Acerocephala ihulena Honsberger & Lorenzo-Elarco, 2024

Honsberger, David N., Honsberger, Maya, Lorenzo-Elarco, J. Hau‘oli & Wright, Mark G., 2024, The genus Acerocephala and observations of the life history of Acerocephala hanuuanamu sp. nov. (Hymenoptera, Cerocephalidae) and its bark beetle host on the island of O‘ahu, Hawai‘i, Journal of Hymenoptera Research 97, pp. 545-589 : 545-589

Acerocephala ihulena Honsberger & Lorenzo-Elarco sp. nov.

Figs 4 , 5

Diagnosis.

Females of this species can be distinguished from other known Acerocephala by the antenna with the first three funicular segments transverse and of similar size and shape, the fourth segment slightly but distinctly larger, the fifth larger and subcircular; internantennal ridge elongate-oval in dorsal view, elevated from face in lateral view; submarginal and marginal veins of forewing join smoothly with no callus; clypeus projected between the mandibles subrectangular and apically mildly bidentate; ovipositor sheaths exerted from gaster, approximately half its length, apically dark brown and basally lighter.

Differential diagnosis.

Females are easily distinguished from A. atroviolacea , A. aenigma , and A. pacifica by the first four funicular segments transverse and much smaller and shorter than the fifth; the overall gracility of the head and body; and the elongate-oval shape of the interantennal ridge ( A. atroviolacea , A. aenigma , and A. pacifica with fifth funicular segment of similar size to preceding segments; body and head thicker and more robust; interantennal ridge in dorsal view triangular or wedge shaped); and from A. atroviolacea and A. aenigma by lack of callus on the forewing ( A. atroviolacea and A. aenigma with callus present).

Females can be distinguished from A. hanuuanamu by the fourth funicular segment slightly but distinctly larger than the first three; interantennal ridge elevated from the plane of the face; ovipositor exerted from the gaster approximately half its length, dark brown apically and lighter basally; scutoscutellar lines sulcate and form a consistent arc to meet the transscutal articulation along the mesal line ( A. hanuuanamu with the fourth funicular segment indistinctly larger than the first three; interantennal ridge flush with the face lateral of the scrobes; ovipositor only slightly exerted from the gaster; scutoscutellar lines foveolate and meet the transscutal articulation lateral of the mesal line).

Females can be distinguished from A. indica by the shape of the interantennal ridge, in A. ihulena elongate-oval and reminiscent of the nose of a proboscis monkey ( A. indica with interantennal ridge blunt wedge-like, widening posteriorly with somewhat straight margins); by the anterior of the face excluding the clypeus distinctly concave in A. indica (nearly straight between the anterolatral corners of the face in A. ihulena ); and by the stigmal vein of the forewing, A. ihulena with the stigmal vein very short, only incompletely separating the stigma from the marginal / postmarginal veins ( A. indica with stigmal vein short, but a distinct vein-like constriction between the marginal / postmarginal vein and the stigma).

Description.

Female (Figs 4 a – d , 5 a – g ; morphometric measurements in Suppl. material 2). Length: Can be expected to vary depending on the size of the host; the 7 individuals collected range from 1.52–1.84 mm (Holotype 1.80 mm).

Coloration: Head brown, slightly lighter anteriorly in lower face, basal part of mandibles, and scape. Mesosoma orange, prothorax somewhat lighter. Gaster dark brown, slightly lighter basally, and ovipositor sheaths yellow in their basal 2 / 3 and brown apically. Legs yellow, pro- and metacoxae sometimes nearly translucent.

Head: Head subrectangular in full face view; more or less parallel sided, but subtly widest across eyes, of nearly equal width at about three quarters its length, and tapers slightly to the anterolateral corner; vertex between posterior ocelli nearly straight, broadly rounded laterally to eyes. Anterior of head excluding mandibles and projected clypeus a mildly concave arc. Clypeus projected medially, somewhat square in shape and mildly bidentate, occupies about half the lateral space between the mandibles when mandibles are closed. Interantennal ridge widest at about half its length, shape elongate-oval, similar to that of a proboscis monkey; carinate laterally, medially between the carinae slightly convexely rounded with light longitudinal striations; slightly elevated from the level of the face lateral of scrobes, in lateral view of head extends tangential to curve of face at its thickest point disregarding the interantennal ridge, and continues straight more or less parallel with ventral profile of the head, giving the head approximately equal thickness over its length while the face lateral of scrobes thins towards the mandibles. Interantennal ridge begins a gradual descent to the mouthparts posterior to the toruli, and at about even with middle of the toruli, somewhat abruptly curves downward, to connect with the projection of the clypeus, at an approximate 60 degree angle to the overall plane of its dorsal surface. Lateral margins of scrobal depression a consistent arc, narrowing posteriorly, and slightly carinate; scrobes decrease in depth posteriorly. Mandibles long and curved with three apical teeth; dorsal and middle tooth of a similar conical shape but with a deep groove between them, middle tooth projecting only very slightly more than dorsal tooth, ventral tooth joined to middle tooth more closely and projecting a short way beyond it. Occipital carina just posterior to posterior ocelli, subcircular. Texture smooth and shiny with sparse setae, on the dorsal side setae concentrated on the gena anterior of the compound eyes and on the face just above the scrobes, a few longer setae dorsally around the occipital carina, ventral side of head posterior of the indented region also with longer setae. Bottom of toruli adjacent to the dorsoventrally inclined arc that comprises the anterior of the face. Ventral side of head generally flat, median suture extends longitudinally from the mouthparts to the occipital carina, basal area near the suture indented slightly for about 2 / 5 the length of the head, cuticle posterior of this indent of rougher texture than cuticle within it.

Antennal flagellum composed of 5 funicular segments and a subconical clava. First three funicular segments transverse cylindrical, progressively becoming subtly conical, and of similar size, 4 th is slightly rounded and slightly but distinctly larger than first three, 5 th is subspherical and larger than the 4 th; clava is approximately 1.4 times as long as the first four flagellar segments together, rounded subconical. All antennal segments with thin setae projecting at approximately 45 degrees, first four funicular segments each with a single whorl of longer setae of length subequal to the width of the segments, setae becoming progressively shorter towards apex of antenna. 5 th funicular segment and claval segments with MPS, clava with 3 whorls of MPS, first four flagellar segments lacking MPS. Mean length / mean width (Ratio) of antennal segments, length and width measured relative to length of F 1 (n = 7): Scape 6.8 / 1.7 (4.0); Pedicel 2.9 / 1.4 (2.0); F 1 1.0 / 1.2 (0.8); F 2 0.9 / 1.2 (0.7); F 3 1.0 / 1.4 (0.7); F 4 1.3 / 1.8 (0.7); F 5 2.1 / 2.6 (0.8); Clava 5.8 / 3.2 (1.8).

Mesosoma: Long prothorax articulates with mesothorax. A light longitudinal carina present laterally on pronotal neck, pronotum otherwise smooth, neck distinct from collar but lacking transverse pronotal carina. Pronotal neck with light transversely ridged texture, and collar smooth; collar with scattered mesal pointing setae except on its dorsal surface near the median line where there are no setae; one or two longer setae on its ventral side, of similar length to those on the ventral side of the head. Pronotum distinctly widest at its middle, but at its widest distinctly narrower than mesonotum in dorsal view, at its maximum width a little less than 3 / 4 the width of mesoscutum not including the tegulae. Transscutal articulation straight, inclines to vertical at its lateral edges. Notauli and scutoscutellar suture both well defined sulci; mesoscutum and scutellum lacking setae medially, a few setae present near the notauli, scutoscutellar suture, and transcutal articulation. On each side of the median line, notaulus curves to meet the transscutal articulation in the vicinity of 60 degrees, and scutoscutellar suture arcs to meet the transscutal articulation along the medial line, so medial lobe of mesoscutum runs up against transscutal articulation but scutellum is separated from it by the axillae except at its middle point (see morphometric measurements). Posterior margin of scutellum evenly convex. Mesonotum smooth and glassy, free of setae except on either side of the notauli and on either side of the scutoscutellar suture, a few scattered setae on the axillae and near the posterior margin of the scutellum. Metanotum visible behind scutellum. Propodeum slightly wrinkled at its anterior margin but otherwise smooth and glassy, flat longitudinally and lightly rounded transversely. Propodeal spiracle slightly recessed in a small depression. Sides of propodeum slightly tapering posteriorly until above the attachment of the rear coxae, where its dorsal margin bends to run nearly transversely except for at the nucha which is bumped out posteriorly, top of petiolar insertion flush with disk of propodeum. Setae near lateral margins of the propodeum: on callus anterior of the spiracle, and on lateral margin where the callus meets the metapleuron. Mesopleuron with light reticulate texture, mesepisternum with setae.

Wings: Forewing: Length of marginal vein approximately 1.1 × length of submarginal vein. Stigmal vein very short, hardly separating stigma from marginal vein, postmarginal vein also very short or absent, appearing together with stigmal vein as a slightly thickened apex of the marginal vein. Marginal vein without dorsal setae. Marginal setae begin at base of marginal vein and continue around apex of the wing with approximately consistent length to trailing edge approximately even with stigmal vein, just apical of the retinaculum. Parastigma not swollen, submarginal and marginal veins joining smoothly and unperturbed with no swollen area or change in pigmentation; lack of any sign of a tuft of setae on the parastigma. Membrane subhyaline with a slightly rippled texture and without setae.

Hindwing: Venation with a concavity at approximately half the distance from the base to the hamuli; anterior margin of wing membrane extends more or less straight in this region so at the apex of the concavity the venation reaches nearly ½ the distance to the posterior of the wing membrane. Venation ends at hamuli but the very anterior of the membrane in line with venation remains somewhat pigmented beyond the hamuli, to approximately 2 / 3 the length of the wing. Marginal setae present from the end of this pigmented region around to the base of the trailing edge of the wing, longest around trailing edge and posterior margin where they are approximately 3 / 4 the maximum width of the wing.

Legs: Fore and rear coxae globose, rear coxa larger than fore coxa, mid coxa smallest. Femora also globose, slightly larger than their respective coxae in front and mid legs, of similar size but more elongate in back legs. Tibiae of similar length to their respective femora, also globose apically, but narrower basally. In all legs, first tarsal segment longest, 2 nd through 4 th segments sequentially decrease in length; 5 th segment not including the claw subequal to 1 st in front legs, subequal to the 2 nd in mid and back legs. Protibial spur deviates from straight with a jog of shape similar to a logistic function, its base set back on tibia and extending near middle of the first tarsal segment, tibia beneath spur with a protibial comb.

Metasoma: Petiole somewhat narrower medially than basally or apically, one or two lateral setae on a small knob near its middle; overall somewhat longer than wide. In dried individuals, 1 st gastral tergite longest; 4 th tergite a little more than half the length of 1 st; 2 nd and 3 rd subequal and about half the 4 th; 5 th very short; 6 th a little shorter than 4 th. Gaster not heavily sclerotized, shrivels slightly on drying. Gaster with few, small setae on first five segments, 6 th with many setae, some short, some longer and reaching about half the length of the exerted ovipositor sheaths. Dried individuals have hypopygium distinct. Ovipositor sheaths substantially exserted from apex of gaster, setae basally present on only ventral side, medially and apically present; longest setae at about half its length, reaching approximately the apex of the sheaths; ovipositor thin and needle-like.

Male (Figs 4 e – h , 5 h – j ; morphometric measurements in Suppl. material 2). Length. Only two individuals collected, allotype 1.14 mm, paratype 0.85 mm.

Similar to female except: Wingless. Antennae with 6 funicular segments, basal four segments of relatively consistent size, fourth through sixth increasing in size; clava with two whorls of MPS. Mean length / mean width (Ratio) of antennal segments, length and width measured relative to length of F 1 (n = 2): Scape 9.1 / 2.7 (3.4); Pedicel 4.3 / 2.3 (1.9); F 1 1.0 / 1.8 (0.6); F 2 1.2 / 1.9 (0.6); F 3 1.1 / 1.9 (0.6); F 4 1.4 / 2.1 (0.7); F 5 2.1 / 3.2 (0.7); F 6 2.9 / 4.1 (0.7); Clava 6.9 / 4.7 (1.4). Ocelli absent and compound eye smaller than in female, face very subtly widest about even with middle of scrobes, morphometrics of face otherwise similar, mandibles similar. Body size relative to head size smaller than in female. Posterior region of pronotum tapers slightly to fit around the narrower medial lobe of the mesoscutum. Notauli meet on medial line anterior of transscutal articulation, scutoscutellar suture meets transscutal articulation lateral to this. Pro- and mesoscutum smaller than in female, scutellum much shorter than in female, metanotum behind it appears in dorsal view with nearly straight posterior margin but partially covered by the scutellum medially. Femora, tibiae, and tarsal segments somewhat stouter than in female, but their relative lengths remain similar. Posterior margin of propodeal disk with scattered setae. Gaster shorter, truncated-looking apically in dry specimens.

Materials examined.

Holotype (Fig. 4 a – d ): ♀; Hawaiian Islands, O‘ahu, Mānoa ; 21.5571 ° N, 157.8783 ° W; 24. vii. 2018; ex Hibiscus tiliaceus branches (deposited in UHIM) .

Allotype (Fig. 4 e – h ): ♂; Hawaiian Islands, O‘ahu, Mānoa ; 21.5571 ° N, 157.8783 ° W; 11. xii. 2019; pupating next to Eidophelus pacificus dessicated larva in E. pacificus tunnel in H. tiliaceus branch (deposited in UHIM) .

Paratypes: 12 ♀, 1 ♂. Hawaiian Islands, O‘ahu, Mānoa ; 21.5571 ° N, 157.8783 ° W; 24. vii. 2018; ex Hibiscus tiliaceus branches; 7 ♀ point mounted, 1 ♀ slide mounted (2 point mounted, 1 slide mounted UHIM; 2 BPBM; 3 NMNH) • Hawaiian Islands, O‘ahu, Mānoa ; 21.5571 ° N, 157.8783 ° W; 17. x. 2018; ex Hibiscus tiliaceus branches; 2 ♀ point mounted (1 BPBM; 1 NMNH) • Hawaiian Islands, O‘ahu, Mānoa ; 21.5571 ° N, 157.8783 ° W; 2. xii. 2019; ectoparasitoid on E. pacificus larva found under bark of H. tiliaceus branch; 1 ♀ ( BPBM) • Hawaiian Islands, O‘ahu, Mānoa ; 21.5571 ° N, 157.8783 ° W; 11. xii. 2019; pupating next to Eidophelus pacificus dessicated larva in E. pacificus tunnel in H. tiliaceus branch; 1 ♂ ( NMNH) • labeled “ Tapatapao // Upolu, Samoa // vii- 13-40 //// 1000 ’ //// Beating dead branches //// EC Zimmerman Collector ”; 1 ♀ ( BPBM) .

Etymology.

The species name is Hawaiian (lit., yellow nose). The nose-like interantennal ridge appears to be a good distinguishing character in the genus. In this species the interantennal ridge is distinguished from that in other Acerocephala by its elongate-oval shape with carinate edges and light longitudinal striations, reminiscent of a banana cut in half longitudinally. Ihulena is a play on ihu (nose) and iholena, a short and rounded variety of banana brought by Polynesian ocean voyagers and grown in Hawai‘i.

Known distribution.

This species is known from the island of O‘ahu in Hawai‘i, and the island of Upolu in Sāmoa. It is likely adventive in Hawai‘i.

Known hosts.

Found developing as an ectoparasitoid of an Eidophelus pacificus (Schedl, 1941) ( Coleoptera : Scolytinae) larva in tunnels under the bark of a Hibiscus tiliaceus L. log in Kahana Bay, O‘ahu, 21.5571 ° N, 157.8783 ° W, 15 m (Fig. 17 b, c ). Also found pupating next to dessicated larvae of E. pacificus in otherwise uninhabited larval feeding tunnels in the same logs (Fig. 17 a, b ).

Key to world known Acerocephala spp. females

(See Figs 1 – 8 for photos)

Note: Males are only known for A. hanuuanamu , A. ihulena , A. indica , and A. aenigma , so males of other species must be found before distinguishing characters are determined with certainty. Males of A. hanuuanamu , A. ihulena , A. indica , and A. aenigma are, however, very similar to females in characters of the head with the exception of the antennae in all four species, and the smaller eye and lack of ocelli associated with aptery in A. hanuuanamu , A. ihulena , and A. indica ; the overall shape of the head and the interantennal ridge very similar between the sexes. If this pattern holds for other species, males may likely be identified using these same characters of the head as well, listed in the key. Males are wingless in A. hanuuanamu , A. ihulena , and A. indica , and winged similar to the females in A. aenigma , so wings and the associated development of the mesosoma, along with characters of the gaster associated with their sex, may be expected to differ.

Behavior of C. brasiliensis and A. hanuuanamu in phloem sandwiches

Cryphalus brasiliensis beetle galleries in F. microcarpa wood

Some of the C. brasiliensis beetles placed in the phloem sandwiches bored into the wood and excavated somewhat irregular subovate chambers, much wider than the entry tunnel, in which they laid their eggs (as in Fig. 12 a , see also Fig. 11 b for the corresponding stage of development in natural wood). Adult beetles were observed to actively clean the inside of these chambers by pushing debris out of the entrance to the tunnel with their elytral declivity. Larvae hatching from their eggs chewed tunnels around the perimeter of the chamber and then progressed generally along the grain of the wood, often moving as a whole in somewhat of a comb pattern but with staggered progress. On one occasion, an adult beetle was observed using its head and prothorax to vigorously and repeatedly push a newly emerged larva, the behavior continuing for a minute or two. The purpose of this behavior was unclear to us, but two possibilities may be (1) an attempt to assist the larva in starting its own feeding tunnel off the main gallery, or (2) a transfer of symbiotic organisms to the larva. If the view boxes were suddenly exposed to high levels of light, adults were observed to become somewhat agitated and occasionally cannibalize their eggs. We did not observe any clear differences in the behavior of immature beetles when exposed to different levels of light.

Moving through the beetle galleries

(Video 1 (https://vimeo.com/717186840 ))

Female A. hanuuanamu entered wood in the phloem sandwiches through holes drilled by the beetles. After entering the tunnels, adult beetles with eggs or larvae deeper in their tunnels were observed to actively attempt to block the wasp from progressing past them deeper into the tunnel. A beetle did so by sharply moving its posterior to block attempts by the wasp to go around. Wasps were observed to grab the elytral declivity or posterior of the abdomen of beetles with their mandibles when attempting to pass them, though this was not observed to actually help them progress past the bark beetle. Physically encountering an adult beetle in an open area, either in a spacious gallery or outside wood, not in the confined space of a tunnel, resulted in the wasp quickly moving away. Such avoidance behavior was potentially for good reason, as when wasps and beetles were placed together in a confined space such as a tube, it was not uncommon to see wasps missing their entire gaster which, though this was never actually observed, was presumably bitten off by the beetles.

The beginnings of some tunnels bored by the larvae were very narrow due to the small size of the larvae in their early development, though small larvae were also observed to sometimes work in small groups of two or three larvae that would move through the wood side by side and create a wider tunnel (e. g. Fig. 12 c and the two larvae on the right side in Fig. 12 a ). Though A. hanuuanamu adults vary substantially in size depending on size of the host on which they develop, even some of the smallest female wasps were not able to fit through these smaller larval tunnels. As time progressed, more space opened up as beetle larvae and adults developed and created tunnels in the wood, and individual, isolated tunnels formed a network of intersecting tunnels. Observations in both naturally infested wood (see Fig. 11 d, e ) and the phloem sandwiches (see Fig. 10 d ) showed that this dynamic would progress until eventually almost all the inner bark had been consumed and the outer bark was only loosely attached to the xylem layer. In a developed tunnel network such as this with a population of host beetles, A. hanuuanamu adults were able to move around with much more ease, encounter more potential hosts and fewer one-way tunnels blocked by adult beetles, and seemed to have the most success.

Such tunnel systems often contained frass and debris from the beetles’ excavation, sometimes packed somewhat densely, and the wasps would dig through it in pursuit of hosts. Acerocephala hanuuanamu has remarkably large mandibles, for which one use was to facilitate this digging behavior. Females progressed through debris-filled tunnels by grabbing a chunk of debris with the mandibles, wrestling it loose, and then passing it within reach of the front legs. Then using all her legs in sequence she would quickly roll it back behind her abdomen in a running motion. In this way, the wasp was able to travel through the tunnels in a way reminiscent of a bubble in a tube of liquid: surrounded on all sides and moving material around its margins to progress through.

Acerocephala hanuuanamu females having entered a tunnel seemed to be able to sense the approximate location of potential host larvae they could not directly see, shown by apparent deliberate motion in its direction, though the sensory mechanisms involved were not clear.

Turning around

(Fig. 13 ; Videos 1 (https://vimeo.com/717186840), 2 (https://vimeo.com/716967741 ))

One amazing feat of agility accomplished by A. hanuuanamu with regularity is their ability to turn around in the tight confines of a tunnel, often not much wider than the wasp itself. The wasp first articulates its neck to face its head down, then bends its prothorax down, and in doing so slides its head under the rest of the as-of-yet unmoved posterior parts of its mesosoma. Bending its body along its points of articulation and pushing on its own body with its legs while doubled over to assist its progress, the prothorax is followed by the rest of the mesosoma and finally by the flexible gaster, ending with the wasp being in more or less the same location but facing the opposite direction. This maneuver is carried out very fluidly and quickly, and usually takes less than a few seconds, but sometimes longer if the wasp appears not to have a strong purpose for its movement. The fluidity in this snake-like maneuver is in part made possible by the head, the prothorax, and the rest of the mesosoma all being subequal in length, and the ability of the prothorax to articulate with the rest of the mesosoma. A similarly long, articulating prothorax is a character found in other unrelated taxa, such as much of the Bethylidae , which also typically attack their hosts in concealed environments. The agility in a tunnel system or tight space that this trait confers could potentially be what has resulted in this convergence.

Parasitism and host feeding on larvae in tunnels and construction of a feeding tube

(Figs 14 – 16 ; Videos 1 (https://vimeo.com/717186840), 2 (https://vimeo.com/716967741 ))

When a host larva was encountered in the tunnels, the A. hanuuanamu female would antennate upon it. Special attention was seemingly placed on the frass and debris around the larva, suggesting that this is a cue for host location, host acceptance, or both. Once a host was identified, the wasp would then turn around using the previously described turning maneuver and back up in the direction of the larva, moving its body backward in pulses and extending its ovipositor. Sometimes the wasp would contact the larva with its ovipositor, but other times it would not, and turn around to reexamine the larva with its antennae. The wasp seemed very cautious when encountering the larva, especially when backing up into it, presumably because of the danger presented by its mandibles. If the ovipositor made contact with the larva, the wasp would fully extend it into the body of the larva (Fig. 14 a ). The wasp would typically remain with its ovipositor inside the larva for approximately 15 minutes. During this time, the larva would gradually cease motion, the wasp apparently having injected it with a paralytic venom.

When finally removing its ovipositor from the larva, the wasp would do so slowly and carefully. In a few observed instances as it withdrew from the larva, tissue from inside the body of the larva was visibly pulled out as a sheath around the ovipositor. This sheath would remain projecting from the larva after the wasp had extracted its ovipositor. The wasp would then turn around in the tunnel and put its mouthparts on that projection, and host feed on the larva through it, evidently having produced a drinking straw (Fig. 15 ). On occasions when construction of such a straw was not observed, the wasp would still put its mouthparts precisely on the place where it had stung the larva. The wasp would remain in this position, host feeding typically for about 15 minutes. The wasp’s body would be still but its mouthparts could be seen moving slightly, and it would periodically expel a clear liquid from its abdomen (Fig. 14 b ). This liquid is likely to be excess fluid excreted for the purpose of concentrating nutrients from the host within the body of the wasp. Eventually the wasp would stop feeding and would then often turn around again, sting the larva for a similar amount of time as it had initially, and then feed again. This process was often observed to be repeated three or four times on the same larva. Stinging was always observed to result in death of the beetle immature. The wasp would most often leave the larva without laying an egg, and sometimes it would eventually lay an egg on the same larva upon which it had host-fed.

Feeding and oviposition through a pupal chamber

(Video 2 (https://vimeo.com/716967741 ))

Cryphalus brasiliensis prepupae were observed to build a hard pupal shell around themselves made from bonded tunnel debris before they pupated (Fig. 12 b ). Upon approaching a pupal chamber, wasps were clearly able to sense the presence of a host and were stimulated to excavate the loose material in the vicinity of the pupal chamber and dig into the hard pupal shell as far as possible. This process of excavating appeared to be very laborious and was observed to last over an hour on occasion. Due to the toughness of the wall, which seemed to increase in strength near its inner boundary, the wasps were not observed to break all the way through the shell to the pupa or prepupa inside. After progressing as far as it deemed practical, the wasp would stop digging, turn around, place its back legs on the wall of the pupal shell with its gaster pointing towards the inside of the chamber, extend its ovipositor, and use its body to push its ovipositor through the shell. This was a lengthy process. After pushing and pushing, the ovipositor would finally break through to the other side (Fig. 14 c ). The wasp would then flex the ovipositor, probing the space in an attempt to contact the larva or pupa inside. Great flexibility and control over its ovipositor was observed (Fig. 14 e ; Video 2 (https://vimeo.com/716967741 )). If able to contact and penetrate the pupa or prepupa inside, it would sting it for about 15 minutes, and then eventually remove its ovipositor, slowly and carefully as it did when stinging a larva, simultaneously pulling the pupa or prepupa into contact with the wall of the chamber. The wasp would then turn around and put its mouthparts on the exact place where it had drilled through the pupal chamber. Physically separated from the pupa by the wall of the pupal chamber, it would then begin feeding, again through a tube it had made (Fig. 14 d, f ). This was observed multiple times. That it was actually feeding successfully was clear: its abdomen would progressively distend and periodically excrete clear liquid, and the pupa would noticeably progressively shrink in size. A feeding straw was never observed directly due to its position within the material, but it was clearly utilizing one as it was feeding from a distance through the wall. Occasionally a small bump was observed on the pupa or prepupa where it was in contact with the wall (Fig. 14 d, f ), presumably comprising the proximal part of the feeding tube. Wasps were observed repeating this drilling and feeding behavior multiple times on the same pupa before oviposition, the egg having also passed through its ovipositor and the narrow hole it had drilled through the wall of the pupal chamber.

Oviposition and development

An egg was never directly observed coming out of the ovipositor but was often later observed on a larva or pupa, often the same one on which a wasp had previously host fed. The surface of the eggs appears to be adhesive, and eggs were either stuck directly to the larva or pupa, or were placed in the tunnel adjacent to it. A wasp larva emerging from the egg would attach to the beetle larva or pupa (Fig. 16 a, b ), and over the next 2 or 3 days the wasp larva would suck the hemolymph out of the beetle larva or pupa, transferring more or less all of it to its own body, leaving the shriveled cuticle of its host (Fig. 16 c ). It would then detach and later pupate (Fig. 16 d , Video 3 (https://vimeo.com/717187250 )). The adult wasp would thus emerge in the tunnel or pupal chamber. Interestingly, though this was never observed in action, emergent adult wasps seemed to be able to make their way out from the inside of the beetle pupal chambers.

Mating

(Video 4 (https://vimeo.com/717187205 ))

A female and male were observed on two occasions to encounter each other and mate. Upon encounter, the wasps slowed their pace and touched antennae to antennae, maintaining that position for a few seconds. The female seemed to relax and the male moved onto the dorsal part of her abdomen, and then curled his abdomen around the female’s abdomen to copulate. The act of copulation took in the vicinity of one minute. The wasps then remained near each other, the male often contacting the body of the female with the antennae, the female remaining calm and slow. This sequence of events, including both copulation that appeared successful and attempts that did not seem mechanically successful, were repeated multiple times until the female moved away.

Both times mating was observed, it was outside the piece of wood in the open area of the observation chamber. This should not necessarily be taken to imply that mating takes place external to the wood in nature, though. The place it occurred both times was in frass / debris piles pushed out by the beetles during construction and maintenance of their tunnels. Mating was likely observed there because the chance of males and females encountering each other in the region outside the tunnels was high given that it was a somewhat open space and there were often wasps present in that region.

Observation in naturally infested wood

In F. microcarpa

(Fig. 11 )

Cryphalus brasiliensis beetles seem to be some of the first colonizers of the environment below the surface of F. microcarpa wood at the location studied on O‘ahu (Fig. 11 a ). They drill through the bark and form tunnels straddling the phloem and xylem layers of the wood, but can also be found entirely in the phloem layer on branches with thicker bark. For healthy wood cut from the tree in the studied environment, initial colonization of wood typically happens 3–5 weeks after it is separated from the tree. As multiple generations of beetle adults and larvae feed and tunnel through the wood, the tunnels become increasingly dense, often forming networks covering nearly the whole phloem layer underneath the bark. The beetles then begin to leave the branch at around this stage, and the bark often begins to separate from the rest of the branch, the phloem layer having largely been removed by the beetles. In sections of branches still connected to the tree that seemed to have been somewhat abruptly cut off from the vascular system of the tree, C. brasiliensis beetles were present in large numbers throughout the phloem layer for a period of time similar to those in the cut branches. They were also found to infest older branches at the junction between the dead section and where sap was still flowing. In these, it was difficult to tell whether the beetles were overcoming the immune defenses of the tree in that area and contributing to death of the wood progressively down the branch, or only progressing to the point where the wood had already stopped sap flow for other reasons.

Where there were reproducing C. brasiliensis in these trees, it was common to find both female and male A. hanuuanamu traveling in their tunnels, especially where there was a high density of C. brasiliensis larvae and pupae. In fact, in branches both disconnected from the tree and still attached to the tree, it was rare to find C. brasiliensis without at least females of these wasps also present. This suggests that the female wasps are efficient at locating wood newly infested by C. brasiliensis , and likely that they can do so by flying. The ectoparasitic larvae and pupae of A. hanuuanamu , reared to confirm their identity, were often found in the tunnels and pupal chambers built by C. brasiliensis beetles. Parasitoid larvae were found developing on both early and late instar larvae, which was reflected in the great size variation found in both female and male A. hanuuanamu . Parasitized larger larvae seemed to be substantially more common, which implies a preference for larger larvae given that these wasps are idiobionts. The parasitism rate seemed very high in some places; many pupal chambers or larval tunnels ending with parasitized beetle immatures or wasp prepupae or pupae (Fig. 11 d, e ). Two other ectoparasitoids were observed also parasitizing the C. brasiliensis beetles. These were an Ecphylus sp. ( Braconidae ) that oviposits into larvae and pupae through the bark, whose pupae could be distinguished morphologically or by the pupal cocoon they build when they pupate, and Cerocephala dinoderi Gahan, 1925 , which was much more rare than A. hanuuanamu or the Ecphylus sp. , and whose ectoparasitic larvae were observed feeding on C. brasiliensis larvae. The cerambycid Pterolophia bigibbera (Newman, 1842) was also common in this wood, and was parasitized by Rhaconotus vagrans (Bridwell, 1920) ( Braconidae ), Sclerodermus immigrans Bridwell, 1918 ( Bethylidae ), and Allobethylus ewa (Bridwell, 1920) ( Bethylidae ).

Additional hosts of Acerocephala hanuuanamu

In addition to C. brasiliensis in F. microcarpa , A. hanuuanamu was also found parasitizing Cryphalus mangiferae Stebbing, 1914 in mango branches ( Mangifera indica L.) in the area of Kahana Bay on O‘ahu island (21.5604 ° N, 157.8765 ° W), and a Cryphalus sp. in breadfruit ( Artocarpus altilis (Parkinson) Fosberg ) branches in Mānoa Valley on O‘ahu island (21.2954 ° N, 157.8145 ° W). Identities were confirmed by rearing of parasitoid larvae, and when bark beetle larvae could be clearly associated with the adult beetles in the tunnels. Acerocephala hanuuanamu was also found emerging from breadfruit fruits with Cryphalus negrosensis Browne, 1979 , which presents another putative host because the beetle was in high numbers and the only scolytid emerging from the fruits, though this was not confirmed through observation and rearing of parasitized beetles. Adults were also found with C. brasiliensis in Trema orientalis (L.) Blume branches in Mānoa Valley (21.2952 ° N, 157.8141 ° W).

Figures

Figure 1. Acerocephala hanuuanamu sp. nov. (paratypes) showing morphometric measurements. Acronyms explained in text. Scale bars: 1 mm (b, c); 500 μm (a).

Figure 3. Acerocephala hanuuanamu sp. nov. a forewing ♀ b hindwing ♀ c antenna ♀ d antenna ♂ e dorsal view ♀ (paratype) f head ♀ (paratype) g anterior view of head showing mandibles ♀ (paratype) h foreleg ♀ i midleg ♀ j hindleg ♀. Scale bars: 1 mm (e); 250 μm (f, g).

Figure 4. Acerocephala ihulena sp. nov. holotype ♀ (a – d) and allotype ♂ (e – h) a, e side view b, f head c, g anterior view of head showing mandibles d, h dorsal view. Scale bars: 1 mm (a, d); 500 μm (e, h); 250 μm (b, c, f, g).

Figure 5. Acerocephala ihulena sp. nov. a head (paratype ♀) b antenna (♀) c side view (paratype ♀) d forewing (♀) e hindwing (♀) f dorsal view (paratype ♀) g anterior view of head showing mandibles (paratype ♀) h head (paratype ♂) i anterior view of head showing mandibles (paratype ♂) j dorsal view (paratype ♂). Scale bars: 1 mm (c, f); 250 μm (a, g – j).

Figure 6. Acerocephala indica. Holotype ♀ (a – e); non-type ♀ (f – j); non-type ♂ (k – m) a, h, l head b antenna c forewing d mesosoma (dorsal view) e gaster f, k side view g, m dorsal view i anterior view of head showing mandibles j head.

Figure 7. Acerocephala pacifica. Holotype ♀ (a – c) and paratype ♀ (d – g) a, g side view b, d dorsal view c, e head f view of head showing mandibles. All photos taken by Lisha Jasper and Jeremy Frank at BPBM.

Figure 8. Acerocephala atroviolacea and Acerocephala aenigma. Acerocephala atroviolacea paratype ♀ (a – e); A. aenigma holotype ♀ (f – h) (photos from NMNH database); A. aenigma allotype ♂ (j – l) a, f, j side view b, h, l head c, g, k dorsal view d anterior view of head showing mandibles e side view of head showing antenna i A. aenigma antenna (paratype ♀) m A. aenigma antenna (paratype ♂).

Figure 10. Acerocephala hanuuanamu, Cryphalus brasiliensis, and the phloem sandwiches a A. hanuuanamu ♀ b C. brasiliensis ♀ c parts of phloem sandwich style observation chamber d phloem sandwich with colonies of C. brasiliensis in F. microcarpa wood e, f closeups of additional phloem sandwiches.

Figure 11. Ficus microcarpa wood where observations of behavior in a natural system were made. Photographs of wood are from naturally infested branches taken after peeling off the bark a the F. microcarpa tree from which most branches were collected and observations were made, in an unmaintained area of the University of Hawai‘i at Mānoa b wood in an early stage of colonization by C. brasiliensis beetles c three C. brasiliensis beetle immatures, each parasitized by a feeding A. hanuuanamu larva d Ficus microcarpa branch with bark peeled off, showing C. brasiliensis tunnel systems and beetles, and A. hanuuanamu larvae, prepupae, and pupae e same as (d) except the bark layer of the branch.

Figure 12. Cryphalus brasiliensis in phloem sandwiches a C. brasiliensis adults in a gallery with eggs and newly hatched larvae feeding and creating tunnels on the margins of the gallery b C. brasiliensis pupa and prepupa in hard pupal chambers they constructed c C. brasiliensis larvae feeding and moving through wood.

Figure 13. Acerocephala hanuuanamu turning around in C. brasiliensis tunnels in a phloem sandwich: sequence showing this agile maneuver frequently used to reverse orientation in the tunnels.

Figure 14. Acerocephala hanuuanamu stinging C. brasiliensis and host feeding in phloem sandwiches a stinging a C. brasiliensis larva. The ovipositor is visible extending nearly the length of the wasp’s gaster within the larva b subsequent host feeding on the same larva as in (a), on the same spot where it stung the larva c stinging a C. brasiliensis pupa through its hard pupal chamber after excavating the loose debris outside the chamber. The ovipositor is visible in the empty space between the chamber wall and the pupa d subsequent to stinging the pupa in (c), host feeding through the wall of the pupal chamber. Part of the feeding tube is marginally visible protruding from the pupa as a small bump between the pupa and the interior of the chamber wall, on the part of the pupa most proximal to the mouth of the wasp e probing with the ovipositor to find the pupa inside the chamber. The ovipositor is visible extending towards the pupa within the chamber. This attempt by the wasp was unsuccessful due to the position of the pupa within the chamber f an additional subsequent feeding event by the same wasp on the same pupa. The gaster is distended and the liquid periodically excreted by the wasp as it feeds is visible just past the end of the wasp’s gaster.

Figure 15. Feeding tube constructed by Acerocephala hanuuanamu after stinging, in phloem sandwiches a close-up of a tube constructed on a larva as the wasp host feeds b tube is visible extending from the posterior end of the larva near the wasp’s mouth. Such tubes constructed for feeding through the hard shell of the pupal chamber are also marginally visible in Fig. 14 d, f.

Figure 16. Development of A. hanuuanamu on C. brasiliensis in phloem sandwiches a small A. hanuuanamu larva on a C. brasiliensis larva b A. hanuuanamu female remaining motionless near a developing larva on its C. brasiliensis host for a long period of time as if guarding it c A. hanuuanamu larva inside a C. brasiliensis pupal chamber having consumed all the hemolymph of its host. This is the same pupal chamber as pictured in Fig. 14 c – f, the wasp having oviposited on the pupa subsequent to stinging and host feeding on it multiple times d A. hanuuanamu pupa inside C. brasiliensis tunnels.

Figure 17. Acerocephala ihulena development. Photos are from naturally infested wood taken after peeling off bark a A. ihulena male (paratype) pupa in an E. pacificus tunnel under the bark of H. tiliaceus wood, having developed on the desiccated E. pacificus larva next to it (indicated by yellow arrow) b A. ihulena female (paratype) pupa found in the same branches, also having developed on the now desiccated E. pacificus larva (indicated by yellow arrow) c, d E. pacificus larvae from tunnels under the bark of H. tiliaceus, with larval A. ihulena males developing as ectoparasitoids (parasitoid larvae indicated by blue arrows).