Boudiennyia Girault

Huber, John T., 2013, Revision of Ooctonus in the Neotropical region and comparison with Boudiennyia (Hymenoptera: Mymaridae), Zootaxa 3701 (1), pp. 1-23 : 11-22

Boudiennyia Girault

( Figs 26 –64)

Boudiennyia Girault, 1937: 2 ; Lin et al., 2007: 11, 15 (key), 27 (diagnosis).

Diagnosis. Boudiennyia is distinguished from Ooctonus and all other mymarid genera by the scutellum with paired, star-shaped internal structures posterior to the placoid sensilla (visible only in cleared slide mounts) ( Fig. 33 ), and the propodeum with a quadrate median areole ( Figs 33 , 50) bounded by distinct carinae and well separated from the dorsellum by slightly diverging sublateral carinae that extend from the dorsolateral corner of the areole and flare laterally to extend parallel with posterior margin of the metanotum to the anterolateral apex of the propodeum. Other features that distinguish the two genera are listed in Table 1 .

Superficially, Ooctonus and Boudiennyia are similar in general appearance ( Fig. 26 , and Huber 2012, figs 2– 5). They have the same body colour, with dark brown body and light coloured petiole (as in various Polynema - group genera), a well sclerotized gaster that does not collapse significantly when air dried, and a wide and apically truncate fore wing with a distinct oblique hair line that extends from near the venation apex toward the posteroapical margin of the wing ( Figs 26 , 36 ). Structural features that appear to unite Ooctonus + Boudiennyia are: 8-segmented female clava; 5-segmented tarsi, pronotum with transverse carina separating collar from neck; hypochaeta next to proximal macrochaeta (as in most mymarid genera); venation parallel with fore wing margin; long and narrow petiole; gt1 much the largest tergum (the latter two features as in most Mymarinae, sensu Annecke and Doutt 1961); and a group or row of setae on gt1 anterolaterally ( Figs 32 , 37, 38). But many features are different between the two genera ( Table 1 ).

Discussion Under their treatment of Gonatocerus Nees (as Lymaenon Walker ), Annecke & Doutt (1961) noted some important differences between Cosmocomoidea Howard , originally proposed as a subgenus of Gonatocerus but synonymized under Ooctonus by two previous workers, and Ooctonus . The differences they noted were position of the hypochaeta and structure of the petiole, which has a different type of junction with gt2. However, they still placed Ooctonus together with Gonatocerus and somewhat similar genera in Ooctonini. They did not mention Boudiennyia at all, which is known only from Australia and New Caledonia and was unknown except as a name to most mymarid taxonomists. Consequently, they could not comment upon its defining features or its relationships. Based on morphology, Schauff (1984) treated Ooctonus as the sister group of Gonatocerus . He did not treat Boudiennyia because his study was limited to the Nearctic fauna. Because the defining features used for Gonatocerus + Ooctonus are plesiomorphies, as Schauff (1984) acknowledged, they are inadequate to support a FIGURES 37–40. Boudiennyia sp. 37, metasoma, dorsal (gt7 detached from gt6); 38, gaster, male; 39, ovipositor (seen through gaster); 40, genitalia, male (seen through gaster). gt = gastral tergum. Scale bars are 100 µm.

FIGURES 47–52. Boudiennyia sp., female except 49, SEM. 47, mesosoma, dorsal; 48, mesosoma, lateral; 49, mesosoma, ventral; 50, frenum–propodeum, dorsal; 51, mesosoma, anterior; 52, mesosoma, posterior. clr = collar, cx = coxal foramen, dc = discrimen, fu2p = mesofurcal pit, mtp = mesotrochantinal plate, nck = neck, ppf = propodeal foramen. Scale bars are 50 µm.

sister-group relationship. Schauff noted only one apomorphy for Gonatocerus , namely, the parastigma (his ‘submarginal’ vein or the ‘marginal’ vein of most other authors) towards the base slants away from the wing margin whereas in other genera, including Ooctonus , it is parallel to the wing margin and at its base usually bends up towards the submarginal vein (‘subcosta’ of Schauff). Noyes & Valentine (1989) also placed Ooctonus and Gonatocerus together in their Gonatocerus group. Based on male genitalia, Viggiani (1989) placed Ooctonus and Gonatocerus each in its own tribe and in different subfamilies. Zeya & Hayat (1995) noted the different structure of the cerci in the two genera and summarized other differences, including the unique structure of the male genitalia. Ooctonus was grouped together with Gonatocerus and two other genera based on molecular evidence (Munroe et al. 2011), and combined molecular and morphological evidence (Heraty et al. 2013). In the latter paper, Ooctonus appeared as sister group to Borneomymar Huber in the morphology-based cladogram (fig. 8) and sister group to Gonatocerus in the combined morphology/molecular-based cladogram (fig. 9). In both cases Ooctonus was indicated as related to other genera with 5-segmented tarsi (not 4-segmented, as inadvertently stated in Munroe et al. 2011, p. 21). Huber (2002) showed putative relationships of major lineages of Mymaridae in an intuitive cladogram (p. 44, fig. 7) and suggested that the sister genus of Ooctonus was Boudiennyia . If stems 7 and 9 on the cladogram in that paper are reversed so that Ooctonus + Boudiennyia are next to Mymarini sensu Annecke & Doutt (1961) the relationships seem to make better sense, as shown here ( Fig. 65). The two clades share a long, narrow petiole of type 2, as described by Schauff (1984), in which the base of the gaster extends slightly into the apex of the petiole. The Ooctonus + Boudiennyia clade would then be defined by two similarities: pronotum with a transverse carina and propodeum with a complete set of carinae. This clade would be separated from the Mymarini clade, distinguished by three apomorphies, the first two of which are homoplasies, as they occur in other lineages within Mymaridae : females with fewer funicle segments (6 instead of 8), both sexes with fewer tarsal segments (4 instead of 5), and venation short, with parastigma short or absent.

Although superficially similar, Ooctonus , Gonatocerus , and Boudiennyia differ from one another in many features, suggesting they are quite distantly related. There are sufficient morphological differences between Ooctonus and Gonatocerus that I agree with Viggiani’s (1989) view in placing Ooctonus and Gonatocerus in different subfamilies. Ooctonus differs sufficiently from Boudiennyia that they should perhaps be placed in different tribes but the same subfamily. The most significant difference is that Boudiennyia ( Table 1 ) has a large fenestra occupying most of the scutellum, as in the Polynema -group of genera within Mymarini. So instead of treating Ooctonus + Boudiennyia as sister genera (Huber 2002), I propose here that Boudiennyia alone is likely the sister group to all remaining Mymarini ( Fig. 66), and is separated from them by two plesiomorphies: funicle 8- segmented, and tarsi 5-segmented. Interestingly, Annecke and Doutt (1961, p.14) had pointed out the similarity of petiole structure between Ooctonus and two genera of Mymarini, Mymar Curtis and Polynema Haliday , but did not realize its significance.

If Ooctonus were placed in its own lineage near Gonatocerus or in a lineage together with Gonatocerus and similar genera, the numerous features it shares with Boudiennyia and other genera within Mymarini, particularly the Polynema group of genera, would then have to be treated as independently derived; they would represent convergences with members of Mymarini. In contrast, if Ooctonus and Boudiennyia are included as sister genera in a sister clade to Mymarini then only two losses (two female funicle segments and one tarsal segment) and one reduction (parastigmal vein length) would be sufficient to explain their relationship to Mymarini Fig. 64). However, if Ooctonus and Boudiennyia are indeed correctly placed each in their own tribe ( Fig. 65) and these are more closely related to groups of genera with 4-segmented tarsi, then the number of tarsal segments can no longer be used to define major lineages within extant Mymaridae because tarsal number reduction would have occurred independently at least three times.

The two other, unrelated clades in which tarsal number reduction occurs are in the Camptoptera -group and Anagrus -group of genera. In the Camptoptera -group, tarsal number reduction occurs in Eofoersteria as a result of fusion of the apical two tarsomeres (Huber & Lin 1999). The second reduction in the Camptoptera group has not been discussed previously. Annecke & Doutt (1961) placed Ptilomymar in Mymarini because it has 4-segmented tarsi, but it is better placed in the Camptoptera group based on at least two features, the short but distinct petiole and the female antenna with at least 7 funicle segments. Within the Camptoptera group, Ptilomymar has the propodeum with translucent flanges, as in Stephanocampta Mathot , but a large, smooth gt1, as in Litus (which should probably also be placed in the Camptoptera group). Tarsal reduction in Ptilomymar is also due to fusion of the last two tarsomeres. In the Anagrus group of genera, reduction occurs from 4 to 3 in Kikiki , due to loss of one tarsomere (Huber 2013). The fact that reduction in tarsal number, either through fusion or loss, occurs at all in the same or different lineages within Mymaridae is not surprising. Such a reduction occurred already among the five Cretaceous genera of Mymaridae , where Enneagmus Yoshimoto has three tarsal segments and the other four genera have five (Poinar and Huber 2011). Reduction in segment number also occurs in flagellar segments of females and/ or males in various genera of Mymaridae through loss or, occasionally, fusion, resulting in the same number of segments in otherwise unrelated lineages.

I consider the generic relationships of Mymaridae not to be resolved well enough to classify genera correctly based on evolutionary relationships into formal tribes and subfamilies. However, I agree with Viggiani (1989) that the best solution for classifying Ooctonus is to place it in its own group that is either related to other genera with 5- segmented tarsi except Gonatocerus , or related to genera with 4-segmented tarsi and a long gastral petiole of type 2. I would classify Boudiennyia in its own group as well, and place it as the basal genus to the rest of Mymarini, whose genera all have 4-segmented tarsi. I exclude Ptilomymar from Mymarini sensu Annecke & Doutt (1961), as noted above, and would exclude Anneckia Subba Rao and Omyomymar Schauff from Mymarini sensu Viggiani (1989). Other genera may also be better placed elsewhere than in Mymarini. If Boudiennyia is correctly placed either as the basal genus within Mymarini or as sister clade to Mymarini it would be the first evidence for linking genera with 8-funicle segments in females and 5-segmented tarsi to genera with 6-segmented funicle, 4-segmented tarsi and a long gastral petiole of type 2. It would be most interesting to obtain fresh specimens of Boudiennyia for molecular study to determine which one of relationships discussed above is correct.

FIGURES 59–64. Boudiennyia sp., apex of gaster. 59, female, dorsal; 60, male, dorsal; 61, female, lateral; 62, male, lateral; 63, male, ventral; 64, male, posterior. adg = aedeagus, dig = digitus, gs = gastral sternum, gt = gastral tergum, osh = ovipositor sheath (3rd valvula), ost = ovipositor stylets (1st + 2nd valvulae), par = paramere. Scale bars are 20 µm.

tarsi 4-segmented petiole long and tubular, type 2 of Schauff (1984) 65 tarsi tarsi 4-segmented 5-segmented

funicle funicle 6-segmented 8-segmented

tarsi 5-segmented

(4-segmented in fenestra large, circular Ptilomymar )

petiole long and tubular, type 2 of Schauff (1984) 66

Figures

FIGURES 26 – 28. Boudiennyia sp. 26, female habitus; 27, head, anterior; 28, mouthparts, anterior. atp = anterior tentorial pit, cl = clypeus, lb = labrum, lbpl = labial palpus, mnd = mandible, mxpl = maxillary palpus. Scale bar for 26 = 1000 µm; for 27 and 28 = 100 µm.

FIGURES 33 – 36. Boudiennyia sp. 33, mesosoma, dorsal, female; 34, antenna, female; 35, antenna, male; 36, wings, female. Scale bars are 100 µm.

FIGURES 29 – 32. Boudiennyia sp. 29, head, dorsal, female; 30, head, lateral (internal), male; 31, head + mesosoma, lateral, male; 32, metasoma, lateral, male. Scale bars are 100 µm.

Tables

TABLE 1. Contrast of morphological features in specimens of Ooctonus and Boudiennyia. Figures with a lower case ‘ f’ are from Huber (2012); otherwise they are in the present paper.

Character	Ooctonus	Boudiennyia
Labral setae	5	4 (Fig. 28)
Setae above clypeus	Usually short	Long (Figs 27, 28)
Anterior tentorial pits	Hidden (barely visible)	Clearly visible (Figs 27, 46)
Ocular apophysis	Longer (fig. 62)	Shorter (Fig. 30)
Claval mps	7 (rarely 8) (Figs 7–11)	10 (Fig. 33)
Propleura	Short median line of junction (fig. 197)	Long median line of junction (Fig. 49)
Prosternum	Large (fig. 197)	Small (Fig. 49)
Notauli	Well defined (Figs 13–17)	Poorly defined (Fig. 47)
Scutellum	Sculptured (Figs 14–17)	Smooth (Fig. 47)
Fenestra	Short and narrow, triangular (fig. 109)	Large, circular (Fig. 33)
Frenum	Clearly separate (Figs 13–17)	Undefined (Figs 47, 48)
Axillula	Not extending to lateral lobe of mesoscutum (Figs 13–17)	Extending to lateral lobe of mesoscutum (Figs 33, 47, 48)
Mesodiscrimen	Distinct (fig. 197)	Indistinct (Fig. 49)
Mesofurcal pit	Large, circular (fig. 197)	Small, linear (Fig. 49)
Parastigma	Shorter; venation 0.35 wing length (Figs 18– 21)	Longer; venation 0.50 wing length (Fig. 36)
Lateral panel of metanotum	With few carinae (Figs 13–17)	With many carinae (Figs 48, 50)
Metapleural pit	Absent (fig. 6)	Present (Fig. 48)
Metadiscrimen	Absent	Present (Fig. 49)
Mesopectus (= mesosternum) in ventral view	Merging smoothly laterally with mesepisternum (Fig. 197)	Separated laterally from mesepisternum by ridge (Fig. 49)
Propodeal stub	Usually present (Figs 13–17)	Absent (Fig. 50)
Petiole	Ventrally with partial longitudinal suture (fig. 198)	Ventrally without longitudinal suture (Fig. 57)
Gt1 setae	Cluster of 3; no setae dorsomedially (figs 195, 196)	Row of 2 to several, and 1–3 dorsomedially (Figs 32, 37, 38)
Gt1 shape in dorsal view Male genitalia	Posterior margin straight medially (fig. 196) Aedeagal apodemes diverging, not united anteriorly (Fig. 12)	Posterior margin strongly V- shaped (Fig. 54) Aedeagal apodemes converging, united anteriorly (Fig. 40)