Larval descriptions of five Oriental bamboo-inhabiting Acroceratitis species (Diptera: Tephritidae: Dacinae) with notes on their biology

Third instar larvae of the genus Acroceratitis Hendel from North Thailand are described for the fi rst time. They belong to A. ceratitina (Bezzi), A. distincta (Zia), A. histrionica (de Meijere), A. incompleta Hardy, and A. septemmaculata Hardy. Short descriptions of eggs, empty egg shells, and puparia are also presented. Acroceratitis larvae infest shoots of bamboo (Poaceae). Larval host plants of the studied species are Bambusa polymorpha Munro, Cephalostachyum pergracile Munro, Dendrocalamus hamiltoni Nees and Arnott ex Munro, D. strictus (Roxbourgh), Dendrocalamus sp. (unidentifi ed) and Pseudoxytenanthera albociliata (Munro). The morphological characters of Acroceratitis larvae are compared with those of other Gastrozonini described so far. A key to Acroceratitis larvae is provided. Acroceratitis ceratitina, A. incompleta and A. septemmaculata are morphologically similar and clearly differentiated from A. distincta and A. histrionica by the lack of additional papillar sensilla on the labial lobe, the arrangement of the spinules on the creeping welts and other characters. The morphological differences between the two groups coincide with the type of substrate utilized by their larvae: A. ceratitina, A. incompleta and A. septemmaculata larvae feed in young and soft internode walls, while A. distincta and A. histrionica utilize harder bamboo tissue of already elongated bamboo shoot internodes. Acroceratitis histrionica larvae are special within the Gastrozonini, because they develop exclusively in cavities formed by the internode surface and the protecting culm sheath. Factors infl uencing spatial utilization of larval resources, preference for upright shoots as breeding substrate, larval behavior, types of bamboo damage caused by different species and attraction to sweat and urine in the adults are discussed.

In the present study we describe the third instar larvae of Acroceratitis ceratitina (Bezzi), A. distincta (Zia), A. histrionica (Meijere), A. incompleta Hardy, and A. septemmaculata Hardy, all collected by D. Kovac in North Thailand. Brief descriptions of the puparia of the fi ve species as well as the eggs of A. histrionica are presented. The larval characters of Acroceratitis species are compared with previously described Gastrozonini larvae. An overview of the hitherto largely unknown habitat and biology of the examined Acroceratitis species is given.
So far, larvae of only six Gastrozonini species have been known: Elson-Harris (1992)

Scanning electron microscopy (SEM)
Larvae were cut in three parts as described above but the fi rst cut was set between the fi rst and second and the second cut was set between the fi fth and sixth abdominal segment. The larval body parts were washed in 80% ethanol three times in order to remove body substances leaking out of the cuts, dehydrated in 96% ethanol for one hour, transferred to hexamethyldisilazane (HMDS; Nation, 1983) for one hour, and subsequently transferred to fresh HMDS for one additional hour. Subsequently, HMDS containing the specimens was allowed to evaporate under a fume hood. Dried specimens were mounted on stubs, coated with gold-palladium and observed under a Hitachi scanning electron microscope (CamScan CS24).
The general description provided below applies to all species. Thereafter A. ceratitina is described in detail, while for A. incompleta, A. septemmaculata and A. distincta only their differences from A. ceratitina are noted; similarly for A. histrionica only its further differences from A. distincta are noted.
The examined material is stored in the collection of the Sencken berg Forschungsinstitut, Frankfurt am Main.
Since none of the morphological characters characterizing Gastrozonini species as Dacinae are present in all Gastrozonini species, the easiest way to discriminate Gastrozonini larvae from other Tephritidae is to include their association with Poaceae as a diagnostic character. This also applies to adult fl ies (Kovac et al., 2006). The only other tephritids breeding in grasses belong to the tribe Acanthonevrini (Phytalmiinae) (Allwood et al., 1999;Dohm et al., 2014). Acanthonevrini larvae can easily be distinguished from Gastrozonini, for example, by their large anal lobes and especially by the typical shape of the facial mask (great number of very long oral ridges; patches of ridges similar to oral ridges anterior to mouth opening) (Elson-Harris, 1992).

Genus Acroceratitis Hendel, 1913 Larval description of third instar larvae
Habitus. Anterior end of the body conical, posterior end truncate.

Study site
Field work was conducted in North Thailand (Province Mae Hong Son) in the mountainous Pangmapha district. Pangmapha has three distinct seasons, the rainy season (mid May to mid October), the cool season (mid October to mid February) and the hot season (mid February to mid May). The daytime temperatures rise up to 35-37°C in March /April with peaks above 40°C. In December / January it is cool at night, when the temperatures drop to around 13°C or lower depending on the elevation. Annual rainfall varies from 1100 to 1500 mm with the peak at the end of the monsoon period in August / September (Müller, 1996;Gardner et al., 2000). Bamboo host plants of Acroceratitis species were common in farming areas at elevations between 600 and 1600 m. Collecting took place between 2010 and 2017. Each year fi eld work was conducted during two periods leaving only the months January and February unsampled.

Collecting, rearing and preservation
Flies were collected by sweeping over bamboo (shoot surface, leaves) or the surrounding vegetation. Larvae were obtained by cutting off pieces of infested bamboo shoots. Infested shoots were recognized by the withered sheaths and blades occurring at the apex of shoots. Shoots were also cut randomly in order to fi nd larvae in seemingly undamaged shoots. Another method for obtaining larvae was to cut bamboo shoots, place them on the ground or in the upright position and check about 2 weeks later if they were infested by tephritid larvae.
Bamboo shoot sections containing tephritid larvae were transported to the lab in plastic bags. In the lab larvae were grouped according to their size and colour and kept in separate plastic rearing containers at room temperature. Rearing containers were provided with original shoot substrate and were covered with fi ne-mesh gauze. They were checked daily and substrate moistened if necessary. If larvae from the same container developed different characters (color, size), indicating a mixture of two or more species, they were separated from each other. Mature larvae left their substrate for pupariation and were then transferred to smaller plastic containers and allowed to pupariate between moistened paper tissues.
For preservation, some Acroceratitis larvae were immersed in boiling water for 2-3 min in order to kill them and keep their cuticle pale and expanded. The larvae were eventually transferred to 70% ethanol for permanent storage. Three days after eclosion the adults were killed with ethyl acetate and pinned or transferred to 70% ethanol. Some adults and unboiled larvae were preserved in absolute ethanol.

Slide mounting
For the study of the cephalopharyngeal skeleton and the posterior spiracles, the preserved larvae were cut transversely in three parts. The fi rst cut was set between the thorax and the abdomen and the second between the 7 th and 8 th abdominal segment. The anterior body part (containing the cephalopharyngeal skeleton) and the posterior body part (containing the posterior spiracles) were macerated overnight in 10% potassium-hydroxide (KOH) solution at room temperature, following the method used by Steck & Wharton (1986) and Steck et al. (1990). Subsequently the cephalopharyngeal skeleton was further cleared at 60°C in 10% KOH for about 1.5 h. The macerated larval parts were washed in distilled water for 1 h in order to remove the KOH. Then, they were dehydrated in 80% and 96% ethanol and fi nally transferred to terpineol for 15 min each. Finally, the cephalopharyngeal skeleton and the posterior spiracles were slide-mounted in Canada balsam, still enclosed by the transparent larval cuticle.

Pseudocephalon.
Facial mask approximately triangular. 7-22 rows of serrated oral ridges of medium length. 8-17 short accessory plates lateral to oral ridges covering a much smaller area than oral ridges. Accessory plates most often very similarly serrated as oral ridges but in some cases teeth of accessory plates shorter, wider, and more irregular than those of oral ridges. Antenna and maxillary sense organs located on moderately to well-developed cephalic lobes. Antenna two-segmented. Maxillary sense organ composed of maxillary palpus and a dorsolateral group of sensilla, both lying close to each other. Maxillary palpus surrounded by a complete or incomplete cuticular fold bearing three papillar sensilla and two knob sensilla. Dorsolateral group of sensilla of maxillary sense organ containing two papillar sensilla. Primary preoral lobe rounded to oval, surrounded by 7-13 non-serrated preoral lobes. Preoral organ located on primary preoral lobe, bearing 2-3 peg sensilla. Labial lobe elongate, widened at apex, either entirely smooth or bearing tubercles at base. Labial organ with two openings, with or without two papillar sensilla near these openings. Five pairs of pseudocephalic pit sensilla present: one pair between the antennae, one pair between the maxillary sense organs, one posteroventral to the antenna, and two laterally to the facial mask.
Cephalopharyngeal skeleton. Total length 0.8-1.1 mm. Indentation between apex of apical tooth and ventral apodeme 0.13-0.23× as deep as length of mouthhook (measurements as shown by Frías et al., 2006;Fig. 2). Mouthhooks each bearing one small preapical tooth. Hypopharyngeal sclerite 2.6-5.1× as long as high. Labial sclerite light brown. Pharyngeal sclerite similar to those of other Gastrozonini: brown, darker in area of tentorial phragma and dorsal and ventral bridges; dorsal and ventral bridges protruding. Length of dorsal cornu 0.8-1.1× that of ventral cornu; dorsal cornu cleft; ventral cornu with a bulge on dorsal margin, and with a lighter brown area below this bulge. Parastomal bars light brown, 0.8-1.1× as long as hypopharyngeal sclerite. Anterior sclerite present in all species (but not in all specimens of A. distincta).
Thoracic segments. Anterior spiracles with 13-21 approximately cylindrical tubules arranged in a single row. Slit-like opening visible at apex of each tubule. Anterior margin of fi rst thoracic segment covered with numerous rows of dentate plates. Anterior margins of second and third thoracic segments lined with spinules. Third thoracic segment with rudimentary spiracle near its anterior margin. As shown in Fig. 17, isolated sensilla form a similar pattern in all examined Acroceratitis species. Each thoracic segment bearing ventrally a pair of Keilin's organs.
Abdominal segments I-VII. As shown in Fig. 17, isolated sensilla similarly distributed in all examined Acrocera titis species. Each segment with rudimentary spiracles near anterior margin and with ventral creeping welts which bear rows of spinules in variable arrangements.
Caudal segment. Posterior spiracles somewhat above midline of caudal area, shortest distance between spiracles 0.3-0.9× length of longest opening. Peritreme hyaline; spiracular chamber light brown, spiracular openings surrounded by broad, dark areas, ecdysial scar visible. In caudal view central opening located more lateral than dorsal or ventral opening; angle between dorsal and central opening up to 25°, shortest distance between dorsal and central opening 0.2-0.7× length of longest opening; angle between central and ventral opening up to 33°, shortest distance between central and ventral opening 0.1-0.5× length of the longest opening. Rima visible; spiracular openings 4.0-9.4× as long as wide. Spiracular hairs emerge from small protuberances and are arranged in four groups; hairs branched up to 7×, up to 0.6× as long as longest spiracular opening; dorsal bundle containing 2-20 hairs, ventral bundle containing 5-20 hairs, lateral bundles containing 2-8 hairs.
Caudal ridges and dark transverse line on caudal area present in all species. Area covered by oval anal lobes approximately as long as wide, surrounded by rows of spinules. Sensilla on caudal segment all present and similarly distributed as in other Dacinae (White et al., 1999). A row of three pit sensilla present on each side between anterior margin of anal lobes and lateral sensillum. Creeping welt of caudal segment similar to creeping welts of abdominal segments II-VII.

Diagnosis of third instar larvae
Facial mask triangular, oral ridges regularly to irregularly serrated, accessory plates slightly to distinctly serrated; preoral organ most often bearing three peg-sensilla; preoral lobes non-serrated. Mouthhooks with one preapical tooth. Tubules of anterior spiracle arranged in single row. Creeping welts bearing broad band of very small spinules between two areas covered with posteriorly directed spinules (in A. histrionica this band is replaced by 3 to 4 rows of anteriorly directed spinules); anterior margins of abdominal segments at least partly lined with bands or patches of very small spinules. Area covered by both anal lobes approximately as long as wide; anal lobes surrounded by spinules. Spiracular slits surrounded by broad, dark areas.

Biology
In north Thailand Acroceratitis adults were found in the fi eld between March and December. During the hot season (March to about mid-May) fl ies stayed in shady places along streams. At the beginning of the rainy season they dispersed to surrounding bamboo areas. Flies fed on juices oozing from cut bamboo shoots or injured mature culms, dabbed on bamboo extrafl oral nectaries, bird droppings, leaves, human urine, human sweat (on backpack, sweeping net, etc.) or imbibed water from vegetation. These food sources provided essential nutrients such as carbohydrates (bamboo nectaries), amino acids, organic acids, sugars (plant exudates) or protein (bird feces) and were also recorded from other tephritids (overview in Drew & Yuval, 2000).
Acroceratitis larvae developed in bamboo shoots of various host species belonging to Bambusa, Cephalostachyum, Dendrocalamus, Gigantochloa, Pseudoxytenanthera and Thyrsostachys (mostly own records, see also Allwood et al., 1999;Dohm et al., 2014). They were oligophagous, but occupied specifi c niches regardless of bamboo species. Thus, A. ceratitina macerated the apex of thicker shoots just protruding from the soil (Fig. 3A, B), A. incompleta and A. septemmaculata mined in walls of taller thick shoots ca. 50 cm below the apex (Figs 12A, B, C, 15A, B), A. histrionica scraped substrate from the surface of partly hardened internodes located in the lower part of thick and tall shoot stems (Fig. 9A, C, D) and A. distincta was confi ned to thin bamboo shoots or thin side branches of larger shoots (Fig. 6A, C).
Eggs were deposited below the edges of the tough internode sheaths (Figs 6G, 12F, 15E). Freshly hatched larvae squeezed between the sheaths and internode surface towards their feeding sites. Young larvae were milky-colored and pale (Fig. 9C, D), but during growth their colors changed. A. distincta and A. histrionica became pale yellow and fi nally bright yellow when mature ( Fig. 6D-F). A. ceratitina, A. incompleta and A. septemmaculata became yellow, then yellow-orange and fi nally spotted orangegreen to greenish (Figs 3C and D,12D,12E,15B and C). Mature larvae readily skipped and abandoned their internodes for pupariation in the soil. Larval development took 4-6 weeks.
Cephalopharyngeal skeleton (Fig. 16). Total length 1.0-1.1 mm (n = 2). Mouthhooks dark brown, indentation between apex of apical tooth and ventral apodeme 0.18-0.20× as deep as mouthhook length; apical tooth dark brown, curved; neck short; preapical tooth about 0.01 mm long; dorsal apodeme posterodorsally directed; ventral apodeme broad, ventrally directed. Hypopharyngeal sclerite about 3.6× as long as high, anterior part black, posterior part brown. Hypopharyngeal bridge 0.02-0.04 mm long, about 0.03 mm wide, located in the middle of hypopharyngeal sclerite. Dorsal and ventral cornua light brown; ventral cornu with hyaline area at ventral margin and large hyaline area at posterior tip; brown part of dorsal cornu about 0.77× as long as brown part of ventral cornu. Parastomal bars straight, slightly shorter than hypopharyngeal sclerite. Anterior sclerite brown.
Thorax. Anterior spiracles with 18-21 tubules (Fig. 1F). Anterior margin of second and third thoracic segments with 5-7 rows of stout spinules; rows on third thoracic segment with a small gap dorsally.
Abdominal segments I-VII. First abdominal creeping welt bearing a narrow band of very small spinules anteriorly, and about 10 rows of large, stout, posteriorly directed spinules posteriorly; creeping welts of abdominal segments II-VII bearing rows of spinules arranged as follows (from anterior to posterior): A band of very small spinules (sometimes reduced to net-like rows), 7-10 rows of large, stout, posteriorly directed spinules, a band of very small spinules, 4-7 rows of larger, more slender, posteriorly directed spinules (Fig. 2D). Laterally and dorsally, anterior margin of fi rst abdominal segment showing a slender band of very small spinules anteriorly and about 5 rows of large, stout, posteriorly directed spinules posteriorly (rows mediodorsally interrupted); anterior margins of second and third abdominal segment covered laterally with very small spinules anteriorly and about 4 rows of large, stout, posteriorly directed spinules posteriorly. Rows of spinules reaching far dorsally in second abdominal segment, extending somewhat less far dorsally in third abdominal segment. Anterior margins of abdominal segments IV-VII laterally covered by very small spinules, extending less far dorsally than on the respectively more anterior segment.
Caudal segment . Shortest distance between spiracles about half as long as longest spiracular opening. Angle between dorsal and central opening 1-9°, shortest distance between dorsal and central opening about one third as long as longest opening; angle between central and ventral opening 16-24°, shortest distance between central and ventral opening about 0.15× as long as long-est opening. Spiracular openings 6.7-7.8× as long as wide. Spiracular hairs up to 0.4× the length of the longest opening, spiracular hairs branched up to 6×; dorsal bundle containing 10-14 hairs, ventral bundle containing 8-11 hairs, dorsolateral bundle containing 3-5 hairs, ventrolateral bundle containing 4-5 hairs (Figs 16 and 2B).
2C). Creeping welt as in abdominal segments II-VII but central band of very small spinules nearly absent.
On two occasions larvae were also reared from freshly cut bamboo shoots of Dendrocalamus sp. bought at a Lahu sales booth in the Pangmapha district, two on 25 June and two on 30 June 2015. They were 30-40 cm long and enwrapped by overlapping internode sheaths. The shoots were transferred to a natural bamboo stand of Dendrocalamus sp. on the day of purchase; one shoot was laid on the ground and the other was placed in the upright position by digging its base into the ground (Fig. 3B).
On the same day females of A. ceratitina were seen to lay eggs under the sheaths, but only in upright shoots. Two weeks later only upright shoots contained A. ceratitina larvae, while shoots lying on the ground contained many larvae of the Gastrozonini Gastrozona spp. and Chaetellipsis spp.
Larvae of A. ceratitina fed inside the walls of the short, white and soft internodes. Some larvae also mined within the thickened base of the internode sheaths. There were more than 50 larvae per shoot (n = 5; two purchased shoots and three shoots found in the fi eld). Younger larvae were yellow (Fig. 3C), older ones sometimes spotted yelloworange and fi nally pale greenish (Fig. 3D). Mature larvae leaving the bamboo shoot readily skipped. First adults emerged 22 and 26 days after the purchased shoots were transferred to a bamboo stand.
Flies were recorded in the fi eld between May and October. Females (Fig. 3E) were seen to oviposit below the protective culm sheaths of upright shoots of two undetermined Dendrocalamus species.
Cephalopharyngeal skeleton (Fig. 16). Total length 0.8-0.9 mm (median: 0.82 mm; n = 6). Indentation between apex of apical tooth and ventral apodeme 0.15-0.18× as deep as mouthhook length; apical tooth brown, neck well developed; preapical tooth about 0.005 mm long. Ventral apodeme triangular shaped. Hypopharyngeal sclerite 2.6-3.4× as long as high. Hypopharyngeal bridge, brown, 0.01-0.02 mm long, about 0.02 mm wide, located between the anterior four sevenths and the posterior three sevenths of the length of the hypopharyngeal sclerite. Dorsal and ventral cornua brown, hyaline areas at posterior tip and ventral margin of ventral cornu; brown area of dorsal cornu slightly shorter than brown area of ventral cornu. Parastomal bars of approximately the same length as hypopharyngeal sclerite, in two specimens hooked at apex (in one of these two specimens only one parastomal bar hooked at apex). Anterior sclerite lacking in four of six examined specimens.
Abdominal segments I-VII. First abdominal creeping welt bearing a band of very small spinules anteriorly and about 7 rows of posteriorly directed spinules posteriorly; creeping welts of abdominal segments II-VII bearing rows of spinules arranged as follows (from anterior to posterior): 5-6 rows of anteriorly directed spinules, a band of very small spinules, being very narrow (about one row) in the middle, 4-7 rows of stout posteriorly directed spinules, a band of very small spinules, 4-5 rows of large posteriorly directed spinules (Fig. 5D). Bands of very small spinules of the creeping welts also covering anterior margins of abdominal segments ventrolaterally.
Caudal segment (Figs 5A-C and 16). Shortest distance between spiracles 0.5-0.8× as long as longest spiracular opening. Angle between dorsal and central opening up to 25°, shortest distance between dorsal and central opening somewhat more than half as long as longest opening; angle between central and ventral opening 1-29°, shortest distance between central and ventral opening one third to half as long as longest opening. Spiracular openings 4.0-6.0× as long as wide. Spiracular hairs up to half as long as longest spiracular opening, hairs branched up to 5×; dorsal bundle containing 5-12 hairs, ventral bundle containing 5-10 hairs, dorsolateral bundle containing 2-4 hairs, ventrolateral bundle containing 4-7 hairs (Figs 16 and 5B).
Anal lobes surrounded by 3-5 rows of large, pointed spinules and anterolaterally by an additional patch of very small spinules (Fig. 5C). Creeping welt bearing rows of spinules arranged as follows (from anterior to posterior): A band of very small spinules, about 5 rows of stout posteri- orly directed spinules, a band of very small spinules, about 4 rows of stout posteriorly directed spinules.
In the smaller bamboo species P. albociliata larvae infested 1-3 internodes located next to each other (Fig. 6A), usually the 3rd-5th internode below shoot apex. The base of the lowest infested internode was located 16-45 cm below shoot apex (n = 4). Infested internodes and apex of the shoot died off and could be easily recognized by the withered sheath blades (Fig. 6A). After a while apical internodes would fall to the ground, whereas the lowest infested internode remained attached to the shoot for a longer time (Fig. 6B). Buds located at the base of the lowest infested internode remained intact.
In larger bamboo species, i.e., C. pergracile, D. strictus and D. sp., larvae of A. distincta only infested thin shoots, which mostly grew in young bamboo stands. Mid-sized to large shoots (diameter about 4 cm or larger) were not colonized, but larvae were found in side branches of large D. strictus shoots (Fig. 6C).
Larvae of A. distincta fed in the basal 1/3-2/3 of the infested internodes. In advanced stage of larval development the feeding area consisted of three clearly distinguishable zones. The basal part of the internode wall was entirely eaten up over a length of up to 6 cm. The adjacent area consisted of small bamboo particles and was of pulpy consistency or turned crumbly after desiccation (Fig. 6E). The upper feeding area was characterized by fi brous vascular bundles, since the larvae apparently fed on the parenchyma cells in which the vascular bundles were embedded (Fig.  6F). Usually, larvae did not enter the internode cavity. Although the basal part of the infested internode wall was Fig. 6. Host plant, larvae and female of A. distincta. A -shoot of Pseudoxytenanthera albociliata infested by A. distincta larvae. The apical part of the shoot has died off due to the larval feeding activity and the leaf-like sheath blades have withered. The arrows point to the two internodes infested by A. distincta larvae. The lower infested internode was 12.2 mm in diameter. B -Pseudoxytenanthera albociliata internode infested by A. distincta. The internode has turned brown, whereas the sheath is still green. The apical part of the shoot has fallen to the ground. C -location of infested internodes in two side branches of A. strictus (arrows). D -larva of A. distincta boring in a thin stem of a P. albociliata shoot, which was cut off and placed on the ground. E -larva of A. distincta feeding in the pulpy basal area of an internode. F -larvae of A. distincta feeding in the fi brous apical area of an internode. G -female ovipositing below the sheath edge of a cut shoot of P. albociliata lying on the ground. totally devoured, the upper part remained attached to the shoot, because it was fi rmly held in place by the enveloping sheath.
Infested internodes belonging to different bamboo species were 5.5-32 cm long (median = 19.4 cm, n = 14) and 4-28 mm wide (median = 12 mm, n = 15). They usually contained 10-20 larvae per internode, depending on internode size. For example, a small internode measuring 6 cm × 4 mm contained 9 larvae and a larger internode measuring 25 cm × 12 mm contained 22 larvae. The number of larvae found in some internodes was low despite the large proportion of eaten bamboo tissue, suggesting that some larvae had already left the internode. Mature larvae (Fig. 6D-F) were dark yellow and readily skipped when disturbed. In four larvae from different shoots the duration between oviposition and emergence of fi rst adults amounted to 25, 27, 30 and 35 days. In the fi eld larvae left their internode for pupariation in the soil. In the rearing boxes most larvae pupariated between moistened tissue, however, some puparia were also found between fi bers inside the internodes.
Larvae developing in side branches of D. strictus shoots were found near the branch apex (Fig. 6C). The infested internodes died off, changed color from green to brown and became soft. After the larvae had left, the apical part of the side branch including the infested internode fell to the ground, whereas the remaining part of the side branch remained intact.
Flies were common in the study area and were collected during all months between March and December. In the lab, some fl ies emerged in January and one fl y on 2 nd February. Females laid eggs by inserting their aculeus below the edge of the internode sheath, usually near the base of the internode (Fig. 6G). In one case a female was observed inserting her ovipositor below a sheath of a larger shoot of Dendrocalamus sp. beside an ovipositing A. incompleta. Flies dabbed on items containing sweat (cloth of sweeping net, surface of backpack, knife handle, walking stick) and on leaves covered with urine. Altogether, 80 males and 5 females were collected from the sweeping net or backpack lying on the ground and 23 males and 1 female from leaves covered with urine on various dates between 2007 and 2017. On one occasion, 27 September 2017, A. distincta was collected from fresh sweat and urine for 1.5 h. The collecting yielded 14 males and 4 females on net/ backpack and 19 males and 1 female on urine.
Cephalopharyngeal skeleton (Fig. 16). Total length 0.9-1.1 mm (median: 1.0 mm; n = 5). Indentation between apex of apical tooth of mouthhooks and ventral apodeme 0.13-0.21× as deep as mouthhook length; apical tooth dark brown, strongly curved; preapical tooth 0.011-0.016 mm long. Hypopharyngeal sclerite 3.6-5.1× as long as high. Hypopharyngeal bridge about 0.02 mm long and wide, located somewhat posterior to middle of hypopharyngeal sclerite. Dorsal and ventral cornua light brown; hyaline areas at posterior tips of both cornua and ventral margin of ventral cornu; light brown areas of both cornua of virtually the same length. Parastomal bars about 0.9× as long as hypopharyngeal sclerite, in two observed specimens with hooked apex. Anterior sclerite light brown, present only in two of fi ve examined specimens.
Abdominal segments I-VII. First abdominal creeping welt bearing about 3 rows of anteriorly directed spinules anteriorly and about 8 rows of posteriorly directed spinules posteriorly; laterally slender, wedge-shaped areas covered with very small spinules inserted in creeping welt; creeping welts of abdominal segments II-VII bearing rows of spinules arranged as follows (from anterior to posterior): 5-7 rows of anteriorly directed spinules, 7-10 rows of posteriorly directed spinules, 3-4 rows of curved anteriorly directed spinules, 4-6 rows of long posteriorly directed spinules; laterally wedge-shaped areas covered with very small spinules inserted in creeping welts (Fig. 8D-F). Anterior margin of fi rst abdominal segment lined with about 5 rows of posteriorly directed spinules, laterally a band of very small spinules located anteriorly to these spinules. Anterior margins of second to fi fth abdominal segments lined with a band of very small spinules anteriorly and few irregular rows of posteriorly directed spinules posteriorly. Anterior margin of sixth abdominal segment bearing very small spinules laterally.
Caudal segment (Figs 16 and 8A-C). Shortest distance between spiracles 0.6-0.9× length of longest spiracular opening. Angle between dorsal and central opening up to 12°, shortest distance between dorsal and central opening somewhat more than half as long as longest opening; angle between central and ventral opening 24-33°, shortest distance between central and ventral opening one fourth to one third as long as the longest opening. Spiracular openings 5.4-7.3× as long as wide. Spiracular hairs up to 0.6× as long as longest spiracular opening; dorsal bundle containing 2-12 hairs, ventral bundle containing 5-9 hairs, dorsolateral bundle with 3-5 hairs, ventrolateral bundle with 2-7 hairs (Figs 16 and 8B).
Dark transverse line on caudal area usually present. Anal lobes surrounded by 3-4 rows of pointed spinules. A group of small spinules located lateral to anal lobes (Fig. 8C). Creeping welt as in abdominal segments II-VII.
Larvae were found in the fi eld between June and beginning of December.
Based on observations of D. Kovac (in Dohm et al., 2014 and unpubl. observ.), A. histrionica larvae developed on the surface of mid-sized to large bamboo shoot stems below the protective culm sheaths (internode diameters 1.2 to 11 cm; Fig. 9A). The larval feeding site was always located in the area above the branch bud, which was situated at the base of the internode. The alternately arranged  In the upper zone (Z1) the shoot is covered by several layers of culm sheaths; in the central zone (Z2) the culm sheaths are shorter than the internodes and protect the softer, lower parts of the internodes; and in the basal zone (Z3) sheaths get loose and then fall to the ground. B -detail of bamboo shoot internodes of D. strictus showing the location of the branch buds and the egg-laying site. C and D -A. histrionica larvae feeding in a triangular area ("feeding triangle") above the branch bud (sheaths removed). E -larval tracks showing the path from the egg-laying to the feeding site. The larvae have already left their feeding site and the sheath has fallen down. F -internode of a mature bamboo culm showing old larval tracks and feeding marks of A. histrionica. G -male of A. histrionica. Abbreviations: bb -branch bud, cs -culm sheath, es -egglaying site, ft -feeding triangle, Int -internode, lt -larval tracks, sb -sheath blade. branch buds were located on the opposite side of the area where the internode sheath was overlapping (Fig. 9B). In D. strictus branch buds occurred on all culm internodes, and A. histrionica larvae were found all along the shoot stem (base of lowermost colonized internode located 30 cm above ground; n = 8), while in D. hamiltoni and D. sp. branch buds formed from mid-culm upwards (base of lowermost colonized internode 3.22 m above ground; n = 3).
Eggs hatched below the upper edge of the sheath, and larvae squeezed their way down between the bamboo surface and the sheath towards their feeding site at the base of the internode. Sometimes, freshly hatched larvae left a permanent track on the bamboo surface enabling the reconstruction of the larval path towards their feeding site (Fig.  9C, E, F). Judging from the tracks the larvae squeezed diagonally downwards, probably along the concealed sheath edge covered by the opposite end of the sheath. The beginning portion of the path was narrow, indicating that initially the larvae followed each other. Further down at the level of the prospective "feeding triangle" (see below) the path became wider and then split into several roughly horizontal, parallel branches, indicating that the larvae independently changed their directions towards their feeding site. The larvae reached their feeding site somewhere between the base and the tip of the prospective "feeding triangle". In several cases indentations leading to the feeding triangle were visible, but feeding marks were lacking.
Young A. histrionica larvae scraped bamboo tissue from the bamboo shoot, thus leaving small, pale feeding marks on the green internode surface (Fig. 9C). At a later stage of larval development the feeding area became larger and the vascular fi bers became visible. The fi nal shape of the brownish feeding area resembled an upward pointing triangle with the base being located at the level of the branch bud and its tip sometimes reaching up to the center of the internode (Fig. 9D). The larvae exploited only the softer substrate of the bamboo surface or substrate between the vascular fi bers and did not penetrate into the already hardened internode wall. They also did not damage the branch buds, which were protected by their own small sheaths (prophylls). The feeding area above the branch bud was moist. Sometimes, water fi lled the space between the bamboo surface and the sheath and larvae appeared to be dead, but they often started to move after a few minutes. There were 3-13 larvae per internode (median = 6, n = 31). Young larvae were pale, while mature larvae turned yellow at the very end of their development. Mature larvae readily skipped when disturbed.
Large shoots of D. strictus and other bamboo species could be divided into three vertical zones regarding the developmental stage of their internodes and the degree of the protection by the culm sheaths (Fig. 9A): In the apical zone (Z1) the newly developing internodes were short and soft. The respective sheaths were much longer than the internodes and covered the shoot surface in several overlapping layers. In the central zone (Z2) the internodes were elongated and their already hardened apical parts were sticking out from their sheaths. Thus, the sheaths enclosed mainly the lower, softer parts of the internodes. In the basal zone (Z3) the bases of the internodes were still more hardened and the protective sheaths gradually loosened and fell off to the ground.
The vertical succession of the internode development was refl ected by the vertical succession of larval developmental stages of A. histrionica. At the base of the apical shoot zone Z1 eggs and freshly hatched larvae were found. In the central zone Z2 the upper internodes contained small larvae with inconspicuous feeding marks (feeding area still green), while in the lower part of zone Z2 the internodes contained larger larvae with a conspicuous large brownish feeding area. In the basal zone Z3 the sheaths of the upper internodes were loosened and just a small number of mature larvae were present, indicating that some had already left the internode for pupariation. In the lower internodes of zone Z3 both sheaths and larvae were lacking, but feeding marks indicated that A. histrionica had been present.
Acroceratitis histrionica fl ies (Fig. 9G) were recorded between March and October. Females oviposited below culm sheaths, about 1-6 mm from the sheath edge. Eggs were laid in clusters of several eggs. In one case eggs were found in fi ve sheaths lying adjacent to each other at a height between ca. 1.50 m and 1.80 m (height of shoot ca. 2.50 m). There were 8; 10; 13; 16 and 20 eggs per sheath. The oviposition site was situated in the area where the culm sheath was overlapping, ca. 8-9 cm below the upper margin of the sheath (Fig. 9B).
Cephalopharyngeal skeleton (Fig. 16). Total length 1.0-1.1 mm (median: 1.0 mm; n = 6). Indentation between apex of apical tooth of mouthhooks and ventral apodeme 0.17-0.23× as deep as mouthhook length; neck well developed; preapical tooth 0.005-0.010 mm long; ventral apodeme posteroventrally directed. Hypopharyngeal sclerite 3.6-4.4× as long as high. Hypopharyngeal bridge about 0.02 mm long, 0.02-0.03 mm wide, located slightly posterior of middle of hypopharyngeal sclerite. Hyaline areas at posterior tips of both cornua and at ventral margin of ventral cornu; brown areas of both cornua of virtually the same length. Anterior sclerite present in four of six examined specimens.
Thorax. Anterior spiracles with 15-16 tubules (Fig.  10F). Anterior margin of second thoracic segment with about 5 rows of spinules dorsolaterally and about 2 rows expanding from the ventral to the lateral side, a narrow area without spinules dorsally. Anterior margin of third thoracic segment with about 3 rows of spinules dorsolaterally and about 2 rows expanding from the ventral to the lateral side, an area without spinules dorsally.
Abdominal segments I-VII. First abdominal creeping welt bearing a band of very small spinules, arranged in patches anteriorly and about 8 rows of posteriorly directed spinules posteriorly; creeping welts of abdominal segments II-VII bearing rows of spinules arranged as follows (from anterior to posterior): A band of very small spinules, arranged in patches, 7-9 rows of posteriorly directed spinules, a band of very small spinules, 4-5 rows of poste- Fig. 10. Pseudocephalon and anterior spiracle of Acroceratitis incompleta (SEM micrographs). A -pseudocephalon, frontal view. Bpseudocephalon and part of the fi rst thoracic segment, lateral view. C -maxillary sense organ. D -preoral organ located on primary preoral lobe, surrounded by remaining preoral lobes. E -labial lobe. F -anterior spiracle. Abbreviations: la org -labial organ, pap-s -papillar sensillum, pit-s -pit sensillum.
riorly directed spinules (Fig. 11D-F). Anterior margin of fi rst abdominal segment with posteriorly directed spinules laterally. Anterior margins of remaining abdominal segments laterally with bands of very small spinules.
Caudal segment (Figs 16 and 11A-C). Shortest distance between spiracles about 0.4× as long as longest spiracular opening. Angle between dorsal and central opening up to 12°, shortest distance between dorsal and central opening about one fourth to one third as long as longest opening; angle between central and ventral opening 7-23°, shortest distance between central and ventral opening about 0.1-0.2× as long as longest opening. Spiracular openings 6.7-9.4× as long as wide. Spiracular hairs up to half as long as longest opening; dorsal bundle containing 8-15 hairs, ventral bundle containing 7-14 hairs, dorsolateral bundle containing 2-4 hairs, ventrolateral bundle containing 3-7 hairs (Figs 16 and 11B). Anal lobes slightly protuberant, with longitudinal ridge (elongate elevation of the surface of the anal lobe, bordering the opening of the anus; Fig. 11C) on each lobe. Few rows of anteriorly directed spinules anterior to anal lobes and rows of dentate scales posterior to anal lobes (Fig.  11C). Creeping welt as in abdominal segments II-VII.
Young larvae fed on white, soft substrate of internodes located about 80 cm below the shoot apex (basal third of the Z1 area, see Fig. 9A). The larvae occupied 1-2 inter- Fig. 12. Host plant, larvae and female of A. incompleta. A -young larvae of A. incompleta feeding on the shoot surface in the basal part of an internode. The upper part of the internode and the sheath were removed. B -apical part of an infested internode seen from below. One side of the wall was largely eaten up (left, white arrow), while the other side is intact and solid (right, black arrow). The larvae did not pierce the pith ring nor enter the internode cavity. C, D and E -body color changes during larval development from milky-pale to yellow (C) and then to yellow-orange (D) and yellow-green (E). F -Acroceratitis incompleta female laying eggs below the sheath edge of a D. strictus shoot.
nodes lying next to each other. They devoured the wall substrate between sheath and pith ring (= the inner side of the culm wall composed of layers of often lignifi ed parenchyma cells (Liese, 1998)) and usually did not enter the internode cavity (Fig. 12B). The largest number of larvae occurred in large upright shoots (more than 70 specimens), while in shoots lying on the ground the numbers were low. Young larvae were milky-pale, older larvae yellow, yellow-orange (orange especially in the front part) and at a later stage spotted yellow-green (Fig. 12C, D, E). Mature larvae readily skipped and left the internodes for pupariation. In three cases the development from egg-laying to eclosion of the adults lasted 20, 28 and 30 days.
Basal segment of antenna approximately cylindrical, apical segment cylindrical with rounded tip. Maxillary palpus surrounded by complete cuticular fold (Fig. 13C).
Thorax. Anterior spiracles with 17-19 (in some cases fl attened) tubules (Fig. 13F). Anterior margins of second and third thoracic segment lined with about 6 rows of spinules, on third thoracic segment all rows with a dorsal gap.
Abdominal segments I-VII. First abdominal creeping welt bearing about 9 rows of dentate plates; creeping welts of abdominal segments II-VII bearing rows of spinules arranged as follows (from anterior to posterior): A band of very small spinules, arranged in patches, 6-10 rows of medium sized, pointed, posteriorly directed spinules, a band of very small spinules, 4-7 rows of large, pointed, posteriorly directed spinules (Fig. 14D). Anterior margins of abdominal segments covered laterally by very small spinules.
Caudal segment (Figs 16 and 14A-C). Shortest distance between spiracles between one third and half as long as longest spiracular opening. Angle between dorsal and central opening up to 9°, shortest distance between dorsal and central opening about one third to half as long as longest openings; angle between central and ventral opening 11-35°, shortest distance between central and ventral opening 0.1-0.4× as long as longest opening. Spiracular openings 6.3-9.2× as long as wide. Spiracular hairs up to somewhat more than one third as long as longest spiracular opening, branched up to 7×; dorsal bundle containing 7-20 hairs, ventral bundle containing 7-15 hairs, dorsolateral bundle containing 2-5 hairs, ventrolateral bundle containing 3-8 hairs (Figs 16 and 14B).
Anal lobes with ridge on each lobe. Patches of very small spinules anterior to anal lobes and 5-8 rows of dentate scales posterior to anal lobes (Fig. 14C). Creeping welt as in abdominal segments II-VII.

Biology. So far, A. septemmaculata was reared from
Dendrocalamus strictus (Dohm et al., 2014). In the present study A. septemmaculata was reared from living or freshly cut upright bamboo shoots of Dendrocalamus strictus (3×) and Dendrocalamus sp. (1×), cut side branches of a D. strictus shoot lying on the ground (1×) and a felled shoot of Bambusa polymorpha (1×). Larvae were found between July and November.
Most larvae were found in large shoots of D. strictus (Fig.  15A) and D. sp., which were about 2-3 m tall. Larvae developed at the base of zone Z1 and in the upper part of zone Z2 (Fig. 9A). They occupied 1-2 internodes lying next to each other. Larvae fed on the internode surface below the culm sheaths and then penetrated into the internode wall (Fig. 15B). One of the larger internodes contained more than 70 larvae. The color of younger A. septemmaculata larvae was pale yellow (Fig. 15B), older larvae turned yellow orange (orange more pronounced in the cephalic part of the body) and then spotted orange-green and greenish (Fig. 15C). Mature larvae readily skipped and left their internodes for pupariation.
Acroceratitis septemmaculata fl ies were recorded in the fi eld between April and November. Eggs were laid side by side below internode sheaths (Fig. 15D). Females were observed to oviposit on D. strictus (Fig. 15E) and D. ham-  iltoni. Occasionally, two A. septemmaculata females laid their eggs side by side without interfering with one another.

Larval morphology
The facial masks of the examined Acroceratitis are approximately triangular and consist of moderately developed, serrated oral ridges, laterally surrounded by short, serrated accessory plates covering a much smaller area than the oral ridges. The facial mask of Gastrozona fasciventris is similar, but possesses dentate instead of serrated oral ridges (Elson-Harris, 1992: Plate 343). In other Gastrozonini the facial mask possesses dentate oral ridges with only a few accessory plates (Cyrtostola limbata; Elson-Harris, 1992: Plate 370), dentate oral ridges entirely lacking accessory plates (Chaetellipsis; Elson-Harris, 1992), possesses mostly accessory plates (Ichneumonopsis burmensis; Kovac et al., 2013) or is signifi cantly reduced (Acrotaeniostola spiralis; Kovac et al., 2017).
The maxillary sense organs of Acroceratitis are similar to those of previously described Gastrozonini species. The antennae are two-segmented as in the third instar larvae of other described Gastrozonini larvae. In the basal area of the antennae there appears to be an additional antennal segment (see Figs 1A, B, 4A, B, 7A, B, 10, 10B, 13A). However, in our opinion this is most likely an artifact. This basal "segment" is, in contrast to the actual two antennal segments, rather amorphous, not well differentiated from the surface of the pseudocephalon, variable in size and shape or even not visible at all. For example, in one specimen of Acrotaeniostola longicauda the basal "segment" of one of the two antennae was not visible and there was just a deep furrow instead. In addition, the basal "segments" sometimes showed a transverse furrow at about the middle of their length. Therefore, the third basal "antennal segment" is probably just the invaginated base of the antenna, which protrudes from the pseudocephalon due to the heating process (see Material and methods). A similar third basal "antennal segment" is also visible in the SEMmicrographs presented by Elson-Harris (1992;Plates 312, 341). Nevertheless, she stated that there are only two antennal segments in the Gastrozonini larvae described by her (Elson-Harris, 1992).
All Acroceratitis species possess identical isolated sensilla on thoracic and abdominal larval segments (Fig. 17). SEM micrographs provided by El son-Harris (1992) suggest that the sensilla of Gastrozona fasciventris are probably similarly arranged as in Acroceratitis. In contrast, Acrotaeniostola spiralis has an additional pit sensillum on the second and third thoracic segments and the sensillum number 11 on the fi rst thoracic segment and number 10 on the third thoracic segment (counted dorsal to ventral) are pit sensilla instead of papilla sensilla, while I. burmensis has a signifi cantly different arrangement (Kovac et al., 2013(Kovac et al., , 2017.
Caudal ridges (Figs 2A, 5A, 8A, 11A, 14A) and a dark transverse line (exception: rare specimens of A. histrionica, Table 1) are present in all Acroceratitis species. They are also present in Chaetellipsis paradoxa, Cyrtostola limbata, and Gastrozona fasciventris (own observations), but absent in Ichneumonopsis burmensis and Acrotaeniostola spiralis (Kovac et al., 2013(Kovac et al., , 2017. Differences among the Acroceratitis species allow a division of the fi ve species broadly into two groups. The fi rst group consists of Acroceratitis ceratitina, A. incompleta, and A. septemmaculata. These three species have an additional pair of papillar sensilla on the labial lobe next to the labial organ (Figs 1E, 10E, 13E), lack large tubercles on the labial lobe, show a nearly identical pattern of spinules on the creeping welts (type 1; Figs 2D, 11D, 14D), and have a shorter distance between the posterior spiracles (up to about 0.5× the length of the longest spiracular opening). Larvae of A. incompleta and A. septemmaculata are very similar, while A. ceratitina shows some differences (labial lobe tubercles present, anal lobe ridges absent, posteriorly directed spinules posterior to anal lobes present, see Table  1).
The second group consists of A. distincta and A. histrionica. These species differ from the fi rst group by the lack of additional papillar sensilla on labial lobe (Figs 4E, 7E), the presence of many large tubercles on labial lobe (Figs 4E, 7E), different arrangements of spinules on the creeping welts and a longer distance between the posterior spiracles (0.5-0.9× as long as the longest spiracular opening). A. histrionica and A. distincta are distinguished by different types of creeping welts. Furthermore, A. histrionica has fewer oral ridges, accessory plates and preoral lobes than A. distincta.

Biology
Acroceratitis larvae occupied specifi c bamboo niches depending on developmental stage and thickness of bamboo shoots, degree of protection by internode sheaths and condition of shoots (living or dead, placed in the upright position or lying on the ground). Thus, A. ceratitina occurred in thick young shoots just protruding from the ground (Fig. 3A, B), A. incompleta and A. septemmaculata in older and taller thick shoots (ca. 1.5 m or higher), A. distincta in thin shoots (Fig. 6A, B, C) and A. histrionica on surface of thick, older shoot internodes (Fig. 9A, C, D). Larvae inhabiting older (elongated) internodes always fed in the soft basal part of the internode, whereas the apical internode area was harder and sticking out from the protecting internode sheath.
Species mining in tall living shoots infested only 1-3 internodes per shoot, which were located in the apical shoot area (Fig. 6A). These internodes were probably particularly suitable for larval development, because they were softer than lower internodes and less protected by tight overlapping culm sheaths than apical internodes. Acroceratitis histrionica larvae were confi ned to living bamboo shoots, while other Acroceratitis developed successfully in living as well as in freshly cut shoots.
Freshly cut bamboo shoots were readily colonized by A. ceratitina, A. incompleta and A. septemmaculata when placed in the upright position, but larvae were lacking or rarely encountered in cut shoots lying on the ground. This is consistent with the observation that females were mainly walking on upright shoots in search of egg-laying sites. These fi ndings indicate that females of these species used the vertical aspect of a bamboo shoot as visual cue for locating egg-laying sites.
The reason for the preference of upright shoots was perhaps the avoidance of competition with other Gastrozonini, especially Gastrozona and Chaetellipsis, because in large freshly cut shoots lying on the ground Chaetellipsis and Gastrozona were very common, while in upright cut shoots they were rarely found. Likewise, Chaetellipsis and Gastrozona were rarely found in thin living or cut shoots, while A. distincta larvae were abundant in living as well as in both vertical or horizontal cut shoots.
Acroceratitis histrionica larvae were special among Acro ceratitis, because they developed on the surface of signifi cantly hardened internode walls, rather than mining in the bamboo wall substrate (Fig. 9C, D). Nevertheless, their mouthparts did not differ from those of other Acroceratitis. Young A. histrionica larvae macerated the uppermost, soft layer of internode surface, while older larvae probably scraped food particles or microorganisms from the surface of underlying hard and fi brous internode wall.
Acroceratitis histrionica larvae developed in a triangular area ("feeding triangle") located close above the branch bud (Fig. 9C, D). Internodes lacking branch buds were not inhabited. The preference for the feeding triangle area was apparently not linked to the food quality at that location, since the internode surface looked identical on all sides of internode base and the branch bud was never eaten. It was rather connected to the fact that the growing branch bud pushed the corresponding culm sheath upward, thus producing a fl at triangular cavity in which larvae could move around easily. Furthermore, moisture collected in the cavity, thus probably creating favorable microclimatic conditions for the larvae. On the other hand, occasional fl ooding of cavities led to periods of low oxygen and temporary immobility or even drowning of larvae.
Occasionally, freshly hatched A. histrionica larvae left indentations on the bamboo surface, thus allowing the reconstruction of the path towards their feeding site (Figs 9E, C, F). Judging from indentations the larvae chose the path of least resistance and squeezed downwards along the concealed sheath edge covered by the opposite end of the sheath. When reaching the basal internode area in the vicinity of the branch bud where the sheath was less tightly enclosing the internode surface larvae dispersed towards the feeding triangle area. In several cases indentations leading to the feeding triangle were present, but feeding marks were lacking, indicating that larvae had died for unknown reason.
Infested shoots showed different patterns of damage depending on Acroceratitis species. Small thick shoots attacked by A. ceratitina always died (Fig. 3A). Shoots infested by A. distincta remained alive and produced side branches, however, the apical part died off and fell to the ground (Figs 6A, B). A. incompleta and A. septemmaculata damaged the walls of 1-3 internodes of a particular shoot (Figs 12B, C, 15B), but it is not certain whether larvae killed the shoot or its apex. A. histrionica larvae damaged only small areas of the basal internode surfaces (Fig. 9F).  A. ceratitina larvae successfully penetrated cut shoots buried in the ground (Fig. 3B), but it is not certain whether they were able to overcome tight sheaths of intact bamboo shoots. According to present observations it is possible that A. ceratitina only infested shoots weakened by disease. Shoots killed by diseases (no traces of insect attacks) were common. They probably possessed slightly loosened shoots due to desiccation, thus allowing easier oviposition and access for Acroceratitis. This may also explain, why A. incompleta and A. septemmaculata larvae were more abundant in cut shoots placed in upright position rather than in intact living shoots. So far, only the Gastrozonini Cyrtostola and Paraxarnuta are known to overcome the apical sheaths barrier by using egg-laying holes made by Cyrtotrachelus beetles (Curculionidae) (Kovac & Azarae, 1994;Dohm et al., 2014).
The favorable response of fl ies to human urine and sweat was remarkable. It was probably triggered by ammonia, a breakdown product of urea which is also present in volatiles from fermentation of protein or bird droppings. Ammonia is a well-known attractant of fruit fl ies and is probably perceived by fl ies as a cue for a protein source (Bateman & Morton, 1981;Mazor et al., 1987;Thomas et al., 2008). In frugivorous tephritids and probably also in Gastrozonini protein feeding is essential for egg maturation, while it seems to be less important for males. However, in some species protein feeding may increase the reproductive success of males, for example, by increasing their pheromone production (overview in Drew & Yuval, 2000).
Studies with Anastrepha showed that urine-baited traps or fermenting protein lures tended to capture signifi cantly more females than males, although the results varied with respect to previous diet, reproductive state, species, season and habitat (Piñero et al., 2002;Aluja & Piñero, 2004;Thomas et al., 2008). In the present study males of A. distincta were more frequent on sweat and urine than females (80 males and 5 females on sweat; 23 males and 1 female on urine). Other Acroceratitis showed the same trend, although the numbers of collected specimens were lower. This result indicates that protein feeding may be an essential factor for males of some Acroceratitis species.
Drew & Yuval (2000) postulated, that the relationship between male diet and reproductive success will be most pronounced in lekking species, where high levels of both protein and sugar are necessary for pheromone production, courtship displays and territory guarding. Lekking is present in Gastrozonini (Kovac, 2015), but the reproductive behavior of Acroceratitis remains to be investigated.
In summary, a comparison of morphological with biological data shows that Acroceratitis groups based on larval morphological characters coincide with differences in habitat or other factors like larval coloration. Thus, larvae of A. ceratitina, A. incompleta and A. septemmaculata developed in soft bamboo internodes and were orange-green to greenish when mature, while larvae of A. distincta and A. histrionica fed on elongated, harder shoot internodes and were bright yellow.