Phylogenetic Analysis of the Subtribe Ageroniina with Special Emphasis on Hamadryas ( Lepidoptera , Nymphalidae ) with an Identification Key to the Species of Hamadryas

A cladistic analysis of the genus Hamadryas was done in order to answer two questions: is the genus Hamadryas monophyletic and, what characters best define the different species groups of the genus? The computer programs WinClada ver. 1.00.08 and Nona ver. 2.8 for phylogenetic analysis were used. The 20 species of Hamadryas were analyzed together with Ectima erycinoides, Batesia hypochlora and Panacea procilla; these four genera together form the sub-tribe Ageronina of the subfamily Biblidinae. These 23 species with a total of 66 characters were included in the data matrix: 43 of external morphology, two from the fore wing spiral organ (responsible for sound production), and 20 of the male genitalia. Three different analyses, including all of the non-Hamadryas, using each species as the out-group, were done in order to compare results. A fourth analysis, using only Ectima erycinoides as an out group, was done in order to verify and compare the species groups of Hamadryas Also, three different phylogenetic attributes were mapped: biogeography, palatability and capacity to produce sound signals. An identification key to all the species of Hamadryas was prepared. All the analysis done combining Hamadryas with the other three genera suggests that Hamadryas as currently defined is polyphyletic.


Introduction
The biodiversity on earth is so enormous that it would not be possible to study it, if it were not classified [1] and phylogenetic analyses of the species permit us to approach their evolutionary relationships in order to establish hypothetical species groups on which the classification can be based.
Hamadryas Hübner (1806) is composed of twenty species ranging from the southern United States to northern Argentina [2].According to DeVries [3] it is distinguished from other butterfly genera by their spotted "calico" pattern on the upper side.Jenkins [4] on the other hand states that Hamadryas is a taxonomically confused genus which has been in great need of revision and highlights the lack of a detailed study to find consistently valid distinguishing characters.
This genus is well known for producing an audible clicking sound, an ability that has been reported for many but not all species.It has been suggested that this sound is used in courtship, defense, and territoriality [5][6][7].It is produced through a modification of the internal walls of the forewing subcostal vein which consists of a spiral organ that ends in a tympanic membrane which in turn releases the sound into the air [7,8].
In contrast to most of other nymphalid butterflies, this genus perches on tree trunks with the wings open against the substrate and the head downward [3,4,7].The reason for this particular behavior may be to increase crypsis because the butterfly's color pattern greatly resembles tree trunks; resting head-downward could make escaping from predators faster since a small jump takes the butterfly into a flying position with little energy cost [7].Only two other genera of Nymphalidae share the exact same perching behavior as that of Hamadryas: Ectima and Panacea [3,4].
The confirmed host plants of all Hamadryas species are in the genera Dalechampia and Tragia, both in the Euphorbiaceae [4].The larval body is spiny and has two upwardly elongated spiky horns on the head.All pupae in the genus bear two flattened head horns that show affinities to Ectima [3].
At present, Hamadryas is classified in the Biblidinae-Biblidini, and the most closely related genera are thought to be Panacea, Batesia, and Biblis according to Hill et al. [9] and Ectima according to Jenkins [4].The goal of this research is to propose a thorough phylogenetic analysis which better represents the evolutionary relationships among Hamadryas species in order to answer three questions: is Hamadryas monophyletic, what is the sister group of Hamadryas, and what characters best define the different species groups of the genus?Also, an identification key to the species of the genus is included in order to make the determination of species easier.
The species H. amphichloe, H. julitta, H. atlantis, H. albicornis, H. alicia, H. rosandra, and H. velutina were analyzed based on photographs of males and females (dorsal and ventral views) and drawings of the male genitalia in Jenkins [4], due to the scarcity of specimens in collections.

Preparation of Material.
Male genitalia were treated with a 10% potassium hydroxide digestion process.To analyze them a stereomicroscope Nikon model 1064565 and dissection scissors where employed, then the material was stored in glycerol vials in the Insect Museum of the University of Costa Rica.

Characters and Terminology.
The character matrix is composed of 66 characters (all binary), of which 41 were derived from the external morphology and color pattern of the wings, 2 from the color pattern of the body, 2 from the spiral organ morphology inside the fore-wing subcostal vein, and 21 from the male genitalia.Also, three phylogenetic attributes where mapped on the resulting tress: one ethological, one biogeographical, and one physiological.
The terminology of the adult external morphology follows Scoble [10].Terminology for male genitalia follows Alayo and Hernández [11].

Phylogenetic Analysis.
The first three analyses were done to test if Hamadryas is monophyletic.For that, combinations of the three genera most closely related to Hamadryas were used.The first analysis used Batesia hypochlora C. Felder and R. Felder 1862 as the outgroup with Panacea, Ectima, and Hamadryas as ingroups.The second analysis used Panacea procilla (Hewitson, 1854) as the outgroup with Batesia, Ectima, and Hamadryas as ingroups; in this analysis information concerning the biogeography, palatability, and sound production was mapped onto the phylogeny in order to find useful taxonomic criteria to support the species groups.The third analysis used Ectima erycinoides C. Felder and R. Felder 1867 as the outgroup with Panacea, Batesia, and Hamadryas as ingroups; a fourth analysis was done including as ingroup only what is considered as Hamadryas in Jenkins [4] and with Ectima as the outgroup since this genus has been considered closely related to Hamadryas [3,4].For all the analyses I used heuristic searches in WinClada ver.1.00.08 with a multiple TBR + TBR (mult.max.), unconstrained search strategy with 100 maximum trees to keep hold and without submitting tress, with one single replication (Mult * N) and one starting tree per replication.Also, the same program was used to identify character changes along the branches.Branch support was estimated with Nona ver.2.8 [12].All characters were analyzed as nonadditive with Fitch parsimony.The data matrix and character list are in Appendices B and C, respectively.
The identification key to the species was prepared using characters that are easy to distinguish with the naked eye butterfly specimen.The key follows a different logic of characteristics than the one employed by Jenkins [4] and is a new option for the identification of adult Hamadryas.

Results
The analyses show the phylogenetic relationships obtained between these four closely related genera using Batesia (Figure 1), Panacea (Figure 2), and Ectima (Figure 3) as outgroups.According to Lamas [2] these three genera together with Hamadryas comprise the subtribe Ageronina.
The first analysis used Batesia as an outgroup, two primary trees were obtained (Appendix D) and a majority rules consensus compromise tree was done (Figure 1).The second analysis used Panacea as an outgroup and a single tree was obtained from a heuristic analysis (Figure 2).The third analysis used Ectima as an outgroup, two primary trees were obtained (Appendix E), and a majority rules consensus compromise tree was done (Figure 3).The fourth analysis gave a single tree obtained from a heuristic analysis using Ectima as an outgroup of Hamadryas (Figure 4).
Trees in Figures 5, 6, and 7 are the result of using the tree in Figure 2 with biogeography, palatability, and sound production, respectively.The biogeographical inferences were done based on the work of Heppner [13].The characters found to form the different species groups (Figure 2) are given in Table 1.
Drawings of different structures which are important for the interpretation of the species key (taxonomic characters) are shown in Figures 8 and 9.

Discussion
Following a single cladistic methodology, four different analyses produced four similar trees, although, some important dissimilarities can be pointed out in order to understand the evolutionary relationships of the groups.The CI (Consistency Index) of all the resulting trees is between 0.39 and 0.50 which is considered appropriate under the scenario presented by [14,15].The use of three other genera together with Hamadryas permits us to test its monophyly [16].The three analyses where other genera besides Hamadryas are included agree that Hamadryas is polyphyletic, since there is a Hamadryas species group that in all the analyses has a different ancestor than the rest of the Hamadryas species.This group is composed by three species: H. laodamia, H. arete, and H. velutina; it is interesting to note that, these species were treated in the past as a different genus, Peridromia, proposed by [17,18].The analyses from Figure 1 to Figure 3 show the relationships between Hamadryas, Panacea, Ectima, and Batesia.Also, in the fourth analysis (Figure 4), all the Hamadryas were treated with only Ectima as the outgroup, and these three species remained together, which is strong evidence for the monophyly of such a species group.
4.1.Species Group Shared in the Four Analyses.The species groups supported by the strongest evidence based on the results of this paper (because they are identical in the four analyses) are as follows.

H. laodamia, H. velutina, and H. arete (Peridromia Subgenus).
The diagnosis of Peridromia was based on characters of the male's forewings, that is, veins R 1 and R 2 arising on a single stalk and branched almost immediately, a character that is shared with H. feronia, H. guatemalena, H. iphthime, and others [4].This character is therefore of questionable validity.Instead, the outwardly convex subcostal cell on the male's hind wing, for instance, seem to be a valid synapomorphy of Peridromia.
Consistently, all the trees obtained in this study grouped these three species apart from all others, and when other related genera (Panacea, Batesia, and Ectima) are included, Peridromia is supported as a valid genus apart from Hamadryas.The synapomorphies for this species group are given in Table 1; the presence of a sex patch in males, red spots around the humeral vein on the VHW, and red markings on the ventral area of thorax are considered to be the most valuable taxonomically.Also, in the trees in Figures 5, 6, and 7 the phylogenetic attributes are mapped, and such attributes are also shared on the three species and also are shared with their closest non-Peridromia-taxa: Batesia, these characteristics are probably carried from their hypothetical common ancestor.
Under the scenario where Peridromia should be treated as a genus, three important characteristics might be controversial: the presence of the spiral organ inside the Forewing subcostal vein (even when no audible sound has been demonstrated) [4], the presence of two lateral rammi on the terminal sternite of the abdomen and similarities with Hamadryas in the early stages and host plants [3].The fact that Peridromia has very distinct characters from all other Hamadryas could instead (and in light of these other characters) be considered evidence for subgenus-level synapomorphies, rather than genus-level synapomorphies; in this scenario the ancestor produced sound but in these species sound (at least that audible to humans) production has been lost even though the organ is still there, as may have happened in species of Satyrinae [19].This species group probably originated in the "Yungas Montane" biogeographical province in South America and then one species, H. laodamia, moved northwards, the subspecies "saurites" through the "Central American woodlands."

H. atlantis, H. chloe and H. albicornis.
Another wellsupported species group is the one composed by H. chloe and its sister species, H. albicornis, plus H. atlantis.The affinity of these three species is also supported by recent morphological studies on the spiral organ [20], which includes synapomorphies such as a modification of the spiral organ located in the subcostal vein of the forewings and the lack of the ability of produce human-audible sound.Evidence for the affinities of these three species is not only morphological and ethological, but also, these are the only three species of Hamadryas that do not have an anterior projection in the male's vinculum (Table 1).H. chloe and H. albicornis probably evolved in the South American "Amazonian Rain Forest" from the same ancestor from which H. atlantis evolved allopatrically in North America, in the "Central American woodlands," and the "Guerreran woodlands" (Figure 5).and H. arinome in the "Amazonian Rain Forest," but its lineage migrated north and colonized Central America, North America and the Northern Caribbean Islands "Cuba" (Figure 5).

H. februa and H. glauconome.
These two species are the sister group of the last species group.They evolved from a continental ancestor which was distributed throughout practically all of the Neotropical region.H. glauconome, instead, evolved from the northern population of this hypothetical ancestor in the "Central American woodlands" or the "Guerreran woodlands."

Conclusions
The close relation between Panacea and Batesia was proposed by Fruhstorfer [21] and then supported by DeVries et al. [22] and Hill et al. [9].However, in their phylogenetic analysis Hill et al. [9] proposed a clade composed of Panacea and Batesia as the sister group of Hamadryas; in the present analysis (Figure 2).another clade, together with Batesia (Figure 2), or alone (Figure 1).The analysis in Figure 1 seems to be more probable because the early stages and host plants of Batesia might be retained from its ancestor, Panacea, which would have evolved only once, leaving the early stages of Peridromia as a homoplasic regression.
Ectima has been considered to be very closely related to Hamadryas by Jenkins [4,23] and DeVries [3], which is consistent with the analyses in Figures 1 and 2. The larvae, pupa and behavior of Ectima are very similar to that of Hamadryas but the adults do not possess the spiral organ used in sound production.This evidence suggests that Ectima could be the sister group of Hamadryas.In the analysis of Figure 4 all the current species of Hamadryas are analyzed with Ectima as the outgroup, and again the species group of "Peridromia" kept their independence as the sister group of all the other species.
In all the analyses done in this work, H. arete, H. laodamia, and H. velutina composed a monophyletic group and showed a stronger relation to Panacea and Batesia than to the rest of the Hamadryas species.
Jenkins [4] also supported the close relation between H. albicornis and H. chloe, but placed them in the H. februa Hübner (1819) species group together with H. atlantis, H. februa, H. amphichloe, H. glauconome, and H. julitta.This species group is recognized by veins R 1 and R 2 arising separately from the radial sector before the branching of R 3 and is supported in the analysis in Figure 4.
Also, in Figure 4 there is another species group composed by H. fornax, H. alicia, H. rosandra, H. amphinome, and H. belladonna.Jenkins [4] was uncertain whether H. alicia and H. rosandra should be placed in this species group, but the new evidence presented here confirms such a group but excludes H. arinome, H. feronia and H. iphtime (which according to Jenkins are included in this group).According to the analysis in Figure 4, similarities such as the dorsal forewings of H. arinome and H. amphinome may be considered as convergences since they have a different ancestor; H. arinome can then be considered as the most basal species of Hamadryas species if Peridromia is excluded from the genus.
More phylogenetic analyses in which early-stage characters are included, in addition to molecular analyses, may provide more evidence as to whether Peridromia should be ranked as a genus in order to make Hamadryas monophyletic.
The tree in Figure 2 is here considered to be the most probable scenario regarding the evolutionary relationships  between Hamadryas and the other three genera included in the analysis.This is because it is the only analysis using the four genera of the subtribe Ageronina in which only one most parsimonious tree was obtained.This scenario suggests the evolution of two different early-stage and host plant lineages: the "Panacea and Batesia-like" which is the root of the tree and the "Ectima-like," which is retained by all species of Hamadryas that evolved from the Ectima-Hamadryas hypothetical ancestor.The "Panacea and Batesia-like" early stages and host plants showed a homoplasic regression in the Peridromia group of Hamadryas since they are more similar to the rest of the "Ectima-like" early stages and host plants [3].
This analysis is the most parsimonious of all the analyses done because it is the one in which the early stages and host plants of Panacea and Batesia evolve only once on a primary most parsimonious tree, with Panacea as the sister genus of Batesia-Peridromia and Ectima-Hamadryas.
The palatability to predators in the genus Hamadryas has been studied by Chai [24,25] and Srygley and Chai [26].They provided important statistical evidence of rejection of certain colorations of models by flying predators.According to the present results and observations of DeVries [3], DeVries et al. [22], and Hill et al. [9], it is possible to presume which hypothetical ancestors might be responsible for this strategy and the way it was retained or lost by the different species groups.For instance, it is possible that due to the scarcity and/or lack of studies on species such as H. arete and H. velutina, no evidence of unpalatability has been documented.Although, it has been proven that the closely related species H. laodamia is distasteful, and since these three species evolved from a probable distasteful ancestor it is possible to infer that these two species are also unpalatable to predators.All Peridromia species together with Batesia probably inherited their unpalatability from Panacea, and their sister clade Ectima-Hamadryas lost it as a synapomorphy, although H. arinome and H. amphinome probably recovered the unpalatability as a homoplasic regression.
The capacity of the studied species to produce human audible clicking noise is also mapped.This ethological attribute is consistent with the unpalatability one.The basal taxa: Panacea, Batesia, and Peridromia lack this ability.It is only in the sister group of Peridromia and Batesia where this character appears.H. julitta seems to be the most plesiomorphic species of Hamadryas which has this ability and it is the sister species of all the rest of Hamadryas.However, the group of H. atlantis, H. chloe, and H. albicornis has lost this ability, and through a different ancestor the same happened to H. alicia, and again from a third different ancestor to H. rosandra.This convergence makes sense since the last four of these five species share the same dark forest habitat of the "Amazonian Rain Forest" bioprovince.For some reason, in this habitat noise production is not as effective as in others.
The key presented in this paper uses new characters obtained from the careful analysis of the color patterns which helps in the identification of the species; some of the characters are taken from the key presented by Jenkins [4], but are arranged in a different order (Figure 9).All the others are original and the interpretation of the key must be done with Figures 8 and 9.The characters selected do not follow any phylogenetic order.(2) Presence of at least two submarginal ocelli on DHW (0), absence of ocelli (1).

Key to Species of Hamadryas
(3) Absence of sex patch on DHW (0), existence of such structure (1).
(13) On males Forewings the discal cell closes right on the bifurcation of M 3 -Cu 1 (0), the discal cell close anteriorly of such bifurcation (1).( 14) Forewing apex closer to the body than distal border (0), Forewing apex more distant to the body than distal border (1).
(15) Absence of white or gray inside the discal cell of DFW (0), presence of such a pattern (1).
(27) Absence of a white marginal macula on VFW from m 2 to m 3 and another from cu 1 to cu 2 (0), presence of both macula (1).
(29) Absence of red spots at the side of the humeral vein on VHW (0), presence of such spots (1).
(31) Females presenting a submarginal white macula on VFW from r 5 to m 1 (0), devoid of such a pattern (1).(32) Absence of a proximal red spot on VHW from Sc + r 1 to r 5 (0), presence of such a spot (1).
(33) Females present white coloration in other parts of the DFW besides the white diagonal band (0), devoid of such a pattern (1).
(36) Absence of a coastal depression at the distal end of the discal cell followed by a bump on the Forewings (0), presence of such a depression (1).
(37) Females DFW with the diagonal white band not fragmented (0), diagonal white band fragmented in many parts or absent (1).
(39) No sexual dimorphism present in the way that discal cell close on the Forewings (0), sexual dimorphism present in such a character (1).

Male Genitalia Characters
(46) Distal point of the arm of the gnathos do not point to the distal point of the uncus (0), distal point of the arm of the gnathos pointing to the uncus (1).
(48) Arm of the gnathos with elbow of more than 90 (49) Point of the arm of the gnathos acute (0), devoid of such a pattern (1).
(50) Arm of the gnathos finishing in any shape except a thick triangle (0), finishing in a thick triangle (1).
(62) Saccus longer than the rest of the genitalia (0), same or shorter than the rest of the genitalia (1).
(66) Anterior portion of tegument less or same long than the projection of the vinculum (0), longer than the projection of the vinculum (1).

Figure 6 :
Figure 6: Palatability of the species mapped on the tree obtained from analyses of the Figure 2. The species that lack information of unpalatability are by default considered palatable.

Figure 7 :
Figure 7: Human audible click noise production mapped on the tree obtained from analyses of the Figure 2.

( 1 ) 19 ( 2 ) 6 ( 3 )
(a) Submarginal ocelli in DHW from M 3 -Cu 2 circular or tear shaped-2 (b) Submarginal ocelli in DHW from M 3 -Cu 2 absent-(a) Submarginal ocelli in DHW enclosing only one blue spot-3 (b) Submarginal ocelli in DHW enclosing other ocelli or moon-shaped spots of any color-(a) VFW with a white spot or macula in R 3 -R 4 anterior of the bifurcation of R 4 -R 5 -4 (b) VFW with no white spot or macula in R 3 -R 4 just anterior of the bifurcation of R 4 -R 5 -H.arinome (4) (a) DFW with a white wide median band-5 (b) DFW without the white wide median band-H.belladonna ♂

(36 : 1 ) 7 ( 7 )
Presence of a depression followed by an outward curve on the costal vein at the end of the discal-cell of forewings (37 : 1) Females without diagonal wide and regular white band on DFW(5) (a) DFW in Cu 1 -Cu 2 with the white macula of the median band enlarged until reaching the point of the bifurcation of M 3 -Cu 1 -H.belladonna ♀ (b) DFW in Cu 1 -Cu 2 with the white macula of the median band finishing two mm distally from the point of the bifurcation of M 3 -Cu 1 -H.amphinome (6) (a) VHW with Submarginal ocelli absent in M 2 -M 3 -H.alicia (b) VHW with Submarginal ocelli or white point present in M 2 -M 3 -(a) VHW with four consecutive ocelli of different sizes from M 1 to Cu 2 -8 (b) VHW with four consecutive ocelli of similar size from M 1 to Cu 2 -12 (8) (a) VHW ocelli from M 1 to Cu 2 , the one in M 2 -M 3 absent or if present then less than a half the diameter of the one in Cu 1 -Cu 2 -9 (b) VHW ocelli from M 1 to Cu 2 , the anterior one large, and the next three small, reduced to white points or absent-11
Two Most Parsimonious Primary Trees Obtained from the Phylogenetic Analysis of Hamadryas, Panacea and Batesia Using Ectima as Outgroup.L : 161, CI : 0.40 and RI : 0.65 See Figures 12 and 13.