Australian gall-inducing scale insects on Eucalyptus: revision of Opisthoscelis Schrader (Coccoidea, Eriococcidae) and descriptions of a new genus and nine new species

Abstract We revise the genus Opisthoscelis Schrader, and erect the genus Tanyscelis gen. n. with Opisthoscelis pisiformis Froggatt as its type species. Species of both genera induce sexually dimorphic galls on Eucalyptus (Myrtaceae) in Australia, with Opisthoscelis subrotunda Schrader also in Papua New Guinea. We synonymise the following taxa (junior synonym with senior synonym): Opisthoscelis fibularis Froggatt, syn. n. with Opisthoscelis spinosa Froggatt; Opisthoscelis recurva Froggatt, syn. n. with Opisthoscelis maculata Froggatt; Opisthoscelis globosa Froggatt, syn. n. (= Opisthoscelis ruebsaameni Lindinger) with Opisthoscelis convexa Froggatt; and Opisthoscelis mammularis Froggatt, syn. n. with Opisthoscelis verrucula Froggatt. We transfer seven Opisthoscelis species to Tanyscelis as Tanyscelis conica (Fuller), comb. n., Tanyscelis convexa (Froggatt), comb. n., Tanyscelis maculata (Froggatt), comb. n., Tanyscelis maskelli (Froggatt), comb. n., Tanyscelis pisiformis (Froggatt), comb. n., Tanyscelis spinosa (Froggatt), comb. n., and Tanyscelis verrucula (Froggatt), comb. n. We redescribe and illustrate the adult female of each named species of Opisthoscelis for which the type material is known, as well as the first-instar nymph of the type species of Opisthoscelis (Opisthoscelis subrotunda) and Tanyscelis (Opisthoscelis pisiformis). We describe four new species of Opisthoscelis: Opisthoscelis beardsleyi Hardy & Gullan, sp. n., Opisthoscelis thurgoona Hardy & Gullan, sp. n., Opisthoscelis tuberculataHardy & Gullan, sp. n., and Opisthoscelis ungulifinis Hardy & Gullan, sp. n., and five new species of Tanyscelis: Tanyscelis grallator Hardy & Gullan, sp. n., Tanuscelis megagibba Hardy & Gullan, sp. n., Tanyscelis mollicornuta Hardy & Gullan, sp. n., Tanyscelis tripocula Hardy & Gullan, sp. n., and Tanyscelis villosigibba Hardy & Gullan, sp. n. We designate lectotypes for Opisthoscelis convexa, Opisthoscelis fibularis, Opisthoscelis globosa Froggatt, Opisthoscelis maculata, Opisthoscelis mammularis, Opisthoscelis maskelli, Opisthoscelis pisiformis, Opisthoscelis recurva, Opisthoscelis serrata, Opisthoscelis spinosa, and Opisthoscelis verrucula. As a result of our taxonomic revision, Opisthoscelis has six species and Tanyscelis has 12 species. We describe the galls of females for all 18 species and galls of males for 10 species of Opisthoscelis and Tanyscelis, and provide photographs of the galls for most species. A key to the adult females of the species of both genera is included.


Introduction
pointed out that gall-inducing insect taxa tend to be species-rich when they specialise on species-rich host groups. Australian gall-inducing scale insects are no exception. A large proportion of the Australian gall-inducing scale insect fauna is associated with the hyper-diverse genus Eucalyptus (Gullan et al. 2005). Scale insect galls on Eucalyptus sensu lato are a conspicuous element of the Australian fauna (at least to botanists) and have even been incorporated into human diets Latz 1995). Gall-induction has evolved repeatedly in scale insects (Gullan et al. 2005), including within the Eriococcidae (Cook and Gullan 2004), the scale insect family with the most gall-inducing species. Th e two most species-rich radiations of gall-inducing eriococcids are parasites of Eucalyptus: (1) the genus Apiomorpha Rübsaamen (ca 40 described species and many undescribed cryptic species) (Gullan 1984a;Cook and Rowell 2007), and (2) a monophyletic group (NBH unpublished data) composed of species currently placed in the genera Lachnodius Maskell, Opisthoscelis Schrader, Sphaerococcopsis Cockerell, Fragorbis Hardy and Gullan, and Floracoccus Beardsley (ca 30 described species). Beardsley (1974a, b) described or redescribed the four species of Sphaerococcopsis and the single species of Floracoccus.  erected the genus Fragorbis for fi ve species that mostly induce blister galls on stems or live under bark. Taxonomic work on a number of species closely related to Lachnodius eucalypti Maskell, initiated by the late J.W. Beardsley, is currently in preparation (NHB, PJG and J.W. Beardsley unpublished data). Schrader (1863) erected the genus Opisthoscelis for O. subrotunda, although O. subrotunda was not offi cially designated as the type of the genus until Fernald's (1903) catalogue. Rübsaamen (1894) described O. globosa, the type material of which appears to have been destroyed in WWII (Gullan 1984a). Froggatt (1894a,b, 1898, 1929), Fuller (1897, 1899 and Kireff er and Jorgensen (1910) described 14 additional species. In his catalogue of the scale insects of Australia, Froggatt (1921) republished his earlier descriptions and illustrations of eight Opisthoscelis species and provided notes on three species described by other authors. Opisthoscelis prosopidis Kireff er and Jorgensen was described from Prosopis adesmiodes (Fabaceae) in Argentina. It is almost certainly not congeneric with Eucalyptus-feeding species in Australia (Miller and Gimpel 2000) and we do not deal with it in this paper. Unfortunately, the whereabouts of the type material of O. prosopidis is unknown (Miller and Gimpel 2000;Hodgson and Miller 2010). Lindinger (1943) chose O. ruebsaamani as a replacement name for O. globosa Froggatt (1929), which is a junior primary homonym of O. globosa Rübsaamen. Hardy and Gullan (2009) transferred O. nigra Froggatt, described from galls on Eucalyptus in New South Wales (Froggatt 1898), to the armoured scale genus Maskellia Fuller (Diaspididae).
Here we revise the genus Opisthoscelis, and transfer a number of Opisthoscelis species into the new genus Tanyscelis. We describe and illustrate: (1) the adult female of each of the previously named species of Opisthoscelis for which the type material is known; (2) the fi rst-instar nymph of the type species of Opisthoscelis and Tanyscelis; and (3) the adult female for each of eight new species. We synonymise a number of species described by Froggatt. We describe the galls of females and, if available, of males of the 18 species of Opisthoscelis and Tanyscelis, and provide a key to the adult females of Opisthoscelis and Tanyscelis.

Methods
Freshly collected specimens were slide-mounted in Canada balsam using the method described in Gullan (1984a). Adult females were mounted one specimen per slide; nymphs were mounted several to a slide. Froggatt prepared very few slidemounts and thus his dry type galls were crucial to confi rming the identity of his species. For most of the named species, adult females were dissected from dry type galls in museum collections and mounted on microscope slides by PJG. Th e dry insects were soaked in ®Decon 90 or ®Contrad 70 detergent for 24 hours and then placed in water to rehydrate the specimen prior to clearing of body contents in 10% KOH and then slide-mounting according to the usual method. Descriptions of all life stages are based on slide-mounted specimens. Measurements of insects were made using an ocular micrometer attached to a compound microscope. All measurements are maximum dimensions (e.g., body width was recorded at the widest point and leg segment lengths were measured along the longest axis), and are expressed as the range. Tarsal length excludes the claw. Spiracle length includes the muscle plate (apodeme). Setal lengths exclude the setal base. Th e length of the macrotubular ducts excludes the fi lament. Th e hind leg of the adult female is described as having a functional articulation between the femur and tibia if intersegmental membrane is present and the leg is capable of fl exing at the joint. Th e number of specimens measured is reported in the heading for each description, although usually many more specimens were examined, as reported in the lists of 'Material examined'. Samples in each 'Material examined' section were sorted in ascending order algorithmically, starting with the fi rst character in each locality, with numeric characters having a higher priority than alphabetic characters. For example '11 km N of Bundoora' would appear before 'Atherton,' and '45 km S of Brisbane' would appear before '9 km E of Townsville. ' Th e morphological terms for Eriococcidae used in the descriptions follow those of Williams (1985), Miller and McKenzie (1967) and Afi fi (1968). We use the term 'vestibule' for the inner end of a tubular duct, where the contents of the inner ductule empty into the main duct (as in Miller and McKenzie 1967), and the term 'cribriform plates' refers to sclerotic areas of cuticle that house aggregations of macrotubular ducts, from each of which, in live insects, a hollow fi lament of wax is produced. Illustrations of the adult females and fi rst-instar nymphs of the new Australian species were prepared by NBH with a drawing tube and the Adobe programs Photoshop CS and Illustrator CS. Following the convention for scale insects, each fi gure displays the dorsal body surface on the left side of the page, and the ventral body surface on the right. In general, translucent pores, small setae, ducts and pores on the main drawing are not drawn to scale because otherwise they would be too small to be seen easily. Enlargements of cuticular features that are placed around the main drawing also are not drawn to scale. We have described the galls of the adult females of each species and, if available, also the gall of the male insects. As far as known, we have indicated the leaf surface that bears the gall orifi ce (the opening to the chamber housing the insect), but many Eucalyptus species have adult leaves that are pendulous and isobilateral (Chippendale 1988;Evans and Vogelmann 2006), and thus often it is diffi cult to distinguish the abaxial (lower) and adaxial (upper) leaf surfaces, especially from isolated leaves in museum collections. In addition, we have provided photographs of the galls of females and, if available, the galls of male insects for all but one species (Figs 1-3). Th ese fi gures are placed together at the beginning of the descriptions to facilitate comparisons of the galls, which are useful for species determination.
Th is study draws extensively from the collections of P. J. Gullan (98 records), L. G. Cook (25 records), H. M. Brookes (25 records) and N. B. Hardy with PJG (19 records). Th e late Helen Brookes (formerly at the Waite Agricultural Research Institute, University of Adelaide, South Australia) acquired and curated a large collection of Coccoidea (including Opisthoscelis and Tanyscelis) during her career and, following her retirement in 1982, deposited this collection and associated notes and photographs in the Australian National Insect Collection (Upton 1997;Taylor and Keller 2008). Brookes also recorded biological and other information on cards fi led by a Specimen Index Number that consisted of a number for the collection and an abbreviation of the year, for example, 44/65 was her 44 th collection for 1965 (Gullan and Williams in press). We have made use of her information on the biology and appearance of some Opisthoscelis and Tanyscelis species.
In the descriptions, we list the slide-mounted specimens that were examined, but most of these collections have associated dry gall material that is too numerous to record and list. Galls and associated slide-mounted insects are stored in the same collections. For many of the species, one to several slide-mounted specimens listed in the 'Material examined' are DNA voucher specimens of either L.G. Cook (e.g.,LGC00101 for O. thurgoona) or NBH (e.g., NH104 for O. thurgoona) and all are deposited in ANIC (see below). Th e reliability of early (1890s to 1960s) identifi cations of eucalypt host plants is uncertain, but more recent records (1970s to present) are considered reliable. Th e classifi cation of Brooker (2000) is used for taxonomic sections within Eucalyptus.
Depositories are abbreviated as follows: ANIC, Australian National Insect Collection, CSIRO, Canberra, ACT, Australia; ASCU, Agricultural Scientifi c Collections Th e International Commission on Zoological Nomenclature (1999) requires lectotypes designated after 1999 to "contain an express statement of deliberate designation" (amended Article 74.7.3). We use the statement 'here designated' to fulfi ll this requirement. Lectotypes have been designated where a name lacks a holotype or  lectotype and unambiguous syntypes have been identifi ed. Th e purpose is to provide stability of nomenclature, and designation is done in a revisionary context in agreement with the amended Recommendation 74G of Article 74.7.3.

Taxonomy
Th e key by  to adult females of genera of Eriococcidae found on Eucalyptus and Corymbia in Australia can be used to distinguish the adult female insects of Opisthoscelis and Tanyscelis from those of all other eriococcid genera on Eucalyptus. Species of Tanyscelis will key to Opisthoscelis in .  Generic diagnosis. Adult female and associated gall. Galls on leaves; each typically globular, conical or hemispherical, rarely pit-like; with small circular, slit-like or fi ssured orifi ce/opening often on adaxial (upper) surface, but surface often diffi cult to determine in mature isobilateral leaves and galls typically opening on same surface on any one leaf. Body outline circular to elliptical; dorsum usually smaller than venter, especially at maturity, with whole of dorsum, a sclerotised part of it, or abdominal apex plugging gall orifi ce; body of mature female usually tightly fi tting gall cavity. Abdomen not tapered. Vulva and anal opening ventral on posterior abdomen, with vulva between abdominal segments VII and VIII. Eyes on margin or dorsad of it. Antennae reduced, with ≤ 7 segments, segmentation often indistinct. Pair of broad frontal lobes posteromedial of antennae. Oral lobes membranous to sclerotic, often forming large circular pad around mouthparts. Tentorial box with aliform anterior extensions. Fore and mid legs varying from well developed to highly reduced, with some segmentation always apparent. Hind legs, on posterior of metathorax, sometimes appearing to be on anterior abdomen, large, always well developed, often elongate; ratio of length of trochanter + femur / length of tibia + tarsus approximately = 1:1; tibia-femur articulation functional; translucent pores present on at least tibia; trochanter with 2 or 3 campaniform sensilla on each side; claw digitules and tarsal digitules either well developed or highly reduced to absent. Anal opening surrounded by anal ring with ≥ 6 (range 6-20) setae; anal ring may be invaginated. Posterior abdominal segments usually with marginal fl eshy projections (absent in O. thurgoona sp. n. and O. tuberculata sp. n.), each projection bearing spines or with a blunt sclerotic point. Marginal fringe of enlarged setae usually present (absent in O. serrata). Dorsal setae bristle-like to fl agellate, minute to large, 4-148 μm long (stout conical setae present in O. subrotunda). Cribriform plates or tight clusters of tubular ducts either present or absent on dorsum. Microtubular ducts usually absent (present in O. ungulifi nis sp. n.). Dorsal quinquelocular pores absent. Venter with macrotubular ducts present or absent; if present, with vestibule thin and sclerotic. Quinquelocular pores present on venter, at least around vulva and spiracles.
Adult male. Antenna 10-segmented. Abdomen not elongated. Gland pouches present, each with pair of setae.
First-instar nymph. Anterior margin of head incised at midline. Each spiracle with one trilocular pore next to opening. One submedial longitudinal row of dorsal setae on each side of body. Antennae 4-segmented, with 4 fl eshy setae on apical segments.
Etymology. Th e genus name is a Latinised combination of the Greek words opisthen, meaning behind, and skelos, meaning leg, and clearly refers to the long hind legs of the adult female because Schrader (1863: 6) diagnosed Opisthoscelis with one brief statement: "Where they have only two long posterior legs." Th e name Opisthoscelis is treated as feminine. Gall ( Fig. 1a,b,c). Female. On leaf. Gall opening circular, 0.3-0.4 mm wide, blocked by white wax and/or part of sclerotised posterior dorsum of female. Gall variable, usually 3-5 mm in diameter, side with opening subconical (Fig. 1a) or hemispherically rounded (Fig. 1b), with truncate apex, typically on adaxial surface; leaf glands enlarged.

Opisthoscelis beardsleyi
Male. On leaf. Similar to gall of female but smaller and more cylindrical, 2-3 mm high, basal diameter 1.5-2.0 mm, with circular opening and truncate apex ca 1 mm across (Fig. 1c), usually on adaxial leaf surface.
Dorsum. Dorsal shield much smaller than venter in old specimens, of variably sclerotic cuticle 1.1-2.0 mm long, 1.0-2.7 mm wide, clearly delineated by marginal fringe of close-set spinose setae, each seta 25-70 μm long. A pair of fl eshy caudal projections on each side of body; each medial projection with ca 4 spinose setae, each lateral projection (probably marginal lobe of abdominal segment VII) with usually 3 spinose setae. Dorsal setae mostly small and robust, 8-23 mm long, scattered over dorsum, a few large spinose setae similar to those in marginal fringe sometimes occur along submargin. Macrotubular ducts occurring in clusters of 1-25 ducts set within heavily sclerotic cribriform plate; shaft of each duct short, < 5 μm long; number of cribriform plates on dorsum varying from ca 10 to ca 40 on each side of body. Microtubular ducts absent. Quinquelocular pores absent.
Venter. Oral lobes membranous. Setae 8-100 mm long, in transverse row or band across each abdominal segment, metathorax, and mesothorax, along margin of head, pro-and mesothorax; setae dense posterodorsal of anal ring. Macrotubular ducts absent. Quinquelocular pores 5-6 μm in diameter, in clusters around each spiracle and along margin of thorax and anterior abdominal segments.
Material examined. Comments. Th e adult female of O. beardsleyi is most similar to that of O. subrotunda in that both species have dorsal cribriform plates, a marginal fringe of spinose setae, and ≥ 2 spinose setae mounted on each caudal fl eshy projection or lobe (these projections poorly developed in some populations of O. beardsleyi). Adult females of O. subrotunda can be recognised by having the dorsal derm densely beset with minute papillae (found in some species of Tanyscelis but no other Opisthoscelis species) and large, spinose dorsal setae (dorsal setae minute in Tanyscelis, or bristle-like in other species of Opisthoscelis).
Galls of adult females of O. beardsleyi may be conical ( Fig. 1a) or rounded with either a truncate ( Fig. 1b) or convex apex. Adult females vary in the number of cribriform plates and enlarged setae on the dorsal submargin as well as the extent to which the caudal projections are developed. It is possible that the material we have listed under the name O. beardsleyi constitutes multiple species, but we have not been able to detect any clear patterns in morphological variation over geography or host use. Th erefore, we have restricted the type material to a single collection. Specimens of O. beardsleyi have been collected only from southeastern Australia, from eucalypt species in the Symphyomyrtus section Maidenaria.
Etymology. Th is species is named in honour of the late Dr. Jack Beardsley, who was an expert on the systematics of parasitic wasps and scale insects. He spent most of his entomological career at the University of Hawaii, but began studies on Australian gall-inducing eriococcids during a research sabbatical spent in Victoria in 1971-1972. He was interested especially in pit-inducing and cryptic bark-living eriococcids and amassed a substantial collection from Victoria, including some specimens of Opisthoscelis and Tanyscelis. Froggatt  Figs 1d,e,f, 5
Dorsum. Head with shield of rugose, sclerotic cuticle, 600-1000 μm long, 450-1100 μm wide; each side of abdomen with segmental pairs of marginal spines, each spine at end of fl eshy projection, these larger on posterior segments; each thoracic segment with a few similar structures along the margin; marginal fringe of elongate setae absent. Derm densely beset with microtrichia. Dorsal setae minute, 4-8 mm long; scattered over dorsum. Macrotubular ducts 12-15 μm long, rim of dermal orifi ce 5 μm wide; ducts in a transverse row across each body segment. Microtubular ducts absent. Quinquelocular pores absent.
Venter. Oral lobes membranous. Ventral setae 5-115 mm long, in transverse row across each abdominal segment, a medial to submedial cluster on each thoracic segment, scattered along margin, longer setae found in medial areas. Macrotubular ducts similar to those on dorsum, scattered along margin, in a transverse row across each of abdominal segments II-IV. Quinquelocular pores 5 mm in diameter, in cluster around each spiracle, along margin of abdomen, in a transverse band across each posterior margin of abdominal segments IV-VIII.
Th e lectotype and associated paralectotype are in poor condition and were mounted from the spirit collection in ASCU by PJG in 1985; additional females, in poor condition, remained in the vial; no original WWF slide-mount was located.
Additional material: AUSTRALIA: New South Wales: 5 adult females, 5 secondinstar females: ex galls on leaves, Eucalyptus largifl orens, Kinchega National Park, 2 July, 1984, J. M. Smith (ANIC).  (1) the shield of rugose sclerotic cuticle on the dorsal surface of the head, and (2) the pair of marginal spine-tipped fl eshy projections on each side of each abdominal segment. Both of these features are unique amongst Opisthoscelis and Tanyscelis species. O. serrata is known only from eucalypt species in the section Adnataria. Th ere is some variation in the shape and size of the gall of the adult female, which is basically globular. Galls on leaves of E. largifl orens and E. polyanthemos protrude as a sphere on one side of the leaf, although some galls from leaves of E. polyanthemos at Warrandyte have a pointed apex (Fig. 1d). In contrast, galls from the leaves of E. melliodora and E. microcarpa from near Nagambie protrude equally from both leaf surfaces and the females are smaller than those from most other localities. Females from all galls look similar, although diff ering in size. Figs 1g,h,i,6,7 Opisthoscelis subrotunda Schrader, 1863: 7. Opisthoscelis globosa Rübsaamen, 1894: 214;synonymy by Froggatt, 1894a: 209. General. Schrader (1863) described and illustrated O. subrotunda based on insects and galls from an unidentifi ed Eucalyptus sp. collected in New South Wales. Schrader's collection was destroyed on July 30, 1943, during WWII according to Ew. Weidner in personal communication to PJG (Gullan 1984a). Rübsaamen (1894) described a number of Australian gall-inducing insects, including O. globosa from Eucalyptus in New South Wales, while his address was "Königlichen Museum für Naturkunde zu Berlin". We contacted Museum für Naturkunde Leibniz-Institut für Evolutions-und Biodiversitätsforschung an der Humboldt-Universität zu Berlin, but were told that there was no Opisthoscelis material from Rübsaamen in that collection, and thus we believe that Rübsaamen's material also was destroyed during WWII, as suggested by Gullan (1984a). Th e synonymy of O. subrotunda and Rübsaamen's O. globosa was proposed by Froggatt (1894a) but not accepted by Hoy (1963); subsequent authors have followed Hoy, although Miller and Gimpel (2000) stated that the status of these two species needed further study. Here we accept Froggatt's (1894a) synonymy of these two species based on the perfect match of Schrader's and Rübsaamen's descriptions and illustrations of the insects and their galls. Th ere is only one species in New South Wales that exactly matches Schrader's and Rübsaamen's descriptions. Th us we are certain of the identity of O. subrotunda and consequently have not designated a neotype.

Opisthoscelis subrotunda Schrader
Gall (Fig. 1g,h,i). Female. On leaf. Gall globose (Fig. 1g), surface punctate, leaf glands enlarged on gall. Gall height 4.0-8.7 mm, width 4.2-9.5 mm, length of basal attachment 3.0-6.8 mm; wall of gall at least 2 mm thick in most mature galls. Gall opening variable, slit-like, cruciform ( Fig. 1h) or round, sometimes with "lips" projecting, 0.1-2.5 mm wide; opening on abaxial or adaxial surface, but usually on same surface on any one leaf or plant. Mature female fi lls gall cavity with her abdominal apex directed towards and plugging gall orifi ce.
Dorsum. Delineated by fringe of spinose setae along margin, each seta similar to those on rest of dorsum; cauda composed of two truncate fl eshy projections on each side of body, each bearing ca 4-6 spinose setae. Derm variously sclerotic, densely beset with minute papillae. Dorsal setae spinose, 20-138 mm long, with stout, or swollen bases. Microtubular ducts absent. Quinquelocular pores absent.
Dorsum. Derm membranous. Dorsal setae ca 3 mm long; arranged in submedial longitudinal row on each side of body, 1 seta on each side of head, each thoracic segment, and each or any of abdominal segments I-III and V. Microtubular ducts 4 μm long, on each side of body: 1 duct on submargin of each side of each thoracic segment plus usually each of abdominal segments I, V and VIII. Marginal setae cylindrical, with hemispherical base with diameter > width of cylinder, distal edge of base fl at, each marginal seta 2-5 μm long; each side of body with ca 6 setae between midline and eyes, 9 on prothorax, 5 on mesothorax, 3 on metathorax, 2 on each of abdominal segments I-VII, and on abdominal segment VIII 1 lateral and 2 medial of apical seta (these 3 probably homologous to anal lobe setae), with setae medial of apical seta robust, each with truncate apex.
Material O. subrotunda also appears to have a much broader host range than is typical for species of Opisthoscelis and Tanyscelis. Specimens have been collected from Corymbia in addition to hosts in the Eucalyptus sections Adnataria, Exsertaria, Latoangulatae, and Similares (all in subgenus Symphyomytrus). A common host of this species is the river red gum, E. camaldulensis, which has the widest natural distribution of any eucalypt species and seven subspecies (Chippendale 1988;McDonald et al. 2009). Th e length of the dorsal setae, and the size of the cribriform plates of the adult females of O. subrotunda vary, but in general, the morphology is remarkably homogeneous across samples. Molecular data from collections of fresh material from across the geographic range of the taxon would be required to determine if cryptic species are present. Gall (Fig. 1j). Female. On leaf. Immature gall (housing yellow nymph with a red dorsal "keel") shallowly conical to hemispherical on orifi ce side (Fig. 1j), with gall surface whitish green to reddish; almost hemispherical on opposite side of leaf, tissue reddish green with distinct oil glands on surface. Gall height on orifi ce side ca 1 mm, ca 3 mm on opposite side, width 4-6 mm. Gall opening on either abaxial or adaxial leaf surface, slit-like, 0.5-1.0 mm long. Mature gall with tissue surrounding orifi ce brown and necrotic.
Dorsum. Delineated by fringe of elongate setae, each seta 50-180 long, with blunt to truncate apex. Derm membranous. Dorsal setae minute, 5-9 mm long, scattered. Macrotubular ducts in two forms: (i) singly or in clusters of 2-8 ducts, cuticle of each cluster variously sclerotic, each duct short, ca 5 μm long, inner ductule not detected, scattered over dorsum, dense on head, posterior abdominal segments and margin of anterior abdominal segments and thorax; and (ii) single, larger ducts, 14-16 mm long, with dermal orifi ce with a rim 3.5-5.0 mm wide, scattered over dorsum, scarce or absent from posterior abdominal segments. Microtubular ducts absent. Quinquelocular pores absent.
Venter. Oral lobes membranous. Setae 13-103 mm long, in a transverse row across each abdominal segment, scattered along margin of head and thorax, on medial areas of meso-and metathorax, and a transverse band of elongate (up to 170 μm long) setae posterior of frontal lobes. Macrotubular ducts with distal (near vestibule) end of shaft constricted, about same size as larger ducts on dorsum. Quinquelocular pores 5 μm in diameter, overlapping in distribution with ventral setae, with a cluster around each spiracle, but most dense medially on posterior abdominal segments.
Material Etymology. Th e species name is taken from the type locality. It is a noun in apposition.
Venter. Oral lobes membranous. Ventral body setae sparse, mostly 50-100 μm long, up to 150 μm long on head behind frontal lobes, shorter (20-50 μm long) in lateral pore band. Macrotubular ducts slender, inner portion dilated, 15-17 μm long, 2.0-2.5 μm wide, with poorly sclerotised rim 2.5-3.5 μm in diameter; ducts distributed in a moderately dense peripheral band, ducts distributed evenly across abdominal segments III-VI, but apparently absent on segments immediately in front of and behind vulva where disc pores are most numerous. Quinquelocular pores 5 μm in diameter, mostly confi ned to abdominal segments V-VIII anterior to anal ring, but a few (12-20) associated with each spiracle. Th e reference to Berlese may refer to Professor Antonio Berlese (dal Laboratorio di entomologia agraria, Portici, Italy), because two nearby entries in the notebook say "Sent to Berlese" and give diff erent numbers. Th e 3 adult females from Victoria were collected dead (post-oviposition), in association with their salmon-pink eggs, under loose bark just above the ground on the main trunk of a stunted tree (about 1.5 m high) growing in the middle of a picnic area car park. Several leaves on the tree had vacated pit galls (Fig. 1k), which are presumed to have been the feeding sites of the females prior to their movement to the bark for oviposition. Th e live eggs were maintained in a small plastic container and started eclosing on 1 Jan., 2009; DNA was extracted from the fi rst-instar nymphs by NBH.
Th ere is only minor variation between the adult female holotype from New South Wales and the three paratype females from Victoria; in particular, there are fewer clusters of macrotubular ducts on the dorsal submargin of the Victorian collection. Etymology Th e species name is a manuscript name of the late J. W. Beardsley who was describing the species as part of a revision of Lachnodius. We assume that his name refers to the sclerotised dorsal duct clusters. Gall (Fig. 1l). Female. On leaf. Raised, rather nipple-like excrescences; height ca 2 mm above leaf and projecting 1 mm below, width 3-5 mm. Gall opening star-like, with fi ssures radiating from central orifi ce, apparently on adaxial leaf surface.
Dorsum. Delineated by marginal fringe of elongate (up to 300 μm) fl agellate setae, fringe extending from head to abdominal segment IV, margin of each of abdominal segments V-VIII with pair of blunt sclerotic projections at end of short fl eshy lamella. Derm densely beset with microtrichia. Dorsal setae fl agellate, 15-148 mm long; scattered over dorsum. Macrotubular ducts 14-15 μm long, dermal orifi ce opening with rim 5 μm wide; ducts in a transverse row across each abdominal segment, scattered over head and thorax. Microtubular ducts 5 μm long, dermal orifi ce with rim 2 μm wide; distributed evenly over dorsum Quinquelocular pores absent.
Venter. Ventral body surface greatly expanded, posterior abdominal segments only visible from dorsal aspect. Oral lobes membranous. Ventral setae 13-250 mm long, in transverse row across each body segment, scattered along margin. Macrotubular ducts similar to those on dorsum, scattered along margin, in a transverse row across each of abdominal segments II-V. Quinquelocular pores 5 mm in diameter, in a cluster around each spiracle, scattered along margin, in a transverse row across each abdominal segment, most dense around vulva.
Comments Etymology. Th e species name is a combination of the Latin words ungula, meaning hoof, and fi nis, meaning end or boundary. It refers to the blunt sclerotic projections at the apex of each marginal lamella. Th e name is a noun in apposition.
General. Th is species was described briefl y by Schrader (1863: 7) as follows: "In another species O. gracilis, which I have observed, the oviparous female is rather slender, and the legs are still longer and thinner than in the species before noticed [O. subrotunda], and the male has no anal setae." Th ere is no illustration and Schrader's specimens were destroyed in WWII (Gullan 1984a Type species: Opisthoscelis pisiformis Froggatt 1894b. Generic diagnosis. Adult female and associated gall. Galls on leaves and stems; each either globular, hemispherical, conical, mammiform or thorn-like, never pit-like; with orifi ce usually slit-like, sometimes circular to oblong; if galls on leaves, orifi ce on either abaxial or adaxial surface, depending on Tanyscelis species, but surface often diffi cult to determine in mature isobilateral leaves and galls typically opening on same surface on any one leaf. Female attached by mouthparts to internal gall wall opposite to opening, with long hind legs projecting forwards over body or towards gall orifi ce; body of female fi lls only part of gall cavity; in some species dorsum often covered with white powdery wax. Body outline ovate to turbinate. Abdomen tapered. Vulva on ventral surface near apex of abdomen between segments VII and VIII. Eyes on dorsum close to margin or a papilliform fl eshy protuberance on each side of head in place of eyes. Antennae reduced, with ≤ 3 segments, usually 1-segmented. Pair of broad frontal lobes posteromedial of antennae (possibly absent in T. villosigibba). Oral lobes membranous to sclerotic, sometimes forming large circular pad around mouthparts. Tentorial box with aliform extensions. Fore and midlegs reduced, segmentation indistinct. Hind legs usually positioned on extreme posterior of thorax and often appear to be located on anterior abdomen; elongate, sometimes nearly as long or longer than body; ratio of length of trochanter + femur / length of tibia + tarsus < 1; tibia-femur articulation either functional or tibia and femur fused to varying degrees; translucent pores present at least tibia or tibiotarsus; trochanter with 2-6 campaniform sensilla on each surface; claw always highly reduced to absent, claw and tarsal digitules often present but reduced. Anal opening apical or dorsal; anal ring absent or small, without pores and no more than 6 minute setae. Posterior abdominal segments usually without marginal fl eshy projections bearing spines (present near anal opening in T. verrucula and T. pisiformis). Margin without linear fringe of diff erentiated setae (though T. convexa with submarginal band of staggered spines on venter extending from anterior portion of head to posterior spiracles). Dorsum may be humped, or have large fl eshy projections. Derm membranous, sclerotic, or bearing minute papillae. Dorsal setae usually fl agellate, minute to very long, 5-220 μm long, sometimes born upon fl eshy spines. Cribriform plates and tu bular duct clusters absent. Microtubular ducts absent from both body surfaces. Macrotubular ducts with vestibule thin and sclerotic, may be present on dorsum and venter, dorsum only, venter only, or absent. Quinquelocular pores on ventral body surface, sometimes throughout dorsum or confi ned to dorsal surface of abdomen.
Adult male. Antenna 9-segmented. Abdomen tapered, elongate in some species. Gland pouches and associated setae absent. Detailed description given by Th eron (1968). Th eron based his description on specimens of two unidentifi ed species collected from Keith, South Australia, ex E. fasciculosa, provided to him by HMB. Exactly two species from this host and locality, and collected prior to 1968, are present in the collection of HMB: O. spinosa and O. maculata, here transferred to Tanyscelis.
In contrast, the known adult males of Opisthoscelis have 10-segmented antennae, a non-tappered abdomen, and have gland pouches, each with a pair of gland setae.
First-instar nymph. Anterior margin of head incised at midline. Each spiracle with one trilocular pore next to opening. One submedial longitudinal row of dorsal setae on each side of body. Antennae 3-or 4-segmented, with 4 fl eshy setae on apical segment(s). Th e known fi rst-instar nymphs of Opisthoscelis share all of these states.
Etymology. Th e genus name is a Latinised combination of the Greek tanymeaning long (from the word tanyo meaning stretch out) and skelos meaning leg, and refers to the long hind legs of the adult female. Our new name is feminine as for Opisthoscelis, the sister genus of Tanyscelis.  a hemisphere." Later Fuller (1899) published a more detailed description with an illustration of three galls of females on a leaf and a hind leg of an adult female (plate XV, fi gs 33-34). Th e galls in Fuller's drawing are very similar to syntype galls housed in the SAMA (see "Material examined" for details), but no insects have been found in the SAMA or elsewhere (e.g., not in the South African National Collection of Insects where some other Fuller material is housed). Th e type locality of O. conica was given as Midland Junction in Fuller (1897) but as Swan River in Fuller (1899), however it is unlikely that there were two localities for this species because Fuller published the second paper from South Africa based on notes that he had made during eight or nine months spent collecting in Perth prior to his 1897 paper. Midland Junction, now the suburb of Midland, is several km east of the Swan River and, in the 1890s, it was a railway station and associated village.
Dorsum. Derm membranous. Dorsal setae robust, often slightly curved, 15-138 mm long; dense over head and thorax, forming transverse band of longer setae on each abdominal segment. Macrotubular ducts 10-15 mm long, dermal orifi ce with a rim 5 mm wide; in transverse row on each of abdominal segments III-VI. Microtubular ducts absent. Quinquelocular pores 6-8 μm in diameter, scattered over dorsum.
Venter. Derm on abdomen with bands of microtrichia. Setae 15-150 mm long, in a transverse row across each abdominal segment, dense along margin and submargin of head and thorax, medial cluster on mesothorax and metathorax. Macrotubular ducts similar to those on dorsum, in sparse transverse rows across each of abdominal segments III-VI. Microtubular ducts absent. Quinquelocular pores similar to those on dorsum, similar in distribution to ventral setae. Comments. Th e adult female of T. conica would be diffi cult to confuse with the adult female of any other known species of Tanyscelis. Th e fusion of the hind tibia and tarsus into a broad, sword-like appendage is unique. Both the form of the dorsal setae (robust and slightly curved) and their dense distribution also separate adult females of T. conica from those of other species. Known populations of T. conica occur in two disjunct clusters, one centered around Perth in southwest Western Australia, and the other in forest reserves in northwestern Victoria and southeastern Australia (e.g., Danggali Conservation Park, Wyperfi eld National Park, Hattah-Kulkyne National Park). Despite the great distance separating the two, ca 2000 km, the morphology is homogeneous. Tanyscelis conica is the only species of Tanyscelis known from Western Australia. Only one species of Opisthoscelis, the type species O. subrotunda, has been collected from Western Australia, in the Kimberley region, which is a great distance (> 1,500 km) north of Perth. T. conica is also the only known species of Tanyscelis or Opisthoscelis known to feed on mallee eucalypt species in the section Dumaria, and along with O. ungulifi nis, is one of only two species in this group known to feed on eucalypts in the section Bisectae. Th e size diff erences he mentions probably relate to the age of the females. Gall (Fig. 2c,d). Female (Fig. 2c). On stem and base of petiole. Height 3.0-7.7 mm, width 4.5-8.6 mm, length of basal attachment 4.6-11.8 mm. Gall opening slitlike to oblong, 0.1-0.4 mm wide, 0.4-1.7 mm long (paralectotype galls with slit-like opening mostly 0.1 x 1.2 mm). Gall globose, surface variable, distal area frequently with 1 or 2 concentric circular scars, distal surface above scars smooth and lighter in colour than surrounding stem tissue.
Venter. Derm with microtrichia on abdominal segments. Oral lobes sclerotic, forming large circular feeding pad. Spines similar to those on dorsomedial areas of thorax and abdomen found in submarginal row extending from anterior margin of head to posterior spiracles. Ventral body surface extremely hirsute, setae similar to those on dorsum, fl agellate, 30-160 mm long. Macrotubular ducts 12-15 μm long, dermal orifi ce with rim 5 μm wide; on medial areas of anterior abdominal segments only. Quinquelocular pores large, 7-10 mm in diameter; small cluster around each spiracle. In the BMNH, there are also two collections of dry galls labelled as "Opisthoscelis convexa n. sp." and apparently from WWF (one from Dandenong, Victoria, and one from Sydney, N.S.W.). Th ese appear to have been sent to London in 1922, prior to Froggatt's description of this species. In 1985, PJG dissected dry galls from the Dandenong collection and slide-mounted two adult females and some fi rst-instar nymphs (listed below). Although WWF presumably identifi ed these as belonging to his new species, with the manuscript name O. convexa, the collection data for these specimens are not listed in Froggatt's (1929) description of the species and thus we do not consider them to be part of the type material.
Comments. Adult females of T. convexa can be recognised easily by the large dorsal spines found on the dorsomedial areas of the thorax and abdomen and in a ventral submarginal row extending from the anterior margin of the head to the posterior spiracles. T. convexa is also the only species of Tanyscelis to have no tubular ducts or multilocular pores on the dorsum. DNA sequence data support a sister relationship between T. convexa and T. maculata + T. maskelli (NBH, unpublished data). T. convexa is one of the more frequently collected, and broadly distributed, species of Tanyscelis. It is known to occur along the coast from Adelaide in South Australia to Cairns in north Queensland. Adult females from near Swan Hill but in New South Wales, and from Canberra diff er from all other collections of T. convexa in having the ventromarginal band of spines poorly developed. Samples from Benalla, Shepparton, and Nagambie (more or less between but about 100 km farther south than the Canberra and Swan Hill samples) have the ventromarginal band of spines well developed. Tanyscelis convexa has been collected exclusively from eucalypt species in the section Adnataria.

Tanyscelis grallator Hardy & Gullan, sp. n.
urn:lsid:zoobank.org:act:F524C2C7-194B-43D4-A10B-D8A716C8D083 Fig. 2e, 13 Gall (Fig. 2e). Female. On leaf. Th orn-like, height 2.2-3.0 mm, width 3.0-4.8 mm, length of basal attachment 5.0-6.0 mm. Gall opening slit-like, 0.5-0.8 mm long; on adaxial leaf surface for several galls examined. Base broad and globose with circular scar; apex tapering to blunt point. Live adult females of T. grallator have white powdery wax in a band across dorsum of each body segment, with intersegmental areas bare of wax. Th e female is small in relation to the size of the gall cavity and eggs are laid into the cavity, which can become fi lled.
Dorsum. Derm weakly sclerotised, densely beset with small papillae; a conical to papilliform protuberance medially on each thoracic segment and on abdominal segment I, and a similar protuberance on each submargin of thoracic segments I and II, smaller protuberances may be present on each submargin of anterior abdominal segments. Dorsal setae fl agellate, 25-212 mm long; arranged in a transverse row across each abdominal segment, scattered over surface of head and thorax; setae increasing in length caudad. Macrotubular ducts 6-12 mm long, dermal orifi ce with rim 4-5 mm wide; in transverse row across each abdominal segment, scattered over head and thorax. Microtubular ducts absent. Quinquelocular pores absent.
Venter Comments. Adult females of T. grallator are most similar to those of T. mollicornuta. Most distinctively, adult females of both species have fl eshy protuberances in place of eyes. Th ese protuberances are much longer on T. mollicornuta (≥ length of hind tarsus) than on T. grallator (<< length of hind tarsus), which has additional protuberances along the dorsal midline plus the submargin of the thorax. Each protuberance on the head of T. grallator has the apex much more diff erentiated than the similar protuberances of T. mollicornuta, and it seems possible that this area is light sensitive. Adult females of T. grallator can be distinguished further from those of T. mollicornuta by having much longer hind legs, each with a cylindrical coxa (more conical in T. mollicornuta), and the ratio of the hind trochanter + femur to tibia + tarsus << 1 (ca 1 in T. mollicornuta). T. mollicornuta is known from only one location, in Queensland, east of Western Creek State Forest, near Milmerran. Samples of T. grallator have been collected from fi ve locations in Queensland, including a collection from the eastern edge of Kumbarilla State Forest, < 50 km to the north of Milmerran. T. mollicornuta and T. grallator have only been found on eucalypts in the subgenus Adnataria.
Etymology. Th e species name is derived from the Latin word grallator, meaning 'stilt walker', in reference to the extremely long, slender legs of the adult female. Th e name is a noun in apposition.

General.
We have synonymised O. recurva with O. maculata. We could fi nd no morphological diff erences between the adult females of each, although type adult females of O. recurva tend to be larger than those of O. maculata. Furthermore, the 'recurved' gall morphology from which the name of the junior synonym was derived is found occasionally in galls of smaller specimens that match the types of O. maculata. Th e type locality for O. maculata is Bendigo, Victoria, with galls reported from E. leucoxylon and E. gracilis (Froggatt 1894b), whereas O. recurva was described from specimens collected on the hills at Warrah Station, a farming property near Willow Tree [not Willowtree as in the description], New South Wales, on an undetermined species of Eucalyptus (Froggatt 1929).
Dorsum. Derm membranous in young females, older specimens may have clusters of small circular sclerotic regions on margin and submargin of head and thorax, or broad areas of sclerotic cuticle on medial portions of head and thorax. Dorsal setae robust, 10-125 mm long; arranged in a transverse band across each body segment and scattered along margin, longer setae present on abdominal segments. Macrotubular ducts 12 mm long, dermal orifi ce with rim 5 mm wide; restricted to posterior abdominal segments. Microtubular ducts absent. Quinquelocular pores, 7-11 μm in diameter, scattered across posterior abdominal segments.
Venter. Setae 20-165 mm long, in a transverse row across each posterior abdominal segment plus a cluster extending from each spiracle to margin, a few mesad to each metacoxa, otherwise ventral body surface naked. Macrotubular ducts absent. Quinquelocular pores large, 7-11 mm in diameter, similar in distribution to ventral setae. Comments. Adult females of T. maculata are very similar to those of T. maskelli. Th e latter can be distinguished by the much smaller eyes, the distinct marginal fringe of setae, and the longer dorsal setae on the abdomen and along the margin. T. maculata is known from many more collections, and appears to be common in the south of Australia, whereas T. maskelli has a more northern distribution. Populations of both species have been collected in New South Wales, although never from the same site. Except for one sample of T. maculata collected by J. W. Beardsley, apparently from E. macrorrhyncha (subgenus Eucalyptus: section Aromatica), and Froggatt's (1894b) unconfi rmed record from E. gracilis (subgenus Symphyomyrtus: section Bisectae), all material of T. maskelli and T. maculata has been collected from eucalypt species in the section Adnataria (subgenus Symphyomyrtus).

Tanyscelis maskelli (Froggatt), comb. n. Figs 2h,i, 15
Opisthoscelis maskelli Froggatt, 1894b: 340-341;Froggatt, 1929: 377. General. In the original description, Froggatt (1894b) states that he collected this species at Cooma, Maitland, Newcastle, and a dozen localities within a radius of 20 miles [32 km] of Sydney, but that a constant locality was Flemington on E. siderophloia, the large-leaved box. Flemington is in western Sydney and now in the suburb of Homebush West Subsequently, Froggatt (1929) treated Flemington as the type locality, as did Miller and Gimpel (2000) in their catalogue of the Eriococcidae.
Dorsum. Derm membranous in young females, sclerotic in older specimens. Dorsal setae ranging from minute and conical to long and fl agellate, 20-175 mm long; arranged in a transverse row or band across each body segment, longer setae present on abdominal segments and forming fringe along entire body margin. Macrotubular ducts 12 mm long, dermal orifi ce with rim ca 5 mm wide; ducts restricted to posterior abdominal segments. Microtubular ducts absent. Quinquelocular pores large, 7-11 mm in diameter, present on posterior abdominal segments.
Male. On stem; similar to gall of female but without apical cone; gall opening irregularly shaped.
Dorsum. Dominated by 3 or 4 large humps. Derm lightly sclerotised, densely beset with minute papillae, these diminishing in size caudad and are replaced by microtrichia on abdomen, cuticle on humps may be more heavily sclerotised than elsewhere. Dorsal setae slender, 7-53 mm long; in transverse row across each abdominal segment, scattered on head and thorax, longest setae on posterior abdominal segments. Macrotubular ducts and microtubular ducts absent. Quinquelocular pores 7-8 μm in diameter, on posterior abdominal segments.
Venter. Oral lobes membranous. Setae as on dorsum, 12-80 mm long, in a transverse row across each abdominal segment, plus a few scattered along margin of head and thorax. Macrotubular ducts absent. Quinquelocular pores similar to those on dorsum, in a transverse row across each abdominal segment plus a few clustered around each spiracle, absent elsewhere.
Comments. Adult females of T. megagibba are most similar to those of T. villosigibba, in the sense that the dorsal surface of both species is dominated by a series of large humps. Th ese are most spectacular in mature females of T. megagibba, but lack the dense covering of setae, each born on a raised fl eshy base, that is characteristic of the adult females of T. villosigibba. Adult females of T. villosigibba are also distinctive in having highly convex eyes, with the base of each eye parallel-sided and perpendicular to the body surface, whereas in T. megagibba the base of each eye is not parallel-sided. T. megagibba is known from only two localities, ca 100 km apart, in South Australia, whereas T. villosigibba is known from four localities in Queensland.
Th e original labels on the slides from two of the three collections of T. megagibba made by HMB have the host as E. odorata. However, her subsequent notes specify that this was a misidentifi cation and the collections were made from E. microcarpa (not E. odorata).
Etymology. Th e species name is derived from the Latin word gibber, meaning hump on the back. Th is species has impressively large dorsal humps. Th e name is a noun in apposition. Gall. Female. On leaves. Gall a circular raised area, up to 5 mm across and 1-2 mm high; opening slit-like.
Dorsum. Derm weakly sclerotised, densely beset with small papillae. Dorsal setae fl agellate, 50-165 mm long; arranged in a transverse band across each abdominal segment, densely scattered over surface of head and thorax; setae increasing in length caudad. Macrotubular ducts 15 mm long, dermal orifi ce with rim ca 7 mm in diameter; in transverse row across each abdominal segment, scattered over marginal and submarginal areas of head and thorax. Microtubular ducts absent. Quinquelocular pores absent.
Venter. Oral lobes membranous to sclerotic, forming circular feeding pad around mouthparts. Setae 63-190 mm long, in a transverse band across each abdominal segment plus across meta-and mesothorax, and along margin of head and prothorax. Macrotubular ducts absent. Quinquelocular pores very large, ca 10 μm in diameter, on abdominal segments and medial area of mesothorax.
Comments. Adult females of T. mollicornuta are most similar to those of T. grallator [see comments under T. grallator].
Etymology. Th e species name is formed from the Latin word mollis, for soft, and cornuta, for horn. It refers to the pair of fl eshy projections on the anterior surface of the head. Th e name is a noun in apposition. Figs 3b,c,18,19 Opisthoscelis pisiformis Froggatt, 1894b: 343-344. General. When Froggatt (1894b) described O. pisiformis, he listed three host species of eucalypts -Eucalyptus melliodora, E. piperita and E. resinifera -from three diff erent localities in New South Wales -Bathurst, Th ornleigh and Sutherland. Even though we designate a lectotype below, the type locality of this species continues to encompass these three localities because the collection locality of the lectotype is not known. Although Miller and Gimpel (2000) listed the BMNH as holding syntypes, the data associated with the four BMNH collections of dry galls labelled as O. pisiformis either clearly show that the material was collected subsequent to description of this species or the data are inadequate to determine type status. Probably only two of the four BMNH collections belong to this species. Insects extracted from dry galls of one collection were slide-mounted by PJG in 1984 and are listed below in Material examined.
Dorsum. Derm membranous. Dorsal setae ranging from minute and conical to long and fl agellate, 8-95 mm long; arranged in a transverse row or narrow band across each body segment, longer setae present on abdominal segments and along margin. Spinose seta, 5-10 μm long, found on margin of each posterior abdominal segment, absent from margin of anterior abdominal segments, head and thorax. Macrotubular ducts 15 mm long, dermal orifi ce with a 5 mm wide rim; in transverse row across each abdominal segment, scattered over thorax, absent from head. Microtubular ducts absent. Quinquelocular pores 5-7 μm in diameter, scattered over dorsum.
Venter. Setae as on dorsum, each 12-175 mm long, in a transverse row or narrow band across each abdominal segment as well as meta-and mesothorax, along margin and submargin of head and prothorax. Macrotubular ducts similar to those on dorsum, restricted to submarginal areas of abdominal segments. Quinquelocular pores similar to those on dorsum, similar in distribution to ventral setae; pores often occurring in pairs.
Dorsum. Derm membranous. Dorsal setae fi ne and minute £1 mm long; arranged in submedial longitudinal row on each side of body, 1 seta on each side of head, prothorax, and each of abdominal segments I-VII. Microtubular ducts 4 μm long, each side of body with 1 duct on submargin of each thoracic segment plus each of abdominal segments I and V; also with 1 duct on submedial area of head, each thoracic segment and abdominal segment VIII. Marginal setae sagittate, 2-5 μm long, each side of body with ca 6 setae between midline and eyes, 4 on prothorax, 3 on mesothorax, 2 on metathorax, 1 on each of abdominal segments I-VII, and on abdominal segment VIII 1 lateral and 2 medial of apical seta, these most likely homologous to anal lobe setae.
Venter. Setae hair-like, each 1-20 mm long, each side of body with 3 setae medial of scape, 1 seta medial of each coxa, 3 longitudinal rows on abdomen, each row with 1 seta on each of abdominal segments II-VII; suranal and ventral lobe setae hair-like, each ca 15 μm long. Trilocular pores 3 μm in diameter, 1 pore near each spiracle. Comments. Adult females of T. pisiformis are very similar to those of T. verrucula. Both species have 4 anal spines and a small but distinct anal ring that often appears horseshoe-shaped and has ca 6 setae. Both also have marginal abdominal spines that are smaller than the anal spines, and they frequently have paired quinquelocular pores.
Adult females of T. pisiformis can be distinguished from those of T. verrucula by having (1) marginal spines restricted to the posterior abdominal segments (marginal spines also occurring along margin on head and thorax on T. verrucula); (2) much smaller eyes (15-25 μm wide, compared to T. verrucula with eyes 30-65 μm wide); (3) elongate fl agellate setae up to 95 μm long on dorsal surface of posterior abdominal segments (shorter than 25 μm in T. verrucula); and (4) no weakly-sclerotised pads or protuberances near the spiracles (present in T. verrucula). Both T. pisiformis and T. verrucula are known from southeastern Australia, although T. verrucula is much more widely distributed and T. pisiformis appears to have a strictly coastal distribution and has not been collected in Victoria. Th e two have never been collected from the same site, although they both occur in New South Wales. T. verrucula is known only from eucalypt species in the section Maidenaria. Most of the T. pisiformis material was collected from unidentifi ed hosts, but one sample was from E. botryoides and another from E. saligna (both in section Latoangulatae). ( siderophloia at Flemington, which is in western Sydney and is now in the suburb of Homebush West, whereas O. fi bularis was described from two localities: Bathurst, New South Wales (which is about 150 km west of Flemington) and Bendigo, Victoria (Froggatt 1894b). We have examined adult female insects ex thorn-like leaf galls from localities in New South Wales, Victoria and South Australia and all specimens are similar.

Tanyscelis spinosa
Gall (Fig. 3d,e). Female. On leaf. Th orn-like, height 3.1-7.3 mm, width 2.7-7.7 mm, length of basal attachment 3.4-8.1 mm, usually on adaxial (upper) leaf surface. Gall opening slit-like, ca 0.1 mm wide, 0.3-1.0 mm long. Base of gall globose and de-  Comments. Adult females of T. spinosa are most similar to those of T. mollicornuta and T. grallator. Adult females of T. spinosa and T. grallator also have the dorsum with weakly sclerotic cuticle and conical to papilliform evaginations. Adult females of T. spinosa can be separated easily from those of T. mollicornuta and T. grallator by having eyes, and lacking the fl eshy projections that are in the place of eyes in T. mollicornuta and T. grallator. T. spinosa has been collected exclusively from eucalypt species in the section Adnataria (also true of T. mollicornuta and T. grallator), with the exception of a collection from a sapling that had been identifi ed tentatively as E. ?goniocalyx (section Maidenaria). Gall (Fig. 3f ). Female. On leaf. Gall circular, 3-4 mm diameter, almost level with lea f surface on orifi ce side, a slightly raised bump on opposite leaf surface; opening round to elongate, surrounded by raised ring of tissue, usually opening on adaxial (upper) surface of leaf. Gall tissue green but highly glaucous due to white waxy exudation on Eucalyptus cephalocarpa; gall not glaucous and becoming brown with age on E. aromophloia.

Tanyscelis tripocula
Male. On leaf, usually opening on adaxial surface. Similar to galls of females; gall surface glaucous if on E. cephalocarpa but gall tissue red, appearing purplish due to white wax covering.
Adult female (Fig. 21) (n = 16). Body outline circular, abdomen tapered and curved dorsad in mature females, vulva and anal opening topologically dorsal,  (1) the manner in which the posterior abdominal segments are directed dorsally in mature females (not in young adult females); (2) the shape of the anal opening, which is an irregular slit in a rugose sclerotic plate; and (3) the complete absence of tubular ducts. Th e phylogenetic position of T. tripocula, as sister to all other species of Tanyscelis based on DNA data (NBH, unpublished data), is not well supported.
Galls of females on E. cephalocarpa are smaller, fl atter and glaucous compared with those on E. aromaphloia. Th e adult females from galls on E. cephalocarpa are smaller and less sclerotised than the females from E. aromaphloia, but this may be an age eff ect. Th e type series has been restricted to specimens from E. cephalocarpa because of the above slight diff erences.
Adult males were observed by PJG emerging from their galls on E. cephalocarpa in early February 1979. Th ey emerged head fi rst from the gall opening, which was plugged by male's fi rst-instar exuviae until gall maturity. Th e males were capable of short jumps of 1-3 cm either forwards or backwards, as well as weak fl ight. Adult females also may bear their fi rst-instar exuviae on their dorsum.
Etymology. Th e species name is derived from the Latin word poculum, meaning cup and tres or trimean three. It refers to the three deep invaginations along the dorsal midline. It is a noun in apposition.

General.
Here we synonymise O. mammularis with O. verrucula. Froggatt (1894b) described them as distinct species on diff erent pages of the same article, but the two are identical based on the morphology of the adult females and their galls. Froggatt claimed that the adult female of O. verrucula had an "anal segment carrying an anal ring with four curved spines forming the tail." (Froggatt 1894b: 338), whereas that of O. mammularis had the "tip of anal segment bearing two small reddish-brown spines curved backwards, forming a tail" (Froggatt 1894b: 344). However all specimens that we examined had one pair of large spines, each with a small spine near its base, on abdominal segment VIII, and one pair of medium-sized spines on abdominal segment VII, with the latter sometimes appearing very close to the spines on segment VIII. Th e type locality of O. verrucula is Napolean Reef, Bathurst, New South Wales, whereas that of O. mammularis is Bendigo, Victoria, and the types of both were collected on unidentifi ed Eucalyptus species.
Dorsum. Derm membranous. Dorsal setae minute, 5-8 mm long; arranged in an irregular transverse row across each abdominal segment, scattered over surface of head and thorax. Stout spinose setae, 5-10 μm long, found on margin of each posterior abdominal segment plus along margin of head and prothorax. Macrotubular ducts 13-15 mm long, dermal orifi ce with rim 5 mm wide; in transverse row across each abdominal segment, scattered over head and thorax. Microtubular ducts absent. Quinquelocular pores 5-7 μm in diameter, scattered over dorsum.
Venter. Membranous lobe-like protuberance immediately posterior of each spiracle, each protuberance having a few slender setae. Setae fl agellate, 30-118 mm long, in a transverse row or narrow band across each abdominal segment, and scattered along margin of head and thorax and mesad to each metacoxa. Macrotubular ducts absent. Quinquelocular pores similar to those on dorsum, in transverse band across each of abdominal segments III-VI, a small cluster near each spiracle, a few scattered among setae along margin of head and thorax, pores often occurring in pairs.
Material examined. Lectotype of Opisthoscelis verrucula (here designated): AUSTRALIA: New South Wales: 1 adult female (3.27 mm long, 2.14 mm wide): ex dry gall on leaf, with printed label: "No. 1788 E / GALL MAKING COCCIDS. / oviposition, the body of the females shrunk to about one-fi fth of their size prior to oviposition. It was estimated that a single female could lay more than 1000 eggs. Gall (Fig. 3i). Female. On stem. Globose, with truncate apex and rugose surface. Mature female almost completely fi lls gall cavity, positioned with three sclerotic dorsal humps positioned just below gall orifi ce.
Dorsum. Dominated by 3 large, sclerotic and setose humps, 1 on each thoracic segment, each hump 450-910 μm long, 520-790 μm wide. Derm elsewhere membranous or weakly sclerotised. Dorsal setae robust and conical to fl agellate, 10-140 mm long; fl agellate setae in a transverse row across each posterior abdominal segment, robust setae mounted on small raised base, scattered over surface of head and thorax, dense on thoracic humps. Macrotubular ducts present or absent, duct 12 mm long, dermal orifi ce with rim 5 mm wide; if present, only 1 to a few present somewhere on abdominal segments III-VI. Microtubular ducts absent. Quinquelocular pores 7-10 μm in diameter, in a transverse row or narrow band across each posterior abdominal segment.
Venter. Oral lobes membranous. Setae hair-like, each 10-130 mm long, in a transverse row across each abdominal segment; a few setae, each on a raised base, scattered along margin of head and thorax. Macrotubular ducts absent. Quinquelocular pores similar to those on dorsum, in a transverse row to narrow band across each posterior abdominal segment plus a few clustered around each posterior spiracle.  of Boomerang Road and Beenleigh-Beaudesert Road, -27. 78°, 153.19°, 2 May, 1993, PJG (ANIC).
Comments. Adult females of T. villosigibba are most similar to those of T. megagibba [see comments under T. megagibba]. Specimens of T. villosigibba from the type locality, just south of Brisbane, lack macrotubular ducts entirely, whereas specimens from the other two localities, the closest of which is about 200 km to the west in Dunmore State Forest, have 1 or a few macrotubular ducts restricted to the dorsal surface of the abdominal segments III-VI.
Etymology. Th e species name is a combination of the Latin words villosus, meaning shaggy, and gibber, for hump. It refers to the setose dorsal protuberances that characterise this species. Th e name is a noun in apposition.
permits to PJG and LGC, and permits were provided via the Entomological Society of Queensland for reserves in southeast Queensland. Th is work was supported in part by the National Science Foundation under PEET Program grant DEB-0118718 awarded to PJG, and an Australian Biological Resources Study Participatory Program research grant  to Lyn Cook and PJG.