Anthomyza gilviventris Roháček & Barber, 2016

Anthomyza gilviventris Roháček & Barber, 2016

( Figs 1, 2 View Figs 1–2 , 4–7 View Figs 3–7 , 10–13 View Figs 8–13 , 16–22 View Figs 14–22 , 25–31 View Figs 23–31 , 33–35 View Figs 32–35 , 37, 38 View Figs 36–38 , 40, 41 View Figs 39–41 )

New material examined. SWEDEN: Ångermanland: Bjurholm parish: river Lögdeälven, at the outlet of the stream Mossavattsbäcken, 64°2′24″N, 18°44′50″E, [river shore, sweeping], 2.vii.2016, 1 J 3 ♀♀ ( SMOC, SHU), 26.vii.2016, 1 ♀ ( SMOC); same locality, river shore, [sweeping] on Carex rostrata and C. acuta , 8.vii.2017, 1 J 1 ♀ ( NHRS, SHU), 15.vii.2017, 12 JJ 15 ♀♀ ( NHRS, SMOC, SHU, including 3 JJ 3 ♀♀ genit. prep., plus 1 J 2 ♀♀ used for molecular study,); Bjurholm parish: river Öreälven at Lagnäset, 63°54′58″N, 19°12′40″E, river shore, [sweeping] on Carex rostrata , 15.vii.2017, 6 JJ 4 ♀♀ ( SMOC, SHU, 3 JJ 1 ♀ genit. prep.); Västerbotten:Vindeln commune: Lerfallet, 64°21′41″N, 19°32′19″E, along a small creek, probably on Carex acuta , 7.vii.2021, 2 ♀♀ ( SHU), all Sven Hellqvist leg.

Morphological analysis. Male. The comparative morphology in the male sex ( Fig. 1 View Figs 1–2 ) has been focused on characters of the male genitalia considered by ROHÁĆEK & BARBER (2016) to best distinguish A. gilviventris and A. tschirnhausi . These mainly include (1) the shape of the epandrium (in caudal view), (2) the form of the anal fissure, (3) the shape and sclerotization of the transandrium and its caudal process, (4) the armature of the apex of the filum of the distiphallus, and, particularly (5) the armature of the saccus of the distiphallus. In addition, the shape and chaetotaxy of the gonostylus has also been compared in both species (including specimens from the Swedish population).

Epandrium. The epandrium (in caudal view) proved to be narrower and dorsally more tapered in A. gilviventris ( Fig. 4 View Figs 3–7 ) than in A. tschirnhausi ( Fig. 3 View Figs 3–7 ). In Swedish male specimens (see Figs 5–7 View Figs 3–7 ), the epandrium shape more closely resembles that of the epandrium in A. gilviventris specimens from Canada and the USA.

Anal fissure. Also, the anal fissure of the epandrium is distinctly narrow in A. gilviventris ( Fig. 4 View Figs 3–7 ) while in A. tschirnhausi ( Fig. 3 View Figs 3–7 ) it is more broadened ventrally. Although the shape of the anal fissure was found to be rather variable in Swedish specimens (see Figs 6–7 View Figs 3–7 ), it was never as wide as that of A. tschirnhausi and, consequently, their anal fissure proved to be more similar to that of Nearctic specimens.

Transandrium and caudal process. These structures should be narrower in A. gilviventris as suggested by ROHÁĆEK & BARBER (2016). Although illustrations of a typical specimen of A. gilviventris from Canada: Ontario ( ROHÁĆEK & BARBER 2016: fig. 491) and of a paratype of A. tschirnhausi from Kamchatka ( ROHÁĆEK 2009: fig. 62) indicate distinct differences, the widths of both the transandrium and its caudal process are, in fact, variable and most specimens from Sweden proved to be rather intermediate between both these extremes. Moreover, these structures can hardly be used for practical identification inasmuch as this posterior part of the hypandrial complex is difficult to correctly prepare to be precisely comparable in all specimens examined.

Filum apex. The apex of the filum of the distiphallus seems to be differently spinose with A. gilviventris having a number (usually> 10) of small subterminal spines (see Figs 13 View Figs 8–13 , 16 View Figs 14–22 ) while that of A. tschirnhausi bears only 3 of them ( Figs 8, 9 View Figs 8–13 ). The Swedish specimens have the terminal part of the filum variously spinose ( Figs 10–12 View Figs 8–13 ) and, rather surprisingly, bearing only 4–7 small spines thus differing in this respect from the Nearctic specimens. On the other hand, the apical part of the filum of the male holotype of A. tschirnhausi has only 3 (very minute) spines (apart from the two forming the bicuspid end on the apex, Fig. 8 View Figs 8–13 ), thus less than have specimens from Sweden. It also should be stressed that both A. gilviventris and A. tschirnhausi have the apex of the filum finely bicuspid ( Figs 8–13 View Figs 8–13 ) although this fact was not given in the original description of the latter species (see ROHÁĆEK 2009: 50). The extent and number of spinulae covering most of the filum, previously considered to be clearly greater in A. gilviventris than in A. tschirnhausi (cf. Figs 9 and 13 View Figs 8–13 ), have been found variable, ranging from relatively sparse to numerous in both Nearctic (usually more numerous, Figs 13 View Figs 8–13 , 16 View Figs 14–22 ) and Swedish populations (usually sparser, Figs 10–12 View Figs 8–13 , 17 View Figs 14–22 ) of A. gilviventris but also in A. tschirnhausi (the least numerous, see Figs 8, 9 View Figs 8–13 ).

Saccus. The distal membranous part of the saccus of the distiphallus is armed with some spines in the terminal part. Their number is somewhat variable as found by ROHÁĆEK & BARBER (2016) for A. gilviventris with 5–8 (most frequently 5, cf. Fig. 16 View Figs 14–22 ) distal spines. The same variability has also been confirmed in the specimens from the Swedish population where 5–7 spines near the apex of the saccus have been documented ( Figs 17–22 View Figs 14–22 ). In A. tschirnhausi , the saccus has been supposed to bear only 3 terminal spines ( Fig. 14 View Figs 14–22 ) (see also ROHÁĆEK 2009) but re-examination of the male holotype revealed that there could also be 4 such spines ( Fig. 15 View Figs 14–22 ). Nevertheless, this number remains less than the minimum number of spines in A. gilviventris . More importantly, we have found that the single spine near the base of the membranous part of the saccus (see Figs 16–22 View Figs 14–22 ) occurring in A. gilviventris (including all Swedish male specimens) is wholly absent in A. tschirnhausi (also confirmed in the holotype, Fig. 15 View Figs 14–22 ) while it is present in the other Nearctic species, A. shewelli , see ROHÁĆEK & BARBER (2016: 289).

Gonostylus. ROHÁĆEK & BARBER (2016) found that the shape and chaetotaxy of the gonostylus are variable in both Nearctic species, viz., A. gilviventris and A. shewelli . Despite this variability, the general outline (in largest extension view) of the gonostylus is distinctly different in these two species, mainly in the curvature of its anterior margin (convex in A. shewelli , concave to straight in A. gilviventris ). Because the formation and setosity of the gonostylus are normally species-specific characters in Anthomyzidae , we attempted to find differences in the gonostylus also in Palaearctic A. gilviventris and A. tschirnhausi . However, examination of the gonostylus of specimens from Sweden ( Figs 29–31 View Figs 23–31 ) and comparison with those of A. gilviventris specimens from the Nearctic Region ( Figs 25–28 View Figs 23–31 ) and those of A. tschirnhausi from Kamchatka ( Figs 23, 24 View Figs 23–31 ) has not revealed any clear differences. Also, the extent of the micropubescence and chaetotaxy of the inner side of the gonostylus proved to be variable and overlapping in both species, including within the Swedish population of A. gilviventris .

Female. Examination of females ( Fig. 2 View Figs 1–2 ) from Sweden revealed that they are (including characters of postabdomen) practically identical with the Nearctic specimens of the typical (= yellow) form of A. gilviventris . The narrow subterminal darkening of the female T7+S7 (best visible in Fig. 41 View Figs 39–41 but also in abdomen of dried specimens, see Fig. 2 View Figs 1–2 ) commonly occurs in the Nearctic specimens with a yellow abdomen (cf. ROHÁĆEK & BARBER 2016: fig. 484) but it can be absent in the lightest specimens having T7+S7 posteriorly almost entirely yellow ( Figs 37 View Figs 36–38 , 40 View Figs 39–41 ). According to ROHÁĆEK & BARBER (2016), females of A. tschirnhausi differ from those of A. gilviventris , apart from the darker abdomen, also by (1) the shape and surface structure of the spermathecae, (2) the shape of T7+S7, (3) the shape of T8 and (4) the shape of S8. However, both species also differ in the shape of sclerites of the 10th abdominal segment (T10, S10) as recognized below.

Spermathecae. In A. tschirnhausi females, the spermathecae are elongately ellipsoid having denser striae covering a larger part of their surface (see Fig. 32 View Figs 32–35 ), in contrast to those of A. gilviventris which are broadly suboval with sparser striae ( Figs 34, 35 View Figs 32–35 ). The latter, more broad, type of spermathecae has also been found in females from Sweden (see Fig. 33 View Figs 32–35 ).

Tergosternum T7+S7. Females of A. gilviventris have T7+S7 more strongly tapered posteriorly ( Figs 37 View Figs 36–38 , 40 View Figs 39–41 ) than have females of A. tschirnhausi ( Figs 36 View Figs 36–38 , 39 View Figs 39–41 ). Swedish females have this tergosternum also strongly narrowed posteriorly, thus very similar to that of Nearctic A. gilviventris , see Figs 38 View Figs 36–38 , 41 View Figs 39–41 ).

Abdominal tergum T8. In A. gilviventris , T8 is extremely long, slender (narrowest at about the middle) and relatively dark-pigmented as seen in both Nearctic ( Fig. 37 View Figs 36–38 ) and Swedish specimens ( Fig. 38 View Figs 36–38 ). Females of A. tschirnhausi also have T8 elongate and narrowed but it is distinctly wider, not further narrowing in the middle, and has paler pigmentation ( Fig. 36 View Figs 36–38 ).

Abdominal sternum S8. This sclerite is longitudinally divided into two sclerites as in other Anthomyza species but is markedly elongate and without micropubescence in all species of the A. tschirnhausi group. However, in A. gilviventris it is more slender and longer than in all other relatives ( Fig. 40 View Figs 39–41 ) as also seen in Swedish specimens ( Fig. 41 View Figs 39–41 ). In contrast, A. tschirnhausi has both parts of the female S8 distinctly the shortest and widest of all species of the group ( Fig. 39 View Figs 39–41 ), hence different from those of A. gilviventris . Thus, the Swedish females agree with typical females of A. gilviventris from the Nearctic Region in all postabdominal characters treated above.

10 th abdominal segment. Both sclerites of this segment are markedly more elongate in A. gilviventris (for T10 see Figs 37, 38 View Figs 36–38 ; for S10 see Figs 40, 41 View Figs 39–41 ) than in A. tschirnhausi ( Figs 36 View Figs 36–38 , 39 View Figs 39–41 ). Moreover, T10 of A. gilviventris normally has some smaller setae in addition to the medial pair of longer setae ( Figs 37, 38 View Figs 36–38 ). We have also found that specimens from Sweden have T10 distinctly narrower ( Fig. 38 View Figs 36–38 ) than specimens from Canada ( Fig. 37 View Figs 36–38 ) indicating certain differences between populations in N. Europe and N. America (as does the differently spinose apex of the filum of the distiphallus as described above).

Molecular analysis. Affinity of the Swedish specimens. We tested the relationships of the specimens from the Swedish population by means of both the Bayesian inference (MrBayes) and the maximum likelihood (RAxML) hypotheses. In both methods, all three specimens from Sweden were clustered together (see Fig. 42 View Fig ) and this cluster was resolved as a sister clade of the Anthomyza gilviventris specimen from Canada (Ontario). Anthomyza shewelli , represented by two specimens (both also from Ontario) in our dataset, is the only other species of the A. tschirnhausi group analysed. With the addition of specimens (of both species) to these analyses, A. shewelli is again recognized as a sister species of A. gilviventris , and consequently, their topology, as recognized by previous molecular hypotheses ( ROHÁĆEK et al. 2019), has been confirmed. However, because of the absence of fresh material of A. tschirnhausi from Kamchatka, it has not been possible to study its relationships by means of the above molecular methods. Inasmuch as this species was found to be the closest relative of A. gilviventris by means of the analysis of morphological characters (see above and also in morphological hypothesis by ROHÁĆEK & BARBER 2016: fig. 605), it can be presupposed that A. gilviventris is also genetically closer (if not conspecific with it) to A. tschirnhausi than to A. shewelli . This question can only be resolved in the future, as and when specimens of A. tschirnhausi become available for molecular study.

Relationships of the A. tschirnhausi group. This group, represented by A. gilviventris and A. shewelli in our dataset, has been confirmed as the sister group of the A. gracilis group with high support. This is in full agreement with both previous morphological ( ROHÁĆEK & BARBER 2016) and molecular ( ROHÁĆEK et al. 2019) phylogenetic hypotheses.

Barcoding of the species of the A. tschirnhausi group. The comparison of COI sequences of A. gilviventris specimens from Sweden and Canada and of A. shewelli specimens from Canada found that their pairwise genetic distances are relatively small (see Table 2). Distances of COI among specimens of A. gilviventris range from 0.0 to 0.5%, with the differences between Canadian and Swedish specimens being greater (0.4–0.5%) than among specimens from the Swedish population (0.0–0.2%). This indicates that the populations in Europe ( Sweden) and N. America ( Canada) are somewhat different. On the other hand, the distances of COI between A. gilviventris and A. shewelli specimens proved to be distinct though relatively small (2.2–3.0%) with, interestingly, the smallest distance occurring between Canadian specimens (2.2%). Biology. In the Nearctic Region, A. gilviventris mainly inhabits various wetland habitats, particularly those with sedges ( Carex and Scirpus species, see ROHÁĆEK & BARBER 2016) which may be host plants of this anthomyzid fly. Most frequently, Carex utriculata and Scirpus microcarpus have been encountered as dominant graminoids in the habitats where A. gilviventris commonly occurs, but this species was also found in growths of several other sedges (e.g. Carex aquatilis , C. stipata ) often intermixed with Equisetum fluviatile .

In Sweden, adults of A. gilviventris have been mainly collected on the shores of rivers ( Fig. 43 View Figs 43–44 ), all specimens in July (in Ontario adults of this species occur in May to September, see ROHÁĆEK & BARBER 2016).

Although various Carex species were swept at the Lögdeälven river locality, A. gilviventris specimens were found only in C. rostrata (4 JJ 11 ♀♀) and C. acuta (8 JJ 4 ♀♀) growing on the river shore (on July 15, 2018). The river bank is rather narrow and stony here, with C. rostrata growing in the water (intermixed with Equisetum fluviatile and Lysimachia thyrsiflora ) and C. acuta at the outlet of a small stream (see Fig. 44 View Figs 43–44 ). In both these sedges, A. gilviventris was found together with A. gracilis Fallén, 1823 and A. dissors Collin, 1944 , while A. paraneglecta Elberg, 1968 (1 J only) was found only in C. acuta . No specimens of A. gilviventris were swept from tussocks of C. nigra ssp. juncella on the river shore or from the growth of C. vesicaria in the adjacent riverine forest. At the second site (at the river Öreälven), A. gilviventris was swept from C. rostrata growing in dense stands on the river bank where it was broader and sandy. In the third locality, at Lerfallet, only 2 ♀♀ were swept at a small creek, most probably on Carex acuta (not precisely recorded) on July 7, 2021.

Distribution. The species is widespread in the Nearctic Region, with numerous records from Canada (Alberta, British Columbia, Labrador, Manitoba, Newfoundland, Northwest Territories, Nova Scotia, Ontario, Quebec, Saskatchewan, Yukon) and the United States of America (Alaska, Colorado, Idaho, Michigan, Montana, New York, Washington, Wisconsin, Wyoming), see ROHÁĆEK & BARBER (2016). The above records of A. gilviventris from NE Sweden (see map, Fig. 45 View Fig ) are the first from the Palaearctic Region. This population is considered native (the species has surely been previously overlooked in northern Europe) and, consequently, A. gilviventris is an additional (one of only a few) naturally Holarctic species in the family Anthomyzidae . Formerly, only Arganthomyza socculata (Zetterstedt, 1847) , Stiphrosoma humerale Roháček & Barber, 2005 and S. sabulosum (Haliday, 1837) were known to be distributed throughout the Holarctic Region (see ROHÁĆEK & BARBER 2005, 2016). However, the latter species could possibly be introduced in North America from Europe only in second half of 20th century ( ROHÁĆEK & BARBER 2005).