Somatiidae, Hendel, 1935

McAlpine, David K., 1997, Gobryidae, a new family of acalyptrate flies (Diptera: Diopsoidea), and a discussion of relationships of the diopsoid families., Records of the Australian Museum 49 (2), pp. 167-194 : 176-178

The Somatiidae

The Somatiidae are a small probably very uniform family of flies, including only the genus Somalia restricted to the Neotropical Region. The species are listed by Steyskal (1970b). Steyskal (1958) made a case for their close relationship to the tephritoid family Richardiidae . 1. McAlpine (1989) and Colless & D. McAlpine (1991) treated the Somatiidae as a family ofDiopsoidea, and a relationship to the diopsoid family Psilidae has sometimes been accepted (e.g., by Hennig, 1971). Griffiths (1972), however, included Somatia in the Periscelididae , and rejected some further historical attempts at determining somatiid relationships. This diversity of opinion is indicative of the difficulty in deciding the systematic position of the Somatiidae within the Schizophora. The antenna of Somatiidae has the essential features described above for the diopsoid families Syringogastridae and Psilidae . Taxa with these antennal features occur also in the superfamilies Tephritoidea , Asteioidea , Ephydroidea , and Muscoidea (= Calyptrata), as used in the classification of Colless & D. McAlpine (1991), so that further consideration of somatiid relationships will focus only on these superfamilies. The Somatiidae have none of the more distinctive features of Muscoidea so that relationships with this superfamily need not be furtherconsidered. TheEphydroidea(syn.Drosophiloidea) probably have as groundplan apomorphies distinct prothoracic precoxal bridges and symmetrical protandrial sclerites. In these features the Somatiidae are too plesiomorphic to have been derived from within the limits of the Ephydroidea , and wide divergence in other characters provides no grounds for suspecting any unusual homoplasy in these characters. My study of comparative morphology indicates that, in the groundplan of the Schizophora, tergites 1 and 2 are imperfectly separated, that there is a linear transverse membranous zone between these tergites centrally, and a separate visible suture on each side running to the lateral margin where there is a notch or incision. This condition exists in at least some taxa of numerous schizophoran families. Also an oblique internal ridge on each side, running from the lateral suture towards the anterior margin of tergite 1, is present either in the groundplan of the Schizophora or of a substantial part thereof. In Somatia there is a long, sharply defined, impressed suture separating tergites I and 2 in the medial region, though the membranous line is indistinct; laterally the suture is obsolete, but there is a sclerotised oblique internal ridge on each side. From the relatively few taxa of the family Richardiidae at present available to me (representatives of subfamilies

Epiplateinae and Richardiinae) it appears that total absence of the suture between tergites 1 and 2 is probably a groundplan apomorphy for the family. Therefore Somatia , which retains a well marked suture, would not seem to belong within the Richardiidae .

The female postabdomen of Somatia ( Steyskal, 1958: fig. 3) includes a definite free sternite behind segment 7, and the terminal parts are not fused into an aculeus. These conditions probably provide sufficient evidence for excluding Somatia from the superfamily Tephritoidea , to which the Richardiidae belong, even though some richardiids have apparently a (? secondarily) divided aculeus ( Steyskal, 1987b). The presence of a large male tergite 6 ( Steyskal, 1958; Griffiths, 1972) alone would negate any close relationship to the Richardiidae and allied tephritoid families. The very elongate distiphallus in the male and annular tergosternite 7 in the female of Somatia are conditions so frequently acquired in the Schizophora that they can no longer be considered to indicate an affinity with the Tephritoidea . The above considerations induce me to discard any hypothesis of close relationship between the Somatiidae and the tephritoid families.

Returning to a comparison of Somatiidae and the Diopsoidea , the unusually large, neck-like pronotum of Somatia is reminiscent of the Diopsidae and Syringogastridae , though detail of the articulation with the occipital part of the head is different The deeply sclerotised metathoracic postcoxal bridge resembles that ofDiopsidae, Syringogastridae , Gobrya , and some psilids, but such a postcoxal bridge has been derived many times in the Schizophora, e.g., several times in each of the families Platystomatidae and Tephritidae . It is particularly frequent in elongate flies with deep thorax and often wasp-like form, and is an element of the megamerinoid character set (D. McAlpine, 1997).

The glabrous arista (apart from its long, bipectinate rays) of the Somatiidae (see Fig. 7 ) contrasts with that of typical diopsoid taxa, which have pubescence on the basal part of segment 6, and on segments 4 and 5 when present. Among the typical diopsoids, only some of the more advanced diopsids show reduction (usually not complete absence) of this pubescence.

The apomorphic loss of the suture between abdominal tergites 1 and 2, apparently in the groundplan of the Diopsoidea , provides a difficulty for inclusion of Somatiidae in this superfamily, just as it does for its placement in the Richardiidae , as mentioned above. The Somatiidae also differ from all more typical diopsoids in the presence of strongly convergent postvertical bristles and a well differentiated cheek bristle (perhaps even to be identified as a vibrissa). The presence of a mesopleural bristle is shared with the Nothybidae and the doubtfully diopsoid family Tanypezidae , but not the Psilidae , the diopsoid family with which the Somatiidae have sometimes been closely associated.

Because the possible synapomorphies of Somatiidae and Diopsoinea are not very persuasive, and because of the above disagreements with the hypothetical groundplan of the Diopsoidea , I think that the hypothesis of close relationship between the Somatiidae and the Diopsoidea should be discarded.

Of the previously floated hypotheses on somatiid relationship, there remains only the question of affinity with the family Periscelididae . The Periscelididae are placed in the superfamily Asteioidea by Colless & D. McAlpine (1991) and in the suprafamily Asteioinea of the superfamily Opomyzoidea by J. McAlpine (1989). The former authors' Asteioidea and the latter's rather similar Asteioinea are not very strongly supported by cladistic evidence, and each should be regarded as a provisional grouping. There is lack of evidence that the diverse family-group taxa included in Opomyzoidea by J. McAlpine cohere in even a few of the postulated groundplan apomorphies (including three autapomorphies) listed by him. A critical analysis of this grouping is beyond the scope of this paper, but I know of no convincing evidence that these taxa of Opomyzoidea (sensu J. McAlpine) share a common ancestral state of, for instance, "face membranized along vertical midline," or "wing contrastingly patterned." Also, the fate of tergite 7 is unknown for these taxa (it is generally absent in likely outgroups to the Opomyzoidea ), and, in the groundplan of some included families, sternite 7 is no more reduced and no more closely fused with sternite 8 than it is in likely outgroups. I therefore cannot recognise any validity in the broad superfamily Opomyzoidea of J. McAlpine, and I adhere provisionally to the separate superfamilies Asteioidea and Opomyzoidea of Colless & D. McAlpine.

Griffiths (1972) regarded the Periscelididae as consisting of the subfamily Periscelidinae (with approximately the same limits as used by Mathis, 1993) plus the genus Somatia . His case for monophyly of this family concept relies on five apomorphic conditions (by inference autapomorphies), numbered (1) to (5), which I review as follows:

(1) Anal vein (culb+1a) abrnptly cut off apically, not reaching margin.

This condition occurs in the groundplans of most families of Asteioidea , as well as Psilidae , Syringogastridae , Diopsinae, Gobrya , and a very large number of other schizophoran taxa.

(2a) Only one fronto-orbital bristle present.

This is the usual condition for Periscelidinae . The absence of such bristles in Somatia could have been achieved either by reduction of a single fronto-orbital or by simultaneous reduction of members of a series. We have no means of deciding which process was the relevant one, unless we first assume derivation from the periscelidine condition, adopting a circular argument.

(2b) ocellar bristles standing near ocellar prominence, not between ocelli.

This description seems to refer to the greater distance between the sockets of the ocellar bristles than that between the posterior ocelli. This condition applies to both Periscelidinae and Somatia , but is less marked in Somatia than in some species of Periscelidinae , in which subfamily there is some variability in the distance between the ocellars. I do not find this degree of resemblance so distinctive, as compared with that in hypothetically related taxa that retain the ocellar pair of bristles, to convince one of synapomorphy. Most of the more diverse acalyptrate families in which ocellar bristles are commonly present show diversity in their placement, and this instability must often result in similarity which is not related to phylogenetic proximity.

(3) 7th abdominal tergum and sternum (female) fused, forming ring which includes the 7th pair of spiracles.

As mentioned above, this condition could be cited to support various relationship hypotheses for Somatia . Fusion of tergite 7 with sternite 7 has arisen many times in the Schizophora, and the superfamilies Nerioidea and Tephritoidea are the only major groups with the condition consistently present.

(4) Pregenital sclerite (male) extending ventrally on either side; 7th abdominal spiracles lying within this sclerite.

Actually, in Somatia the dorsal pregenital sclerite (apparently fused sternites 7 and 8) extends further downwards on the left side than on the right, because of inclusion of the laterally placed sternite 7. Such a condition is frequently found in the Schizophora as a stage in an often repeated reduction series (see D. McAlpine, 1985; 1988). In Gobrya and in numerous similarly reduced taxa the condition occurs with inclusion of one or more spiracles enclosed in the compound sclerite ( Fig. 8 ).

(5) Aedeagus (male) slender and ribbon-like, supported by broad strip of flexible sclerotisation.

The "ribbon-like" or tubular aedeagus of the Periscelidinae is now known to show some diversity (e.g., Mathis, 1993; Mathis & Papp, 1992). As an elongate, flexible tubular or strap-like aedeagus occurs in many families of acalyptrate Schizophora (in addition to most tephritoid families), this is not a very particular point of resemblance between Periscelidinae and Somatia . Some groups containing species with flexible, strap-like aedeagus also include others with quite different aedeagal structure (e.g., Clusiidae , Teratomyzidae , the heleomyzid tribes Allophylopsini and Gephyromyzini ).

I consider that, viewed in the broader field of schizophoran morphology, the supposed evidence for monophyly of the taxon Periscelidinae + Somatia does not hold up well.

In the broader concept of Periscelididae accepted by D. McAlpine (1983) and Mathis (1993) there is a distinctive apomorphy ofthe arista. Segment 5 (primitively the middle segment of the arista) is rather short, very asymmetrical, and reduced on the outer side, and the base of segment 6 is asymmetrically oblique. This structure can be traced through the periscelidid genera Scutops , Periscelis , Cyamops , and Stenomicra . In examined material of Stenomicra , segment 5 has disappeared, but the oblique base of segment 6 is retained. The aristal structure of Somatia ( Fig. 7 ) does not fit into this sequence. Segment 5 is more elongate, ovoid, and almost symmetrical; segment 6 is not noticeably oblique at the base and lacks the general tendency seen in periscelidids (also in Nothybidae but not in Syringogastridae ) for the dorsal rays to be crowded towards the base of the segment.

For these reasons I believe that the Somatiidae should not be merged with the Periscelididae . The Somatiidae have several strongly developed apomorphies, which have had a transforming effect on their general morphology, so that it is difficult to establish what the morphology of its lineage must have been before it reached such a degree of specialisation. I am unable at present to adduce a superfamily placement for this family.

Figures

Figs 5-12. 5, Nothybus decorus, left antenna, vestiture and arista (segment 6) simplified. 6, Gobrya cyanea, antennal segments 3-6 (note absence of segments 4, 5). 7, Somatia aestiva (Fabricius), antennal segments 3-6. 8, G. cyanea, protandrogram (sclerites of segments 5-8 of male abdomen shown as if split along median ventral line and spread flat). 9, G. cyanea, female postabdomen, ventral view. 10, 11, G. cyanea, epandrium and associated structures, posterior and right lateral views. 12, G. cyanea, aedeagus, right lateral view. c, cercus. ds, compound dorsal sclerite probably including tergite 6 and stemites 6 to 8. ep, epiproct. s5 -s7, stemites 5-7. ss, surstyli. t5-t8, tergites 5 to 8. ts6, tergostemite 6.