4 April 2005 PROC. ENTOMOL. SOG. WASH. 107(2), 2005, pp. 398-416 PHYLOGENETIC ANALYSIS OF CYANOPHRYS CLENCH, A SYNOPSIS OF ITS SPECIES, AND THE POTENTIALLY THREATENED C. BERTHA (JONES) (LYCAENIDAE: THECLINAE: EUMAEINI) ROBERT K. ROBBINS AND MARCELO DUARTE (RKR) Department of Entomology, RO. Box 37012, NHB Stop 127, Smithsonian In- stitution, Washington, DC 20013-7012 U.S.A. (e-mail: RobbinsR@si.edu); (MD) Cole??o de Lepidoptera, Museu de Zoolog?a, Universidade de Sao Paulo, Av. Nazar? 481, 04263-000 Sao Paulo SP Brasil (e-mail: mduartes@usp.br) Abstract.?A species level phylogenetic analysis of Cyanophrys Clench was performed using 14 characters of the frons, wing venation, wing shape, androconia, and genitalia. There were 15 most parsimonious cladograms, and a successive weighting iteration re- sulted in three of these 15. Cyanophrys is characterized in this paper by two hypothesized synapomorphies (an expanded hindwing anal lobe that is tan-brown and a pair of dorsal brush organs), and adults can be readily identified in the field. Cyanophrys has been divided into Plesiocyanophrys K. Johnson, Eisele and MacPherson, Apophrys K. Johnson and Le Crom, Antephrys K. Johnson, Eisele and MacPherson, Mesocyanophrys K. John- son, and Cyanophrys (as delimited in Johnson and Le Crom 1997a), but only the mono- typic Antephrys is monophyletic on the most parsimonious cladograms. A synopsis of Cyanophrys species includes notes on their distribution, habitat, identification, nomencla- ture, larval food plants, and male behavior. Cyanophrys bertha, which occurs in moist evergreen and seasonally deciduous forests in the mountains of southern Brazil from 800 to 1,400 m elevation, has been proposed for threatened status. It is "cladistically distinct" {sensu Vane-Wright et al. 1991) in that its sister is a lineage of five montane species or seven primarily lowland species. Key Words: Callophrys, phylogeny, venation, genitalia, androconia, brush organs Generic taxonomy of the primarily Neo- tropical Cyanophrys Clench (Theclinae: Eumaeini) has been unstable (Robbins 2004b). Clench (1961) described it as a subgenus of the Holarctic Callophrys West- wood, but since 1993, Cyanophrys has been divided into five genera and subgenera, pri- marily on the basis of differences in geni- talia and color of the frons (Johnson et al. 1993, Johnson and Le Crom 1997a). While most of the 16 Cyanophrys species are common and widespread, the Brazilian C. bertha (Jones) is exceedingly rare and has been proposed for "vulnerable" status (Brown 1993, Brown and Freitas 2000 and references therein. Otero et al. 2000). More recently, it has been listed as "almost threatened" (Mielke and Casagrande 2004). Little is known about C. bertha, and it is unclear which species, or group of species, is its closest relative. Such information might provide clues to its biology. Further, phylogenetic position is a factor to be con- sidered in assessing the status of threatened species (Vane-Wright et al 1991). The first purpose of this paper is to assess the monophyly of the genera and subgenera into which Cyanophrys has been parti- s A l^As^'^.d^ ?X ?'-.'?A,--.' VOLUME 107, NUMBER 2 tioned. To accomplish this goal, we code interspecific morphological variation to in- fer phylogenetic relations among Cyano- phrys species. Besides frons color and gen- italia, we code characters of the male fore- wing venation, androconia, brush organs {sensu Eliot 1973), and wing shape. Johnson and co-authors have proposed 433 new species names in the Eumaeini during the past three decades, including 31 in Cyanophrys (Robbins 2004a,b), but have not been careful. The adults and genitalia of many types are different sexes or species (Robbins and Lamas 2002). Data labels on types appear to have been switched (Rob- bins and Nicolay 1999, Robbins and Lamas 2002). In one case, the two adult types of one species from different collections were similarly glued parts of species in different genera (Robbins and Lamas 2002). Finally, a number of types cannot be found (Rob- bins and Nicolay 2002, G. Lamas, personal communication). The publications of Johnson and co-au- thors on Cyanophrys contain similar prob- lems. For example, Johnson and Le Crom (1997b:23) designated a neotype for Papi- lio amyntor Cramer 1775 from Surinam and deposited it in the Natural History Museum (London). However, the labeled neotype de- posited in this collection has data labels from French Guiana. Another specimen with their neotype label in the American Museum (New York) has a Surinam label, but is a different species than the designated neotype in London. A second representative example of their lack of care is outlined in the synopsis below under Cyanophrys ro- raimiensis K. Johnson and D. S. Smith. Al- though we plan to address the neotype des- ignation elsewhere, correcting all problems created by Johnson and co-authors in Cy- anophrys would be a Sisyphean task that could delay publication of our current re- sults for years. We follow the classification in Robbins (2004b). The second purpose of this paper is to present a synopsis of Cyanophrys species, with special emphasis on C. bertha. Al- 399 though identification of Cyanophrys species was not a goal of this study, the coded char- acters can be used to identify most species. And in the species synopsis, we note infor- mation on the distribution, habitat, identi- fication, nomenclature, behavior, larval food plants, and published illustrations of adult Cyanophrys species. We summarize and assess what we know about C. bertha in light of the phylogenetic results. MATERIALS AND METHODS Genitalia, androconia, and venation were examined by standard techniques (Robbins 1991) using the approximately 1,075 spec- imens of Cyanophrys in the National Mu- seum of Natural History (USNM) as well as others borrowed from the American Mu- seum of Natural History (AMNH) in New York, the Illinois Natural History Survey (INHS) in Champaign-Urbana, The Natural History Museum (BMNH) in London, and the Universidade Federal do Paran? (UFPR) in Curitiba. Wing venation illustrations were made by digital scanning of wing slides. We examined 82 male and 62 female genitalic preparations. Genitalic terms fol- low those in Klots (1970), wing vein names follow Nicolay (1971, 1977), and andro- conial terminology follows Robbins (1991). The terminal taxa are C. goodsoni (Clench), C. argentinensis (Clench), C ber- tha, C acaste (Prittwitz), C amyntor (Cra- mer), C fiisius (Godman and Salvin), C. herodotus (Fabricius), C. miserabilis (Hew- itson), C. velezi K. Johnson and Kruse, C. crethona (Hewitson), C. longula (Hewit- son), C. pseudolongula (Clench), C agri- color (Butler and H. Druce), C banosensis (Clench), and C remus (Hewitson). Cyano- phrys roraimiensis K. Johnson and D.S. Smith is omitted because it is known only from the holotype female, but we discuss its likely phylogenetic placement on the ba- sis of structures in the female type. As out- groups, we chose three species from three other subgenera of Callophrys recognized by Clench (1961); Callophrys (Callophrys) 400 PROCEEDINGS OF THE ENTOMOLOGICAL SOCIETY OF WASHINGTON Table 1. Data matrix for Cyanophrys. The three Callophrys species are the outgroups, and the characters and their states are detailed in the text. Characters Species 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Callophrys rubi 1 0 0 0 0 0 0 0 0 0 ? ? 7 ? C. gryneus 0 0 0 0 0 0 0 0 0 0 ? ? 7 ? C. niphon 0 0 0 0 0 0 0 0 1 0 ? ? 7 7 Cyanophrys goodsoni 1 0 0 0 0 0 0 0 0 0 0 0 C. argentinensis 1 0 1 0 0 0 0 0 0 0 0 C. bertha 0 0 0 0 0 2 0 0 1 0 0 C. acaste 1 0 1 1 0 2 0 0 1 0 0 C. amyntor 1 0 0 1 0 2 0 0 1 0 0 C. fusius 1 0 2 7 2 0 0 1 0 0 C. herodotus 1 0 1 1 2 1 0 0 0 0 C. miserabilis 0 0 1 1 2 1 0 0 0 0 C. velezi 0 0 1 1 2 1 0 0 0 0 C. crethona 0 0 1 1 2 1 0 0 0 0 C. longula 0 0 0 0 2 0 0 C. pseudolongula 0 0 0 0 2 0 0 C remus 0 0 0 0 2 0 0 C. banosensis 0 0 0 0 2 0 0 C. agricolor 0 0 0 0 1 0 0 rubi (L.), C. (Mitoura) gryneus (H?bner), and C. (Incisalia) niphon (H?bner). We coded 14 characters (Table 1) and used a question mark (?) for inapplicable states and for one case of an intraspecific dimorphism. All multi-state characters were treated non-additively (unordered). We used the implicit enumeration option of Hen- nig86 software to derive most parsimonious cladograms. A strict consensus tree was de- termined. To test the assumption of equally weighted characters, a successive weighting iteration was performed (Farris 1969), and a consensus of the resulting trees was de- termined. Mapping of characters on trees was done with Winclada software (Nixon 2002) using the "unambiguous changes only" optimization option. Jackknife sup- ports were determined in Winclada using Nona (1000 replications with mult* 10, memory 1000 trees). The synopsis contains additional detail for C. bertha because of its conservation interest. The distribution of C bertha is based on 13 museum specimens (Appen- dix). Its morphology is based on 6 individ- uals in the USNM collection (Appendix). Distributional information for the other spe- cies is based on specimens from many mu- seums and from the literature. Records of larval food plants without citation are based on Janzen and Hallwachs (2004) or data compiled by Robbins. Plant family names follow those in Willis (1973). Records of "territorial" behavior are based on unpub- lished data compiled by Robbins. Type lo- calities are listed in Robbins (2004b), and citations for original descriptions can be found in Lamas (1995). MORPHOLOGY AND CODED CHARACTERS Head.?Presence or absence of iridescent green scales on the frons is intraspecifically consistent. Character 1: Iridescent green scales on frons (0) absent, (1) present. This character needs to be used with cau- tion because exposure to humidity, partic- ularly during preparation, or to physical abrasion can change scale color from green to brown. Iridescent butterfly scales return to their original color after drying (Nijhout 1991), but for reasons currently unknown, this is generally untrue for hairstreaks VOLUME 107, NUMBER 2 401 Fig. 1. Dorsal (top) and ventral wings of male (left) and female Cyanophrys bertha. Scale 1 cm. (Robbins, unpublished). Also, green scales on the frons may be rubbed off of worn individuals. Wing pattern.?We were unable to code wing pattern variation in Cyanophrys for a number of reasons. Presence of a postme- dian line on the ventral forewing varies in- terspecifically and intraspecifically among species. A dark maroon band along the sub- margin of the ventral hindwing and dark brown scales at the base of the ventral hindwing are conspicuous wing pattern el- ements in some Cyanophrys species, but the dark maroon and brown scales occur in all Cyanophrys and the extent of their expres- sion shows almost continuous variation among species and sometimes within a spe- cies. Male wing venation.?Male forewing venation varies within Cyanophrys, and we code it as two characters. Character 2: Male forewing veins R3 and Ml (0) connected by vein UDC, which may be poorly developed (Figs. 3, 5-7), (1) fused at the origin of vein Ml (vein UDC absent) (Fig. 4). Males of C. pseudolongula aie coded with a question mark because they are geographically variable for fore- wing venation, with both states occurring in males from Ecuador and northern Peru east of the Andes. Character 3: Male forewing veins R2 and R3 (0) arise from the discal cell (Figs. 3-4, 6), (1) are stalked (Fig. 5). Androconia.?There are three major an- droconial patterns. Males of some Cyano- phrys species have two clusters of andro- conia; one at the base of veins R2, R3, and Ml and one along the base of vein M3 (Figs. 4-5). Other species have androconia only at the base of veins R2, R3, and Ml (Figs. 3, 6-7), although this cluster is some- times small, as in C. remus. Finally, males of C. fusius are unique among Cyanophrys 402 PROCEEDINGS OF THE ENTOMOLOGICAL SOCIETY OF WASHINGTON Fig. 2. 60? 50? 40? 30? Distribution of Cyanophrys bertha in southern Brazil from 13 museum specimens (Appendix). in having brown dorsal forewings (irides- cent blue in the others) that lack androcon- ia. Because it lacks androconia, C. fiisius was coded ? (inapplicable) for characters 5-6. Character 4: Male dorsal forewing (0) with androconia only at the base of veins R2, R3, and Ml (Figs. 3, 6-7), (1) with androconia at the base of veins R2, R3, and Ml and along the base of vein M3 (Figs. 4?5), (2) lacking androconia. Character 5: Dorsal forewing androcon- ial cluster at the base of veins R2, R3, and Ml (0) extends into cell R2-R3 (Figs. 6-7), (1) absent from cell R2-R3 (Fig. 3-5). Character 6: Basal edge of dorsal fore- wing androconial cluster at the base of veins R2, R3, and Ml (0) located at, or just basal to, the origin of vein R2 (Figs. 3, 5- 7), (1) extending basally well beyond the origin of vein R2 to, or almost to, the origin of vein Rl (Fig. 4). Wing shape and hindwing tails.-Despite variability in wing shape and number of hindwing tails, we code only one such char- acter We did not code number of tails be- cause it can vary between males and fe- males (C. languid), geographically (C Her- odotus, C. acaste), or at one locality (C. amyntor, C. herodotus). We also did not code wing shape because there appears to be a continuum of variation from the round- ed wings of C. goodsoni to the angular wings of C agricolor. The character below, however, is distinct and unambiguously scored. Character 7: Hindwing anal lobe (0) not produced posteriorly (<0.1 mm beyond the outer margin of cell Cu2-2A) (Fig. 7), (1) produced posteriorly, at least 0.25 mm be- yond the outer margin of cell Cu2-2A (Figs. 3, 6). Male genitalia.?We found considerable interspecific variation in the male genitalia and its associated dorsal brush organs. Var- iation of some structures, such as length of the saccus (Table 2), phallobase (internal part of the penis), entire penis, and tegu- men, was not phylogenetically useful be- cause differences in means between species were small (typically less than one standard deviation), which made it difficult to rec- VOLUME 107, NUMBER 2 403 R3 R2 RI Ml R3 R2 RI Figs. 3-5. Forewing and hindwing venation witli detail of androconial clusters. 3, Cyanophrys bertha. Arrow at top points to the upper discocellular vein (UDC), which arises distal of the origin of vein R2. Arrow at bottom points to the posteriorly produced hindwing anal lobe. 4, C. herodotus. Detail shows anastomosis of veins R3 and Ml; there is no upper discocellular vein. Arrow points to androconia along the base of vein M3. 5, C. acaste. Veins R2 and R3 are stalked. Arrow points to androconia along the base of vein M3. The upper discocellular vein arises basal of the origin of vein R2. ognize distinct states. We were able to code Character 8: Dorsal male genitalia brush other aspects of interspecific variation. Be- organs (0) absent, (1) present (Figs. 8-10). cause they lack brush organs, the outgroup Character 9: Male genitalia valvae (0) species were coded? (inapplicable) for char- with a terminal thickening that extends acters 11-14. along the inner margin of the valva (Fig. 3 404 PROCEEDINGS OF THE ENTOMOLOGICAL SOCIETY OF WASHINGTON Table 2. Mean length in mm of the saccus mea- sured in ventral aspect along the sagittal plane. Species are listed in ascending rank order with differences be- tween succeeding species typically less than a standard deviation. Standard Species Mean Length Deviation Sample Size C. acaste 0.53 0.050 7 C. remus 0.54 0.036 5 C. amyntor 0.61 0.086 5 C. agricolor 0.64 0.042 4 C. pseudolongula 0.68 0.036 3 C. longula 0.69 0.035 4 C. bertha 0.72 0 2 C. herodotus 0.74 0.062 7 C. velezi 0.79 0.051 2 C. goodsoni 0.79 0.204 2 C. fusius 0.81 0.091 4 C. argentinensis 0.82 0.062 4 C. banosensis 0.83 0 1 C. miserabilis 0.86 0.030 5 C. crethona 1.35 0.040 2 in Wairen and Robbins 1993), (1) with a strongly demarcated terminal thickening that does not extend along the inner margin of the valva (Fig. 10, Fig. 3 in Warren and Robbins 1993), (2) with no terminal thick- ening ("not capped" in Clench 1961) (Figs. 8-9). Clench (1961) reported valva tips with a terminal thickening in subgenera Incisalia, Sandia Clench and Ehrlich, and Xamia Clench. This valva tip was illustrated by Warren and Robbins (1993), and also oc- curs in C. agricolor. A less prominent thickening that extends along the inner mar- gin of the valvae occurs in subgenera Cal- lophrys and Mitoura (Warren and Robbins 1993) as well as C. goodsoni and C. argen- tinensis. Character 10: Posterior-ventral process of vinculum-tegumen (0) less than 0.15 mm in length (Figs. 8, 10), (1) with a long (> 0.15 mm) pointed process (Fig. 9). Character 11: Anterior end of the setae comprising the dorsal male genitalia brush organs oriented (0) dorso-posteriorly (Figs. 8, 9), (1) horizontally (Fig. 10). Character 12: Distance from anterior to posterior ends of the dorsal male genitalia brush organs in lateral aspect (0) less than 1.3 mm long (Figs. 9-10), (1) more than 1.4 mm long (Fig. 8). Character 13: Anterior end of the dorsal male genitalia brush organs originates in lateral aspect (0) near or below the origin of the vinculum strut (Figs. 8-9), (1) pri- marily above the origin of the vinculum strut (Fig. 10). Character 14: Width of the dorsal male genitalia brush organs (dorsal aspect) at its widest point (0) less than 0.5 mm (Figs. 8- 9), (1) more than 0.5 mm (Fig. 10). Female genitalia.?Although length and width of the ductus bursae, development of sclerotized plates surrounding the ostium bursae, and length of the spines of the signa vary in Cyanophrys, we did not code this variation. For the ductus bursae, mean lengths and widths for species were similar, which made it difficult to delineate distinct states. Similarly, we were unable to char- acterize distinct states among species in de- velopment of the lamellae vaginalis or spines of the signa. In brief, interspecific variation in the female genitalia was not phylogenetically informative even though it was sometimes useful for distinguishing species, such as C. herodotus and C. amyn- tor. The bursa copulatrix of C. bertha is illustrated (Fig. 11). PHYLOGENETIC ANALYSES AND RESULTS Analysis of the character matrix with the Hennig86 "ie*" option, which searches ex- haustively for the most parsimonious clad- ograms, yielded 15 equally parsimonious 23-step trees with a consistency index of 0.69 and retention index of 0.86. The equal- ly parsimonious trees differed primarily in the placement of C. bertha and C. fusius, with C. bertha being the sister of either the C. longula group (C. longula, C pseudo- longula, C agricolor, C. banosensis, and C. remus) or C. herodotus group (C acaste, C amyntor, C. fusius, C. herodotus, C. mis- erabilis, C. velezi, and C. crethona). This result is represented in the consensus tree VOLUME 107, NUMBER 2 405 ("nelsen" option) as a trichotomy among C. bertha, the C. longula group, and the C. herodotus group, but the latter group is not strongly supported (Fig. 12 with jackknife values). The consensus tree is not one of the 15 most parsimonious trees. A successive weighting iteration (Farris 1969) produced three trees, each of which was one of the 15 original most parsimo- nious ones. Wing venation and frons color (characters 1-3) were weighted four or less out of 10 while the other characters (4-14) were fully weighted (10 out of 10) except for "capped" valvae (character 9) and brush organ length (character 12). In other words, androconia and male genitalia char- acters were less likely to be homoplastic than those of wing venation and frons color Carpenter (1988) argued that these trees are the best phylogenetic hypotheses. The con- sensus of the three successively weighted trees (24 steps) shows a sister relationship between C. bertha and the C. herodotus lin- eage, and there is greater resolution within the C. herodotus group (Fig. 13) than in the original consensus tree (Fig. 12). MONOPHYLY OF GENERA AND SUBGENERA Hypothesized synapomorphies for Cy- anophrys are (1) paired dorsal brush organs on the intersegmental membrane between the male genitalia vinculum and the 8* ab- dominal tergum (Character 8, Figs. 8-10) and (2) an elongate hindwing anal lobe with tan-brown coloration (Character 7, Figs. 3, 6). These synapomorphies are unique in the Callophrys Section (Robbins 2004b), but occur in other sections of the Eumaeini. However, the green underside of the wings coupled with the elongate hindwing anal lobe allow adult Cyanophrys to be identi- fied in the field. These results do not pro- vide data on whether the Holarctic mem- bers of Callophrys {sensu Clench 1961) form a monophyletic group. Cyanophrys as characterized in this paper has been partitioned into subgenera (some- times treated as genera) Cyanophrys, Ple- siocyanophrys K. Johnson, Eisele and MacPherson, Antephrys K. Johnson, Eisele and MacPherson, Apophrys K. Johnson and Le Crem, and Mesocyanophrys K. Johnson (Johnson et al. 1993, Johnson and Le Crom 1997a). The species that belong to each is noted (Fig. 12). Subgenus Cyanoprhys con- tains the C. longula group (as denoted above), C. miserabilis, C. velezi, C. cre- thona, and C roraimiensis (the latter omit- ted from the phylogenetic analysis) (John- son and Le Crom 1997a). This grouping is polyphyletic in the 15 original most parsi- monious cladograms, as summarized by the consensus tree (Fig. 12). Subgenus Plesio- cyanophrys contains C goodsoni and C. ar- gentinensis (Johnson et al. 1993), a group- ing that is paraphyletic in the 15 most par- simonious trees. Subgenus Antephrys con- tains only C. fusius (Johnson et al. 1993, Johnson and Le Crom 1991 SL, Johnson and Amarillo 1997, Robbins 2004b), making it a monotypic genus. Subgenus Mesocyano- phrys contains C acaste and C. bertha (Johnson and Le Crom 1997a), a grouping that is paraphyletic or polyphyletic in the original 15 most parsimonious trees. Final- ly, subgenus Apophrys contains C. herod- otus and C. amyntor (Johnson and Le Crom 1997a), a grouping that is not monophyletic in the 15 most parsimonious cladograms. SYNOPSIS OF SPECIES Cyanophrys goodsoni (Clench, 1946) Distribution, habitat, and abundance.? Usually uncommon in seasonally dry hab- itats from southern Texas (United States) to Guanacaste (Costa Rica). Larval food plants.?Blossom buds of Rivina (Phytolaccaceae) in Texas and on Vernonia (Compositae) in Costa Rica. Illustrations of adults.?Both sexes in Scott (1986). Cyanophrys argentinensis (Clench, 1946) Distribution, habitat, and abundance.? Uncommon to rare in both dry and wet for- ests from sea level to over 1,000 m eleva- tion from eastern Panama (Darien) and 406 PROCEEDINGS OF THE ENTOMOLOGICAL SOCIETY OF WASHINGTON s'^? ?^ \(f -y. "^ J Figs. 6-7. Forewing and hindwing venation with detail of androconial clusters. 6, Cyanophrys goodsoni. Arrow at top points to the upper discocellular vein. Arrow at bottom points to the posteriorly produced hindwing anal lobe. 7, Callophrys rubi. Arrow at top points to the upper discocellular vein. Arrow at bottom points to the hindwing anal lobe that is not posteriorly produced. northern Venezuela (Aragua) south along (Panama, Venezuela) to 0.25 mm (Peru) to the Andes to northwestern Argentina, then 0.20 mm (Argentina). eastward to Paraguay and southern Brazil. Male behavior.?A "territorial" male was Variation?Width of the ductus bursae in recorded at about 1400 hours on a hilltop, the middle in ventral aspect appears to vary Illustrations of adults.?Both sexes in geographically, from more than 0.30 mm D'Abrera (1995). VOLUME 107, NUMBER 2 407 Figs. 8-9. Male genitalia in lateral (left) and ventral aspect. 8, Cyanophrys bertha. B-brush organs, P-penis, S-saccus, T-tegumen, V-valvae, Vi-vinculum (which is fused with the tegumen in all eumaeines). 9, C. herodotus. The arrow points to the ventro-lateral process of the tegumen, which is foreshortened in this aspect. Fig. 10. Mdie gea&aXia of Cyanophrys agricolor in lateral (left) and ventral aspect. The arrow (right fig- ure) points to terminal thickening of valvae. ViS-vin- culum strut, which is an internal ridge. Cyanophrys bertha (Jones, 1912) Distribution.?Coastal mountains of southern Brazil from about 800 to 1,400 m elevation in the states of Minas Gerais, Sao Paulo, Paran?, and Santa Catarina (Fig. 2, Appendix). There are literature records from Rio de Janeiro (Brown 1993) and Rio Grande do Sul (Draudt 1919-1920), but we Icnow of no extant specimens to verify the occurrence of C. bertha in these states. The elevation of the locality where one speci- men was supposedly collected is 75 m, but it is probably mislabeled (discussed in Ap- pendix). Habitat.?The habitat of C. bertha is var- ied. Holdridge's (1947) ecological life zones of the localities in the Appendix in- clude subtropical lower montane moist for- est, subtropical moist forest, and warm tem- perate moist forest (Tosi 1983). Vegetation zones include dense evergreen forest, mixed evergreen forest, and seasonally de- ciduous forest (IBGE 1993). Although C. bertha appears to be primarily a resident of 408 PROCEEDINGS OF THE ENTOMOLOGICAL SOCIETY OF WASHINGTON Fig. 11. Female genitalia bursa copulatrix of Cyanophrys bertha in ventral aspect. moist coastal mountains, three specimens found at Barbacena (Minas Gerais) indicate that it also occurs, at least on occasion, in semi-deciduous forest and may possibly be more widespread inland than is currently realized. Identification.?The unique ventral wing pattern of C. bertha (Fig. 1) has distin- guished this species since it was described. Larval food plants.?The larval food plant and immature stages of C. bertha are unknown. Larvae of other Cyanophrys have been reared on flowers, fruits, or young leaves of plants in families Anacardiaceae, Boraginaceae, Dipsacaceae, Compositae, Leguminosae, Malvaceae, Phytolaccaceae, Sambucaceae, Sterculiaceae, Ulmaceae, and Verbenaceae. Male behavior.?Brown (1993) observed males of C. bertha setting up mating terri- tories in the crowns of trees on hilltops at Serra do Japi (cf. Appendix) in the early afternoon. Other than C. velezi, whose males set up mating territories on hilltops before 0930 hours, the afternoon male "ter- ritorial" behavior of C. bertha is typical for the genus Illustrations of adults.?Fig. 1. Cyanophrys acaste (Prittwitz, 1865) Distribution, habitat, and abundance.?A very common species in a variety of habi- tats and elevations from southern Brazil to eastern Bolivia, south to central Argentina and Uruguay. Variation.?Variation of the ventral wing pattern, particularly expression of the post- median line and a white hindwing discal bar, accounts for this species being named repeatedly, usually as a subspecies. Individ- uals from the southern temperate parts of the range are tailed (C. acastoides pheno- type). Larval food plants.?Chuquiraga (Com- positae) in Brazil. Male behavior.?Males "perch" on hill- tops from about 1100 to 1500 hours. Illustrations of adults.?Both sexes in D'Abrera (1995). Cyanophrys amyntor (Cramer, 1775) Distribution, habitat, and abundance.? Widespread and common from sea level to about 1,100 m from northern Mexico to southern Brazil. Recorded from Texas, United States (Kendall and McGuire 1984), based on a female in the Illinois Natural History Survey. Johnson and Le Crom (1997b) stated that it was a misidentified female of C herodotus. Robbins (unpubl.) confirmed their identification. Variation.?Individuals in the southern parts of its range may have or lack tails, but do not differ otherwise. The name Cyano- phrys caramba (Clench) refers to tailless individuals. Identification.?Slightly larger on aver- age than sympatric C. herodotus, but males are distinguished by different androconial VOLUME 107, NUMBER 2 409 1 h-O- 0 Callophrys (Callophrys) rubi Callophrys (Mitoura) gryneus ? Callophrys (Incisalia) niphon 1 -?-? (75) 1 1 Cyanophrys C. (Plesiocyanophrys) goodsoni 3 r-0- 1 (54) C. (Plesiocyanophrys) argentinensis 1 1 C. (Mesocyanophrys) bertha 2 1?0 C. (Cyanophrys) longula 0 2 11 13 14 ? ? ? (64) 1 1 3 r-O C. (Cyanophrys) pseudolongula C. (Cyanophrys) remus C. (Cyanophrys) banosensis ? C. (Cyanophrys) agricolor 1 ? C. (Mesocyanophrys) acaste C. (Apophrys) amyntor 2 4 1 2 2 6 10 12 1 1 .?-?- C. (Antephrys) flisius C. (Apophrys) herodotus ? C. (Cyanophrys) miserabilis ? C. (Cyanophrys) velezi 1 L_0_ 0 C. (Cyanophrys) crethona Fig. 12. Strict consensus of the fifteen equally most parsimonious cladograms (23 steps, ci = 0.69, ri = 0.86) for Cyanophrys species. Callophrys rubi, C. gryneus, and C niphon (top) are outgroups. Character numbers are placed above nodes and character state numbers below nodes. Open circles represent reversal or convergence of the character state at that node. Jackknife values are noted in parentheses. Subgeneric placements for Cyano- phrys from Johnson et al. (1993), Johnson and Le Crom (1997a), and Johnson and Amarillo (1997) patches on the dorsal fore wing. Females of the two have different genitalia. Larval food plants.?Reared from Ul- maceae {Celtis, Trema) in Mexico, Ecuador, and Brazil (Kendall 1975, Hoffmann 1937), and from Verbenaceae (Clerodendron) in Trinidad. Male behavior.?Males set up mating ter- ritories along trail edges and on hilltops from about 1200 to 1600 hours. Illustrations of adults.?Male in D'Abrera (1995), but identity of the female that he illustrated as this species is unclear. Cyanophrys fiisius (Godman and Salvin, 1887) Distribution, habitat, and abundance.? Seasonally dry habitats from northern Mex- ico to Colombia and Venezuela (Clench 1946). Identification.?Males are brown above, not blue, and lack forewing androconia, both unique traits in Cyanophrys. Clench (1946) noted the similarity of females of this and the previous species, but female C. fusius have a short white-tipped tail at the end of vein Cul while Central American 410 PROCEEDINGS OF THE ENTOMOLOGICAL SOCIETY OF WASHINGTON longula pseudolongula Fig. 13. Strict consensus of the three most parsimonious successively weighted trees for Cyanophrys species. Characters of wing venation and frons color were under-weighted because they were homoplastic. females of C. amyntor have only a black stub at the end of vein Cul. Larval food plants.?Reared from Arra- bidaea (Bignoniaceae) in Costa Rica. Male behavior.?A "territorial" male was collected on a ridge top at 1500 hours. Illustrations of adults.?Male in D'Abrera (1995), but identity of the female that he illustrated as this species is unclear. Cyanophrys herodotus (Fabricius, 1793) Distribution, habitat, and abundance.? The most common, widespread, variable, and weedy species in Cyanophrys, occur- ring in almost all habitats under 1,500 m throughout the Neotropics. Variation.^Size and wing pattern of this species are highly variable, with the largest individuals being most frequent in the southern part of the range. Individuals from northwestern Peru, western Ecuador and Colombia, and northern Colombia lack a tail (to which the name C. detesta refers), but do not otherwise differ. Specimens with and without a tail occur sympatrically and synchronically in El Salvador and southern Brazil, but again, do not otherwise differ. Despite substantive genitalic variation, in- cluding clinal differences, we find no con- vincing evidence of sibling species. Larval food plants.?A polyphagous spe- cies that has been reared from plants in the VOLUME 107, NUMBER 2 411 Anacardiaceae, Boraginaceae, Compositae, Dipsacaceae, Malvaceae, Sambucaceae, and Verbenaceae (Lima 1928, 1930, 1936; San- tos 1933; Monte 1934; Biezanko et al. 1974; Robbins and Aiello 1982). Male behavior.?Males set up mating ter- ritories on hilltops and ridges in the middle of the afternoon from about 1300 to 1600 hours at slightly greater heights above the ground, on average, than males of C. amyn- tor. Illustrations of adults.?Male in D'Abrera (1995), but identity of the female that he illustrated as this species is unclear. Cyanophrys miserabilis (Clench, 1946) Distribution, habitat, and abundance.? Occurs in seasonally dry habitats from sea level to about 2,000 m elevation from southern United States (Texas) to the Pacif- ic side of Costa Rica. Identification.?It is easily distinguished from the sympatric C. herodotus by the brown frons (green in C. herodotus) and by the greater amount of dark maroon scaling along the outer margin of the ventral hindwing. Cyanophrys miserabilis appears to form a superspecies (monophyletic line- age in which no two species are sympatric) with the next two species (Figs. 12, 13). Besides the presumed evolution of a brown frons in the ancestor of these species, they also possess "thin, stringy" brush organs. Larval food plants.?Recorded larval food plants are Compositae (Eupatorium) and Leguminosae (Caesalpinia, Parkinson- ia). Illustrations of adults.?Both sexes in D'Abrera (1995). Cyanophrys velezi Johnson and Kruse 1997 Distribution, habitat, and abundance.? Occurs in Panama, where it has been re- corded only during the dry season, to west- ern Ecuador in seasonally dry habitats. Identification.?Unlike C miserabilis, it lacks a hindwing tail and dark maroon scal- ing along the outer margin of the ventral hindwing. Male behavior.?Males "hilltop" in the morning before 0930 hours, which is unique in the genus, so far as is known, but which may be shared by the preceding and succeeding species. Illustrations of adults.?This species has been illustrated, so far as we are aware, only in the original description. Cyanophrys crethona (Hewitson, 1874) Distribution, habitat, and abundance.?A Jamaican endemic. Identification.?The male genitalia of C. crethona is similar to the preceding two species, but is distinguished by its larger size (i.e.. Table 2) and a very lightly scler- otized ventral cornutus. Nomenclature.?Cyanophrys hartii Turner and J. Y. Miller differs from C. cre- thona by ventral brown coloration. Because exposure to humidity or physical abrasion can change scale color from green to brown, as mentioned in the discussion of Character 1, this name was synonymized with C. crethona (Robbins 2004b). Illustrations of adults.?A male is illus- trated in D'Abrera (1995). Cyanophrys roraimiensis Johnson and Smith, 1993 Distribution, habitat, and abundance.? Known only from the holotype female, which was collected on the Brazilian side of Mt. Roraima. We have seen a picture of a female collected in the adjoining tepui re- gion of Venezuela that may be this species, but have not had the opportunity to exam- ine it. Identification.?The ventral hindwing pattern and size of this species is exceed- ingly similar to that of C. crethona (Huntington 1933, Comstock and Huntington 1943), with which it also shares a brown frons, for which reason we suspect that it is closely related to the preceding three species. Nomenclature.?Johnson and Smith VOLUME 107, NUMBER 2 413 It probably occurs in Bolivia, but we know of no records. Illustrations of adults.?A female is il- lustrated in D'Abrera (1995). Cyanophrys remus (Hewitson, 1868) Distribution, habitat, and abundance.?A very common species in the mountains of southern Brazil, Argentina (Misiones), Uru- guay, and Paraguay. In the subtropical parts of its range, it occurs at lower elevations. Identification.?Males have brilliant dor- sal blue color, similar to that in C. longula and C. pseudolongula, but the underside wing pattern has extensive brown markings, similar to the preceding two species. Larval food plants.?It has been reared from Calliandra (Leguminosae) in Uruguay and from Abutil?n and Pavonia (Malva- ceae) in Brazil (Zik?n 1956, Biezanko et al. 1966, Silva et al. 1968). Male behavior?Males "hilltop" in the early afternoon, with records from 1145 to 1500 hours. Illustrations of adults.?Both sexes in D'Abrera (1995). DISCUSSION Building on the work of Clench (1961), Robbins (2004a) distinguished the Callo- phrys Section of the Eumaeini by valve tips that are ?attened and without setae, but sometimes with a terminal thickening (il- lustrated in Warren and Robbins 1993). The form of the anal lobe (Character 7) and presence of brush organs (Character 8) dis- tinguish Cyanophrys from other members of the Callophrys Section. Although the ge- nus could be split into smaller genera on the basis of the phylogenetic results (Figs. 12 and 13), the anal lobe character allows individuals to be recognized in the field. It is unclear whether the remaining Holarctic members of Callophrys, or the many genera into which they have been divided, are monophyletic. The subgeneric nomenclature used by Johnson and colleagues (Johnson et al. 1993, Johnson and Le Crom 1997a, John- son and Amarillo 1997) is inconsistent with the phylogenetic results (Fig. 12). Four of these subgenera are not monophyletic and one is monotypic. In some cases, such as the aptly named Plesiocyanophrys, they ap- pear to have been characterized with sym- plesiomorphies. In others, such as the char- acterization of Cyanophrys (a brown frons), they were delimited by a homoplastic char- acter The C. longula and C herodotus groups are monophyletic on the 15 most parsimo- nious trees. The C longula group consists of species with a brown frons (Character 1, which is homoplastic), brush organs wider than 0.05 mm (Character 14, Fig. 10), brush organs that are oriented horizontally at the anterior end (Character 11, Fig. 10), and brush organs that arise primarily above the origin of the vinculum strut (Character 13, Fig. 10). The latter three character states are unique in the Callophrys Section. These species occur in montane habitats (above 800 m) except in the subtropical parts of southern South America. The C. herodotus group consists of species with two andro- conial clusters on the dorsal forewing (Character 4, Figs. 4-5, both lost in C fu- sius according to our results). All seven species occur in the lowlands, unlike the previous lineage, but most are also found in montane habitats. Cyanophrys bertha is the sister of the seven-species C herodotus lineage in the successively weighted cladogram (Fig. 13), but in the equally weighted original most parsimonious trees, it was sometimes the sister of the five-species C. longula lineage. The genitalia of C. bertha are nearly iden- tical to those of C. amyntor, but C. bertha shares frons color, androconial structure, and restriction to montane habitats with C. longula (Table 1). Cyanophrys bertha has been proposed for "vulnerable" status (Brown 1993, Brown and Freitas 2000 and references therein. Otero et al. 2000) and listed as "al- most threatened" (Mielke and Casagrande 2004). Its relatively basal position in the 414 PROCEEDINGS OF THE ENTOMOLOGICAL SOCIETY OF WASHINGTON cladogram increases its priority as a species of conservation concern (Atkinson 1989, Vane-Wright et al. 1991). ACKNOWLEDGMENTS We are grateful to Keith Brown for his kind hospitality, for generously sharing his knowledge of C. bertha, and for providing us with study specimens. We thank Mary Sangrey and the NMNH Research Training Program for providing an internship to Duarte; the Women's Committee of the Smithsonian Institution and the National Science Foundation (award #BIR-9531331) for supporting this program; Olaf Mielke and Mirna Casagrande for obtaining CNPq support for Robbins to curate the UFPR collection in Curitiba; Karl Ebert for allow- ing us to examine his father's collection; the curators and collections managers at the AMNH, BMNH, UFPR, INHS for allowing us to examine specimens under their care; the many biologists who shared larval food plant information with us; Carl Hansen for technical help in digitizing wing venation directly from slides; George Venable for making the figures; Astrid Caldas for pro- viding a smooth interface between English and Portuguese; and especially Andrew Brower, John Brown, Keith Brown, Astrid Caldas, Larry Gall, Jason Hall, Donald Har- vey, Gerardo Lamas, James Miller, Carla Penz, J. B. Sullivan, and Andy Warren for commenting, sometimes extensively, on the manuscript. The junior author has been fi- nancially supported by the State of Sao Paulo Research Foundation?FAPESP (02/ 13898-0 and 03/05895-4). LITERATURE CITED Atkinson, I. 1989. Introduced animals and extinctions, pp. 54-69. In Warren, D. and M. Pearl, eds. Con- servation for the twenty-first century. Oxford Uni- versity Press, New York, xv + 365 pp. Bailowitz, R. A. and J. E Brock. 1991. Butternies of Southeastern Arizona. Sonoran Arthropod Stud- ies, Inc., Tucson, ix + 342 pp. Biezanko, C. M. De, Ruffinelli, A. and C. S. Carbonell. 1966. Lepidoptera del Uruguay. III. Notas com- plementarias. Bolet?n Facultad de Agronom?a (Montevideo) 91, 53 pp. Biezanko, C. M. De, Ruffinelli, A. and D. Link. 1974. Plantas y otras sustancias alimenticias de las oru- gas de los lepid?pteros uruguayos. Revista do Centro de Ciencias rurais 4: 107-148. Brown Jr, K. S. 1993. Selected Neotropical species, pp. 146-149. In New, T. R., ed. Conservation Bi- ology of Lycaenidae (Butterflies) lUCN, Gland, Switzerland. Brown Jr, K. S. and A. V. L. Freitas. 2000. Diversidade de Lepidoptera em Santa Teresa, Espirito Santo. Boletim do Museu de Biolog?a Mello Leit?o (N. S?r.) 11/12: 71-118. Carpenter, J. M. 1988. Choosing among multiple equally parsimonious cladograms. Cladistics 4: 291-296. Clench, H. K. 1946. Notes on the amyntor group of the genus Thecla (Lepidoptera, Lycaenidae). En- tomologist 79: 152-157, 185-191. . 1961. Tribe Thechni, pp. 177-220. In Ehrlich, P. R. and A. H. Ehrlich, eds. How to know the butterflies. Brown Company, Dubuque, Iowa, 261 pp. Comstock, W. P. and E. I. Huntington. 1943. Lycaen- idae of the Antilles. Annals of the New York Academy of Sciences 45: 49-130. D'Abrera, B. L. 1995. Butterflies of the Neotropical region. Part VII. Lycaenidae. Hill House, Black Rock, i-xi + 1098-1270 pp. Draudt, M. 1919-1920. Theclini F, pp. 744-812. In Seitz, A. ed. Macrolepidoptera of the World, Vol. V, The American Rhopalocera, Alfred Kernen Verlag, Stuttgart, 1139 pp., 194 pi. Eliot, J. N. 1973. The higher classification of the Ly- caenidae (Lepidoptera): A tentative arrangement. Bulletin of the British Museum of Natural History (Entomology) 28(6): 371-505. Farris, J. S. 1969. A successive approximations ap- proach to character weighting. Systematic Zoolo- gy 18: 374-385. Hoffmann, E 1937. Beitr?ge zur Naturgeschichte bras- ilianischer Schmetterlinge. Entomologisches Jahr- buch (Frankfurt) 1937: 126-139. Holdridge, L. R. 1947. Determination of world land formations from simple climatic data. Science 105: 367-368. Huntington, E. I. 1933. Erycinidae and Lycaenidae (Lepidoptera, Rhopalocera) from the regions of Mt. Roraima and Mt. Duida. American Museum No vitales 611: 1-5. IBGE, Diretoria De Geoci?ncias. 1993. Mapa de ve- geta??o do Brasil. Rio de Janeiro, Brazil. Janzen, D. H. and W. Hallwachs. 2004. http://janzen. sas.upenn.edu Johnson, K. and A. R. Amarillo S. 1997. Review of Antephrys with description of a new Colombian species. Revista de Theclinae Colombianos, Vol. 1, 16 pp. Johnson, K., R. C. Eisele, and B. N. MacPherson. VOLUME 107, NUMBER 2 415 1993. Genera for the small green hairstreak but- terflies of the Neotropical realm (Lepidoptera, Ly- caenidae, Theclinae). Reports of the Museum of Natural History, University of Wisconsin (Stevens Point) 39: 1-16. Johnson, K. and J. F. Le Crem. 1997a. The "brown- frons" complex of Cyanophrys sensu lato and new sister species from Colombia and Brazil. Re- vista de Theclinae Colombianos, Vol. 1, 28 pp. . 1997b. Clench's Cyanophrys "amyntor Group" in Colombia with descriptions of several outlying endemics from elsewhere in South Amer- ica. Revista de Theclinae Colombianos, Vol. 1, 21 pp. Johnson, K. and D. S. Smith. 1993. A remarkable new butterfly species from Jamaica (Lepidoptera: Ly- caenidae), with notes on Jamaican endemics and their sister species. Reports of the Museum of Natural History, University of Wisconsin (Stevens Point) 24: 1-13. Kendall, R. O. 1975. Larval foodplants for seven spe- cies of hairstreaks (Lycaenidae) from Mexico. Bulletin of the AUyn Museum 24: 1-4. Kendall, R. O. and W. W McGuire. 1984. Some new and rare records of Lepidoptera found in Texas. Bulletin of the AUyn Museum 86: 1-50. Klots, A. B. 1970. Lepidoptera, pp. 115-130. In Tux- en, S. L., ed. Taxonomist's Glossary of Genitalia in Insects. Munksgaard, Copenhagen, 359 pp. Lamas, G., R. G. Robbins and W. D. Field. 1995. Bib- liography of butterflies. An Annotated Bibliogra- phy of the Neotropical Butterflies and Skippers (Lepidoptera: Papilionoidea and Hesperioidea). In Heppner, J. B., ed. Atlas of Neotropical Lepidop- tera. Gainesville, Scientific Publishers, Vol. 124, xiv + 463 pp. Lima, A. M. C. 1928. Segundo cat?logo systematico dos insectos que vivem ?as plantas do Brasil e ensaio de bibliografia entomol?gica brasileira. Ar- chivos da Escola superior de Agricultura e Med- icina veterinaria. (Niter?i) 8: 69-301. . 1930. Suplemento ao 2? cat?logo system?tico dos insectos que vivem ?as plantas do Brasil e ensaio da bibliografia entomol?gica brasileira. O Campo (Rio de Janeiro) 8: 84-91. . 1936. Terceiro cat?logo dos insetos que vivem ?as plantas do Brasil. Rio de Janeiro, Ministerio da Agricultura, 460 pp. Mielke, O. H. H and M. M. Casagrande. 2004. Bor- boletas, pp. 713-739, In S. B. Mikich and R. S. B?rnils, eds., Livro vermelho da fauna amea?ada no Estado do Paran?. Instituto Ambiental do Pa- rana, Curitiba, 763 pp. Monte, O. 1934. Borboletas que vivem em plantas cul- tivadas. Publica??es. Secretaria da Agricultura. Estado de Minas-Gerais (Serie agr?cola) 21, 221 pp. Nicolay, S. S. 1971. A new genus of hairstreak from Central and South America (Lycaenidae, Thecli- nae). Journal of the Lepidopterists' Society 25, Supplement 1, 39 pp. . 1977. Studies in the genera of American hair- streaks. 4. A new genus of hairstreak from Central and South America (Lycaenidae: Eumaeini). Bul- letin of the AUyn Museum 44: 1-24. Nijhout, H. F. 1991. The development and evolution of butterfly wing patterns. Smithsonian Institution Press, Washington, DC, 297 pp. Nixon, K. C. 2002. WinClada version 1.00.08. Pub- lished by the author, Ithaca, New York. Otero, L. S., K. S. Brown, O. H. H. Mielke, R. E Monteiro, J. M. Costa, M. V. Mac?do, N. C. Ma- ciel, J. Becker, N. G. Salgado, S. B. Santos; G. E. Moya, J. M Almeida, and M. Duarte. 2000. In- vertebrados terrestres, pp. 53-62. In Bergallo, H. G., C. E D. Rocha, M. A. S. Alves, and M. Van- sluys, eds. A fauna amea?ada de extin??o do es- tado do Rio de Janeiro. Editora da Universidade do Estado do Rio de Janeiro, Rio de Janeiro. Robbins, R. K. 1991. Evolution, comparative mor- phology, and identification of the eumaeine but- terfly genus Rekoa Kaye (Lycaenidae: Theclinae). Smithsonian Contributions to Zoology, No. 498, 64 pp. . 2004a.Introduction to the checklist of Euma- eini (Lycaenidae), pp. xxiv-xxx. In Lamas, G., ed. Checklist: Part 4A. Hesperioidea?Papilionoidea. In Heppner, J. B., ed. Atlas of Neotropical Lepi- doptera, Vol. 5A. Association for Tropical Lepi- doptera, Scientific Publishers, Gainesville, xxxvi + 439 pp. . 2004b. Lycaenidae. Theclinae. Tribe Euma- eini, pp. 118-137. In Lamas, G., ed. Checklist: Part 4A. Hesperioidea?Papilionoidea. In He- ppner, J. B., ed. Atlas of Neotropical Lepidoptera, Vol. 5A. Association for Tropical Lepidoptera, Scientific Publishers, Gainesville, xxxvi + 439 pp. Robbins, R. K. and A. Aiello. 1982. Foodplant and oviposition records for Panamanian Lycaenidae and Riodinidae. Journal of the Lepidopterists' So- ciety 36: 65-75. Robbins, R. K. and G. Lamas. 2002. Nomenclatural Changes in the Neotropical Eumaeini (Lepidop- tera: Lycaenidae: Theclinae). Revista brasileira de Zoolog?a 19: 197-214. Robbins, R.K. and S.S. Nicolay. 1999. Taxonomy of Strymon toussainti, S. andrewi, S. amonensis, and S. rhaptos (Lycaenidae: Theclinae). Journal of the Lepidopterists' Society 52: 328?334. . 2002. An overview of Strymon H?bner (Ly- caenidae: Theclinae: Eumaeini). Journal of the Lepidopterists' Society 55: 85-100. Santos, E. 1933. Inimigos e Doen?as das Fruteiras. Rio de Janeiro, Biblioteca Agr?cola d'O Campo. 81 pp. Scott, J. A. 1986. The butterflies of North America. A 416 PROCEEDINGS OF THE ENTOMOLOGICAL SOCIETY OF WASHINGTON Natural History and Field Guide. Stanford Uni- versity Press, Stanford, xv + 583p. Silva, A. G. d'A. e, C. R. Gon?alves, D. M. Galvao, A. J. L. Gon?alves, J. Gomes, M. do N. Silva, and L. de Simoni. 1968. Quarto cat?logo dos insetos que vivem ?as plantas do Brasil seus par?sitos e predadores. Edi??o ampliada do "3? cat?logo dos insetos que vivem ?as plantas do Brasil" de au- tor?a do Prof. A. M. da Costa Lima. Parte II. In- setos, hospedeiros e inimigos naturals. ?ndice de insetos e ?ndice de plantas. Rio de Janeiro, Min- isterio da Agricultura. 1: xxvii -I- 622 pp.; 2: [viii] + 265 pp. TOS?, J. Jr. 1983. Provisional life zone map of Brazil at 1:5,000,000 scale. Publication of the Institute of Tropical Forestry, Rio Piedras, Puerto Rico, 16 pp. + map. Vane-Wright, R. I., Humphries, C. J. and R H. Wil- liams. 1991. What to protect??Systematics and the agony of choice. Biological Conservation 55: 235-254. Warren, A. D. and R. K. Robbins. 1993. A natural hybrid between Callophrys (Callophrys) sherida- nii and C. (Incisalia) augustinus (Lycaenidae). Journal of the Lepidopterists' Society 47: 236- 240. Willis, J. C. 1973. A Dictionary of the Flowering Plants and Ferns. 8th edition revised by H. K. Airy Shaw. University Press, Cambridge, 1,245 pp. Zik?n, J. E 1956. Beitrag zur Biologie von 12 Thecli- nen-arten. Dusenia 7: 139-148. APPENDIX The data for the 13 museum specimens of C. bertha from 7 localities in southern Brazil that we have examined are listed be- low (museum acronyms listed in Methods). If the elevation was not recorded on a spec- imen's data label, we parenthetically note the elevation for that site as it is Hsted in gazetteers. So far as we are aware, there are no other known specimens of C. bertha. 1. Minas Gerais, Barbacana, 900 m. 1? (Private collection of Karl Ebert, Santa Clara, S?o Paulo, Brazil); 12,11 May 1969 (USNM); l?, 27 May 1986 (USNM). 2. Minas G?rais, Po?os de Caldas, 1,300- 1,400 m. 1 (J and 1 9, March (Ebert). 3. S?o Paulo, Serra do Japi, 1,050-1,250 m. 1 c?, 24 March 1990 (USNM); 1 S, 25 March 1990 (USNM); \ S, lA May 1990 (USNM); 1 c?, 2 June 1990 (USNM). 4. Paran?, Curitiba, (934 m). IS, Decem- ber 1945 (UFPR). 5. Paran?, Ponta Grossa, (969 m). 1$, March 1948 (UFPR). 6. Paran?, Castro, (999 m). 1 ? (BMNH) 7. Santa Catarina, Corup?, 19 October 1975. 1 ? (UFPR). (Corup? is located at 75 m elevation at the base of the moun- tains on the road leading to S?o Bento do Sul, a well-known butterfly locality above 800 m elevation that is the likely collection place for this specimen.)