THE CLASSIFICATION AND INTERRELATIONSHIPS OF THORONINI (HYMENOPTERA: PROCTOTRUPOIDEA, SCELIONIDAE)

Abstract The tribe Thoronini Kozlov, 1970 as interpreted here in its broader sense includes also the tribes Pseudanteridini and Tiphodytini as proposed by Kozlov (1970), i.e. the following genera: Thoron Haliday, 1833, Neothoron n. gen., Pseudanteris Fouts, 1927, Thoronella n. gen., Microthoron n. gen., Tanaodytes n. gen., and Tiphodytes Bradley, 1902. The new taxa described are: Neothoron lautus n. gen. and n. sp., Ecuador; Thoronella elegans n. gen. and n. sp., Brazil; Microthoron baeoides n. gen. and n. sp., Malaya; Tanaodytes longipes n. gen. and n. sp., Malaya; Tanaodytes soror n. sp., Ceylon; Tiphodytes godavari n. sp., Nepal; Tiphodytes mymar n. sp., Malaya; Tiphodytes selangor n. sp., Malaya.The genera Thoron, Pseudanteris, and Tiphodytes are recorded for the first time in Canada. The name Calliscelio Ashmead, 1893 is considered a synonym of Calotelea Westwood, 1837 (new synonymy).The species of Thoronini, the biology of which is known, parasitize eggs of water Heteroptera (Gerridae, Nepidae) and it is assumed that most if not all species of this tribe either live in, or are closely associated with, aquatic or semi-aquatic habitats. This is inferred, in the case of species for which definite information is lacking, from the sites in which the adults have been collected, all of which were near water.

Classification of the Scelionidae at the tribal level is a subject that has been ignored for many years. The groups originally proposed by Ashmead (1893Ashmead ( , 1903 were accepted by most of the subsequent students. Seventy-three years later Szab6 (1966) established the new subfamily Gryoninae. Later Masner (1968) emphasized the importance of tibia1 spurs and palpi in the suprageneric classification of Scelioninae. Kozlov ( 1970) proposed a new tribal reclassification of the whole family. He also accepted the four major subfamilies of Ashmead ( 1903) viz. Scelioninae, Teleasinae, Baeinae, and Telenominae. At the same time he erected a number of new tribes, treating Gryoninae only as a tribe within Scelioninae. HellCn (1971 ), apparently unaware of Kozlov's classification, introduced a rather aberrant system of Scelionidae. The latter he divides into three subfamilies: Scelioninae, Telenominae, and Platygasterinae. The subfamily Scelioninae is subsequently divided into five tribes viz. Baeini, Teleasini, Scelionini, Anteridini, and Gryonini. Kozlov (1970) tried to employ some new aspects in the tribal classification of Scelionidae. One of those characters used is a peculiar area situated in front of mesoscutum in some Scelioninae ('1 TP@ y roJr&Hoe ~S I T H O " ) . This character was mentioned and employed first by Nixon in 1933 ('a specialized mesonotal area') and, much later, by Szab6 (1958Szab6 ( , 1969 ('dreieckiges Spezialfeld'), who misinterpreted the area for prescutum. It seems that this structure (Figs. 1,5,8 ) is quite peculiar to several genera of Scelioninae and, to my knowledge, is not encountered anywhere in the Hymenoptera. The dissection of this area shows that it is well separable from prothorax by suture, yet well fused with mesoscutum in one solid sclerite (e.g. in Psilanteris sp.). Usually, a well defined dorsal keel and a corresponding internal apodeme delimit this area posteriorly from the central lobe of mesoscutum. As a rule, it also differs strikingly from mesoscutum in sculpture, being usually smoother and perfectly bare. Hence it appears as a shiny mirror or a cup in front of mesoscutum. Since no special term has been applied to this structure I suggest calling it, for the sake of brevity, skaphion (Greek orta#~ov = concave mirror, cup). The latter is intended as an auxiliary term since my previous effort to homologize this area with prescutum failed (cf. Matsuda 1970).
Skaphion is reported by Kozlov as occurring only in the tribe Psilanteridini of the Scelioninae. In fact, it is also well developed in Pseudanteridini and two other tribes classified by Kozlov outside of the Scelioninae, namely Thoronini (Baeinae) and Tiphodytini (Telenominae). A closer examination of this fact led me to conclude that Thoronini, Pseudanteridini, and Tiphodytini form a natural unit within Scelioninae with closest ties to Psilanteridini. I propose therefore to treat these as a single tribe for which I have chosen the name Thoronini and consequently suppress the names Pseudanteridini and Tiphodytini. The presence of the skaphion alone, however, is not sufficient evidence for bringing these tribes into synonymy. There are other characters, both morphological and biological, that stress the relationship. Some of these characters seem to reflect the aquatic habits of Thoronini, e.g. the strong semierect bristles on the vertex of the head and dorsum of the mesosoma, presumably used in formation of an air plastron for diving. Because of some other important characters in common the above mentioned bristles are not considered a convergence due to similar habits but true relationship within Thoronini.
The taxonomic value of skaphion as a character in classification of higher categories (tribes) presents some problems. There are several undescribed tropical species of Opistacantha Ashmead in which this character seems to have only a specific value, i.e. some species may have or may not have the skaphion well developed. Another problem is that skaphion is missing in some genera closely related to Psilanteridini (e.g. Anteris Foerster) but present in some peculiar unrelated ones (e.g. Parascelio Dodd) . This might indicate a parallel evolution of this character in Scelioninae. With respect to facts mentioned above, skaphion is a character of problematic value.
Thoron was traditionally classified in Baeinae and Tiphodytes in Telenominae yet to me they represent strange elements in the respective groups. In order to learn more about the structure of the mesosoma and metasoma, individuals representing species of Tiphodytes and Thoron as well as species of various genera of Baeinae and Telenominae, were dissected. The results confirmed the original assumption that both Thoron and Tiphodytes were misclassified. The well marked skaphion of Thoron does not occur in Idris Foerster nor in any of the genera assigned to Baeinae (sensu Kieffer 1926). The metasoma of Idris males is composed of 7 tergites and 7 sternites whereas that of Thoron of 8 tergites and 7 sternites. The cylindrical flagellar segments of the male antenna in Thoron differ strikingly from the peniciliate flagellar segments of Idris males and from those of allied genera. Moreover, there is no fusion between antenna1 segments 11 and 12 in males of Thoron. Finally, the smooth unsculptured body in Thoron is quite strange among other genera classified in Baeinae (sensu Kieffer 1926). The latter group was not recognized as an independent subfamily (Masner 1956;Muesebeck and Masner 1967;Masner and Muesebeck 1968) because it was believed that it represents but a highly apomorphous and specialized branch of the Scelioninae. Reflecting Kozlov's (1970) system of tribes I suggest treatment of Baeini and Idrini tentatively as independent tribes within Scelioninae.
Tiphodytes resembles superficially Telenominae yet dissection of metasoma indicates a closer relationship with Scelioninae (notably Psilanteridini) than with Telenominae. Tiphodytes has distinct broad sternopleurites which are absent in the Telenominae, 7 tergites and 6 sternites in the female and 8 tergites and 7 sternites in the male contrasting with 7 tergites and 7 sternites in the female and 8 tergites and 8 sternites in the male of Telenominae. The most typical feature of the telenomine metasoma, the U-shaped sternite 7 in the female ('U formiges Stiick'of Oeser, 1961 ) and the similarly reduced sternite 8 in the male, is absent in Tiphodyles. Moreover, no genus of Telenominae has developed the skaphion or antennal clava with semifused segments in the female sex. The large segment 3 of the metasoma in Tiphodytes also points towards a relationship with Scelioninae rather than with the Telenominae. In this and other respects Tiphodytes seems closest in relationship to Pseudanteris (Masner 1964) and Thoronella not only in morphological but also in biological characters (see below). Marchal (1900b) compared Tiphodytes (as Lirnnodytes) with Thoron as the closest genus. Szab6 (1957Szab6 ( , 1969 classified Tiphodytes (as Hclngaroscelio) correctly in Scelioninae.
I think that the lack of an impressed submarginal groove in the metasoma of a scelionid genus alone is not necessarily proof of its relationship with the Telenominae. Similarly, this character will fail to prove any relationship of some genera of platygasterid subfamily Inostemminae (Allotropa Foerster, Fidiobia Ashmead, Amitus Haldeman) with Telenominae. These similarities are considered to be a convergence and I do not therefore hesitate to classify Tiphodytes and the related genus Tanaodytes in Scelioninae rather than in Telenominae.
Body perfectly smooth and glossy, with no matt microsculpture, highly polished, sometimes with slight metallic shade; numerous hard semierect bristles scattered on vertex of head and dorsum of mesosoma; skaphion well developed and sharply delimited; metasoma distinctly elongate (except in Microthoron), often obpyriform, segments 2 and 3 largest and broadest, tergites 1 and 2 longitudinally costate, following tergites smooth; 7 tergites and 6 sternites in female, 8 tergites and 7 sternites in male; apical tergite ( P T7, 8 T 8 ) rather large, triangular, with 2 pairs of long upcurved bristles, not extrusible with ovipositor but articulating with tergite 6 in female; submarginal impressed groove of metasoma either developed or absent; antenna 12-segmented (secondarily 6-segmented in Microthoron 9 ), with articles of club poorly differentiated in female, in male with cylindrical funicular segments; lateral ocelli very close to eyes; vertex of head bluntly rounded, not carinate or excavate; tibia1 spurs 1, 1,. 1.
Confined to aquatic habitats, two genera known to parasitize eggs of water Heteroptera.
Genera included: Thoron Haliday, Neothoron n. gen, Pseudanteris Fouts, Thoronella n. gen., Tiphodytes Bradley, Tanaodytes n. gen., and Microthoron n. gen.  Three groups of genera comprise the tribe Thoronini. In the first group TJiormn and Neothoron are closely related to each other, the latter being more plcsiomorphous than the former. Pseudunteris and Tl~orot?ella may also be classified in the group above as they bridge tlicse genera with Tiplrorlvtes and Tanaodytes. Indeed. Psertiiatrteris and Thoror~ella are midway, combining characters of both Tl~oron and Tiphodvres. In the second group Tanndyfes represent a highly apomorphous branch of Tiphodytes. Microthoron, however, represents a rather lonely element within Thoronini and is treated as the only representative of the third group. Its superficial resemblance with Raeini is considered as a convergence and it is hoped that the biology, once known, will confirm this assumption.
One species recorded from eggs of the water scorpion (Nepa L.).
Ferrikre ( 19 16) and Henriksen ( 19 18) reported the' species being bred from eggs of the European water scorpion, Nepa cinerea L. (Nepidae: Heteroptera); however, Kozlov (1970) speculates that the species might be parasitic upon eggs of spiders confined to aquatic habitats.
Mesosoma only slightly arched dorsally (lateral aspect); skaphion well developed, forming an acute angle with horizontal axis of mesosoma (lateral aspect); notaulices almost percurrent, abbreviated in extreme anterior part of mesoscutum; scutellum separated from mesoscutum by series of small pits, similar pits bordering its posterior margin; metanotum unarmed; propodeum rather large, at least as long as scutellum, i.e. not excavated posteromedially as in Tllororr in which the first segment of metasoma appears to be almost touching metanotum; wing with long marginal ciliae all around; subcosta with row of strong bristles well surpassing the front margin of wing; marginalis almost triangular, slightly longer than wide, stigmalis much shorter than long postmarginalis; basalis spurious, indicated as a cloud (Fig. 10); tibia1 spurs 1, 1, 1.
Metasoma distinctly elongate, pedunculate, broadest in apical two-thirds, composed of 8 tergites and 7 sternites; submarginal impressed groove well developed; first three segments the longest, almost subequal in length, first segment almost twice as long as wide; tergite 8 with 2 pairs of long sensory bristles.

TYPE-SPECIES. Neothoron lalrtus n. sp. (dcscribcd below).
Neothoron is closely related to Thoron and it is believed that this will be further substantiated by subsequent discovery of the female of the former genus. The long postmarginal vein in Neotlzoron makes this genus readily distinct from Thoron. The difference in mesosomal structures between the two genera discussed is somewhat more difficult to appreciate yet offers another good distinguishing character.
Using Kieffer's ( 1926) key this new genus may be brought under Calliscelio Ashmead. The examination of the type of the latter genus (Calliscelio laticinctus Ashmead, in USNM) reveals it to be congeneric with Calotelea Westwood, and consequently Calliscelio is hereby considered a junior synonym of Cabtelea (new synonymy).
First  (1964) and mentioned also by Kozlov (1970). The morphological similarity of Pseudanteris with Tiphodytes is SO great that some biological observations of American authors (e.g. Matheson and Crosby 1912) of the latter could also refer accidentally to the former genus. The chance of misinterpretation of one genus for another is increased by the same habits, same niche as well as minute dimensions of the adult wasps (approx. 1 mm).

Pseudanteris insignis Fouts, 1927
Fig. 3 -New to Canada. The first Canadian record extends the range of distribution further to the north (Oxford Mills, Little Rideau River, Ont., 20-IX-1970, 1 9 , L. Masner collr.). In fact, the above mentioned record is the first recovery of the species since the description. The male remains unknown.
The biology is not known but the Canadian specimen was taken on a lily pad (Nymphaea sp.) approximately 3 ft from the bank, in community of several males of Tiphodytes gerriphagus. Kozlov's ( 1970) assumption that water striders (eggs) are potential hosts of Pseudanteris is thus strongly supported.
Mesosoma shining, with scattered bristles, moderately arched when viewed from side, rather slender when seen dorsally; skaphion well marked, trisinuate posteriorly; parapsidal furrows percurrent, dilated posteriorly; mesoscutum shining yet with very delicate net-like microsculpture; scutellum with strong, scattered bristles, perfectly smooth, unarmed, margined posteriorly with crenulate rim; metanotum unarmed, not particularly protruding at meson; propodeum unarmed, sculptured at sides, perfectly smooth and shining medially where its sloping central part forms opposite side to metasomatic horn; wings rather slender, with very long marginal fringes throughout; submarginal vein in fore wing with long bristles along its whole length, not "broken" in distal two-thirds, attaining almost half of wing length; marginalis almost punctiform, stigmalis relatively short, knobbed apically, postmarginalis very short almost rudimentary, shorter than stigmalis; basalis clearly indicated as a spurious vein; hind margin of fore wing without angular process; hind wing with complete submarginal vein dilated apically; tarsi 5, 5, 5; tibia1 spurs 1, 1, 1.

TYPE-SPECIES. Thoronella elegans n. sp.
Thoronella combines many characters of Tiphodytes, Pseudanteris, and Thoron. Habitually it reminds one much of Pseudanteris yet has entirely different wing venation. It is also reminiscent of Tiphodytes but the metasoma in the latter genus has no submarginal groove due to broad laterotergites. So in metasomatic structure Thoronella is veri close to Thoron yet in antenna1 formula refers back to some species of ~i~h o d~i e s .
It may therefore be concluded that Thoronella (together with ~seudanteris) bridge the seemingly wide gap between Thoron and Neothoron on one side and Tiphodytes and Tanaodytes on the other. In my assessment Thoronella demonstrates the subordinate role of submarginal groove of the metasoma as a character in the higher classification of ~celioninae.
New to Canada. Two species were known in the Holarctic region but the present paper extends the distribution of the genus to the Oriental region. Dissolcus flavipes Ashmead, classified in Tiphodytes by Kieffer (1926), belongs to Trimorus Foerster (Masner and Kozlov 1965;Masner and Muesebeck 1968). Notilem Brkthes, reportedly similar to Tiphodytes (cf. Brkthes 1913), was shown to be a synonym of Gryon Haliday (De Santis 1967).
Hosts include Gerris and Trepobates spp. (eggs). The wasp is confined to aquatic habitats; it dives and swims actively under water, walks on water plants (e.g. upper and lower side of lily pads), or flies around the edges of water. The specimens from Ontario were picked from lily pads (Nymphaea sp.) and other plants growing near muddy river banks.

Tiphodytes setosus (de Stefani, 1902)
Described from Sicily from eggs of Gerris sp. but not recorded or studied since, as de Stefani's collection is feared lost or destroyed (Bin, pers. comm.). There are some doubts that the species really belongs to Tiphodytes. Several characters of the laconic description both support (bristles on apical tergite) and contradict (hind margin of fore wing not angulate) the placement of this species in Tiphody tes.

Tiphodytes godavari n. sp.
FEMALE. Body dark; head black; mesosoma, funicle, and clava dark brown; radicle, scape, pedicel, mandibles, and legs including coxae bright yellow; wings very slightly infuscate, with little dark patch under knob of stigmalis and a darker band indicating basalis.
Notaulices abbreviated, indicated posteriorly as short impressions; metanotum with a bunch of decumbent hairs at meson; marginalis, stigmalis, and postmarginalis in proportions 4:12:6; longest marginal ciliae only as long as % of maximal width of fore wing.

MALE.
Like female but antenna thread-like, 12-segmented, flagellar segments with long bristles; fifth segment modified, with an acute tooth inwardly, all segments of flagellum distinctly elongate, 4-5 times as long as wide, apical segment even longer.
At first I hesitated to include this and the following species (T. selangor) in Tiphodytes because of the postmarginal vein in the fore wing. This vein is much better developed in T. godavari than in T. selangor. T. godavari seems to be the most generalized and least specialized species of Tiphodytes.

Tiphodytes selangor n. s p .
FEMALE. In colour, sculpture, and pilosity extremely like preceding species but differs as follows: Head slightly less transverse (18:28); antennal segment 6 twice as long as wide (3:1.5), so that there are only two short segments before clava; notaulices almost entirely wanting; wings narrower, angular process more prominent, stigma1 vein shortened, considerably closer to wing margin (i.e. smaller angle) than in T. godavari; longest marginal ciliae in fore wing as long as ?h of the maximal width of the wing.

Tiphodytes mymar n. sp.
Notaulices wanting, replaced by shallow impressions posteriorly; fore wing extremely attenuate, strip-like, with long marginal ciliae, the longest ones slightly longer than maximal width of the wing (12: 10); marginal vein thick, almost as long as short marginalis; hind wings virtually strip-like, longest ciliae 4 times as long as the maximal width of wing (12:3).

MALE. Unknown.
This is the most remarkable species of Tiphodytes because of its narrow striplike wings that are reminiscent of some minute Mymaridae.

Tanaodytes n. gen.
FEMALE. Head transverse; vertex with numerous semierect bristles; ocelli very close to eyes yet not contiguous; eyes very large with dense, minute hairs; malar groove very deep; antenna seemingly 9-but actually 12-segmented, very stout, short, clavate, segments of funicle and particularly of clava very closely adpressed and difficult to distinguish (Fig. 8 ) .
Sides of prothorax partly sculptured but with a smooth field in the middle; skaphion well developed, smooth, forming angle slightly less than 90" with the horizontal axis of mesosoma (Fig. 8 ) ; mesoscutum arched, smooth, with numerous semierect bristles, notaulices abbreviated anteriorly, reaching about halfway; scutellum smooth, with scattered bristles; in lateral aspect metanotum bluntly bulging at meson, with few bristles; propodeum very long, June 1972 rugose, with no pilosity, margined by sharp lateral carinae and two keels medially forming an inverted V-shaped field; wings remarkably attenuate, heavily infuscate, with long marginal ciliae; suhcosta running close to front margin to form its edge, with no bristles but a few minute hairs in anterior %; marginal vein short, almost point-like, stigmalis short and knobbed apically, postmarginalis absent (Fig. 7 ) ; legs extremely long, coxae and femora thickened, tarsal segments 1 and 5 extremely long, claws strong (Fig. 9 ) ; tibia1 spurs 1, 1, 1, the middle and hind ones very short and diminished.

TYPE-SPECIES.
Tanaodytes longipes n. sp. (described below). Tanaodytes is closely related to Tiphodytes but seems to be more specialized than the latter. The bionomics of the genus are probably similar to those of Tiphodytes. The elongate propodeum and first segment of metasoma together with the peculiar structure of antennae and legs and the lack of bristles on the subcosta readily distinguish this genus from Tiphodytes. The male is unknown.
Sides of prothorax (under scutal suture) with several rough longitudinal wrinkles but with a smooth field medially; skaphion almost in 85" angle with horizontal axis of mesosoma, trisinuate posteriorly.

Tanaodytes soror n. sp.
FEMALE. Extremely similar to T. lorlgipes but differs in the few following characters. Tibiae and scape infuscate, antennal segments 3-8 light brown so that the colour contrast with dark clava is less distinct; sides of prothorax with very large central smooth field, no longitudinal wrinkles under the scutal suture, fine granulose sculpture along the anterior edge and in the upper corner in front of tegula; skaphion not trisinuate posteriorly; first segment of metasoma slightly shorter (25: 18).
Mesosoma short, as long as high, arched dorsally, with numerous adpressed hairs but no distinct bristles (Fig. 5); skaphion clearly cut, forming an obtuse angle with horizontal axis of mesosoma; notaulices wanting; scutellum semicircular, unarmed; metanotum bluntly projecting medially; propodeum widely excavated medially, only the lateral corners visible and protruding; wings with long marginal ciliae; subcosta with long bristles along its entire length, slightly broken in distal three-quarters (Fig. 6); marginal vein moderately elongate, slightly shorter than stigmalis and as long as postmarginalis; basalis indicated as a shadow, other veins absent; legs short, tibia1 spurs 1, 1, 1.
Metasoma composed of 7 tergites, very short, almost subsessile, slightly longer than wide, all segments transverse, first without projection dorsally; laterotergites broad, not tight to sternites; submarginal impressed groove absent; third segment the largest; apical tergite with two pairs of long bristles curved upwards. TYPE-SPECIES. Microthoron baeoides n. sp. (described below). The higher classification of Microthoroa presents a problem. Superficially it reminds one very much of Idris and related genera of 1drini (antenna, ocelli, shape of body) but differs from them on the basis of the distinct skaphion, lack of impressed submarginal groove on the metasoma, presence of a postmarginal vein, etc. I would prefer to classify it in Thoronini yet it lacks some features of this tribe (no distinct bristles on vertex of head and dorsum of mesosoma, gaster not distinctly elongate, antennal clava with sutures almost completely obliterate). The genus Microthoron points up again the problematic value of thi submarginal groove of the metasoma as a character in the cla.ssification of the higher categories of Scklionidae. On the basis of this character alone Microthoron should be classified in Telenominae. The resemblance to Baeini can be explained as convergence. In the latter group there is no skaphion and the subcosta is never broken in its distal part. In wing venation Microthoron reminds me of Opistncantha Ashmead and some other genera of Psilanteridini, yet it cannot be classified in the latter because of the fused segments of the antennal clava. The 6-segmented antennae distinguish Microthoron u from all genera of Scelionidae; the mesosomal and metasomal characters are also different from the Ethiopian Paraneurobaeus Risbec, the only other genus in which the antenna is 6-segmented. The male of Microthoron is unknown.