gross-Schmetterlinge der Erde, vol. 5. Die amerikanischen Tagfalter. Stuttgart, Alfred Kernen. Duarte, M. & R. K. Robbins. In press. Description and phylogenetic analysis of the Calycopidina (Lycaenidae, Theclinae, Eiimaeini): a subtribe of detritivores. Revista Brasileira de Entomologia. Duarte, M., R. K. Robbins, & O. H. H. Mielke. 2005. Immature stages of Calycopis caulonia (Hewitson, 1877) (Lepidoptera, Lycaenidae, Theclinae, Eumaeini), with notes on rearing detritivorous liairstreaks on artificial diet. Zootaxa 1063: 1-31. Eliot, J. N. 1973. The higher classification of the Lycaenidae (Lepidoptera): a tentative arrangement. Bulletin of the British Museum (Natural History) Entomology 28(6): 371-505. Feinstein, J., S. Mori, & A. Berkov. 2007. Saproflorivory: a diverse insect community in fallen flowers of Lecythidaceae in French Guiana. Biotropica 39(4): 549-554. Findiay, S. & K. Tenoril 1982. Nitrogen source for a detritivore: detritus substrate versus associated microbes. Science 218: 371-37.3. Goi)m.-\n, F. D. & O. Salvin. 1887-1901. Biologia Centrali-Americana. Insecta. Lepidoptera-Rhopalocera. Fam. Lycaenidae. 2: 1-112. Hohn, F. M. & D. L. Waoner. 2002. Larval substrates of herminiine 42: 64-73, 2003 (2010) 73 noctuids (Lepidoptcra): niacrodecoiiiposers of teni|)erate leal litter. Eiivironnieiital Eiitoniolog)' 29: 207-212. Johnson, K. 1993. Neotropical hairstreak butterflies: new sister genera of the “Elerlroslrymon/ Angulopis" grade of Enniaeini (Lepidoptera, Lycaenidae, Theclinae) . Reports of'the Museum of Natural History. University of Wisconsin (Stevens Point) 32: 1-76. Johnson, K. 8c K. R. Krof.ni.kin. 1993. Revision of Angiilopis (l.epidoptera, Lycaenidae, Theclinae). Reports of the Museum of Natural History. University of Wisconsin (Stevens Point) 33: 1-38. JoHN.soN, S. A. 1985. Uultnringadetritivore, Calycopis isobeon (Butler & Druce). News of the Lepido|>terists’ Society 3: 41-42. Kaye, W. J. 1921. A catalogue of the Trinidad Lepido]3tcra Rhopalocera. Memoirs of the Department of Agriculture of Trinidad and Tobago 2: i-xii, 13-163. Kiors, A. B. 1970. l.epidoptera, pp. 1 15-130, In: S.L. Tuxen (Ed.), Taxonomist’s glossary of genitalia in insects. Munksgaard, (Copenhagen. Lamas, G., R. K. Roisisins & D. J. Harvey. 1997. Mariposas del alto Rio Napo, Loreto, Peru (Lepido|Ytera: Pajtilionoidea y Hesperioidea). Revista Peruana de Entomologfa 39: 63-74. Mark, H. -G. 1993. Erste .Vlitteilung fiber Tagfalter- und Zyganenzuchten mit semisynthetischem Kunstfutter. Nachrichten des Entomologischen Vereins Ajtollo, N.F., 14: 27.5-280. Mark, H. -G. 1995. Tagfalter- und Zygaenenzuchten mit semisynthetischem Kunstfutter, zweite Mitteilung: Lepidojttera, Lycaenidae. Nachrichten des Entomologischen Vereins Apollo, N.F., 16: 263-274. Morion, R. S. 1981. Rearing butterflies on artificial diet. Journal of Research on the Lepidoptera 18: 221-227. Powiti.i , J., G. Muter & B. D. Farrei i.. 1998. Evolution of larval food preference in Lepidojttera, pp. 403-422, In: N. P. Kristensen (Ed.). Band/ Volume IV Aithropoda: Insecta. Lepidoptera, Moths and Butterflies: Evolution, .Systematics, and Biogeography. In: M. Fischer (Ed.). Handhuch der Zoologie. Handbook of Zoology. Berlin, Walter de Gruyter. Rinor.i.v, R. S. 1976. A guide to the birds of Panama. I’rinceton Univ. Press, Princeton, NJ. 394 i5|t. RoBiiiNS, R. K. 1991. Evolution, comparative morphology, and identification of the eumaeine butterflv genus Rckoa Kaye (Lycaenidae: Theclinae). Smithsonian Gontribution to Zoology 498, 64 pp. RoiiiiiNS, R. K. 2004a. Introduction to the checklist of Eumaeini (Lycaenidae), pp. xxiv-xxx, in: Lamas, G. (Ed.), Ghecklist: Part 4A. Hesperioidea - Papilionoidea. In: Heppner, J. B. (Ed.), Atlas of Neotropical Lepidoptera. Volume .5A. (ktinesville. Association fbrTro|)ical Lepido|)tera; Scientific Publishers. Robbins, R. K. 20()4b. Lycaenidae. Theclinae. Tribe Eumaeini, jtp. 1 18-137, In: Lamas, (k (Ed.), Ghecklist: Part 4A. Hesperioidea -Papilionoidea. In: Heppner, J. B. (Ed.), Atlas of Neotropical l.epidoptera. Volume .5A. Gaine.sville, A.ssociation lor Tropical LepidojYtera; Scientific Publishers. Robbins, R. K. & Ck Lymas. 2008. Nomenclature, variation, and the biological species concept in Lnmasina (Lycaenidae: Theclinae: Eumaeini). Revista Brasileira de Zoologia 25(1): 1 16-127. Robbins, R. K., G. L.ymas, O. H. H. Miei.ke, D. J. Harvey & M. G.\.s.\gr.\ni)E. 1996. Taxonomic composition and ecological structtire of the species-rich butterflv community at Pakitza. Parque National del Mann, Peru, |)|). 217-252, In: Mann: fhe biodiversity of Southeastern Peru (D. E. W'ilson and A. Sandoval, (Eds.). Smithsonian Institution Press, Washington DG. 679 ]tp. Robbins, R. K. & G. B. Smai.i,. 1981. Wind dispersal of Panamanian hairstreak butterllies (Lejtidoptera: Lycaenidae) and its evolutionary significance. Biotropica: 13(4): 308-315. SNonoR.\.ss, R. E. 1935. Principles of insect inoiphologY’. McGraw- Hill Book Go., New York, 667 jtp. Weeks, A. G. 1911. Illustrations of diurnal l.epidoptera. With flescri]3tions. Boston, The L niversity Press. 2: xvi + 37 p. Journal oj Research on the Lepidoptera 42: 74-78, 2003 (2010) Ag-gregated oviposition in Actinote pellmea pellenea Hubner (Lepidoptera? Nymphalidae) Ronaldo Bastos Francini' and Andre Victor Lucci Freitas-* 'Universidade Catolica de Santos, Campus D. Idiliojose Soares, Av. Conselheiro Nebias, 300, 11015-200, Santos, Sao Paulo, Brazil. fra ncini@unisan las. br “Departamento de Bioiogia Animal and Museu de Zoologia, Instituto de Biologia, Universidade Estadual de Campinas, CP 6109, 13083-970, Campinas, Sao Paulo, Brazil. bak u@ii nka mp. br Abstract. The oviposition pattern of Actinote pellenea pellenea on Austroeiipatorium inulaefolium was investigated in two coastal sites in SE Brazil to test if there is any pattern of preference for host plant traits. At both sites, host plants were marked and measured for height, distance of the nearest plant, number of leaves with A. p. pellenea immatiires (eggs and first instar larvae), number of groups of immatures, and total number of eggs per plant. An apparency index was calculated relating plant size to distance from its three nearest conspecific neighbours. Total leaf area, orientation and inclination were recorded for each leaf with a group of immatures. There was no significant correlation between the number of ovipositions on each plant and habitat and plant characteristics as plant height and apparency index. The number and density of eggs by oviposition was not correlated with leaf characters as area, orientation or inclination. At botli sites clusters of immatures showed a grouped distribution, with some plants having more immatures than predicted by chance. In the only case of a double oviposition on the same leaf, the later cluster was significantly smaller. The results show that distribution of eggs - reflecting female choice - was not related w'ith the above measured plant and leaf traits. However, the results do suggest that females probably choose plants where immatures are already present, resulting in the observed grouping pattern. Keywords: Actinote, Austroeupatorium, host plant selection, oviposition. Introduction For most holometabolous insects in general, and with herbivores in particular, adults are more mobile than immatures, with the decisions of ovipositing females often critical for the fitness of the offspring (Doak et at, 2006). For most herbivorous insects, host plant acceptability and quality vary strongly among plant species, populations and individuals and even among different plant parts (Strong et al, 1984; Price, 1997; Kerpel et at, 2006). The ability of females to choose among different liosts or host parts has important consequences on their offspring’s performance, with the females usually using plant cues to identify the most suitable food resource. Many different cues are used by females during the oviposition process. These include secondary compounds, visual signals (plant and leaf shape), presence of natural enemies or mutualists, presence of conspecific immatures, microclimate (Raiisher, 1978; Williams & Gilbert, 1981; Freitas & Oliveira, 1996) H’.orrespon den ce a u thor Received: 22 June 2009 Accepted: 25 June 2009 and plant vigour (Price, 1991, 1997). Furthermore, ovipositing females can also respond to indirect cues such as leaf age and size, internode length (Price et al, 1987; Freitas et al, 1999), plant apparency, grouping and position of host plants across the landscape (Feeiiy, 1976; Courtney & Courtney, 1982; Mackay & Singer, 1982). The Neotropical genus Actinote Hiibner, [1819] (Nymphalidae: Helicoiiiinae: Acraeiiii) has 31 described species distributed through Central and South America, reaching maximum diversity in the montane regions in the Andes and Southern Brazil (Francini et al, 2004; Lamas, 2004; Paluch, 2006; Paluch et al, 2006; Silva-Brandao et al, 2008). Al known species feed on Asteraceae and are gregarious during all stages (Francini, 1989, 1992; Paluch et al, 2005; Freitas et al, 2009). The widespread Actinote pellenea Hiibner, [1821] has 17 recognized subspecies distributed across all South Aiierica from Colombia to Northern Argentina found over a wide variety of secondary and open habitats (Francini, 1989, 1992; Paluch, 2006). In coastal Southern Brazil, A. pellenea pellenea Hubner, [1821] is very common with four or five generations per year. Here larvae of A. p. pellenea feed on three species of Asteraceae: the vines Mikania micrantha and Mikania cordifolia, and the shrub Austroeupatorium 42: 74-78, 2003 (2010) 75 inulaefolium (Francini, 1989, 1992) (Fig. 1). The following study describes the pattern of oviposition distribution of A. p. pellenea on patches of A. inulaefolium, and discusses the factors that may explain the observed patterns. Methods Study sites The study was carried out at two coastal sites of Sao Paulo State, Southern Brazil: 1) Xixova-Japui State Park (JAPUI), Sao Vicente, Sao Paulo (23°59’ S, 46°23’ W), in July and 2"‘‘ , 1991, and 2) the valley of the Cubatao river, (VRCUB), Cubatao, Sao Paulo (23°53’ S, 46°27’ W) during November 12, 1992. Both sites are covered by lowland subtropical forest (Ururahy el ai, 1984). Annual rainfall reaches 2500 mm and the average annual temperature is 21°C (Setzer, 1949; Nimer, 1989). Field work was conducted along open trail edges and in early succession stages secondary vegetation. Sampling procedures At each site all individuals of A. inulaefolium j^resent along a previously defined linear transect of 200 m were sampled, including all plants to a 5 meters distance on both sides of the transect, including all nearby plants in the area. Immatures of A. p. pellenea of each host plant were recorded and all ovipositions were collected for laboratory work. Each individual plant was tagged and recorded for height (five classes of 100 cm), distance of the nearest plant (in meters), orientation in relation to the nearest plant (in degrees), number of leaves with immatures, total number of immature groups and total number of immatures. For each leaf with immatures, the area, inclination (to the horizontal), orientation (in degrees to the central axis of the plant), amount of herbivory and number of immatures per group were recorded. An “apparency index” (AI, follow'ing Feeny, 1976) was calculated for each individual plant, as: AI = [(Ho - Ha + DA) + (Ho - Hb + DB) + (Ho - He + DC)] / 3; where Ho = plant height, Ha, Hb, He = height of the nearest three plants of the species, and DA, DB, DC = distance of the nearest three plants of the same species present along the linear transect. This index is lowest (including negative values) when the distance between a pair of plants is less than the lower plant of the pair. The index increases with increasing distance between the two plants. To determine the distribution of the plants, the study area was included in a 108 m- rectangle, and this was subdivided in 27 squared sectors of 2x2 m (three rows and nine columns). The distribution pattern of plants was determined using the index of dispersion (Ludwig & Reynolds, 1988) with values < I indicating a uniform distribution, = 1 a random distribution and > 1 a grouped distribution. Inclination and orientation of each leaf was estimated with compass and protractor with a precision of 1°. Each oviposition was assigned to one of three developmental stages based on predominant egg colour following Erancini (1989), where Y = yellow (0-1 day after oviposition), O = orange (1-3 days after oviposition), R = red (more than 3 days after oviposition), B = black (1 or two days before eclosion). First instars were also considered in the present study, since larvae usually remain, together with their empty egg shells, on the same leaf for the first days following eclosion. All measured leaves were scored into four categories of herbivore damage: (0) 0%; (1) up to 10%; (2) 11% to 25%; (3) 26%\o 50%; (4) more than 51%. Since the leaves of A. inulaefolium are nearly rhombus-shaped, leaf area (in cm-) was estimated using the formula (L*W)/2, where L = length and W = width. Egg density for each oviposition was estimated from the average of hve independent counts of eggs over different sections of each oviposition event, giving the number of eggs per cm'*^. The number of eggs in each oviposition was estimated as the product of the average density by the area of the oviposition. Results Host plant density. At JAPUI 23 |)lants of A. inulaefolium \were inspected showing a density of 0.21 plants/m‘‘^ with a non-random distribution (Poi.s.son test, x- = 4.2031, DF = 2, p < 0.05). The dispersion index was 1.331, indicating that the plants are grouped. At VRCUB 1 1 plants were inspected with a density of 0. 10 plants/m- with a random distribution (Poisson test; y- = 0.7365, DF = 1, p > 0.70) and dispersion index of 0.8042, which indicated a homogenous distribution. Host plant height. AtJAPUl A. inulaefolium height ranged from 100 to 500 cm ( J - 213.0 cm, SD = 109.98, n = 13), significantly lower than from VRCUB, that ranged from 300 to 700 cm ( x = 472.7 cm, SD = 1 10.37, n = 1 1 ) (t test = -6.4339, DF = 32, p < 0.001 ) . At JAPUl the AI (apparency index) ranged from -100 to 1760 while at VRCUB the range was greater, from 800 to 3040, indicating that plants with high apparency co¬ occur with plants with low apparency. There was no correlation between the AI and plant height at either site (Spearman, r = 0.0689 in JAPUI, t = 0.3163, DF = 21, p > 0.75, and r = 0.1908 in VRCUB, t = 0.5831, DF 76 J. Res.Lepid. = 9, P>().57). Distribution of immatures. AtJAPUI 43 groups of ininiatures (eggs or first iiistar larvae) were recorded on 42 leaves of all plants with the number of groups per plant ranging from 0 to 9 {X= 1.9 ovipositioiis/ plant, SD = 2.40, n = 23) and a dispersion index of 3.078. At VRCUB 41 groups of immatures were recorded on 41 leaves of all plants, with a range from 0 and to 1 1 (X = 3.7 ovipositioiis/ plant, SD = 12.82, n = 23) and a dispersion index of 3.439. The values of the dispersion indexes indicate that distribution of immatures on the plants of both sites was grouped. Ovipositions were recorded on leaves ranging from 15 to 450 cm above ground atJAPUI and from 130 to 500 cm at VRCUB. When grouped into intervals of 100 cm, the data showed a concentration of 27 ovipositions within the interval 101 - 200 cm in JAPUI, and of 16 ovipositions in the interval 201 - 300 cm in VRCUB. A double oviposition on the same leaf was observed once during the study. Oviposited leaves. The average surface area of leaves that showed oviposition atJAPUI ( X = 2553.21 mm2, SD =1929.36, n = 58) was significantly lower than at VRCUB (x = 4549.3 mm2, SD = 2653.82, n = 25) (t test = -3.8755, p < 0.0002, DF = 81). There was no preferential orientation of oviposited leaves at either site (JAPUI X-' = 2.00, p = 0.57, DF = 3; VRCUB X’ = 3.37, p = 0.34, DF = 3) . With respect to leaf inclination, most oviposited leaves were between 10° and 20° from the horizon at both sites. Considering only leaves showing oviposition, most were leaves with low herbivoiy damage (0 to 10% herbivoiy) at both JAPUI (67.6%) and VRCUB (92.6%). However, because the patterns of leaf orientation and inclination, and herbivory, were not evaluated for all plants, these data are not useful in sliowing tendencies or preferences by ovipositing females. Eggs X plants. AtJAPUI the number of eggs per plant ranged from 0 to 3850, with a total of 20,100 eggs from 14 of 23 plants (X = 873.9 eggs/plant, SD = 1154.42) and with a dispersion index of 1525, a significant grouped pattern. No significant correlation was determined among the following parameters: number of eggs and plant height (Spearman, r = 0.1179, t = O.sW, DF = 21, p > 0.59), number of ovipositions and plant height (Spearman, r = 0.0982, t = 0.4521, DF = 21, p > 0.66), the AI and number of eggs per plant (Spearman, r = -0.1485, t = -0.6879, DF = 21, p > 0.49), or the AI and number of ovipositions per plant (Spearman, r = -0.3571; t= -1.7521, DF = 21, p < 0.09). At VRCUB the number of eggs per plant ranged from 338 to 4236 (all plants with at least one oviposition) giving a total of 15,880 eggs on 1 1 plants (X= 1443.6 eggs/plant, SD = 1399.91) and with a Figure 1 . A female Actinote pellenea pellenea ovipositing in Austroeupatorium inulaefolium. dispersion index equal to 1357.55. These data also revealed a significant grouped pattern. Again, as at JAPUI, no significant correlation was demonstrated between the following parameters: number of eggs and plant height (Spearman, r = 0.1908, t = 0.5830, DF = 9, p > 0.57), number of ovipositions and plant height (Spearman, r = 0.2435, t = 0.5392, DF = 9, p > 0.6) , the AI and number of eggs per plant (Spearman, r = 0.3184, t = 0.31, DF = 9, p > 0.76), or the AI and the number of ovipositions per plant (Spearman, r = 0.3184, t = 1.0078, DF = 9, p > 0.33). Eggs X leaves. AtJAPUI the average density of eggs ranged from 1 1 1 to 470 eggs/ cm- ( x = 259.5 eggs/ cm-, SD = 68.32, n = 57) with the number of eggs per oviposition ranging from 86 to 1266 (x = 479.2 eggs, SD = 228.93, n = 56). There was no significant correlation among the following parameters: mimber of eggs and leaf area (Spearman, r = 0.1054, t = 0.7934, DF = 56, p < 0.43) , number of eggs and leaf orientation (Spearman, r = -0.01 14, t = -0.0856, DF = 56, p > 0.93), or number of eggs and leaf inclination (Spearman, r = 0.0144, t = 0.1078, DF = 56, p > 0.91). At VRCUB the average density of eggs was from 95 to 374 eggs/ cm'*^ ( X = 238.5 eggs/cm^; SD = 66.81, ii = 27) with the number of eggs per oviposition ranging from 100 to 883 ( X = 422.31 eggs, SD = 199.01 eggs; ii = 26). And again there was no significant correlation among the parameters: number of eggs and leaf area (Spearman, r = -0.2736, t = -1.3644, DF = 23, p > 0.18), number of eggs and leaf orientation (Spearman, r = -0.0415, t = -0.1991, DF = 23, p > 0.84), or number of eggs and leaf inclination (Spearman, r = 0.0327, t= 1.1571, DF = 23, p < 0.87). Correlations between leaf area and 42: 74-78, 200,8 (2010) egg density were also not significant at either JAPUI (Spearman, r = -0.718, t = -0.339, DF = 55, p > 0.9) and in VRCUB (Spearman, r = -0.1327, t = -0.0419, DF = 23, p> 0.52). Discussion Oiir study was not conclusive in revealing any consistent pattern of oviposition in A. p. pellenea, except for clearly indicating that ovipositions tend to be grouped. The morphological plant traits investigated by our study were apparently not used by the females when selecting oviposition sites. There are of course several additional factors that would be important in selection by oviposition sites by A. p. pellenea females that were not evaluated in our study, such as: 1) other plant features, such as secondary compounds, nutritional c|uality and/or vigor (Kerpel et ai, 2006), 2) presence of alternative host plants in the same area (the common scandent vines M. micranlha And M. cordifolia) that might influence the patterns we found, and 3) a strong preference for plants previously oviposited by females (Ulmer et ai, 2003). Despite of which factors are influencing female choice, it is worth noting that the grouped pattern of immatures was revealed at both sites. As a result, many plants were not used for oviposition females at all, by contrast to a few that received up to nine ovipositions (> 3000 eggs). The advantages of grouped eggs are well known for many species of Lepidoptera, including protection against desiccation and predation (Stamp, 1980; (dark & Faeth, 1998). Gregarious larvae from egg clusters also benefit from increased develojiment rates and survival as well as reduced predation and parasitism (Lawrence, 1990; Clark & Faeth, 1998; Denno & Benrey, 1997). On the other hand, as pointed out above, super- oviposition results in high mortality of small larvae which will not get enough food as we frequently observed in the field. The reasons for this grouped pattern of immatures in A. p. pellenea require further investigation to reveal if the advantages in many larvae feeding in the same plant are higher than the risk of death by starvation. Additionally comparisons with other species of Actinote are also needed to reveal if the pattern of group immatures can be generalized in this genus. Acknowledgements We would like to thank Renato Rogner Ramos tor help in field work, and Carla Penz, Daniela Rodrigues and Rudi Mattoni for valuable suggestions in the manuscript. This study was funded by FAPESP (grants 84/04.82-.8, 86/0618-5, 88/;t069-8 and 93/0097-9 to RBF, and grants 00/01484-1 and 04/05269-9 for AVl.F) and is jtart of the project “Lepidoptera of Siio Paulo State” (BIOd'A-FAPFSP program - grant 98/05101-8). Andre V. L. Freitas thanks also the Brazilian CNPq (fellowship 300282/2008-7) and the National Science Fotmdation (DEB-0527441). Literature cited CiARK, B. R. & S. H. Faeth. 1998. The consequences of larval aggregation in the hutterflv Chlosync taciiua. Ecological Eiuomology 22: 408-415. Courtney, S. P. & S. Coiirtney. 1982. The ‘edge-effect’ in butterfly oviposition: casuality in Anthocharis rardamines and related species. Ecological Entomology 7: 131-137. Denno, R. F. & B. Benrey. 1997. Aggregation facilitates larval grouth in the neotro|5ical nymphalid hutterflv Cldos\ne junais. Ecological Entomolog)’ 22: 13.3-141. Do,\k, P., P. Kvreiva & |. Kinc'.soi.n'er. 2006. Fitness consequences of choosy oviposition for a time-limited btitterfly. Ecology 87: 395-408. Ft;ENY. P. 1976. Plant apparency and chemical defense. Recent advances in Phytochemistry 10: 1-40. Fr.\n(;ini, R. B. 1989. Biologia e ecologia das horholetas Actinote (Le])ido|ttera: Nymjrhalidae, Acraeinae) na transi(;ao subtropical no stideste do Brasil. Ms Thesis. Universidade Estadual de Campinas, Campinas, Sao Paulo, Brazil. Francini, R. B. 1992. Ecologia das taxocenoses de Actinote (Lepidoptera: Nym|)halidae) em Asteraceae (Angiosperma, Magnoliatae) no sudeste do Brasil: stihsKlios para conservat^iio. PhD Thesis. Universidade F/stadtial de Campinas, Cam|>inas, Sao I’atilo, Brazil. Fr.\n(;im, R. B., a. \'. L. Frei evs & C. Penz. 2004. Two new' species of Actinote (Lejtidoptera, Nymphalidae) from Southeastern Brazil. Zootaxa7l9: 1-10. Freitas, A. \'. 1,. & P. S. Oi,i\tir-\. 1996. Ants ;ls selecti\e agents on herbivore hiolog)': effects on the behaviour of a non-mymiecojrhilous hutteifly. Jotirnal of Animal Ecology 65: 20.5-210. Freitas, A. V'. L., I. R. I,e.\t & S. O. Ferreir,\. 1991. Selection of oviposition sites by a lejridopteran community of a tropical forest in southeastern Brazil. Biotropica 31: 372-375. Freitas, A. V. L., L. A. Kaminski, R. C. M.vros & K. L. Sitva-Brand.ao. 2009. Immature stages of the Andean butterfly Actinote rn/ina (Nymphalidae: lleliconiinae: Acraeini). Tropical l.ejtidoptera Research 19: 18-21. Keri’ET, S. M., E. S<)I>r.\no &• G. R. P. Moreir.\. 2006. Effect of nitrogen on Passijlora stiherosa L (Passifloraceae) and consequences for larval performance and oviposition in Heliconius erato phyllis (Fahricius) (f.epidoptera: Nymphalidae). Neotro|)ical Entomology 35: 192-200. L. am.as, G. 2004. Checklist: Part lA. He.s|rerioidea-Pa|5ilionoidea. In: Heppner, J.B., ed.. Atlas of Neotrojrical I,epido|3tera, Vol .5A. Association for Tropical Lepidoptera/Scientific Publishers, Gainesville, 439 jr. l,.\w'RENCE, W. S. 1990. The eff ects of group size and host sjrecies on develo|)ment and survivorship of a gregarious caterpillar Halisidota caryae (Le|3idoptera: Arctiidae). Ecological Fhitomolog)' 15: 5.3-62. Lunwio, j. A. & J. F. Reynotd.s. 1988. Statistical ecology. A primer on methods and computing. 1st ed. John Wiley and Sons, Inc., New York, 337 p. M. vckay, D. a. & M. C. SiNCiER. 1982. The basis of an apparent preference for isolated host plants by ovipositing Enptychia lihye btitterflies. Ecological Entomology 7: 299-303. Nimer, E. 1989. Climatologia do Brasil. IBGE, Rio de Janeiro, 421 p. P.M.ucH, M. 2006. Revisao das espccies de dc/wo/c Uiihner, [1819] 78 J. Res.Lepid. (Lepidoptera, Nyniphalidae, Heliconiinae, Acraeini). PhD. Thesis. Universidade Federal do Parana, Curitiba, Parana, Brazil. Pai.uch, M., M. M. Casagrande & O. H. H. Mielke. 2005. Coinportaniento de agrega<;ao noturna dos machos de Actinote surima surinia (Schaus) (Lepidoptera, Heliconiinae, Acraeini). Revista brasileira de Zoologia 22: 410-418 Pai.uch, M., M. M. C.a,sagr.-\nde & O. H. H. Mielke. 2006. Tres especies e duas siibespecies novas de Actinote Hiibner (Nyniphalidae, Heliconiinae, Acraeini). Revista brasileira de Zoologia 23: 764-778 Price, P. W. 1991. The plant vigour hypothesis and herbivore attack. Oikos 62: 244-251. Price, P. W. 1997. Insect Ecology. 3rd ed. John Wiley & Sons, Inc., New York, 874pp. Price, P. W., H. Roininen &J. Tahvanainen. 1987. Wliy does the bud- galling sawfly, Euura mucronata, attack long shoots? Oecologia 74: 1-6. Rausher, M. D. 1978. Search image for leaf shape in a butterfly. Science 200: 1071-1073. Setzer, J. 1949. Contribuitiao para o estudo do clima do estado de Sao Paulo. Edit. Escolas Profissionais Salesianas, Sao Paulo, 239 p. Silva-BrandAo, K. L., N. W.ahi.berg, R. B. Francini, A. M. L. Azeredo- Espin, K. S. Brown Jr., M. Pai,uch, D. C. Lees & A. V. L. Freitas. 2008. Phylogenetic relationships of butterflies of the tribe Acraeini (Lepidoptera, Nymplialidae, Heliconiinae) and the evolution of host plant use. Molecular Phylogenetics and Evolution, 46; 515-531. Stamp, N. E. 1980. Egg deposition patterns in butterflies: why do some species cluster their eggs rather than deposit them singly? American Naturalist 115: 367-380. Strong, D. R.,J. H. Lawton & R. Solithw'ood. 1984. Insects on plants: community patterns and mechanisms. Harvard University Press, Cambridge, Massachussetts, USA, 330pp. Ulmer, B., C. Gillott & M. Eriandson. 2003. Conspecifc eggs and bertha armywonn, Mamestra configurata (Lepidoptera: Noctiiidae), oviposition site selection. Environmental Entomology 32; 529-534. Ururahy, J. C.,J. E. R. CoLiARES, M. M. Santos & R. A. A. Barreto. 1984. 4. Vegetaijao, p. 553-61 1. /n: Projeto RADAMBRASIL, levantamento de Recursos Naturais. Vol. 32. fls. SF 23-24. Rio de Janeiro e Vitoria. Fundapao IBGE, Rio de Janeiro. Williams, K. S. & L. E. Gilbert. 1981. Insects as selective agents on plant vegetative morphology: egg mimicry reduces egg laying by butterflies. Science 212: 467-469. journal of Research on the Lepidoptna 42: 79-83, 2003 (2010) Notes A new subspecies of Argynnis nokomis from the Sacramento Mountains of New Mexico (Nymphalidae) The most vexing question in New Mexico butterfly lore long was, “Did the Sacramento Mts. ever support the Nokomis Fritillary?” At Paul Grey’s instigation, I first started searching for colonies in 1963. The next 44 years produced nothing. I still believed that colonies had once existed at Bent, Otero Co., and at Ft. Stanton, Lincoln Co., but I had only utter frustration to show for it. Eventually, I reasoned where any museum specimens would most likely be, and had the AMNH collection searched accordingly. Eureka^two male Argynnis nokomis from Bent, Otero Co. This success inspired John Rawlins to search the Carnegie. Result: more specimens, including two of the magnificently colored dark yellow-green females (Holland, 2008). Within the current concept of A. nokomis ssp., the Sacramento Mts. population is distinctive. This case deals with a most likely extinct, high-profile organism that may come to be called by an English name in town meetings. I would prefer that name be easily translated into Latin, with the result being near its actual Latin name. “The Tularosa Fritillary” seemed reasonable in this context. My use of Argv«»Afollows the recent phylogenetic placement by Simonsen et al. (2006). Argynnis nokomis tularosa R. Holland, new stibspecies Diagnosis: The general shape of all markings, black and silver resemble t\pical A. n. nokomis. Two known female spec imens lack the fnlviis spot on DHW costa, DHW with redticed eycspots on the PM band, silvered spots on disc redticed, DFW and DHW with redticed black scaling along the veins. Holotype: Female, Mescalero, Tularosa River, Otero County, New' Mexico, ca. 7000', Atig. 13, 1931, leg. W. Huber, CMNH collection, ex. Philadelphia Academy of Sciences Collection. Paratypes: 2 males. Bent, Otero County, New Mexico, ca. 6000', Ang. 12, AMNl 1 collection, ex Paul Grey coll., ex. Paul Ehrlich coll. Year is not specified, but Paul Ehrlich was born in 1932, and Paul Grey donated his collection to the AMNH in 1948. 1 female, tlata as |ter hokitype. Identification: The key below will distingtiish A. n. tularosa h am all other subspecies based on females characters and provide my diagnosis. Vlales are harder to separate. Etymology: The name is feminine, as that of the nearest town and oldest Eitropean settlemetit in the Tularosa Valley or Tularosa Basiti. Tularosa itself is not a Spatiish root, btit is Nahuatl (Aztec), tneaning “cattail” (|ulyan, 1996). One is tempted to spectilate it has the satne etvmolog)' as the English word “toolies.” Habitat: The past and present habitats are described iti detail by Holland (2008). Comments: The locality where the specitnens are from give eticouragemetit to the possibility A. n. tularosa is tiot extinct. y\h ktiown specimens were taken on the Mescalero A|iache Itulian Reservation. The .VIescaleros were borti xenophobic and live in alpine meadow's an hoitr outside El Paso which has done little to tnake them trust clem city slicker Texan strangers. The Mescaleros Key to the subspecies of Argynnis nokomis. 1. la 2. 2a 3. 3a 4. 4a 5. 5a 5b 5c Sexually dimorphic . 2 Not .so . ivenona (dos Passos & Grey) Dorsally yellow' green in the lighter wing portions . 3 These areas veity blue, discal cell of VEW black aticl silver . coerulesceiis (W. Holland) DHW discal cell yellowish, silver spots large, dorsally the eyespots iti the PM batid reduced, fulvus spot on costa ol DHW absent . apacheana (Skinner) Not so . 4 Very black wings, terminal regions of bctth wings, bcjth stirfaces, ahnost solid black. DI IW wanting a fulvous sjtot, VI IW' silvering of disc reduced . nitocris (W. 11. Edwards) Not so . 5 The typical nokomis cluster i DHW with ftilvtis .s|)ot on costa abotit 80%. DHW with large eyespots in p.m. band, silvered spots in disc large, DEW atid VEW with heavy black scaling alotig veins, occtirs iti few large colotiies, fore wing > 41 turn, fulvous scaling not always presetit in DFW' yellow-green areas, yellow-green does not invade DFW cell . Sangre de Gristo typical nokomis DHW with ftilvtis spot on costa nearly always, DHW w’ith large eyespots in the p. m. band, silvered spots in disc large, DFW and VFM' with heai'y black scaling along the veins, occtirs in a myriad of tiny colonies, fore wing > 40 mm, noticeable ftilvtis scaling in DFW yellow-green areas, yellow-green may invade DFW' cell Chuska Mts., Navajo Res . nigrocaerulea (W. & T. Gockerell) Two known specimens lack fulvits sjtot on DHW costa, DHW with redticed eyespots on the p.m. baiul, silvered spots on disc reduced, DFW and DHW with reduced black scaling along the veins; not seen in 70 years and feared extinct, wingspread closer to typical nokomis than nigrocaerulea, Sacramento Mts., Mescalero Res . Speyeria nokomis tularosa 80 /. Res.Lepid. Figure 1. The Sacramento Mountains Argynnis nokomis tuiarosa population. Top two rows, males, Bent, Otero County, New Mexico, ca. 6000’, Aug. 12, AMNH collection, ex Paul Grey coll., ex. Paul Ehrlich coll. Year is not specified, but Paul Ehrlich was born in 1 932, and Paul Grey donated his collection to the AMNH in 1 948. Bottom two rows, females, Mescalero, Tuiarosa River, Otero County, New Mexico, ca. 7000’, Aug. 13, 1931, leg. W. Huber, CMNH collection, ex. Philadelphia Academy of Sciences Collection. The specimen in the third row is designated the female holotype. Of the two known females, it is the only one with both antennae intact. 42: 79-83, 2003 (2010) 81 patrol ever)' inch t)f road like they were providing the Coliseum with virgins; merely stopping is forbidden. Permits to collect, stirvey, or just watch the wildlife are very nearly unobtainable: they don’t want yon spotting anything endangered, the existence of which could be cause for limiting the absolute control the tribe enjoys on its land. The endemic Euphydryas anicia doudcrofti Ferris and R Holland, is known right up to the reservation line, hut there is not one report from on the re.servation anywhere in the |)ublic domain. The only person I ever knew to negotiate sticcessftilly to collect on the Mescalero Reservation posed au natural for a tribal art class in exchange. Acknowledgements My sincerest thanks to Jocelyn Gill of the Canadian National Museum for the magnibcent photo work that can protect an irreplaceable national asset from pointless handling. All specimens of Argynnis nokomis from Otero Co. in either the AMNH or the Carnegie are paratypes, the holoty|je in the latter instittition. Literature Cited DOS P.-\s,sos, C. F. & L. P. Grkv. 1947. Systematic catalog oi Spryeria (Lepidoptera, Nymphalidae) with designations of types and fixations of type localities. Amer. Mus. Nat. 1 list. Novitates 1297: 1-30. Holiand, R. 2008. The holv grail of New Mexico Lepidoptera: Sacramento Motintains Speyeria nokomis (W. H. Edwards) (Nymphalidae). j. Lepid. Soc. 62(3): 171-173. Jui.VAN, R. 1996. The place names of New Mexico. Albuquer(]ue SiMttN'SEN, T. J., N. W.AHI.BliRG, A. V. Z. BrOWKR &: R. [)E JttNG. 2006. Morphokygv', molecules and Fritillaries: a|j]jroaching a stable ]>hylogenv for Argrnuini (Lepidoptera: Nvmphalidae). Insect Systematics and Evohition 37:40.6-418. Richard Holiand, 1625 Roma NE, .-Mbuqnerque, NM 87106 Use of Hippuris, an emergent aquatic plant, as a larval host by the buckeye, /Mwonm coenia, in Northern California Recent advances in DNA-seqnence-based phylogeny have radically altered botanists’ concepts of the relationships within the old family Scrophnlariaceae and between the now-disaggregated components of that family and others previously classified in a variety of ways (Olmstead et al, 2001; Kadereit in Kubitzki & Kadereit, 2004). In addition to DNA evidence, the distribution of characteristic secondary phytochemicals affords a partially-independent indication of plant relationships. In that vein, host-plant choices by oligophagous insects may suggest underlying chemical, and thus potentially phylogenetic, affinities among the taxa involved. The chemical basis for host selection in various Melitaeiui (Nymphalidae) is the presence of the bitter compounds called iridoid glycosides (Bowers &: Puttick, 1986; Gardner & Sternitz, 1988). Shapiro and Hertfelder (2009) recently reported the iridoid- selecting variable checkerspot, Euphydryas clialcedona, feeding spontaneously, repeatedly and successfully on the exotic garden shrub butterfly bush, Buddleja davidii, historically placed in the Loganiaceae or its ow'ii family Buddleiaceae but now incorporated into Scrophnlariaceae. The common buckeye, Junonia coenia is also a Nymphalid but not a Melitaeine, and its host range in California embraces Scrophnlariaceae, the very Figure 1. The Biggs garden pond. Emergent stand of mare’s tail at right. Figure 2. Two buckeye larvae, Junonia coenia, feeding on mare’s tail in situ. 82 /. Res.Lepid. closely-related Plantaginaceae, and the genus Phyla {=Lipj>ia) in the Verbenaceae (Shapiro & Manolis, 2007). The chemical basis for host selection in this species has been shown to involve the presence of iridoid glycosides (Bowers, 1984) although the story must be more complicated insofar as some Verbenaceous genera known to produce iridoids, such as Lantana (Rimpier & Sauerbier, 1986) are common in buckeye environments but never utilized. Mare’s tail, Hippiiris vulgaris, is an emergent aquatic flowering plant with a superficial resemblance to a horsetail (Equisetaceae); it is widely distributed in the cooler parts of both North and South America but rather rare and local in California where, however, it is occasionally grown in garden ponds. It has been classified in the monotypic family Hippuridaceae, whose affinities have been obscure until recently although several authors placed it near the Scrophulariaceae. Iridoid glycosides were reported in mare’s tail by the pioneering phytochemist Hegnauer in the 1970s and confirmed by Damtoft et at (1994). Their importance for plant systematics was emphasized by Jensen et al. (1975) and El-Naggar and Beal (1980). Grayer et at (1999) noted the convergence of phytochemical and molecular-phylogenetic data in the group of families around Scrophulariaceae, and subsequent autliors have treated Hippuridaceae as a member of the “Scroph” dade (Kadereit, 2004). Given these facts it was not completely surprising when one of us (KB) found common buckeye larvae feeding on mare’s tail in her home pond at Sebastopol, CA (see photo) — twice in the pond’s 12-year existence. We subsequently learned that Mr. Michael Koslosky found buckeye larvae on the same plant “about ten years ago while shopping at Gonnie’s Pond Supply in Castro Valley [CA]” and reared them out successfully on it (M. Koslosky, pers. comm.). This is the only emergent aquatic plant known to be a buckeye host. It is not dear whether a larva can complete development on a single shoot or has to access an adjacent one at least once in its development; KB has seen them use downed stems as “bridges.” The stems are tall enough to permit pupation and edosion well above the water line. The distribution of iridoid glycosides is such that many otlier plants not known to be buckeye hosts are potentially usable. One of us (AMS), based on the confirmed presence of iridoids in princess tree, Pauloumia (Bignoniaceae) foliage (Lino von Poser A ai, 2000) , has on several occasions confined buckeye females on it, obtained eggs easily, and reared the larvae througli to the adult on it. The same is true on both Catalpa speciosa3.nd C. hignonioides (currently but shakily placed in Bignoniaceae or Scrophulariaceae), which also produce iridoids (Sha’ban et al, 1980; Iwaga et al, 1991). All of these are trees, and there are no records of the common buckeye using any tree as a host. However, the tropica! buckeye, Junonia genoveva, feeds on black mangrove, Avicennia (Avicenniaceae, sometimes put in Verbenaceae) and occasionally on Lippia (Scott, 1986), and these are iridoid producers. There is a suggestion that host selection by these butterflies is mediated by both apparency (growth form) and, at dose range, iridoid chemistry. Clearly, we have a lot to learn. Acknowledgements We thank Mr. Jeffrey Caldwell for getting us in touch, and Mr. Michael Koslosky for sharing his record with us. Literature cited Bowers, M. D. 1984. Iridoid glycosides and host-plant specificity in larvae of the buckeye butterfly, /Mtionta coenia (Nymphalidae). J. Chem. Ecol. 10: 1567-1577. Bower.s, M. D. & G. M. Plittick. 1986. Fate of ingested iridoid glycosides in Lepidopteran herbivores. J. Chem. Ecol. 12: 169-178. DM'Itoft, S., S. R.Jensen,J. Thorsen, P. Molgard & C. E. OitiEN. 1994. Iridoids and verbascoside in Callitrichaceae, Hippuridaceae and Lentibulariaceae. Phytochemistry 36: 927-929. E[,-N.4GG.4R, L.J. &J. 1.. Beal. 1980. Iridoids: a review. J. Nat. Prod. 43: 649-707. G.ardner, D. R. & F. R. Sternitz. 1988. Hostplant utilization and iridoid-glycoside sequestration by Etiphydryas (Lepidoptera: Nymphalidae). J. Clieni. Ecol. 14: 147-168. Grayer, R. J., M. W. Chase & M. S. J. Simmonds. 1999. A comparison between chemical and molecular characters for the determination of phylogenetic relationships among plant families: an appreciation of Hegnauer’s “Chemotaxonomie der Pflanzen.” Biocheni. Syst. and Ecol. 27: 369-393. Iwaga, T., T. H.amada, S. Kurogi, T. H.ase, T. Okubo & M. Kjm. 1991. Iridoids from Catalpa bignonioides. Phytochemistry 30: 4057- 4060. Jensen, S. R., B.J. Nielsen & R. D/Wlcren. 1975. Iridoid compounds: their occurrence and systematic importance in the Angiosperms. Bot. Not. 128: 148-180. Kadereit, J. W. 2004. Lamiales: introduction and conspectus, hr. K. Kiibitzki and J.W.Kadereit, eds., The families and genera of flowering plants, Vol.7. Springer-Verlag. Lino von Poser, G., J. Sghripsema, A. T. Henriques & S. R. Jen.sen. 2000. The distribution of iridoids in Bignoniaceae. Biochem. Syst. and Ecol. 28: 351-366. Oemstead, R. G., C. W. DePamphilis, A. D. Wolfe, N. D. Young, J. W. El. ISONS & P. A. Reeves. 2001. Disintegration of the Scrophulariaceae. Am. j. Bot. 88: 348-361. Rimpler.H. and H. Sauerbier. 1986. Iridoid glucosides as taxonomic markers in the genera Lantana, Lippia, Aloysia and Phyla. Biochem. Syst. and Ecol. 14: 307-310. Sc:oTr,J. A. 1986. The butterflies of North America. Stanford. Siia’ban, R., S. F. Ei.-N.aggar & R. K. Doskotch. 1980. Spedoside: a new iridoid glycoside from Catalpa speciosa. J. Nat. Prod. 43: 524-526. 42: 79-83, 2003 (2010) 83 SiiAPiRfi, A. M. & K. Uertfki.dkr. 2009. Use of Buddleja as host plant by Euphydryas chalcedonn in the Sierra Nevada foothills, California. News Lep. Soc. 51: 27, 39. Shapiro, A. M. & T. Manoi.is. 2007. Field guide to the butterflies of the San Francisco Bay and Sacramento Valley Regions, California. Univ. of Calif. Fre.ss. Arthur M. Shapiro, Center for Population Biology', University of California, Davis, California 95616 a mshapi ro@u ala vh.edii Kathy Biggs, 308 Bloomfield Road, Sebastopol, California 95472 big;snesl@sonk. net Journal of Research on the LepiclojHera 42: S5, 2003 (2010) Editorial Auf Wiedersehen Gutenberg The recent trends of disappearing print media - words and images on paper - is disturbing. Many newspaper are already extinct, and excepting parts of the financial press, more are seriously endangered. Scientific print publications are also threatened by the global financial crisis as budgets everywhere are being drastically cut back. More distressing is serious economic analysis that indicates the situation is likely to become much worse, or not much better for quite some time. University and research institution libraries are shiny targets as bureaucratic managers scramble to save their jobs by demonstrating “waste.” After all we can get many things on line for almost free, and if Google gets its way we might be able to abandon whole libraries with attendant savings. Who needs books? Especially as one analyst claims $130 billion a year will be saved by going wholly on line. Elimination of transitory matter, novels, magazines and so forth will save many trees, yet all print media share one basic issue that cannot be overlooked - permanence. A good friend of mine made a documentary film a decade ago. Slow fires: on the preservation of the human record, which focused on the issue of questioniug the data loss in the storage media. We all know papyrus texts survived thousands of years. There is no test of reliability of any other information storage system. As Prof. J. E Blanchette of UCLA put it “Imagine of the only copy left of Imaging in Oncology were the Kindle version, with its garbled tables and lost color coding? Or, a more likely scenario, if several copies of the book existed in different formats, each with a different visual presentation?” Then there is long range political stability. A luddite dictatorship would have a grand following for a data farm auto de fe. So science presses face an increasing problem of how to pay for printed copy when there is increasing resistance concerning cost. Our subscription base does not begin to cover costs and the base is shrinking. Instittitional sub.scribers are also canceling. No funds. We are investigating on line publishing, whereby a small print run by a Docutecb system will provide a limited press run for archival pitrposes, along the lines of “Zootaxa.” Lastly, and perhaps most disturbing is an observation by an academic colleague in Europe that students no longer are interested in joining scientific societies. Now if this becomes the postmodern world view, where are we headed and what is to be done? If rape is inevitable, should we relax and enjoy it? Rudi Mattoni, Editor , >i'.H J- w jSi*--* ■'i^/ ' '< ^ineci pellenea}r{nhi\ev (Lepidoptera: Nymphalidae) Ronaldo Bustos Francini and Andre Victor Liicci Freitas 74 NOTES A new subspecies of Argynnis nokomis from the Sacramento Mountains of New Mexico (Nymphalidae) Richard Holland 79 Use of Hippuris, an emergent aquatic plant, as a larval host by the buckeye, coenia, in Northern California Arthur M. Shapiro and Kathy Biggs 81 EDITORIAL Auf Wiedersehen Gutenberg Riidi Mattoni 85 Cover: adult of Eacles impericdis, 7 August, 2004. © M. W. Nelson/MtissachusetLs Natural Heritage & Endangered Species Program.