Abstract
Agromyza torta Eiseman & Lonsdale, sp. nov. (Figs. 16, 103, 104, 249–255) Holotype. USA. NORTH CAROLINA: Wake Co., Raleigh, Durant Nature Preserve, 2.v.2024, em. 21–23.iii.2025, T.S. Feldman, ex Celtis laevigata leaf roll, # CSE9336, CNC2138621 (1♂). Paratypes. NORTH CAROLINA: Same collection as Holotype, CNC2138622 (1♀); Scotland Co., Laurinburg, St. Andrews University, 11.iv.2022, em. 17–18.iii.2023, T.S. Feldman, ex Celtis laevigata (leafroller), # CSE8013, CNC2105108, CNC2105109 (2♀); same but 11.iv.2023, em. 27.iii.–2.iv.2024, # CSE8705, CNC2105128, CNC2105129 (2♀). Etymology. The specific epithet is derived from the Latin “ tortilis,” for “twisting” and “winding,” most notably referring to the leaf rolls formed by the larvae, but also applicable to the unusual and unique twisting basiphallus of the male. Host. Cannabaceae: Celtis laevigata Willd. Leaf mine. (Figs. 103, 104) Each of the specimens listed above was reared from a tender young leaf that was tightly rolled downward into an oblique, cigar-like roll that occupied the whole leaf (or nearly so) and entirely concealed the mine. When we unrolled some of these leaves, each was found to have a single mine, which appeared brown and full-depth (we were unable to determine on which leaf surface the mines originated, although one unaffected leaf collected along with the rolled leaves was found to have an early-instar larva forming a lower-surface mine; this may, however, have been a larva of A. aristata, one specimen of which was also reared from a collection of rolled leaves). One similarly rolled leaf had no mine inside, leading us to suspect that the rolling was somehow caused by oviposition rather than by larval feeding. The 11 April 2022 collection consisted of two rolled leaves, which yielded six puparia. Puparium. Reddish-brown; posterior spiracles large, C-shaped, with ~16 marginal pores on C-shaped plate; formed outside the mine. Phenology and voltinism. Larvae are active from late March to mid-April, with pupae overwintering and adults emerging the following spring. Distribution. USA: NC. Adult description. (Fig. 16) Wing length 2.2 mm (♂), 2.6 mm (♀). Length of ultimate section of vein M 4 divided by penultimate section 0.7. Eye height divided by gena height: 4.1–4.3. First flagellomere well-developed, slightly elongate with longer distal marginal hairs. Epistoma present, much broader than high, nearly as high as first flagellomere, deeper at middle. Body subshining, but thorax with dense gray-blue pruinosity. Orbital plate and parafacial visible laterally, modestly projecting, forming narrow ring around eye. Clypeus broad and straight anteriorly. Chaetotaxy: Two ori, two ors; relatively short and decreasing in length anteriorly, shorter than ocellars and larger postverticals. Four dorsocentrals, decreasing in length anteriorly, with anterior two thinner and fourth slightly longer than setulae; acrostichal setulae in six irregular rows. Coloration: Setae dark brown to black. Mostly dark brown; frons and anterior region of head light yellow with brownish spots at base of ors; broad semicircular dark spot around tubercle; posterolateral corner of frons dark brown to base of inner vertical; clypeus brown and first flagellomere deeper yellow. Thorax pruinosity dense grayish-blue; pigment slightly paler in faint irregular pattern from postpronotum to wing base. Wing veins light yellow with irregular brownish spots on R 1; calypter with whitish margin, hairs light brown to brown. Apices of femora light yellow for distance about equal to apex width; base and ends of tibiae light yellow (more extensive on fore leg, less so on hind), tarsi light yellow with apical segments slightly darker. Genitalia: (Figs. 249–255) Epandrium shallow, dome-like, fully fused to shallow, incurved, subtriangular surstylus that only bears short setae. Cercus length nearly equal to epandrium height, much narrower on distal ¼, and with several short, stout, darker setae on anterodistal surface. Hypandrium subtriangular with sides slightly bowed and apex with long, dark apodeme; inner lobe thick and curved with inner-basal region darker, with several short setae scattered along margin with weaker outer region. Postgonite with large apical lobe that is strongly incurved with minute setae along ventral surface; base of lobe bumpy, irregular, and covering base of long, dark, smooth band that fuses to phallapodeme. Ejaculatory duct wide, with dorsal and ventral sclerotized plates before ending at distiphallus, with the ventral plate very likely representing the mesophallus. Phallapodeme elongate, bent at midpoint, more weakly sclerotized on distal half (which is possibly homologous to base of basiphallus), open along most of right lateral surface, and with weakly sclerotized “window” on left distal half; basal half very dark along venter, bridge to epiphallus, and left ventrolateral strip. Basiphallus dark, curling around phallus shaft, with narrower base originating on left lateral surface where it is fused to dorsal and left lateral margins of phallophorus; basiphallus becomes much broader dorsally, and at right lateral surface, it splits into two branches; left branch apically paler, pointed and minutely spinulose; right branch flat, ribbon-like, with desclerotized line along midline, and inner face closely adjoining distoventral surface of ejaculatory duct. Distiphallus approximately half as long as basiphallus, relatively pale, and flared apically; with raised vental suture; seen ventrally, about as long as wide, narrowest medially, and with broad ventrobasal extensions that adjoin distal margin of ventral plate of ejaculatory duct. Ejaculatory apodeme subtriangular, very dark, especially along nearly straight distal and lateral margins, and with relatively broad, stout base; sperm pump broad, clear. Comments. Most observations of the leaf rolls on Celtis laevigata associated with this species have been found at a single locality in Scotland Co., North Carolina, but they have also been found in Wake Co. (iNat 213021473). TSF originally collected them on 13 April 2021 with the assumption that they were created by lepidopteran larvae, and when the first puparium appeared in a vial of leaf rolls (SA1962), it was transferred to a vial containing mined (but not rolled) leaves collected from the same tree on 6 April (SA1954). Three puparia were ultimately obtained from the leaves collected on 6 April, and from the four total puparia in the vial, three adults of Agromyza fission emerged (CSE6887, CSE7339). The fourth puparium, from which nothing emerged, probably belonged to A. torta. However, the only adult to emerge from the puparia that remained in the SA1962 vial was a male of A. aristata (CSE7378). We suspect that this male originated from a larva that happened to be mining a leaf that was rolled by A. torta, because every subsequent adult we have obtained from a rolled C. laevigata leaf belonged to this species. Similar leaf rolls that appear to be associated with mining have been found on C. reticulata in New Mexico (iNat 160440484). When lepidopteran larvae create superficially similar leaf rolls, they invariably do so using silk, which causes the leaves to curl because it contracts as it dries. We found no trace of silk in any of the leaf rolls that we opened, nor was there any external frass or feeding damage as there would be in a lepidopteran leaf roll (Eiseman & Charney 2010). We are therefore confident that Agromyza torta is responsible for the rolls, and although the mechanism is unknown to us, the rolling is evidently induced by the ovipositing female (see above under Leaf mine). The roll can be considered a type of gall, and it is noteworthy that the related species A. deserta also induces galls on Celtis. In addition to the distinctive leaf roll in which the larva forms its mine, the adult of Agromyza torta is also quite diagnostic. The body is dark with relatively dense bluish-gray pruinosity covering the thorax, and the head is contrastingly pale with a narrow ring around the eye, a dark semicircular spot around the tubercle, broad dark corners in the posterolateral corner of the frons and spots at the base of the ors. The male genitalia associate this species with the other Celtis feeders, but beyond this, most of the components are highly distinct and not readily confused for those of the others. The cerci are quite narrow apically with inner-apical spine-like setae; the basiphallus is dark, bifid and twisted; the distiphallus is smooth, short and broad, with the base adjoining a broad, pale mesophallus on the ventral surface.