A REDEFINITION OF ARADOPHAGUS (HYMENOPTERA: SCELIONIDAE), WITH A. KEY TO DESCRIBED SPECIES

Abstract A new, broader concept of the genus Aradophagus Ashmead is presented to include Abuko Masner and Huggert and Ladora Masner and Huggert (new synonyms). The extent of the tribe Aradophagini is discussed and some character states re-evaluated. Ladora maura Masner and Huggert is synonymized under Apegus squamosus Dodd (new synonym). The distribution of some species of Aradophagus is discussed. The first fossil member of the genus is reported from late Tertiary amber of Hispaniola (Dominican Republic). Aradophagus diazisp.n ov. is described from Venezuela as the first member of the genus and the tribe in South America. A key to described world species of Aradophagus is presented. Résumé Le genre Aradophagus Ashmead est redéfini selon une conception plus large et inclut maintenant les genres Abuko Masner et Huggert et Ladora Masner et Huggert (nouvelles synonymies). La définition de la tribu des Aradophagini fait l’objet d’une discussion et certains caractères sont réévalués. Ladora maura Masner et Huggert devient synonyme d’Apegus squamosus Dodd (nouvelle synonymie). La répartition de certaines espèces d’Aradophagus est examinée. Le premier fossile du genre à être mentionné a été trouvé dans de l’ambre d’Hispaniola (République dominicaine) de la fin du Tertiaire. Aradophagus diazisp.no v. du Vénézuéla, dont on trouvera ici la description, est la première espèce de ce genre et de cette tribu à être découverte en Amérique du Sud. Une clé d’identification permettra de séparer toutes les espèces d’Aradophagus du monde. [Traduit par la Rédaction]


INTRODUCTION
The genus Aradophagus was proposed by Ashmead (1893) for A. fasciatus Ashmead described from Florida and presumed to parasitize the eggs of Aradidae (Heteroptera). Ashmead classified the genus in the subfamily Telenominae (actually the tribe Telenomini), a placement followed by most subsequent authors (Kieffer 1926;Muesebeck and Walkley 195 1;Masner and Kozlov 1965;Kozlov in Medvedev 197 1;Kozlov and Kononova 1983). Kozlov (1970) erected the tribe Aradophagini in the Telenominae to contain the single genus Aradophagus. Masner (1976~) classified Aradophagus and the Aradophagini in the subfamily Scelioninae; this concept was reaffirmed by Masner and Huggert (1979) after dissection of the metasoma of A. fasciatus. The genera Abuko Masner and Huggert (one species), Ladora Masner and Huggert (three species), and two new species of Aradophagus also were described (Masner and Huggert 1979). Galloway and Austin (1984) reported three species of Ladora from Australia, originally described by Dodd (1914Dodd ( , 1915Dodd ( , 1926, the first two of' these species in Microteleia Kieffer and the last in Apegus Foerster. Mineo and Caleca (1992) -described Aradophagus nicolai from Zimbabwe. The tribe Aradophagini, as interpreted by Masner and Huggert (1979), contained three genera: Aradophagus, Abuko, and Ladora. The concepts of the three genera were based on fragmentary knowledge of the world fauna; members of the Aradophagini are very rare in collections. During the past 14 years one of us (L.M.) focussed on accumulating additional material of this tribe. New evidence failed to support the continued existence of Abuko and Ladora as separate genera. Character states believed to be of generic value appeared to be only of specific value (e.g. ocellar configuration, presence or absence of striae on cheek, presence or absence of notaulus, relative length of T2 and T3, etc.). Thirty-four undescribed species ofAradophagus were studied (Canadian National Collection of Insects, CNCI) from Africa and Madagascar (21 spp.), Australia (11 spp.), the Orient (1 sp.), and North America (1 sp.). These species bridge the morphological gaps between all three genera. As a result both Abuko and Ladora are treated here as junior synonyms of Aradophagus. Mineo and Caleca (1992) similarly expressed an opinion that Abuko and Aradophagus could be the same genus.
The revised generic diagnosis (see below) takes into consideration all described (I 1) as well as undescribed (34) species known to us. The widely separated notauli deserve special attention. When notauli are present in Aradophagus, they are usually abbreviated, parallel, and situated near the posterolateral comers of the mesoscutum. The unusual position of notauli in some Aradophagus is interpreted as a result of flattening of the mesoscutal plate (similar to Platyscelio Kieffer). The parapsidal lines (Menke 1993), located laterad of the notauli, occur in several undescribed species of Aradophagus as mere traces. Parapsidal lines are present in some Scelionidae (e.g. in Leptoteleia Kieffer, Masner 1978), and appear as smooth, glabrous, elevated ridges. The telescoped ovipositor tube and extruded apical tergite of the Aradophagini are homologous with these structures in the tribe Calliscelionini. Aradophagini can be distinguished from Calliscelionini by the unusually high position of the foramen magnum and the consequently long hypostomal bridge, the long bidentate mandibles, the trapezoidal, densely striate TI, and the rather wide laterotergites, which are only loosely attached to the sternum of the metasoma.

MATERIALS AND METHODS
This report is based on examination of 474 specimens; 45 species were recognized, of which 11 are keyed. Almost all specimens belong to the Canadian National Collection of Insects (Ottawa, Ontario), but several were borrowed from the Australian National Insect Collection (ANIC, Canberra) and from the Museo del Instituto de Zoologia Agn'cola (MIZA, Maracay, Venezuela).

ARADOPHAGUS ASHMEAD Aradophagus
The above definition applies simultaneously to the tribe Aradophagini. Character states such as short, squat, convex body, and metasoma subcircular with T1 strongly transverse, refer to a group of undescribed species from Australia, Madagascar, South Africa, and the United States (Texas). These forms, which are similar in habitus to species of Gryon Haliday or Trissolcus Ashrnead, are considered more primitive than the flattened, elongate forms of the described species of Aradophagus (see key below). The shorter, convex members display sculpture that is generally rougher than in more flattened members. However all intermediate states exist between these two groups. We expect that further exploration of the world Aradophagini will support our new, broader concept of Aradophagus. Presently the following species are transferred to Aradophagus (all new combinations), viz. sarotes Masner and Huggert (from Abuko), brunneus Masner and Huggert and trjapitzini Masner and Huggert (from Ladora), and pulchricorpus Dodd and pulchripennis Dodd (from Microteleia).
Currently 11 species of Aradophagus have been described worldwide; an additional 34 undescribed species are known to us (CNCI). Information based on all species indicates that members of Aradophagus prefer dry to arid habitats, especially on the African continent, but also in Madagascar, western and southern Australia, and one species (only males known) in desert habitats in Texas. Three species display a wide distribution: A. fasciatus is Holarctic (Masner and Huggert 1979); A. squamosus is known from Australia (Queensland), Africa (South Africa, Zimbabwe), and the United States (Florida; new record); and A. microps Masner and Huggert is known from South India, Italy(Sicily), and Spain. Prior to this study, Aradophagus was not recorded south of the United States and no fossils were known. Aradophagus diazi sp.nov., from Venezuela, is the first extant member known to us to occur in South America. We also examined a female of Aradophagus sp. in a piece of amber from Hispaniola (Dominican Republic) (CNCI) presumed to be from the late Tertiary. This species resembles A. squamosus in having only one dark band in the fore wing.
The host and the habits of Aradophagus are not known. Ashmead's (1893) original assumption of Aradidae as the host was never confirmed; so far only members of the genera Telenomus Haliday and Aradoctonus Masner (Johnson 1992) were positively reared from eggs of Aradidae. Masner and Kozlov (1965) mentioned dead wood as a possible habitat and Masner (1976b) reported A. fasciatus intercepted in large numbers over several years inside a house. Mineo and Gatto (1982) reported A. microps in large numbers from yellow pan traps operated in vineyards and citrus orchards. The record (Masner and Huggert 1979) of an armoured scale as host for A. squamosus (= A. maurus) remains highly doubtful. However, it is plausible to assume that at least the strongly flattened members of Aradophagus may attack flattened eggs of a host associated with dead wood. Finding the first positive host record as well as habits for any species of Aradophagus remains a challenge.  Ladora squamosus (!): Galloway and Austin, 1984: 63 Ladora maura Masner andHuggert, 1979: 1097 Q (new synonym) We compared three females (ANIC) from Australia (Queensland: Brisbane, March 1930;Queensland, Gogango, Sep. & Dec. 1931, identified by A.P. Dodd [Dodd's handwriting]), the paratype female of L. maura (CNCI) from South Africa (Humansdorp, Dec. 1960), a female (CNCI) from Zimbabwe (Chishawasha, May 1990, A. Watsham, pan trap), and five females (CNCI) from the United States, Florida (Alachua Co., Gainesville, Feb.-Nov. 1987& 1988. Wahl, Malaise trap and interception trap in regrown oak forest). All individuals are identical except that Dodd's specimens are somewhat faded due to age. The fragmentary distribution pattern spanning localities in three continents indicates the possible cosmopolitan nature of A. squamosus. It is only logical to assume that the host of this parasitoid is a very widespread insect, possibly a tramp, or A. squamosus is polyphagous. sp.nov. (Figs. 1,2,3) Diagnosis. FEMALE. Body considerably depressed dorsoventrally, distinctly flattened dorsally (lateral view); head in frontal view subcircular, wider than high, frons and vertex with deep punctures, punctures sparser on frons than on vertex; process on oral carina (near mandibular condyle) prominent (Fig. 3, arrow); cheek not striate, but with rudiments of irregular striation near mandibular condyle; hypostomal carina not developed; OOL distinctly shorter than LOL; notaulus strongly impressed but not exceeding basal half of mesoscutum; parapsidal lines faintly indicated as shallow depressions; fore wing predominantly infuscate with 2 lighter zones; metasoma elongate, strongly depressed dorsoventrally, almost foliaceous; T2 slightly shorter than T3; T3 with 2 dark, circular spots posterolaterally; laterotergites (T3) subequal in width to gap between them.

Aradophagus diazi
Description. HOLOTYPE FEMALE. Length 1.04 mm (measured from anterior ocellus to apex of metasoma). Body predominantly pale yellow, T3 with 2 dark circular spots near posterolateral comers; antenna varicoloured, radicle and scape pale yellow as body, A2-A3 dirty yellow, A4-A7 whitish, AS-A12 black; legs completely yellow, slightly more pale than body; fore wing predominantly infuscate, with basal third portion hyaline and with 1 transverse hyaline band near the middle of the wing below marginal vein.

Distribution. Venezuela (Lara).
Discussion. Aradophagus diazi combines some character states of the former genera Abuko and Ladora; with the former it shares the absence of striae on cheeks and with the latter the T2-T3 ratio. This new species differs from A. sarotes by the absence of a specialized tuft of black bristles below the submarginal vein, and in the T2-T3 ratio; from A. nicolai it differs principally by sculpture of the head, length ratio of T2-T3, and colour of the metasoma. Aradophagus diazi differs from all species described in Ladora by the colour of the body and absence of striae on the cheek. The main diagnostic characters of A. diazi are the sculpture of the head, the spots on T3, the T2-T3 ratio, and the relatively short notaulus.

Etymology. The new species is named in honour of Dr. Francisco Diaz (Universidad Centro
Occidental, Venezuela), a good friend, and enthusiastic hymenopterist, who collected this new species.