SYSTEMATIC RELATIONS AMONG Philornis MEINERT , Passeromyia RODHAIN & VILLENEUVE AND ALLIED GENERA ( DIPTERA , MUSCIDAE

Passeromyia Rodhain & Villeneuve and Philornis Meinert are the only known Muscidae whose larvae are parasites of birds. Passeromyia is known from the Old World and Philornis from the New World. Opinions on the relations between these two genera and their systematic positions among the Muscidae have varied. This survey aims to clear the discussion on the relations of Passeromyia, Philornis, and of some allied genera and give an overview of the classification of the Muscidae based on cladistic methodology. Thirty-two terminal taxa (2 of them outgroups) were analysed based on 54 characters. The cladistic analysis, carried out using Hennig86, resulted in 1 minimal tree (length 373), with a consistency index of 71 and a retention index of 85. Philornis and Passeromyia belong to a monophyletic group, supported by a synapomorphy, the presence of a cocoon, enclosing the pupa. The phylogenetic relationships found in this group are: (Muscina (Philornis (Phaonina ((Fraserella, Passeromyia) (Synthesiomyia (Calliphoroides, Reinwardtia)))))). Other probable monophyletic muscid groups, like Muscinae (with Stomoxyini and Muscini) and Coenosiinae (with Limnophorini and Coenosiini) are also discussed. Phylogenetic patterns within Reinwardtinae and Dichaetomyiinae could be explained by a Gondwana distribution.


INTRODUCTION
Passeromyia Rodhain & Villeneuve and Philornis Meinert are two of the most interesting muscid genera, found in larvae.Biology present knowledge shows that the larvae are parasites of birds, associating with then in ways ranging from scavenging in their nests to subcutaneous bloodfeeding.Pont (1974) revised Passeromyia and provided data on the biology of the species.Couri (1999) and Teixeira (1999) summarized the available data on Philornis relations with birds.Carvalho & Couri (2002a) presented a key to species, modified from Couri (1999).
Passeromyia is distributed throughout the Old World, Afrotropical and Oriental regions, Australasia, and the Western Pacific (Pont, 1974) and Philornis is known from the New World, mainly in the Neotropical region, with 2 species occurring in the southern United States.Besides larval behaviour similarities, both genera also share some morphological characters like the presence of a tuft of bristles on the post-alar wall (present in all Philornis species and in some Passeromyia) and very similar ovipositor morphology.
More recent publications and catalogues also show divergence on the Passeromyia systematic position.Pont (1980Pont ( , 1989) ) placed the genus among the Muscinae, tribe Hydrotaeini and Reinwardtiini, respectively.
Passeromyia is known from 5 described species: P. heterochaeta (Villeneuve), P. indecora (Walker), P. steini Pont, P. longicornis (Macquart), and P. veitchi Bezzi.The larvae of these species are known to be scavengers in bird nests (P.steini); external parasites of nestlings, remaining on the body surface and piercing the skin to suck blood (P.heterochaeta); or subcutaneous parasites of the nestlings (P.indecora).In the last case, if the host dies the subcutaneous larvae can feed on the carcass until ready to pupate (Pont, 1974).The life history of P. longicornis and P. veitchi is unknown.
The main objectives of this paper are to clarify the relations of Passeromyia and Philornis and the discussion on the evolutionary line of larvae habits (see also Dodge, 1971) based on cladistic methodology.

MATERIAL AND METHODS
The material used in this study belongs to Museu Nacional, Rio de Janeiro, Brazil; The Natural History Museum, London; Canadian National Collection of Insects, Ottawa; and the Australian National Insect Collection, Canberra.
The terminal taxa are represented by Philornis, Passeromyia and the following related genera considered by different authors as in different positions among the Muscidae, but close to those two: Aethiopomyia, Alluaudinella, Calliphoroides, Charadrella, Fraserella, Muscina, Ochromusca, Phaonina, Reinwardtia, and Synthesiomyia.
The cladistic analysis was carried out using Hennig86, version 1.5 (Farris, 1988), running in "Tree Gardener", version 2.2 (Ramos, 1997), a program designed for running Hennig86 in a Windows environment.Minimum-length trees were calculated using the options "mhennig" associated with "successive weighting".
Thirty-two terminal taxa were analyzed based on 54 characters.Characters were polarized by the outgroup method (Watrous & Wheeler, 1981;Maddison et al., 1984).The outgroups were represented by two Anthomyiidae genera: Coenosopia Malloch and Phaonantho Albuquerque.These two genera are the only ones among the Anthomyiidae where the anal vein does not reach the wing margin (as found in Muscidae).Michelsen (1991) proposed Anthomyiidae as the sister-group of the Muscidae.
Characters were coded as binary and multistate.The latter was considered as additive or non-additive, depending on available information on the contiguity of states in the outgroups.Information not available was coded as a question mark (?).The character distributions were examined using Tree Gardener and Clados (Nixon, 1995).
Table 1 shows the matrix and the characters used in the analysis.

RESULTS AND DISCUSSION
The cladistic analysis resulted in 1 minimal length (373) tree, with a consistency index of 71 and retention index of 85 (Fig. 1).The monophyly of the Muscidae was confirmed by the loss of the postabdominal spiracles in the female, as pointed out by Hennig (1965Hennig ( , 1973)).Among the Muscidae, only Acanthiptera Rondani and Cariocamyia Snyder have independently re-acquired spiracle 6 (Carvalho & Couri, 2002a).
The classification shows that Philornis and Passeromyia belong to a monophyletic group supported by one synapomorphy, the presence of a cocoon, enclosing the pupa.This clade includes the following genera: Muscina, Philornis, Phaonina, Fraserella, Passeromyia, Synthesiomyia, Calliphoroides, and Reinwardtia.Skidmore (1985) placed Philornis close to Passeromyia because, according to him, "it appears inconceivable that many similarities between these two genera should be due merely to convergence.The same range of larval habits is found in both genera, but elsewhere in the Muscidae parasitism of birds is known except in Muscina and Synthesiomyia where it is of purely casual nature".The present analysis shows this character (larvae associated with birds) as having appeared independently in both genera.
The close relations between Calliphoroides and Reinwardtia, corroborated in this analysis, were long discussed by Hennig (1965) who placed them among the Muscinae.The strong curvature at the end of the vein M, used by many authors to characterize this subfamily, was a homoplasic character state in this analysis.Hennig (op. cit.) also approximated Calliphoroides to Muscina and Synthesiomyia.
In this analysis, Alluaudinella, Aethiopomyia, and Ocrhomusca appear to form a monophyletic group, defined by the presence of remarkably short stubby spines on the upper side of the palpi.Pont (1980) following Hennig (1965) considered Ocrhomusca, Aethiopomyia, and Alluaudinella in the Dichaetomyiini.Emden (1939) placed them in the Muscinae based on their Musca-type thoracic calypter.Skidmore (1985) mentioned that the larval mouthparts and larval spiracle of Ocrhomusca suggest a close relationship with Muscina and Synthesiomyia.According to Skidmore (op.cit.) these three genera are clearly closely related.The larva of Ocrhomusca and Alluaudinella feed on dead snails and, although the larval habits of Aethiopomyia are not known, the final larval instar of Aethiopomyia closely resemble those of Ochromusca, Alluaudinella, Synthesiomyia, and Muscina (Skidmore, 1985).Charadrella appears as the sister group of (Alluaudinella (Aethiopomyia, Ocrhomusca)) forming with them a monophyletic group based on the association of their larvae with snails.Skidmore (1985) considered that, although it shares with the three previously mentioned genera an unusual mode of life and some structural affinity, it may not be particularly closely related to them nor to Muscina and Synthesiomyia.Skidmore (op. cit.) provisionally placed Charadrella close to Ocrhomusca on the strength of their biology.A more complete analysis will certainly better define this relation.Charadrella together with the three preceding genera were considered to be in Reinwardtiini (Reinwardtinae) by Skidmore (op. cit.).
Cyrtoneurina and Cyrtoneuropsis appeared close to these groups of genera.Skidmore (1985) examined Cyrtoneuropsis gemina Wiedemann in his study, considering it as a Hemichlrorini (Reinwardtiinae).He mentioned that the species is apparently "archaic combining some features of the Reinwardtinae, Muscinae and Azeliinae".
The paleotropical genus Dichaetomyia (two species in the Palearctic, Skidmore, 1985), seems closely related to this group in the analysis.According to Skidmore (1985) Dichaetomyia appears to be closely related to Phaonia, "but further work may reveal that some of the species included under this genus belong to the Ochromusca group".
Although using few genera representing the other muscid subfamilies, the resulted cladogram corroborated the data in the literature, showing some monophyletic muscid groups such as Muscinae.This subfamily appears in the base of the cladogram and is here represented by Stomoxys, Polietina, Morellia, and Musca.The "Azeliini" (as termed by Carvalho, 1993) was represented by Micropotamia and Hydrotaea.The genus Micropotamyia was constructed by Carvalho (1993) who also presented a discussion of some character-states of phylogenetic importance to Azeliini (as termed by Carvalho, 1989), a tribe in which Micropotamyia was included.Carvalho (op. cit.) called attention to the striking configuration of the male distiphallus, with juxta spinulose, shared by the Azeliini genera.He also mentioned that although there is no published phylogenetic hypothesis for intergeneric relationships in Azeliini, there is evidence that the tribe is monophyletic (Michelsen, 1978;Carvalho, 1993).
The other monophyletic group, and one of the most solidly based among the Muscidae -Mydaeinae + Coenosiinae (Hennig, 1965;Carvalho, 1989;Couri & Pont, 2000) appears at the apex of the cladogram supported by five sinapomorphies.Scutellomusca and Mydaea, considered today among the Mydaeinae, did not form a monophyletic group.The Coenosiinae represented by Limnophora, Coenosia, and Neodexiopsis confirmed their monophyly (six sinapomorphies).The last two genera, known as well as the tribe Coenosiini, also confirmed their monophyly, supported by one sinapomorphy (see also Couri, 2000).

BIOGEOGRAPHICAL ANALYSIS
Historical biogeographical studies on the Muscidae family are few, and rare are the hypotheses about the distribution patterns of the muscid species.In the sixties, Hennig (1965) explained such patterns of some muscid species based exclusively on dispersion.Carvalho (1999) and Carvalho & Couri (2002b) analyzed distribution patterns of muscid species using vicariance, but both studies were based on genera occurring exclusively in the Neotropical region.Although Hennig (1965) theorized on this subject, nothing is actually known about the transoceanic patterns of relationships of muscid species.On the other hand, the intercontinental relationships of the basic groups of Diptera are well known (for a summary, see Amorim & Tozoni, 1994).At the same time, few papers exist on Schizophora (Griffiths, 1972;Papavero, 1977;Matthis, 1977Matthis, , 1978;;Barnes, 1981;Cortéz, 1983).
Using the methodology of cladistic biogeography, by replacing the taxa nomenclature with names of areas where the species occur, a biological area cladogram can be generated (for methodological procedures, see Morrone & Crisci, 1995).Based on this, sister group relationships of some genera could be analyzed based on vicariance biogeography.
The distribution pattern of the species among the genera of Reinwarditiinae (Fig. 1) resembles a Gondwana pattern, suggesting that the ancestor of these genera could have appeared before the Upper Cretaceous, the hypothetical age previous to Muscidae (Hennig, 1965).As pointed out in this paper, the breakup of the Gondwana could have strongly affected the pattern distribution of these genera, suggesting an older age for this family.Pont & Carvalho (1997) described the first muscids, found in Dominician ambar (about 15-20 On the other hand, the distribution pattern here shown for another monophyletic group within the Dichaetomyiinae subfamily formed by Charadrella + Alluaudinella + Aethiopomyia + Ochromusca (Fig. 1) (the first, an exclusively Neotropical genus and the three others Afrotropical) also corroborates a hypothetical older age for the Muscidae.
This older age for Muscidae, as here suggested, is hypothesized to be based on a partial phylogenetic analysis of this family, although some of the resulting monophyletic groups inside this family corroborated previous classifications.Therefore, the hypothetical age of this family remains undefined and more fossil evidence and historic biogeographic studies are required.The geographical distribution of the genera included in this analysis is shown in Table 2. ?0100?0000000?1010?010100100???????????10????

TABLE 2 TABLE 2 TABLE 2 TABLE 2 TABLE 2
Grimaldi & Cumming (1999), eastwards to Australia and the West Pacific years ago, the minimum age of those species).Evenhuis (1994)identified a muscid fossil from the Eocene, andGrimaldi & Cumming (1999)recorded the oldest definitive cyclorrhaphan larvae in Cretaceous ambar, but did not position them in a family (uncertain family). million