1067 A new phylogeny and phylogenetic classification for the Canthyloscelidae (Diptera: Psychodomorpha) Dalton de Souza Amorim Abstract: A new phylogeny and phylogenetic classification for the Canthyloscelidae (Diptera: Psychodomorpha) is presented. A phylogenetic analysis of the Scatopsoidea is performed. A sister-group relationship between the Canthyloscelidae and Scatopsidae is accepted and the monophyly of the Canthyloscelidae is corroborated, including the genera Exiliscelis, Synneuron, Hyperoscelis, and Canthyloscelis. An earlier phylogenetic analysis of the group is considered, in which Synneuron was accepted as the sister-group of the Scatopsidae and Exiliscelis was considered the sister-group of Synneuron + Scatopsidae. Some apomorphic similarities between the larvae of all genera of Canthyloscelidae, especially the reduction of the head capsule, are considered true synapomorphies. Exiliscelis is considered the sister-group of the rest of the family and is placed in a new subfamily, Exiliscelinae. In the Canthyloscelinae, Synneuron is the sister-group of Hyperoscelis + Canthyloscelis. A phylogenetic classification of the group is proposed. Prohyperoscelis rohdendorfi Kovalev, 1985, from the Middle Jurassic in Russia, is accepted as the sister-group of Canthyloscelis. Résumé : Une nouvelle classification phylogénique et phylogénétique pour les Canthyloscelidae (Diptera: Psychodomorpha) est présentée. Nous avons procédé à une analyse phylogénétique des Scatopsoidea. Les Canthyloscelidae et les Scatopsidae sont déclarés groupes-soeurs et le monophylétisme des Canthyloscelidae est confirmé; le groupe comprend Exiliscelis, Synneuron, Hyperoscelis et Canthyloscelis. Une analyse phylogénétique antérieure a été réexaminée; Synneuron y est considéré comme le groupe-soeur des Scatopsidae et Exiliscelis, comme le groupe-soeur de Synneuron + Scastopsidae. Certaines similarités apomorphiques entre les larves de tous les genres de Canthyloscelidae, particulièrement la réduction de la capsule céphalique, semblent être de véritables synapomorphies. Exiliscelis est considéré comme le groupe-soeur du reste de la famille et classifié dans une nouvelle sous-famille, les Exiliscelinae. Chez les Canthyloscelinae, Synneuron est le groupe-soeur de Hyperoscelis + Canthyloscelis. Une classification phylogénétique du groupe est proposée. Prohyperoscelis rohdendorfi Kovalev, 1985, du milieu du Jurassique de Russie, est reconnu comme le groupe-soeur de Canthyloscelis. [Traduit par la Rédaction] Introduction 1077 Amorim There has been virtually no dispute that the Scatopsoidea constitutes a monophyletic group within the Diptera. Most authors have placed the group close to other families in the Bibionomorpha (Enderlein 1912; Edwards 1925; Hennig 1948, 1954, 1973; Amorim 1993). Hutson (1977) did not particularly address the problem of the placement of the Scatopsoidea within the Diptera, but his comparisons of larval features included only families in the Bibionomorpha. Recently, Wood and Borkent (1989) proposed a quite new approach based on larval features, and suggested that the Scatopsoidea should be transferred to the Psychodomorpha. This position was later accepted by Amorim (1994). During the nineteenth century and at the beginning of the last century the first known genera of Canthyloscelidae were considered part of the Scatopsidae. The inclusion of canthyloscelid genera in a taxon with family status was first proReceived May 20, 1999. Accepted December 9, 1999. D.S. Amorim. Departamento de Biologia, Faculdade de Filosofia, Ciências e Letras de Ribeirão Preto, Universidade de São Paulo, Avenida Bandeirantes 3900, 14040-901 Ribeirão Preto SP, Brazil (e-mail: dsamorim@usp.br). Can. J. Zool. 78: 1067–1077 (2000) posed by Enderlein (1912), who gave subfamilial status to Corynoscelis (=Hyperoscelis). Later, Enderlein (1936) ranked the taxon as a family and erected Synneuridae. Considerable confusion concerning name priority has affected the historical taxonomy of the group (see Hutson 1977; but also Nagatomi and Saigusa 1984: 463–464) at the generic level and family level. Corynoscelidae is an invalid name. Other junior synonyms or invalid names have been proposed for the group (Hyperoscelidae, Hyperoscelididae, Synneurontidae, etc.). Hutson’s (1977) review solved most of these problems and indicated that Synneuridae Enderlein, 1936 would be the oldest valid family-group name and Canthyloscelidae Rohdendorf, 1951 would also be a valid name if one separates these genera into two taxa with family rank. However, Nagatomi and Saigusa (1984), in their review of Japanese Hyperoscelis, followed Dr. Curtis Sabrosky’s opinion (personal communication in a letter) that the oldest valid family-group name proposed for the taxon is not Synneuridae Enderlein, 1936 but Canthyloscelidae, proposed by Shannon (1927), which was largely overlooked in the literature. Two main taxonomic groups other than the Scatopsidae have been identified and given family status in Scatopsoidea. One is composed of Hyperoscelis Hardy and Nagatomi, 1960 (a new name for the preoccupied generic name Corynoscelis © 2000 NRC Canada 1068 Fig. 1. Proposed phylogeny for relationships among the genera of Canthyloscelidae. Characters are numbered according to the list of transformation series in the text. Can. J. Zool. Vol. 78, 2000 sheina (1969), Peterson and Cook (1981), and Evenhuis (1994). Hennig (1954) and Cook (1963) did not segregate the two groups into separate families. More recently, Wood and Borkent (1989) considered it more likely that the genera included in Synneuridae and Canthyloscelidae belong to a monophyletic group, which should be included in a single family, Synneuridae. Larvae of Canthyloscelis, Hyperoscelis, and Synneuron have a greatly reduced, membranous head capsule, with largely unsclerotized mouthparts. The larva of Exiliscelis is still unknown. Wood and Borkent’s (1989) argument is straightforward: the Scatopsidae have larvae with a fully developed head capsule, so if we accept Hutson’s (1977) cladogram we must admit two or three (assuming that Exiliscelis conforms to the other genera in this aspect) independent reductions in the sclerotization of the larval head capsule. I agree with Wood and Borkent (1989) that it is quite unlikely that Synneuridae sensu Wood and Borkent (1989) is paraphyletic. Hutson’s (1977) discussion is nevertheless very useful in furnishing much information on character evolution, as well as for his solutions for the nomenclatural problems of the group. This paper is an analysis of the Scatopsoidea as a whole, with particular emphasis on the generic relationships of the Canthyloscelidae. Even when two separate families are considered, Hutson’s (1977) identification key is still the best available, comprising all four genera. Other published keys exclude one or two of the presently known genera. Material and methods A very large number of the species of Scatopsidae, belonging to most genera of the family, have been examined, together with species belonging to Canthyloscelis (Araucoscelis), as well as Synneuron decipiens. The taxa used as outgroups were mainly Psychodidae, Tanyderidae, Tipulidae, Anisopodidae, Bibionidae, Sciaridae, and in some cases, species of different families of Mecoptera. The discussions of Hennig (1973), Matile (1990), and Collucci (1995) concerning the ground plan of the Diptera were very useful for solving some problems of homology and character polarity. List of characters Boheman) and Canthyloscelis Edwards, 1922, collectively named Canthyloscelidae (referred to as Canthyloscelididae by some authors). The other includes Synneuron Lundström, 1910 and Exiliscelis Hutson, 1977, named Synneuridae. The best review of the taxonomic history of the group was made by Hutson (1977) and does not need to be repeated here. Hutson (1977) proposed a phylogeny for the Scatopsoidea and concluded that Canthyloscelis + Hyperoscelis would correspond to the sister-group of Synneuron + Exiliscelis + Scatopsidae. In his classification, Canthyloscelidae includes the first two of these genera, while Synneuridae includes Synneuron and Exiliscelis. Hutson’s (1977) classification, however, has Synneuridae as a paraphyletic group. The inclusion of Synneuron and Canthyloscelis in different families was proposed earlier by Enderlein (1936) and accepted by Rohdendorf (1938, 1964), Hennig (1960), Mamaev and Krivo- In the following list of transformation series, the plesiomorphic condition of each transformation series is followed by the apomorphic condition or conditions. When a linear transformation series has more than two conditions, letters are used to designate successive apomorphic conditions. A discussion follows each character or group of characters modifying the same structures, justifying decisions concerning homology, character polarity, and other problems. The taxonomic nomenclature used in the discussion follows the classification proposed below for the family. Character numbers are the same as those used on the cladogram (Fig. 1). Figures 2, 3, and 4 present the wings and male and female terminalia, respectively, for the genera of the family on a cladogram. Adult features 1. Median ocellus well developed / (a) reduced / (b) absent. There is some variation in the size of the median ocellus in the Scatopsidae, but there seems to be no doubt that it is well developed in the ground plan of the family. In the © 2000 NRC Canada Amorim 1069 Fig. 2. Pictorial cladogram of the Canthyloscelidae, showing the wings of the genera and subgenera of the family. Those of Exiliscelis californiensis and Synneuron decipiens are modified from Peterson and Cook (1981), those of Hyperoscelis eximia are modified from Mamaev and Krivosheina (1969), those of Canthyloscelis (Araucoscelis) antennata are modified from Edwards (1930), those of Canthyloscelis (Canthyloscelis) brevicornis are modified from Nagatomi (1983), and those of Prohyperoscelis rohdendorfi are modified from Kalugina and Kovalev (1985). Canthyloscelidae we find the median ocellus well developed in Synneuron and Exiliscelis, reduced in Hyperoscelis and Canthyloscelis (Araucoscelis), and absent in Canthyloscelis (Canthyloscelis) (Hutson 1977). It is assumed here that the reduction is synapomorphic for Hyperoscelis + Canthyloscelis, with the loss of the median ocellus as a synapomorphy for Canthyloscelis s.str. 2. Eye bridge absent / (a) incomplete / (b) complete. In the Scatopsidae ground plan the eyes are completely holoptic, although some secondary reductions to a nearly holoptic condition occurred. The fully holoptic condition in Canthyloscelidae occurs in Canthyloscelis (Araucoscelis) and Synneuron, the other two genera possessing eyes that nearly meet above the antennae (Hutson 1977). It is quite clear that © 2000 NRC Canada 1070 Can. J. Zool. Vol. 78, 2000 Fig. 3. Pictorial cladogram of the Canthyloscelidae, showing the male terminalia of the genera of the family. Those of Exiliscelis and Synneuron are modified from Hutson (1977) and those of Hyperoscelis are modified from Nagatomi and Saigusa (1984). Gc, gonocoxite; Gs, gonostyle; Ae, aedeagus; Su, surstylus; S9 and S10, sternites 9 and 10; T8 and T9, tergites 9 and 10. the incomplete eye bridge is a synapomorphy of the Scatopsoidea. It is possible to argue either that the fully holoptic condition was achieved independently in Synneuron, Canthyloscelis (Araucoscelis), and Scatopsidae, or that it is a scatopsoid ground-plan feature that has reversed to an incomplete condition in Exiliscelis and Hyperoscelis (which implies that the number of steps is the same). The first of these possibilities is accepted here. 3. Well-developed clypeus separates eyes below antennae / eyes closer to each other below antennae. The Scatopsidae have a well-developed clypeus and the compound eyes are not close to each other below the antennae. In Synneuron, Hyperoscelis, and Canthyloscelis the eyes nearly meet below the antennae (Hutson 1977), but in Exiliscelis they are well separated. The ventrally contiguous eyes are synapomorphic for the Canthyloscelinae. 4. Antennal flagellum with 16 articles / (a) 14 articles / (b) 10 articles (Hutson 1977). A reduction from 16 flagellomeres (the Diptera ground plan condition) to 14 is apomorphic, but whether it is a synapomorphy of the Scatopsoidea is difficult to determine. The Psychodidae, for example, retain 16 flagellomeres, the number also seen in the Tanyderidae, but in other basal families of Diptera the number of articles in the flagellum is reduced. I tentatively accept here the reduction to 14 flagellomeres as synapomorphic for the Scatopsoidea. Within the group, Synneuron has 10 flagellomeres, the Scatopsidae have 10 at most, and the remaining genera of Canthyloscelidae retain 14. The reduction from 14 to 10 articles in the group is considered homoplasious between Synneuron and Scatopsidae. 5. Gena short / well developed (Hutson 1977). I agree with Hutson’s (1977) position that the well-developed gena in the Scatopsoidea is apomorphic. In the ground plan of the Psychodidae the gena is quite small. Hutson (1977) indicates that this would be an autapomorphy of Canthyloscelis s.str. In Canthyloscelis (Araucoscelis), on the other hand, there is some development of the gena, although it is not as conspicuous as that in the other subgenus. The apomorphic condition is here restricted to Canthyloscelis (Canthyloscelis). 6. Maxillary palpus with sensory pit / sensory pit lost. A sensory pit is present on the third article of the maxillary palpus of different families of Diptera (e.g., Tipulidae, Sciaridae) and is present in that condition in Exiliscelis. The Scatopsidae retain a sensory pit on the single article of the maxillary palpus, so it is reasonable to accept it as part of the Scatopsoidea ground plan. Synneuron, Hyperoscelis, and Canthyloscelis do not have the pit on any of the articles. The loss of the sensory pit is considered here to be a synapomorphy of the Canthyloscelinae. 7. Distal three palpomeres elongate / short and round. This reduction in size and modification of the shape of the distal palpomeres in Synneuron (Hutson 1977) is an autapomorphy of the genus. 8. Thorax elongated / stout. Hutson (1977) indicates that the elongated thorax would be a synapomorphy of his “Synneuridae” plus the Scatopsidae. However, an elongated thorax is a plesiomorphic condition, found in different basal groups of Diptera. The inversion of character polarity would show the apomorphic condition shared © 2000 NRC Canada Amorim 1071 Fig. 4. Pictorial cladogram of the Canthyloscelidae, showing the female terminalia of the genera of the family (all drawings are modified from Hutson 1977). AnM, anal membrane; Ce, Ce1, and Ce2, cercus and articles 1 and 2; S7, S8, S9, and S10, sternites 7–10; T8, T9, and T10, tergites 8–10. by Hyperoscelis + Canthyloscelis and different subgroups of Scatopsidae. 9. Anterior thoracic spiracle within membrane / on anepisternum. 10. Anterior thoracic spiracle within membrane / on epimeron I. Hutson (1977; his character 9 in Fig. 26 and on p. 99) proposed that the anterior spiracular sclerite “on a separate sclerite” would be a synapomorphy of Exiliscelis + Synneuron + Scatopsidae. This seems to be an incorrect interpretation. The Scatopsidae clearly have the spiracle on a dorsal extension of epimeron I in the Aspistinae and Ectaetiinae; this sclerite is partially divided in the Psectrosciarinae (as correctly interpreted by Hutson 1977) and completely divided in all Scatopsinae. In Canthyloscelis the spiracle is in its original position, but on a large plate that is more likely an extension of the anepisternum. This condition is found in Synneuron, but the plate around the spiracle seems to be more sclerotized than the rest of the anterior margin of the anepisternum. This may be the reason for Hutson’s (1977) interpretation of the genus as having the spiracle on a separate sclerite. Insofar as I can interpret Hutson’s (1977, Fig. 1) drawing of the thorax of Exiliscelis, the genus has the plesiomorphic condition, with the anepisternum restricted to a more posterior position, epimeron I not developed dorsally (as in the Scatopsidae), and the spiracle situated on the membrane. I do not have this information for Hyperoscelis. I prefer two different transformation series. One is in the ground plan of the Scatopsidae, in which the development of epimeron I is dorsal, involving the spiracle. The other would be a synapomorphy of the Canthyloscelinae, with the anepisternum including the spiracle. 11. Suture of fusion of meron to thoracic sclerites clear, sclerites clearly separated / suture not conspicuous, meron incorporated into metathorax. I disagree with Hutson’s (1977) interpretation of the evolution of the meron in the group. He proposes that a small meron would be a synapomorphy of (Exiliscelis + Synneuron + Scatopsidae). However, the size of the meron in Exiliscelis is very similar to that seen in Canthyloscelis, and the meron in Synneuron is even larger than that in Canthyloscelis. An apomorphic condition is found in the Scatopsidae, in which the © 2000 NRC Canada 1072 suture of the fusion of the meron to the thoracic sclerites is much less evident than in the Canthyloscelidae. 12. Posterior femur and tibia unmodified / posterior femur swollen, tibia curved (Hutson 1977). The apomorphic condition of this transformation series is exclusive to Hyperoscelis and Canthyloscelis within the Canthyloscelidae. 13. Tibial spurs reduced / absent. The reduction of the tibial spurs seems to be a feature shared by the Scatopsoidea and Psychodidae. The spurs are short but present in Exiliscelis, Synneuron, and Hyperoscelis and completely absent in Scatopsidae and Canthyloscelis. Two independent losses of the spurs in the superfamily appear likely. 14. Tarsal claw simple / with basal prominent tooth. 15. Tarsal claw simple / with large basal lobe with small teeth. Two modified conditions of the tarsal claws are seen in the Canthyloscelidae: in Exiliscelis there is a basal prominent tooth, while in Canthyloscelis there is a basal lobe on which a group of small teeth is seen (Hutson 1977). Even if we code these two modifications as identical in a data matrix, they would appear as homoplasies in the cladogram. Because they are not identical in shape they are coded differently and assumed to be different autapomorphies for the genera. 16. Wing membrane completely clear / maculate. The wing membrane of all scatopsids is completely clear except for some entirely yellowish or brownish wings. Considering the variation within the Scatopsidae and most outgroups, it is possible to infer with certainty that in the ground plan of the family the wing is completely clear. This condition is also seen in Exiliscelis and Synneuron. Hyperoscelis and Canthyloscelis have dark markings on the wing that correspond to a synapomorphy of Hyperoscelis. This refers to a transverse distal band; an additional medial band may appear. Figure 2 illustrates the wings of the genera of the family. 17. Wing membrane with macrotrichia and microtrichia / with only microtrichia, macrotrichia entirely absent. The evolution of macrotrichia on the wing membrane in Diptera is still unclear. Basal Bibionomorpha groups (such as the Pachyneuridae and Cramptonomyidae) have macrotrichia on the wing membrane, as do the Ptychopteridae, some Tipulomorpha (especially the Eriopterini), some Culicomorpha (such as Thaumaleidae, Ceratopogonidae, and Chironomidae), and Brachycera families (such as Stratiomyidae). Within the Scatopsoidea, most Canthyloscelidae do have macrotrichia on the wing membrane, as well as in the basal scatopsid stems. It does not seem possible now to determine whether or not there were macrotrichia in the Diptera ground plan. Aside from this question, however, macrotrichia are quite widely distributed among members of the Scatopsoidea. An optimization analysis points to the presence of macrotrichia on the membrane as a feature of the Scatopsoidea ground plan, so its absence within the group would be apomorphic. In the Canthyloscelidae this loss is Can. J. Zool. Vol. 78, 2000 known to have occurred in Canthyloscelis and on different occasions in the Scatopsidae (in which the loss occurred gradually). 18. Subcostal vein (Sc) complete / incomplete. This is a synapomorphy of the Scatopsoidea. None of the scatopsid or canthyloscelid genera have a complete Sc. Hutson (1977) proposed that the faint Sc would be a synapomorphy of his “Synneuridae” + Scatopsidae. I disagree, since Hyperoscelis has a very faint Sc and the condition in most Canthyloscelis species is similar to the plesiomorphic condition found in Exiliscelis and some Scatopsinae. 19. Vein R1 long, reaching C quite close to R4 / R1 shorter, reaching C in a more basal position. In the Diptera ground plan, R4 reaches C very distally on the wing, quite close to the wing apex, so the apomorphic condition of the character is certainly that the vein ends in a more basal position. In the Canthyloscelidae we find R4 reaching C quite separate from R1 in Canthyloscelis but closer to it in Exiliscelis and Hyperoscelis. Synneuron is not comparable in this feature, nor are the Scatopsidae. This could represent an apomorphic feature shared by Exiliscelis and Hyperoscelis. However, more detailed observation of the wing shows that the condition in Canthyloscelis may not be plesiomorphic: there is probably a retraction of R1 in this genus (in which it ends quite close to the medial fork). 20. Vein Rs free from R1 distally / Rs and R1 partially fused distally (Hutson 1977). This is an autapomorphy of Synneuron. 21. Vein R4 present / lost (Hutson 1977). Synneuron is the only genus of Canthyloscelidae that lacks R4. This feature is considered here to have been achieved independently from the Scatopsidae. It is possible that the feature is somehow related to the distal fusion between Rs and R1. 22. Vein Rs completely separate from R1 in the basal third, r1 cell angled / (a) r1 cell not too wide, angled / (b) r1 cell slender, Rs parallel to R1 from the beginning. Exiliscelis is the only genus of Canthyloscelidae in which Rs is largely separated from R1 in its basal third, which is similar to the shape seen in other basal families of Diptera (see Anisopodidae, Trichoceridae, Cramptonomyiidae, basal Brachycera, some Culicomorpha families, etc.). In Synneuron the r1 cell is still angled anteriorly, although it is not particularly slender. Hyperoscelis and Canthyloscelis have a slender r1 cell, with Rs parallel to R1 right from its base, and the very basal sector of Rs is nearly transverse. 23. Vein M1+2 not fused to Rs / (a) fused for a short distance / (b) fused for at least three times r–m length. Fusion of the base of M1+2 to Rs is seen in all Canthyloscelidae. The fusion is quite short in Exiliscelis but longer in all remaining genera of the family. There is a similar short fusion in some genera of Scatopsidae, such as Aspistes, Ectaetia, and part of Psectrosciara, but in the ground plan of the family there is certainly no fusion. Hence, this is an additional synapomorphy of the Canthyloscelidae. The longer fusion is a synapomorphy of the Canthyloscelinae. © 2000 NRC Canada Amorim 24. Vein R2+3 present / absent. Because of the inclusion of the Scatopsoidea in the Psychodomorpha, the absence of R2+3 must be seen as a synapomorphy of the group. It is present in the Psychodidae and Tanyderidae (and even in other supposedly closely related families) but absent in the Scatopsidae and Canthyloscelidae. I must agree with Hutson (1977) that the vein branching from R5 is R4, not R2+3, a conclusion that differs from Mamaev and Krivosheina (1969), Nagatomi (1983), Nagatomi and Saigusa (1984), and Cook (1981). This comes almost necessarily from the position of the radial-sector fork, which is very distal in the wing, even in Exiliscelis, which is supposed here to be a genus with quite conservative wing venation (that nevertheless presents its own apomorphies). 1073 (1977: 78) in his key to the genera of Canthyloscelidae confirm that the basal interruption of M2 is a key feature of the genus. 31. Vein m-m present / absent. 32. Vein M3 present / absent. Veins m-m and M3 are absent in all members of the Scatopsoidea and must be taken as synapomorphies of the group, since it is placed within the Psychodomorpha, most members of which have them. The inclusion of the Scatopsoidea in the Mycetophiliformia is partially based on the absence of M3 and m-m. 25. Vein R5 unmodified / thickened. In all Canthyloscelidae, R5 is unmodified in relation to the ground plan of the family (which is virtually the same as the ground plan of the Diptera in its shape). Canthyloscelis (Araucoscelis), however, has a thickened R5 (Hutson 1977), an obvious apomorphic condition exclusive to the subgenus. 33. Contact between M1+2 and M3+4 well sclerotized / hardly sclerotized mesally. With the loss of bM, the base of M1+2 and the base of M3+4 (the latter is called “m-cu” by most authors) are aligned. The vein is completely sclerotized in Exiliscelis and Synneuron, but is rather interrupted in Hyperoscelis and Canthyloscelis. The condition seen in the latter two genera is certainly apomorphic. 26. Basal sector of Rs (before contact with M1+2) long (longer than R4) / basal sector of Rs very short. This is an additional synapomorphy of the Canthyloscelinae in the sense used here. M1+2 contacts Rs quite distally and the shape of the basal sector of Rs is conservative. In the remaining genera of the family, M1+2 connects to Rs very basally. 34. Vein M3+4 (m-cu) connects with CuA1 / connects with CuA. This is an autapomorphy of Canthyloscelis (Canthyloscelis). Even the Scatopsidae, in which the cubital fork is in a very basal position in the wing, seem to show a “m-cu” connection to CuA1, not to CuA. 27. Vein bM present / absent. Like the loss of R2+3, the absence of bM must be seen as a synapomorphy of the Scatopsoidea. 28. Free sector of M1+2 (basal to fusion with Rs) short / long. Exiliscelis possesses a short free basal M1+2 (before the fusion between bM and Rs), which is also seen in Synneuron and Canthyloscelis. In Hyperoscelis this basal sector of M1+2 is more elongated, possibly because the cubital fork is displaced to a more basal position, without a corresponding shift in the base of Rs. Synneuron also has a very basal cubital fork, but a short M1+2 base. In the Scatopsidae, bM is long and longitudinal in position because of strong displacement of the cubital fork to the base of the wing. However, this condition seems to have been acquired independently. Other outgroups, such as the Psychodidae, would show in their ground plan the plesiomorphic condition seen in the Diptera. This can be considered a synapomorphy of Hyperoscelis. 29. Vein M1 curved basally / M1 stems straight from fork. This is a modification seen in Synneuron and Canthyloscelis. The typical anterior basal curve of M1, seen in most Diptera families, is absent in these two groups; this is possibly related to the basal interruption of M2. 30. Vein M2 complete basally / incomplete basally. The basal interruption of M2 is certainly apomorphic and shared, homoplastically, in the Canthyloscelidae by Synneuron and Canthyloscelis. Edwards (1930, text and Fig. 4) depicts the wing of Canthyloscelis antennata Edwards with a complete M2, but Edwards (1930: 90) in the text and also Hutson 35. Vein M3+4 (m-cu) nearly horizontal / oblique or vertical. In Exiliscelis the medial connection to CuA1 is nearly horizontal. This is due to displacement of the cubital fork to the base of the wing. As is described below, this cubital displacement goes even further in the Canthyloscelinae and Scatopsidae. However, there is parallel displacement of the origin of Rs to the base of the wing, so this CuA1–Rs connection moves from a nearly horizontal to an effectively transverse position. 36. Cubital fork short, shorter than CuA / (a) cubital fork longer, CuA about half length of CuA2 / (b) cubital fork very long, CuA at most a third of CuA2. If the Psychodidae is considered the sister-group of the Scatopsoidea, the displacement of the cubital fork to the base of the wing would be an apomorphy shared by the Scatopsidae and Canthyloscelidae. If any other family in the Psychodomorpha is considered the sister-group of the Scatopsoidea, the displacement would be synapomorphic for the group. I here tentatively place the feature as a synapomorphy for the Scatopsoidea. The second step, where the cubital fork is still more basal in the wing, seems to be a synapomorphy of the Canthyloscelinae within the family. Obviously the condition in Scatopsidae is yet more apomorphic than that found in any genus in the Canthyloscelidae, but would have arisen independently under this scheme. 37. Vein CuP present / virtually nondiscernible. The CuP vein (“Cu2”) is indicated by Hutson (1977) as a synapomorphy of his “Synneuridae” + Scatopsidae. However, CuP can be clearly seen in Exiliscelis. Even though CuP is quite faint in some basal members of the group, it © 2000 NRC Canada 1074 should be considered present in the ground plan of the Scatopsidae. On the other hand, in Hyperoscelis and Canthyloscelis the vein seems to be indistinguishable. I prefer to state that the nearly complete reduction of the vein is a synapomorphy of the Canthyloscelinae. 38. Vein A1 distally complete / (a) very short / (b) absent. Hutson (1977) interpreted the distal interruption of A1 as a synapomorphy of Exiliscelis + Synneuron + Scatopsidae. This seems to follow Hennig’s (1954) statement that the loss of A1 would be synapomorphic for the Scatopsidae. However, Aspistes, Ectaetia, and part of Psectrosciara do have a complete A1 and it should be designated a complete vein in the Scatopsidae ground plan. The first apomorphic condition in the Canthyloscelidae is seen only in Exiliscelis. Complete loss of the vein is known only in Synneuron. A number of other apomorphic features indicate that Synneuron constitutes a monophyletic group with Hyperoscelis and Canthyloscelis, so the reduction of A1 must be seen as a homoplasious derivation in Synneuron and Exiliscelis. Can. J. Zool. Vol. 78, 2000 43. Gonocoxites separated / fused mesally. 44. S9 situated anteriorly on the terminalia / S9 displaced to a posterior position on the terminalia. 45. Gonocoxites rather small / gonocoxites well developed posteriorly and laterally. 46. Gonostylus finger-shaped / slender and hardly sclerotized posteriorly. 47. Gonostylus rather straight / folded mesally. 48. S9 rather small / projects posteriorly between the gonostyli. 49. S9 rather small / present as a flat, well-developed plate ventrally on the terminalia. 41. Sperm pump attached to male terminalia / free in abdomen (Hutson 1977). A sclerotized male sperm pump attached to the male terminalia is supposedly present in the Diptera ground plan (Wood 1991), and is also seen in the Mecoptera and Siphonaptera. The sperm pump, which is present in the Tipulinae, Ptychopteridae, Anisopodidae, Scatopsidae, and Canthyloscelidae, was certainly lost many times in the Diptera. Within the Scatopsoidea, the sperm pump, a sclerotized structure, of the Scatopsinae is detached and free in the abdomen, connected to the terminalia only by the ducts. Hutson (1977) indicates that in Synneuron also, the sperm pump is detached from the terminalia. In the male S. decipiens I examined it seemed that the sperm pump was attached to the terminalia, although the connection between the two structures was not as strong as in other genera. I take Synneuron to be apomorphic for the transformation series. It is very obvious, however, that the condition in Synneuron, if its apomorphic condition is confirmed, was achieved independently from that seen in the Scatopsinae. 50. Surstyli present, short / present, well developed. Some distinctive features of the male terminalia can be found in the genera of Canthyloscelidae in relation to the Diptera ground plan, but they are not easy to interpret. It is possible that mesal fusion of the gonocoxites is a synapomorphy of the Canthyloscelidae. This is certainly found in Exiliscelis and seems to be present in the terminalia of the other three genera also (Fig. 3). In Exiliscelis the gonocoxites are well developed (see Peterson and Cook 1981, Figs. 4 and 5), are quite rounded, and displace the gonostyli to a more posterior position in the terminalia. In Synneuron there is a mesal plate between the gonostyli that seems to be an extended sternite 9. A posteriorly displaced and elongated sternite 9 plate is also found in Exiliscelis. Hyperoscelis has a large mesal plate on the terminalia ventrally, with the gonostylus placed lateral to it. It seems likely that posterior displacement of sternite 9 is a synapomorphy for the Canthyloscelidae. If this is correct, Peterson and Cook’s (1981) indication of the position of sternite 9 (“hypandrium”) would be incorrect (unless sternite 9 is extended laterally at the anterior extremity of the terminalia and posteriorly between the gonocoxites). The mesally projected sternite 9 in Synneuron and the large flat sternite 9 in Hyperoscelis would be autapomorphies. The gonostylus of Exiliscelis, on the other hand, is easily identifiable, having a large base and a slender, wellsclerotized distal end, certainly an apomorphic condition. In the other three genera, the gonostyli are more difficult to identify. They are rather elongated, but they have a mesal fold with a different shape in each genus. This modification seems to be synapomorphic for the Canthyloscelinae. The question of the homology of the “surstyli” is a difficult one to answer. Matile (1990: 64–65) stated that these sclerites are metameric appendages of segment 10 and would be present in the Diptera ground plan. If this is correct, their presence in Exiliscelis, for example, would be archeomorphic. Hence, the apomorphic condition here would be a secondary development. It can be also seen in Canthyloscelis, but is not as well characterized in Synneuron or, especially, Hyperoscelis. Here I tentatively place the feature as a synapomorphy of the Canthyloscelidae. 42. Sperm pump encapsulated / not encapsulated (Hutson 1977). Hutson (1977) describes the sperm pump in Canthyloscelis as not encapsulated, supposedly an apomorphic condition. 51. Female spermatheca simple / modified (Hutson 1977). An oval or spherical female spermatheca is seen in different families, including the Scatopsidae. This condition is modified in both subgenera of Canthyloscelis, which have an elongated spermatheca with a medial constriction and an 39. Vein A1 complete basally / A1 incomplete basally. This quite unusual modification in A1 is apomorphic only for Canthyloscelis (Canthyloscelis) (Hutson 1977). 40. Unmodified male pregenital segments 1–7 / (a) 1–6 / (b) 1–4 (Hutson 1977). Male abdominal segment 8 is reduced in the Scatopsidae and Canthyloscelidae. In the abdomen of Canthyloscelis (Araucoscelis), segment 7 is strongly modified, while in that of Canthyloscelis (Canthyloscelis), segments 5–7 are strongly modified (Hutson 1977). The modification of segment 7 is understood as synapomorphic for Canthyloscelis and the modifications of segments 5 and 6 as autapomorphic for Canthyloscelis (Canthyloscelis). © 2000 NRC Canada Amorim apical extension that continues into the duct. This is certainly a synapomorphy of the genus. 52. Female tergites 9 and 10 separate / fused. 53. Female sternite 8 with short posterior notch separating gonapophyses 8 / deep mesal notch separating short gonapophyses 8. 54. Female cercus two-segmented / one-segmented. Few features can be found in the female terminalia of the Canthyloscelidae (Fig. 4). The presence of two segments in the female cercus is a feature of the Diptera ground plan (Hennig 1973). In the Canthyloscelidae this is modified only once, in Synneuron, the only genus with a one-segmented cercus. In the Scatopsidae, a one-segmented cercus is present only in the Aspistinae. The structure interpreted as a cercus in the Scatopsinae is probably a secondarily divided tergite 10. Fusion between tergites 9 and 10 seems to be a synapomorphy of the Canthyloscelinae. The deep mesal notch on sternite 8 of the female terminalia is an apomorphy seen in Synneuron and Hyperoscelis but not in Exiliscelis or Canthyloscelis. Here the notch is interpreted as a synapomorphy of the Canthyloscelinae, reversed to a condition similar to the plesiomorphy in Canthyloscelis. Immature features 55. Head capsule well developed and sclerotized / (a) head capsule weakly sclerotized and reduced / (b) greatly reduced and unsclerotized (Hutson 1977). 56. Antennae short but normally developed / reduced, with a characteristic posteriorly ovoid shape (Hutson 1977). 57. “Enigmatic organ” absent / present (Tonnoir 1927 in Hutson 1977). 58. Cuticle with normal setae / setae on body surface reduced and covered with spinulose areas (Hutson 1977). 59. Posterior end of body without hooks / with a pair of heavily sclerotized hooks set on a single adanal plate (Hutson 1977). Hutson (1977) listed a number of features of the larvae of Synneuron, Hyperoscelis, and Canthyloscelis that are modified, and in which they differ from the larvae of Scatopsidae, Cecidomyiidae, and Mycetophilidae. The differences are not restricted to the most conspicuous modification, the reduction of the head capsule. Hutson (1977) listed other features known to occur in the larvae of the three genera with known immatures. However, he did not consider that these modifications were sufficient to corroborate the hypothesis that Synneuron and Exiliscelis constitute a monophyletic group with Hyperoscelis and Canthyloscelis. I concur with Wood and Borkent’s (1989) position that collectively the larval features strongly indicate that the three genera compose a monophyletic group separate from the Scatopsidae. Unfortunately, the larva of Exiliscelis is still unknown and generalizations are merely hypothetical. We assume here provisionally that the apomorphic conditions proposed above for the three genera are part of the Canthyloscelidae ground plan. 1075 Discussion The hypothesis proposed here to account for the relationships between members of the Scatopsoidea differs from that of Hutson (1977) in two main aspects. First, in his system Exiliscelis and Synneuron compose a monophyletic group with Scatopsidae, while here these two genera compose a monophyletic group with Canthyloscelis and Hyperoscelis. In this respect the monophyly of the Canthyloscelidae follows Wood and Borkent (1989). Second, some similarities shared by Exiliscelis and Synneuron that were considered apomorphic by Hutson (1977) are considered here to be homoplasies or plesiomorphies. I shall address in more detail Hutson’s (1977: 98–99, Fig. 26) discussion of the characters at these levels of his cladogram. The synapomorphies that he proposed as a basis for including Synneuron with the Scatopsidae are as follows (only apomorphic conditions are given): head rounded and somewhat flattened, eyes meet above the antennae, prothorax stoutly developed, R4 lost, and cercus one-segmented (referred to as ovipositor one-segmented). Regarding this set of features, I would disagree that the shape of the head is an apomorphy of these two groups and I cannot understand Hutson’s (1977) description of the modification of the prothorax. The other three features (2b, 16, and 36 here) are not unusual in outgroups. Hence, based on the available data set for larvae and adults, it seems more parsimonious to consider the similarities between Synneuron and Scatopsidae as homoplastic. To accept that the considerable modifications of the larvae shared by Synneuron and two Canthyloscelidae genera originated homoplastically, it would be necessary to find considerably more consistent apomorphies shared between Synneuron and Scatopsidae than are presently known. The eye bridge is certainly homoplasious between Synneuron and Canthyloscelis (Araucoscelis), and is also known in many other groups of Diptera (some Cecidomyiidae, Sciaridae, Perissomatidae, etc.); R4, as discussed below, is possibly present in the sister-group of the Scatopsidae, Mesoscatopse rohdendorfi Kovalev; the one-segmented cercus is seen in numerous other groups of Diptera. The features proposed by Hutson (1977) to unite Synneuron and Exiliscelis with the Scatopsidae also deserve some attention. His proposed synapomorphies for this group are as follows: thorax elongated, spiracle on a separate sclerite, CuP (his Cu2) absent, anal vein absent, and Sc weakly present. I believe that thoracic shape has the opposite character polarity to that which he presents, the stout thorax being an apomorphy shared by Hyperoscelis and Canthyloscelis, and has originated several different times within the Scatopsidae. Hutson’s (1977) discussion concerning the position of the anterior thoracic spiracle, as mentioned above, seems incorrect. In Canthyloscelis and Synneuron the spiracle is actually on a large plate formed by the anepisternum. In the Scatopsidae ground plan the spiracle is on a dorsal extension of epimeron I and in Exiliscelis it is in its plesiomorphic position, on the membrane between epimeron I and the anepisternum. Hence, two independent apomorphic conditions can be defined, one for the Scatopsidae and the other for the Canthyloscelinae. The reduction of CuP is also not a synapomorphy of this group: it is very faint in the © 2000 NRC Canada 1076 basal Scatopsidae but visible in Synneuron and effectively present in Exiliscelis. It would be better to restrict the very faint condition of CuP to the Scatopsidae. The reduction of A1 is indeed an apomorphic condition shared by Synneuron and Exiliscelis, but is most certainly not a feature of the Scatopsidae ground plan. As mentioned above, the vein is complete in different basal groups of Scatopsidae, hence the plesiomorphic condition would have to be attributed to the ground plan of the family. Finally, the faint Sc would be better seen as a synapomorphy of the Scatopsoidea. There is indeed some variation in the extension of the vein and the degree of sclerotization in the group. In Canthyloscelis, as well as the Scatopsinae, it seems to be slightly more developed. In Exiliscelis the vein is longer but not very well sclerotized, and in Synneuron and Hyperoscelis the vein is actually hardly visible, so I would not conclude that there is an apomorphic condition shared exclusively by Exiliscelis, Synneuron, and Scatopsidae. Again, it is difficult to accept that the features proposed by Hutson (1977) as synapomorphies of his “Synneuridae” + Scatopsidae could be more reliable indications of a common exclusive ancestry than the modifications of the larvae and the other adult features shared by these genera with Canthyloscelis and Hyperoscelis. On the other hand, Hutson’s (1977) conclusions about the monophyly of the genera and subgenera Hyperoscelis, Canthyloscelis, Canthyloscelis (Canthyloscelis), Canthyloscelis (Araucoscelis), Exiliscelis, and Synneuron and the group Hyperoscelis + Canthyloscelis are completely corroborated in this analysis. Known fossils of Canthyloscelidae There is an undescribed fossil species from the Lower Cretaceous of Australia referred to by Eskov (1992) as belonging to Synneuron that will not be considered here. If its inclusion in the genus is confirmed, this would extend the distribution of the genus to the southern hemisphere, at least during the Mesozoic, which would be expected, considering the age of the group (Nagatomi 1983). A fossil species from the Middle Jurassic of Siberia, Prohyperoscelis rohdendorfi, has also been described as belonging to the Hyperoscelidae by Kovalev (Kalugina and Kovalev 1985) and reassigned to the “Canthyloscelididae” by Evenhuis (1994) (actually an incorrect form of Canthyloscelidae; see Hutson 1977). As can be seen from the discussion above, I accept the Canthyloscelidae as a whole as monophyletic and I keep it as a single family-rank taxon. However, I agree that Hyperoscelis + Canthyloscelis compose a monophyletic group (that corresponds to Hutson’s (1977) Hyperoscelidae), and Prohyperoscelis indeed seems to belong to this taxon. The long fusion of M1+2 with Rs, the long cubital fork, the slender r1 cell, the alignment between the base of M1+2 and the base of M3+4 (m-cu), the incomplete contact between M1+2 and M3+4, and the shorter R1 are some of the apomorphies seen in the wing of Prohyperoscelis that are shared with the remaining extant Canthyloscelinae, the Canthyloscelini, and the Canthyloscelina. A1 is present and complete, reinforcing the conclusion that the position of this fossil species is not in Exiliscelis or Synneuron. Actually, the wing venation is very similar to that of Canthyloscelis. Some important differences are that M2 is complete at the base, CuP seems to be well developed, M1 is not very curved at its base, although it is not straight as in Canthyloscelis, and M1+2 and M3+4 are not Can. J. Zool. Vol. 78, 2000 completely aligned, the base of M1+2 occupying a slightly more distal position. These are actually plesiomorphies of the fossil species in relation to modifications shared by Canthyloscelis (Canthyloscelis) and Canthyloscelis (Araucoscelis) and it would be very reasonable to consider that Prohyperoscelis corresponds to the sister-group of Canthyloscelis. This adds a northern hemisphere representative to a monophyletic group previously regarded as having only a southern temperate distribution. Classification proposed The classification below is a sequenced phylogenetic classification (Nelson 1972), with “plesion” used as a substitute category for Linnaean ranks of extinct taxa (Patterson and Rosen 1977). Square brackets are used to indicate redundant nominal taxa in the classification, a convention proposed by Christoffersen (1988). Scatopsoidea Scatopsidae Newman, 1834 Canthyloscelidae Shannon, 1927 Exiliscelinae, subfam.nov. [Exiliscelis Hutson, 1977 [E. californiensis Hutson, 1977]] Canthyloscelinae Shannon, 1927 Synneurini Enderlein, 1936 [Synneuron Lundström, 1910] Canthyloscelini Shannon, 1927 Hyperoscelina Hardy and Nagatomi, 1960 [Hyperoscelis Hardy and Nagatomi, 1960] Canthyloscelina Shannon, 1927 Plesion Prohyperoscelis Kovalev, 1985 [P. rohdendorfi Kovalev, 1985] Canthyloscelis Edwards, 1922 Canthyloscelis (Canthyloscelis) Edwards, 1922 Canthyloscelis (Araucoscelis) Edwards, 1930 Acknowledgements The manuscript benefited from corrections and suggestions made by Dr. Art Borkent, to whom I am sincerely indebted. An anonymous reviewer also helped with comments and corrections. This work was supported by the Brazilian federal science agency, Conselho Nacional de Desenvolvimento Científico e Tecnológico, with a research fellowship. References Amorim, D.S. 1993. A phylogenetic analysis of the basal groups of Bibionomorpha, with a critical examination of the wing venation homology. Rev. Bras. Biol. 52: 379–399. Amorim, D.S. 1994. A new suprageneric classification of the Scatopsidae (Diptera: Psychodomorpha). Iheringia Ser. Zool. No. 77. pp.107–112. Christoffersen, M.L. 1988. Genealogy and phylogenetic classification of the world Crangonidae (Crustacea, Caridea), with a new species and new records for the southwestern Atlantic. Rev. Nordestina Biol. 6: 43–59. Collucci, E. 1995. Morfologia e filogenia dos Antliophora (Insecta: Holometabola: Mecopteroidea). M.Sc. thesis, Universidade de São Paulo, Ribeirão Preto, Brazil. Cook, E.F. 1963. Guide to the insects of Connecticut. Part VI. The Diptera or true flies of Connecticut. Fascicle 8. Scatopsidae and © 2000 NRC Canada Amorim Hyperoscelidae. Bull. Conn. State Geol. Nat. Hist. Surv. No. 93. pp. 1–37. Cook, E.F. 1981. 20. Scatopsidae. In Manual of Nearctic Diptera. Vol. 1. Coordinated by J.E. McAlpine, B.V. Peterson, G.E. Shewill, H.J. Teskey, J.R. Vockeroth, and D.M Wood. Can. Dep. Agric. Res. Branch, Monogr. No. 27. pp. 313–319. Edwards, F.W. 1922. Preliminary note on a new genus of scatopsid flies from New Zealand. Ann. Mag. Nat. Hist. 9: 267–269. Edwards, F.W. 1925. A synopsis of the British Bibionidae and Scatopsidae. Ann. Appl. Biol. 12: 263–275. Edwards. F.W. 1930. Bibionidae, Scatopsidae, Cecidomyiidae, Culicidae, Thaumaleidae (Orphnephilidae), Anisopodidae (Rhyphidae). In Diptera of Patagonia and South Chile. Vol. 2. Part 3. pp. 77–119. Enderlein, G. 1912. Zur Kenntnis der Zygophthalmen : über die Grupierung der Sciariden und Scatopsiden. Zool. Anz. 40: 261– 282. Enderlein, G. 1936. Ordnung Zweiflügler, Diptera. In Die Tierwelt Mitteleuropas. Vol. 6. Part 3. Section 16. Eskov, K. 1992. Archaeid spiders from Eocene Baltic amber (Chelicerata: Arachneida: Archaeidae) with remarks on the socalled “Gondwanan” range of recent taxa. Neues Jahrb. Geol. Palaeontol. Abh. B, 185: 311–328. Evenhuis, N.L. 1994. Catalogue of the fossil flies of the world (Insecta: Diptera). Backhuys Publishers, Leiden, the Netherlands. Hardy, D.E., and Nagatomi, A. 1960. An unusual new Nematocera from Japan (Diptera) and a new family name. Pac. Insects, 2: 263–269. Hennig, W. 1948. Die Larvenformen der Dipteren. Part 1. AkademieVerlag, Berlin. Hennig, W. 1954. Flügelgeäder und System der Dipteren, unter Berücksichtigung der aus dem Mesozoikum beschriebenen Fossilien. Beitr. Entomol. 4(3/4): 245–388. Hennig, W. 1960. Die Dipteren-fauna von Neuseeland als systematisches und tiergeographisches Problem. Beitr. Entomol. 10: 221–329. Hennig, W. 1973. Ordnung Diptera (Zweiflügler). Handbuch der Zoologie. Vol. 4. Part 2. 2/31. Hutson, A.M. 1977. Revision of the families Synneuridae and Canthyloscelidae (Diptera). Bull. Br. Mus. (Nat. Hist.) Zool. 35: 67–100. Kalugina N.S., and Kovalev, V.G. 1985. Dipterous insects of Jurassic Siberia. [In Russian.] Paleontological Institute, Akademiya Nauk, Moscow. 1077 Lundström, C. 1910. Beiträge zur Kentnnis der Dipteren Finlands. V. Bibionidae. Acta Soc. Fauna Flora Fenn. 33: 1–16. Mamaev, B.M., and Krivosheina, N.P. 1969. New data on the morphology and ecology of the Hyperoscelididae (Diptera, Nematocera). [In Russian.] Entomol. Obozr. 48: 933–942. [Translated in Entomol. Rev. Wash. 48: 594–599.] Matile, L. 1990. Recherches sur la systématique et l’évolution des Keroplatidae (Diptera, Mycetophiloidea). Mem. Mus. Natl. Hist. Nat. Ser. A Zool. 148: 1–682. Nagatomi, A. 1983. A new Canthyloscelis from New Zealand (Diptera, Canthyloscelididae). Mem. Kagoshima Univ. 3: 163–168. Nagatomi, A., and Saigusa, T. 1984. The Japanese Hyperoscelis (Diptera, Canthyloscelidae). Kontyu, 52: 463–471. Nelson, G. 1972. Phylogenetic relationship and classification. Syst. Zool. 21: 227–231. Patterson, C., and Rosen, D.E. 1977. Review of ichthyodectiform and other Mesozoic teleost fishes and the theory and practice of classifying fossils. Bull. Am. Mus. Nat. Hist. 158: 81–172. Peterson, B.V., and Cook, E.F. 1981. 21. Synneuridae. In Manual of Nearctic Diptera. Vol. 1. Coordinated by J.E. McAlpine et al. Monogr. No. 27, Research Branch, Agriculture Canada, Hull, Que. pp. 321–324. Rohdendorf, B.B. 1938. Mesozoische Dipteren aus Karatau. I. Brachycera und einige Nematocera. [In Russian.] Tr. Palaeontol. Inst. Akad. Nauk SSSR, 7: 29–67. Rohdendorf, B.B. 1951. Organs of locomotion in Diptera and their origin. [In Russian.] Tr. Palaeontol. Inst. Akad. Nauk SSSR, 35: 1–180. Rohdendorf, B.B. 1964. The historical development of Diptera. [In Russian.] Tr. Palaeontol. Inst. Akad. Nauk SSSR, 100. [Translation, University of Alberta Press, Edmonton, 1974.] Shannon, R.C. 1927. Some new Diptera from Argentina. Rev. Soc. Entomol. Argent. 4: 31–42. Wood, D.M. 1991. Homology and phylogenetic implications of male genitalia in Diptera: the ground plan. In Proceedings of the Second International Congress of Dipterology, Bratislava, Slovakia. Edited by L. Weismann, I. Orszagh, and A.C. Pont. SPB Academic Publishing, The Hague. pp. 255–284. Wood, D.M., and Borkent, A. 1989. Phylogeny and classification of the Nematocera. In Manual of Nearctic Diptera. Vol. 3. Edited by J.F. McAlpine, B.V. Peterson, G.E. Shewill, H.J. Teskey, J.R. Vockeroth, and D.M Wood. Can. Dep. Agric. Res. Branch, Monogr. No. 32. pp. 1333–1370. © 2000 NRC Canada