Prorops umiehu Honsberger, Lorenzo-Elarco & Magnacca, 2024

Honsberger, David N., Magnacca, Karl N., Lorenzo-Elarco, J. Hau‘oli & Wright, Mark G., 2024, Life history of two new species of Prorops (Hymenoptera, Bethylidae) ectoparasitic on adult Hypothenemus eruditus beetles (Curculionidae, Scolytinae) in Hawai‘i, Journal of Hymenoptera Research 97, pp. 1221-1256 : 1221-1256

Prorops umiehu Honsberger, Lorenzo-Elarco & Magnacca sp. nov.

Figs 4 , 5

Diagnosis.

Females can be distinguished from other described Prorops spp. by the combination of: head and mesosoma orangish-brown to reddish-brown; fore wing without vein 2 r-rs + Rs; snout bifid and with mesal sulcus; head including snout approximately 1.3 times as long as wide; metapectal-propodeal disc approximately 1.2 times longer than wide in dorsal view, and without distinct lateral carina. Males can be distinguished by the same set of characters as the females except for coloration and morphometrics of the head: body darker in color, brownish; head including snout approximately 1.1 times as long as wide.

Differential diagnosis.

Prorops umiehu can be differentiated from Prorops maya by lack of vein 2 r-rs + Rs in the fore wing (2 r-rs + Rs distinct in P. maya ); metapectal-propodeal complex and mesoscutellum overlapping medially, dividing metanotum in dorsal view (metanotum continuously visible posterior to mesoscutellum in P. maya ); female head 1.3 times as long as wide, male head 1.1 times as long as wide (head 1.1 times as long as wide in both sexes of P. maya ); mandible wide and tridentate, though dorsal tooth small and inconspicuous (mandible more narrow and bidentate in P. maya ); clypeus with anterior margin broadly rounded, posterior margin slightly emarginate (anterior margin acutely rounded, posterior margin sinusoidal in P. maya ); metapectal-propodeal complex without distinct lateral marginal carina (lateral marginal carina more distinctly conspicuous in P. maya ); female with head and mesosoma of similar orange-red-brown color (dark brown head and metasoma contrasting with orange mesosoma in P. maya female).

Female (Figs 1 , 4 a – d , 5 ). Length range: 1.18–1.41 mm (n = 11); Holotype: 1.38 mm. Head (Figs 1 , 4 a – c , 5 c, e ). Face, gena, vertex, and occiput orange to red-brown in color. Snout slightly lighter orange-brown, antenna yellow-brown basally, fading to brown apically. Frons and gena with lightly reticulate texture, density of reticulations increasing towards snout, snout itself with bumpy texture. Compound eye with short, sparse setae between ommatidia. Vertex incurved medially. Head with sides more or less parallel and only slightly outcurved, eye protruding slightly so widest part of head is across eyes and about half-way along VOL. Snout clearly bifid apically, ending in two distinct lobes apically curved dorsally to form two small teeth, and with dorsal median groove extending from where lobes meet to approximately even with middle of eye. Torulus located ventral to lateral margins of snout at approximately half its length, visible in dorsal view. Clypeus in anterior view extending laterally past torulus, tapers laterad; medially with dorsal margin broadly arched against snout and torulus, ventral margin shallowly incurved; overall appearing as handlebar mustache not, or only slightly, turned dorsad at its lateral corners. Antenna with scape curved ventrally and widened apically, apical surface forming cavity; pedicel subovate. Ten flagellomeres, 1 st smallest and cone shaped, 2 nd through 9 th shaped like apically truncated spheres, and similar in size, shape, and structure. Terminal flagellomere ovate. Setae of similar length and density on all flagellomeres, less dense on scape and pedicel. Mandible wide, ventral margin thickened with blunt ventral knob at about half its length; apically tridentate, ventral tooth largest, middle tooth smaller, dorsal tooth small and inconspicuous. Anterior region of head, including snout, clypeus, and mandible, distinctly more densely setose than rest of head. Morphometrics (range, n = 7 for all measurements): LH: WH = 1.24–1.31; LH: LE = 3.26–3.64; LH: VOL = 2.19–2.34; LHBE: LHAE = 0.35–0.38; VOL: LE = 1.41–1.66; WF: LE = 1.70–1.87; LE: BEM = 5.71–7.30; WOT: OOL = 0.40–0.45; POL: AOL = 1.10–1.54; AOT = 66 ° – 75 °; LH \ S: WH = 0.97–1.03. Antennal ratio approximately Scape: Pedicel: F 1: F 2: F 3: F 4: F 5: F 6: F 7: F 8: F 9: F 10 = 4.0: 2.6: 1.0: 0.9: 1.2: 1.2: 1.4: 1.4: 1.5: 1.6: 1.6: 2.9.

Mesosoma (Figs 4 a, d , 5 d ). Pronotum and metapectal-propodeal complex yellow to reddish-brown; mesothorax slightly darker in some individuals; legs apical of femora yellow-brown. Nota shiny with very light reticulate texture, difficult to see except under high magnification with the right lighting. Texture on dorsal surfaces slightly strongest on pronotal flange and metapectal-propodeal disc. Pronotal flange conspicuous; pronotum with sparse setae, longer setae at posterior margin. Anteromesoscutum with scattered short setae; mesoscutellum with pair of short setae on its lateral margin approximately even with posterior of axillae, another pair of longer mesally pointing setae at its posterior; tegula with few setae, denser than on surrounding sclerites; metapectal-propodeal disc without setae, declivity with few setae just mesal of transition to lateral surface of metapectal-propodeal complex. Transscutal suture such that posterior margin of anteromesoscutum is straight mesal of axillae. Anterior margin of anteromesoscutum somewhat visible under translucent pronotum. Mesosoma relatively flat along dorsomedian line. Neither notaulus nor parapsidal signum manifest on cuticle surface. Mesocutellum reaches metapectal-propodeal complex medially, metanotum very thin or not visible medially in dorsal view, distinct laterally. Mesopleuron projects from side of mesosoma, flattened subcylindrical in shape, and with few setae only on its ventral side. Mesopleural pit somewhat centrally located. Metapectal-propodeal disc approximately 1.25 times as long as wide; lateral marginal carina indistinct or absent, transition from disc to lateral surface rounded, though cuticle may appear thickened in dorsal view at this transition; disc somewhat outcurved transversely. Declivity when viewed perpendicular to its surface has vaguely subtriangular raised region, but appears overall subrectangular: posterior margin of metapectal-propodeal complex straight except for small bump above petiolar foramen. Lateral surface of metapectal-propodeal complex flat or slightly convex.

Legs (Fig. 5 f – h ). All tibiae with scythe shaped apical spur with comb-like setae on inner edge. Mesotibia with apical spines more numerous and stronger than in pro- and metatibiae, and with additional row of strong spines over its length on side opposite tibial spur; pro- and metatibiae lacking this row of spines. First tarsomere longest in each leg, most notably so in pro- and metaleg where it is subequal to 2 nd through 4 th tarsomeres combined. 2 nd through 4 th tarsomeres subequal in length in proleg, and sequentially decrease in length in meso- and metalegs. Length of 5 th tarsomere not including claw subequal to combined length of 2 nd and 3 rd tarsomeres in proleg, and subequal to 2 nd segment alone in meso- and metalegs.

Fore wing (Fig. 5 a ). Sc + R vein, prestigmal abcissa of radial 1, and pterostigma present; 2 r-rs + Rs absent, reduced to fold. Rs + M, M + Cu, and A veins lightly indicated. Sc + R vein with 1–3 setae. Prestigmal abcissa of radial 1 slightly inset from wing margin, pterostigma borders margin. Prestigmal flexion line present as hyaline stripe that separates prestigmal abcissa of radial 1 and pterostigma; thickness of this hyaline stripe and shape and size of prestigmal abcissa of radial 1 and pterostigma somewhat variable among individuals, though prestigmal abcissa of radial 1 and pterostigma typically subequal in size. Wing membrane overall subhyaline, slightly infuscate basal of prestigmal abcissa of radial 1 and apical of imaginary line between pterostigma and apex of retinaculum. Cubital and median flexion lines not visibly present; small hyaline spot projects into wing membrane on posterior margin at apex of retinaculum. Marginal setae present from prestigmal abcissa of radial 1 around wing apex, abruptly ending at beginning of straight trailing margin of wing. Hind wing (Fig. 5 b ). Wing membrane subhyaline, slightly infuscate apical of hamuli. Marginal setae absent on leading edge, present around apical margin and on trailing edge where length is about half maximum width of wing. Leading edge with slight projection culminating in dark spot with three hamuli at about half wing length.

Metasoma (Figs 4 a, b , 5 d ). Petiole and gaster dark brown to black, distinctly darker than orangish-brown mesosoma. First tergite constricted anteriorly to form distinct petiole, segment as a whole somewhat wider than long; constricted petiolar region of subequal length and width, with fine bumpy texture and shallow median dorsal groove extending its length. Remainder of metasoma shiny dark brown to black with smooth texture, each segment dorsally with sparse setae in a somewhat transverse row, these setae increasing slightly in length on posterior segments, last segment before sting with many setae on dorsal and lateral surfaces. In dorsal view, 1 st and 2 nd gastral segments subequal in length on median line, 3 rd and 4 th also subequal but shorter than 1 st and 2 nd. Metasoma widest at approximately 4 th segment. Sting often visible projecting slightly from apex of abdomen in dried or alcohol preserved specimens.

Male (Figs 4 e – g , 5 i ). Length range: 0.87–1.38 mm (n = 3); Allotype: 1.38 mm

As in female but with the following differences: Head more square than in female, length to width ratio approximately 1.1; eye larger and more bulging; ocelli more widely placed; vertex only weakly concave; 2 nd through 9 th antennal flagellomeres longer and more cylindrical; coloration typically darker with head and metasoma brown, mesosoma slightly lighter brown. Genitalia: See Fig. 5 i . Small relative to body size. Genital capsule broad, gonostipites and harpes combined only slightly longer than wide; harpe short, quadrate, truncate distally, shallowly concave medially. Penis valvae equal to or slightly exceeding volsella, distinct in ventral view. Morphometrics (range, n = 3 for all measurements): LH: WH = 1.07–1.12; LH: LE = 2.48–2.63; LH: VOL = 2.71–3.11; LHBE: LHAE = 0.33–0.38; VOL: LE = 0.80–0.96; WF: LE = 1.42–1.58; LE: BEM = 6.34–7.72; WOT: OOL = 0.63–0.74; POL: AOL = 1.35–1.68; AOT = 75 ° – 82 °; LH \ S: WH = 0.87–0.88

Materials examined.

Holotype (Fig. 4 a – d ): ♀; Hawaiian Islands, O‘ahu, Kahana Bay; 21.5573°N, 157.8781°W, 15 m; 27.viii.2021; ex Trema orientalis branches; D. Honsberger ( UHIM) .

Allotype (Figs 4 e – g , 9 ): ♂; Hawaiian Islands, O‘ahu, Mānoa Valley ; 21.3288°N, 157.7930°W, 154 m; 12.ii.2020; ex H. eruditus tunnel in T. orientalis branch; D. Honsberger ( UHIM) .

Paratypes: 18 ♀, 5 ♂. Hawaiian Islands, O‘ahu, Kahana Bay ; 21.5573°N, 157.8781°W, 15 m; 27.viii.2021; ex Trema orientalis branches; D. Honsberger (1 ♀, 1 ♂ BPBM) • Hawaiian Islands, O‘ahu, Mānoa ; 21.3009°N, 157.8196°W, 39 m; 20.v.2021; ex Delonix regia seed pod; D. Honsberger (1 ♀ BPBM) • same data as previous except 13.iv.2021 (1 ♀ UHIM; 1 ♀ BPBM; 1 ♀ CNC) • Hawaiian Islands, O‘ahu, Mānoa ; 21.3009°N, 157.8196°W, 39 m; 23.iv.2021; reared from Hypothenemus eruditus adult in Delonix regia seed pod; D. Honsberger (1 ♀ CNC) • Hawaiian Islands, O‘ahu, Waimānalo ; 21.3341°N, 157.7113°W, 28 m; 19.ii.2021; reared from Hypothenemus seriatus adult in Macadamia integrifolia husk; D. Honsberger (1 ♀ UHIM) • Hawaiian Islands, O‘ahu, Wahiawā ; 21.5151°N, 158.0423°W, 296 m; 11.i.2020; ex Hypothenemus eruditus tunnel in Spathodea campanulata branch; D. Honsberger (1 ♂ CNC) • Hawaiian Islands, O‘ahu, Wahiawā ; 21.5143°N, 158.0419°W, 301 m; 1.iii.2019; ex Spathodea campanulata branches; D. Honsberger (1 ♀ CNC) • Hawaiian Islands, O‘ahu, Pearl Harbor ; vi.1954 (1 ♀ BPBM) • Hawaiian Islands, O‘ahu, Waipi‘o ; ix.1957; light trap; J. W. Beardsley (1 ♂ BPBM) • Hawaiian Islands, O‘ahu, Waipi‘o ; ii.1960; light trap; J. W. Beardsley (1 ♀ BPBM) • Hawaiian Islands, O‘ahu, Pearl Harbor, West Loch , el. 3 ft; 13–24.vi.1998; yellow sticky board trap; W. D. Perreira (2 ♀ BPBM) • Hawaiian Islands, Moloka‘i, Kamalō Bridge , 3 ft.; 19.viii–2.ix.1994; yellow sticky board trap; W. D. Perreira (1 ♀ BPBM) • Hawaiian Islands, Moloka‘i, Kualapu‘u in coffee field , el. 750 ft.; 27.x–10.xi.1995; yellow sticky board trap; J. W. Beardsley and W. D. Perreira (1 ♀ BPBM) • Hawaiian Islands, Maui, Kahului Airport ; 4.x.1999; Malaise trap site # 1, wet spot nr. bike path, nr. water amongst kiawe & palm trees; F. G. Howarth, D. J. Preston, & J. Dockall (1 ♀ BPBM) • Hawaiian Islands, Maui, Kahului Airport ; 20°54'22"N, 156°25'42"W; 3–16.xii.1999; Malaise trap site # 2; F. G. Howarth, D. J. Preston, F. Starr, & K. Martz (1 ♂ BPBM) • Hawaiian Islands, Maui, Kahului Airport ; 20°54'22"N, 156°25'56"W; 1.ii.2000; Malaise trap site # 1; F. G. Howarth, D. J. Preston, J. E. Dockall, F. Starr, & K. Martz (1 ♂ BPBM) • Hawaiian Islands, Hawai‘i, Honomalino ; iv.1987; carob fruits; HY 87–14; G. Shaner (3 ♀ BPBM) • Hawaiian Islands, Hawai‘i, MacFarms ; 9.iii.1995 (2 ♀ BPBM) .

Etymology.

The species name is Hawaiian, ‘ umi‘ehu (lit., blonde mustache). When the head is viewed anteriorly (Fig. 5 c ), the clypeus appears as a blonde (‘ ehu) handlebar mustache (‘ umi‘umi) between its snout and mouth. This small, cryptoparasitic wasp also appears like a mist (‘ ehu) in the environment, often faintly perceptible and then evaporates from view. The name is to be treated as a noun in apposition.

Known distribution.

This species is known from the islands of O‘ahu, Moloka‘i, Maui, and Hawai‘i in the Hawaiian Islands, where it is likely adventive, and from the United Arab Emirates near Al Ajban, Emirate of Abu Dhabi ( Vargas 2017). It has long been present in the islands, with the earliest Hawai‘i specimen dating back to 1954. This suggests it may have arrived from the southwest Pacific during World War II or shortly afterward.

Known hosts.

Hypothenemus eruditus and Hypothenemus seriatus ( Coleoptera : Scolytinae); see Biology section.

Key to the known world species of Prorops

Note that males of P. rakan , P. mandibularis , P. “ sp. 23 ”, and P. “ sp. 24 ” are currently unknown.

Biology

Known hosts

Prorops maya has been found parasitizing H. eruditus adults in Trema orientalis (gunpowder tree) branches in Mānoa Valley at the foot of the Ko‘olau Mountains on O‘ahu island (21.3288°N, 157.7930°W, 154 m) (Fig. 6 ).

Prorops umiehu has been found parasitizing H. eruditus in T. orientalis branches and D. regia seed pods in Mānoa, O‘ahu island, and Hypothenemus seriatus in macadamia nut husks in Waimānalo, O‘ahu (Figs 6 , 7 ). It has been found emerging from Spathodea campanulata P. Beauv (African tulip) branches near Wahiawā, O‘ahu and Ceratonia siliqua L. (carob) pods in South Kona, Hawai‘i, but its development was not observed. While we cannot confirm the host relationship in these S. campanulata branches, we presume it also to be attacking H. eruditus because this was the only Scolytinae found to be present in the collections from which P. umiehu also emerged. Thus there seems to be overlap in host tree and host beetle use between these two species. Notably, P. umiehu has not been found parasitizing the related Hypothenemus spp. which commonly co-occur with H. eruditus in D. regia seed pods. Since H. eruditus uses many more trees than the three listed above as hosts in Hawai‘i, it seems reasonable to assume that both species are associated with more trees than the few listed here.

Life cycle

Development of P. maya and P. umiehu immature stages has been observed to occur on H. eruditus adult beetles in chambers and galleries the beetles excavate below the surface of the plant material (Figs 6 – 8 ). Both species follow a similar pattern. Eggs are laid on the adult beetle on the ventral side of the membranous region of articulation between the prothorax and mesothorax, and the emerging larvae feed on the beetle through the same location. As the developing larva feeds and grows, much of the larva remains outside the beetle and wraps around it, as if the beetle were wearing a necklace, and the anterior of the wasp's body extends further inside the beetle. When feeding by the wasp larva on its host has completed, the larva disconnects from the remains of the beetle and spins an off-white ovoid pupal cocoon in which it pupates.

The growth of the larva, positioned as it is, forces the two halves of the beetle apart. If development occurs in a tunnel, there is nowhere for this extra volume to go except to expand along the length of the tunnel. Thus between the growth of the wasp larva, the increased brittleness of the beetle after having been desiccated by the feeding of the larva, and the activity involved in creation of a pupation area by the wasp prepupa, the beetle splits apart and pupating wasps are often found in-between the two parts of the beetle, with the beetle's head and prothorax on one side of the pupating larva and the rest of the beetle on the other (Fig. 8 ).

Field collections have indicated that oviposition and larval development occurs exclusively on the adult stage of the beetle, and the laboratory tests subsequently described that present P. maya and P. umiehu adult females with a variety of life stages of H. eruditus have resulted in parasitism of only adult beetles. Larval development of laboratory reared P. umiehu from egg to adult eclosion is pictured in Fig. 9 .

Behavior observed in field collected plant material

When parasitized beetles were found in T. orientalis branches, nearly all the beetles in the gallery were either paralyzed or parasitized (see Fig. 6 b – d for examples). In such circumstances, parasitoids on the beetles were close to the same stage of development, and most of the beetles were dead or paralyzed but not obviously parasitized. This was observed for gallery systems containing developing larvae that yielded P. maya (n = 5) and P. umiehu (n = 1). Because the density of beetles in these galleries was in the range of that typically observed for surrounding, unaffected gallery systems, this suggests that P. maya often attacks nearly all beetles in a host patch. The proportion of adult individuals that were clearly parasitized in a gallery versus dead or paralyzed was recorded on four occasions, with 5 / 19, 5 / 6, 3 / 6, and 5 / 6 beetles parasitized. Adults of the two Prorops spp. were only sometimes found in the galleries with parasitized beetles, implying that adult females of these species do not necessarily remain with their young as do some other bethylids [see for example Sclerodermus harmandi ( Hu et al. 2012) , and Goniozus nephantidis ( Hardy and Blackburn 1991) ]. In the T. orientalis branches found to contain these Prorops spp. , there were often many distinct, unconnected beetle gallery systems in a branch. Wasps or paralyzed beetles were found in very few of these galleries, with the vast majority of galleries inhabited by healthy beetles. This implies that while P. maya tends to use nearly the whole host patch within a beetle gallery, most galleries were not utilized by these wasps, and the resulting overall percent parasitism and host mortality in this particular environment was low.

In D. regia pods, the pattern of patch use within a gallery system seems to be different. It seems to be much more sparse than in the T. orientalis branches, with only a small proportion of individuals in a gallery section either paralyzed or parasitized. This was observed only for P. umiehu (n> 15); P. maya has thus far only been found in T. orientalis branches. While this could be a result of differences in behavior between the two species, it appears more likely that this could be because the geometry of H. eruditus gallery systems tends to be different in these two plants. In T. orientalis branches, H. eruditus use only the thin phloem layer, and tend to construct a somewhat round chamber that extends in two dimensions under the bark, though this chamber eventually branches into a network of tunnels as the second and subsequent generations of beetles develop in the wood. Thus, at least in the earlier stages of beetle activity in the wood, their population tends to be somewhat localized in an uncomplex shape. In D. regia pods, the tissue the beetles use as a food source is thick enough relative to the size of the beetles to accommodate movement in three dimensions. Instead of forming a chamber, the beetle galleries take on a topologically more complex pattern, eventually creating a sponge-like network of tunnels spread through the material. The population of beetles inhabiting these tunnels tends to be more spread out within this maze of tunnels. Such variability in the gallery system created by H. eruditus among different host plants has been previously reported by Wood (1982) and Browne (1961). The geometry of the host's tunnels may contribute to this difference in patch use patterns, possibly due to the wasps' ability to locate their hosts within them.

Observation chambers in a laboratory setting

Observed behaviors were largely similar between P. maya and P. umiehu . While differences in behavior presumably exist, none of the general aspects of behavior and observations recorded here were distinct enough to be clearly associated with one species or the other. Typical observed behavior was as follows, and unless noted otherwise, the below observations apply similarly to both species.

General searching behavior (Video 1: https://vimeo.com/688211081, Video 2: https://vimeo.com/691136279): The wasps quickly moved through the tunnels in the wood, holding their antennae straight, vibrating and feathering them over the surfaces. This position of the antennae is in contrast to Allobethylus ewa (Bridwell, 1920) , another bethylid observed in separate studies using the same apparatus, which holds its antennae curved while antennating the surface of a wood substrate or beetles within it, which may function to increase the contact area of the antenna against surfaces with pits or other irregular texture (D. Honsberger, pers. obs.). The Prorops spp. showed clear interest upon finding a beetle, and when immature and adult beetles were present together, the wasps seemed to take preferential interest in the adults, and subsequent stinging, malaxation, chewing, and host feeding behaviors were initially focused on them.

Upon encountering an active H. eruditus adult beetle, a female wasp was observed to typically examine and antennate it, often climbing somewhat on top of the beetle in doing so. This was most often followed by an attempt at stinging the beetle, or more rarely, the wasp would either move away and explore elsewhere or repeatedly bite the beetle with its mandibles seemingly with the objective of attempting to move it. Stinging was typically followed by a quick exploration of the beetle and the area around it, and then often by a chewing behavior and host feeding, these actions covered in more detail below. The beetle was typically then abandoned and the wasp moved through the arena and was arrested by the presence of additional beetles on which it performed similar behaviors. Beetles that had previously been stung and paralyzed were often re-encountered by the wasps, who would examine them, occasionally sting them again, and often perform additional chewing or host feeding. Prorops umiehu , if taking interest in a larva or pupa, was observed to exhibit a similar progression of stinging and host feeding as with an adult.

The wasps were observed to adeptly turn themselves around in the tight space of a tunnel. Similar behavior has been observed in other parasitoids living in concealed tunnel environments using the same apparatus, such as Acerocephala hanuuanamu ( Honsberger et al. 2024) . As in that species, this maneuver seems to be made possible by the flat shape of the head and the long, articulating prothorax, presumably evolutionary adaptations to moving in tunnel environments. In this maneuver, the wasp ducks its head under its thorax, and follows it with the prothorax, the rest of the mesosoma, and then the flexible abdomen, smoothly sliding over its own body to switch the position of its head and metasoma (Fig. 10 e , Video 1: https://vimeo.com/688211081).

Stinging (Figs 10 a, b , 11 b, c , Video 1: https://vimeo.com/688211081, Video 2: https://vimeo.com/691136279): Initial exploration of an active adult beetle was most often followed by an attempt to sting the beetle, in which the wasp would climb fully on top of the beetle, grip the beetle's elytra or abdomen with its legs, and elongate and arch its metasoma around the beetle and search with its ovipositor for a location on the beetle susceptible to its sting. Crevices in the beetle's morphology encountered by the apex of the abdomen seemed to draw the focus of the exploration, though whether the wasp was able to contact an acceptable part of the beetle with its stinger often seemed more a matter of luck than of planning. A particular preference for aiming at the junction between the pro- and mesothorax was observed, especially evident for P. umiehu for which the number of observed stinging events was higher, but attempts were also observed to be made at stinging the ventral side of the abdomen, between the elytra, or the apex of the abdomen for both species. An attempt at stinging was either over quickly and often repeated more than once (though it was unclear if these attempts were successful), or in other cases the behavior persisted for a longer period of up to 5 minutes. Hypothenemus eruditus larvae and pupae were also similarly explored and occasionally stung by both species.

An adult or immature beetle that was successfully stung slowed down over the next few minutes before becoming more or less motionless, making only marginal twitchy movements. After a few days, such subtle twitching movements were still observed, including in beetles having been oviposited on. This implies that the chemicals injected by the wasp are paralytic and do not necessarily kill the beetle. This may maintain the integrity of the nutrition and water content of the beetle as its young develops on it ( Vinson and Iwantsch 1980), while preventing the host from dislodging the egg or larva through its movements ( Quicke 2015).

Chewing, oviposition, host feeding, and function of the snout (Figs 10 c, d , 11 e, f , 12 ; Video 1: https://vimeo.com/688211081, Video 2: https://vimeo.com/691136279, Video 3: https://vimeo.com/688588477): Stinging was often followed by the wasp climbing on top of the lateral or ventral side of the beetle and chewing on the membraneous region between the pro- and mesothorax. This behavior was either abandoned quickly or persisted for extended periods of time, typically in the vicinity of 5 minutes but occasionally as long as 15 minutes. While performing the chewing behavior, the wasp gripped the beetle's abdomen with its legs and pushed the prothorax of the beetle forward using its head, widening the separation between the pro- and mesothoracic sclerites and opening up the membranous region of articulation. The chewing was performed on the stretched membrane at a position as anteriorly advanced as the wasp was able to achieve. In this maneuver, the wasp was able to push on the prothorax by locking the groove formed between the projecting snout and mandibles with the edge of the prothoracic sclerite of the beetle, and in doing so was able to push the sclerite forward without it sliding over the wasp's face while maintaining use of its large mandibles for chewing. This chewing action functioned as the method of host feeding on adult beetles and was also performed in preparation for oviposition.

All eggs observed both in the laboratory observation chambers and field collected wood from both species were placed at the same location on the adult beetles: transversely oriented on the ventral side of the membraneous region of articulation between the pro- and mesothorax. Thus one use of the chewing behavior was apparently to prepare the host for oviposition, though its exact function was not clear: it could possibly be to cut the membrane so that the gap between the pro- and mesothorax would remain open and present an accessible area for oviposition and egg development; to cut open the tissue so that the emerging larva would be able to penetrate the membrane and feed; to cut the ventral nerve cord; for the adult to taste the beetle's hemolymph to assess its quality as a host; or alternatively, simply for the adult to host feed on the beetle. Adult Sclerodermus harmandi (Buysson, 1903) (Hym.: Bethylidae ) have been observed to chew holes through the cuticle of their host larvae through which their offspring feed as larvae ( Hu et al. 2012)

The act of oviposition (Video 3: https://vimeo.com/688588477) was only observed once, for P. umiehu . An H. eruditus adult previously stung, paralyzed, and likely chewed on as evidenced by the separation between the pro- and mesothorax greater than normal for a beetle that had simply been stung, was approached by the wasp and its actions were observed for the 2.5 hours leading up to oviposition. In this time, the wasp initially explored the beetle and the surrounding area, focusing much of its attention near the junction of the pro- and mesothorax, and attempted to subtly manipulate the beetle's position in the tunnel by grabbing the body of the beetle with its mandibles. The wasp then remained motionless in the tunnel, its body oriented opposite that of the beetle, the two touching head to head. This was followed by a short chewing interval, after which the wasp reassumed its position motionless in the tunnel with the beetle for approximately 1.5 hours. The wasp then resumed its exploration of the beetle, short chewing intervals, and subtle manipulations of the beetle's position, followed by an extended chewing event on the membrane along the midline of the beetle. It then explored the area and the beetle, repeating short but vigorous chewing intervals. The wasp then after a few attempts grabbed the beetle and repositioned it approximately a body length away in a slightly wider section of the tunnel. This was followed by one more vigorous chewing event, after which the wasp turned around and reached the apex of her metasoma into the gap between the pro- and mesothorax. The wasp remained in this position for about 10 minutes, her metasoma pulsating slightly. Then, over about 30 seconds, the wasp slowly moved the apex of her metasoma transversely across the crevice and the egg was visible emerging. The wasp slowly withdrew and became active again, exploring the vicinity for a few minutes and then left the area. The next day, the wasp was observed again in the tunnel in the same position as that in which it had remained motionless for extended intervals prior to oviposition, this time for at least 30 minutes as if host guarding, but did not maintain that position.

The chewing behavior was commonly observed on paralyzed adult beetles for both Prorops spp. , and in the vast majority of observed instances were not followed by oviposition. Many of these interactions seemed to be for the purpose of host feeding, but it was unclear if some also were for the purpose of oviposition but on a host that was eventually rejected. Chewing by both species was also observed to occur on immature stages that had previously been stung. Since no eggs or developing larvae were ever observed on an immature stage, chewing on pupae and larvae was presumably not for the purpose of oviposition, but instead for host feeding.

Intraspecific interactions were also observed. When encountering each other in open space, the wasps typically ignored each other. Females in the action of stinging or chewing were typically not interrupted by passing conspecifics, which might explore the beetle with their antennae. They were observed to be somewhat affected, though not to show any overt aggression, if both were exploring the same beetle adult or immature: if the wasps came in contact with each other, they would move slightly apart.

Defense against stinging by a Hypothenemus sp. (Fig. 13 , Video 4: https://vimeo.com/688212175): An unidentified Hypothenemus sp. male, not H. eruditus but similar in size to H. eruditus females, was placed in the observation chamber along with H. eruditus beetles. In a behavior not observed in H. eruditus beetles under the same circumstances, when a P. umiehu adult female attempted to sting the Hypothenemus sp. adult dorsally between the pro- and mesothorax, the beetle clamped down at this junction on the apex of the wasp's abdomen as soon as it made contact. The wasp did not seem to become agitated, but when appearing to attempt to withdraw after 1.5 minutes, a time in the range of a normal stinging interval, seemed to struggle and was unable to extract its metasoma from the grasp of the beetle until it relaxed 3.5 minutes later. The beetle seemed to be unaffected by the stinging attempt and remained active through the next day. It was unclear if the beetle was not paralyzed because P. umiehu is not physiologically able to paralyze this species, or was a result of this behavior which resulted in an unsuccessful stinging attempt. This behavior was only observed once. The wasps for the most part showed little interest in this species when they were placed in the arena and focused their attention on H. eruditus . Aside from this observation, neither this Hypothenemus sp. nor H. eruditus were observed to present other active forms of defense, such as running away or biting, against the wasps.

Patch use: Almost all beetles entered into the observation chambers with P. maya or P. umiehu females were motionless within 12 hours (n = 3 for P. maya , n = 3 for P. umiehu ), helping to confirm the observation in field collected wood that P. maya , and potentially P. umiehu as well, typically attacks more or less the whole host patch at least in geometrically simple environments such as those with which they were presented. In one instance where the details were more precisely recorded, five H. eruditus and the one unidentified Hypothenemus sp. male which showed the defensive behavior were inserted into the observation chamber in naturally bored H. eruditus galleries in D. regia pods with two P. umiehu adult females. All five H. eruditus were stung and paralyzed within two hours.

Interest in Hypothenemus hampei , the coffee berry borer?

Given that the congeneric Prorops nasuta is a parasitoid of the coffee berry borer H. hampei , and of the known species of Prorops it seems close at least morphologically to P. maya females, and H. eruditus and H. hampei are also congenerics, we investigated whether P. maya would take any interest in the coffee berry borer. To test this, wasps were released into the same apparatus described above that successfully resulted in stinging, chewing, host feeding, and parasitism of H. eruditus , with H. eruditus adults and immatures switched out with H. hampei adults and immatures (Video 5: https://vimeo.com/691136424). No interest was observed, and the wasps even seemed to show slight repulsion, upon encounter briefly antennating the beetles and then quickly moving on. No stinging, chewing, host feeding, or parasitism was observed. Given this lack of interest, no further tests were done.

Figures

Figure 1. Prorops umiehu sp. nov. head (paratype) showing morphometric measurements. Acronyms explained in text. Scale bar (upper left): 100 μm.

Figure 4. Prorops umiehu sp. nov. holotype ♀ (a – d) and allotype ♂ (e – g) a, e side view b, f head c, g anterior view of head d dorsal view. Scale bars: 500 μm (a, b, d, e); 250 μm (c, f, g).

Figure 5. Prorops umiehu sp. nov. a fore wing ♀ b hind wing ♀ c anterior view of head ♀ (paratype) d dorsal view ♀ (paratype) e antenna ♀ f proleg ♀ g mesoleg ♀ h metaleg ♀ i male genitalia, ventral view. Scale bars: 250 μm (c); 500 μm (d).

Figure 6. Hypothenemus eruditus beetles paralyzed and parasitized by P. maya or P. umiehu in T. orientalis branches. All photographs are of naturally occurring situations, taken while peeling bark from branches found in a forested region of Mānoa Valley on O‘ahu (21.3288 ° N, 157.7930 ° W, 154 m). Some beetles, such as each of the three in (a), are clearly parasitized with a P. maya or P. umiehu larva feeding through the membranous region of the beetle ventrally between the pro- and mesothorax, the posterior of the larva wrapped around the beetle. All larvae of P. maya and P. umiehu found in this study were observed to feed in this way. Prorops pupae, empty pupal cocoons, and more developing larvae are also visible in (b, c, d). Eggs of a thrips species that may scavenge on the remains can also be seen placed on top of beetles in (d). All beetles in these pictures were not moving, either paralyzed or killed presumably by P. maya or P. umiehu. Note that, in contrast to Fig. 8, these photographs are all of H. eruditus beetles in unconfined, wide chambers, and thus the development and construction of a pupal cocoon by wasp prepupae does not typically split the beetle into two pieces, and parasitoid pupae are located adjacent to or near the host beetle. It is unknown which of the developing parasitoids in these photographs correspond to which species of Prorops, as immatures collected from these branches yielded both P. maya and P. umiehu.

Figure 7. Prorops umiehu developing immatures. Photographs are of naturally occuring situations, taken while dissecting plant material collected from field environments a P. umiehu larva on H. eruditus from D. regia pods collected from the campus of UH Mānoa b pupating P. umiehu having completed its larval stage on H. eruditus in D. regia pods from UH Mānoa c, d P. umiehu larvae on H. eruditus from D. regia from UH Mānoa e macadamia husk from Waimānalo, O‘ahu with its inner layer peeled, showing pupating P. umiehu having developed on the H. seriatus adults next to them.

Figure 8. Typical placement of pupae when development occurs in a narrow tunnel environment. The parasitoid larva feeds on the adult beetle through the ventral membranous region between the pro- and the mesothorax. The growth of the larva and its subsequent construction of a pupal chamber forces these two sections of the beetle apart, and pupating larvae are typically found concealed between them. In these pictures, the white material near or around the parasitoid immature is what remains of the pupal cocoon it had constructed after the tunnel was broken open by peeling apart the plant material. Photographs are of naturally occurring situations, taken of plant material collected outdoors a, b Prorops sp. prepupa (a) and pupa (b) between two halves of an H. eruditus beetle in an H. eruditus tunnel in a T. orientalis branch collected from Mānoa Valley on O‘ahu; c: Prorops umiehu pupa in a D. regia pod from a tree on the campus of the University of Hawai‘i at Mānoa. Note the position of the pupal cocoon relative to the beetle in (c), not the larva itself which was moved as the pod was peeled apart. Such placement was more common in T. orientalis branches where scolytid tunnel systems were often more linear, but atypical in D. regia pods possibly due to the less linear and more confused organization of beetle feeding in these pods, but in (c) developed in this way due to the topography in the certain section where the beetle was parasitized.

Figure 9. Development of P. umiehu on H. eruditus. The beetle was parasitized by a P. umiehu adult female entered into the observation chamber described in the text containing H. eruditus beetles on a piece of D. regia pod with channels cut into it with a knife. The parasitized beetle was then moved into a small hole carved into a wood substrate and covered with a piece of glass slide cover to mediate humidity and to create an enclosed environment to facilitate construction of a pupal cocoon by the prepupa. Time elapsed after presence of an egg was first observed on the beetle are noted in the photographs. The emerged adult is the allotype.

Figure 10. Aspects of the behavior of P. maya in laboratory observation chambers. The apparatus is the same as in Fig. 11 a except using bark peeled from T. orientalis branches with channels cut into it with a knife. Beetles are all H. eruditus. a, b stinging c chewing behavior on adult beetle, it was unclear if (c) was host feeding or the eventually abandoned preparation of an oviposition site or both d host feeding on a larva e turning around, reversing the orientation of its body in the tight space of a tunnel. These actions are also shown in Video 1: https://vimeo.com/688211081.

Figure 11. Stinging and chewing behavior of P. umiehu in laboratory observation chambers a the observation chamber used in this study with a piece of D. regia pod tissue having been naturally infested by H. eruditus sandwiched between the aluminum and plexiglass b, c P. umiehu stinging H. eruditus adult females d profile view of a P. umiehu female next to H. eruditus, showing the groove formed between the snout dorsally and the mandibles ventrally which acts as a mechanism for holding the edge of the prothoracic sclerite of its host e the wasp grasps the abdomen of the beetle with its legs and pushes forward on the sclerite with this structure to expose the membranous region between the pro- and mesothorax of the beetle, while maintaining use of the mandibles for chewing on the stretched membrane. This was observed, as in (e), for the purposes of host feeding on adults having previously been stung and paralyzed, and was also observed during preparation for oviposition (shown in Fig. 12) f host feeding on an H. eruditus pupa previously stung and paralyzed. These actions are also shown in Video 2: https://vimeo.com/691136279.

Figure 12. Oviposition by P. umiehu on H. eruditus in a laboratory observation chamber. The apparatus in which this occurred is the same as that pictured in Fig. 11 a a H. eruditus beetle approximately 30 minutes subsequent to oviposition, showing the newly laid egg between the beetle’s pro-and mesothorax b, c use of the groove between the snout and mandibles in P. umiehu to push forward on the prothoracic sclerite of the beetle while chewing on the exposed and tensioned membrane, in preparation for oviposition d the wasp then placed the apex of its metasoma between the pro- and mesothorax, into the crevice that remained open as a result of the chewing as in (b) and (c), and maintained that position for approximately 10 minutes e the wasp slowly draws her metasoma across the crevice while the egg exits f the host, wasp, and egg subsequent to completion of oviposition. These actions are also shown in Video 3: https://vimeo.com/688588477.

Figure 13. Defense by Hypothenemus sp. male against a stinging attempt by P. umiehu. The identity of the beetle is unknown, but is not H. eruditus a position immediately prior to contact of stinger. Note that the beetle is in a relaxed state, with the gap between the pro- and mesothorax slightly open b The beetle quickly clamps down after contact, trapping the apex of the metasoma of P. umiehu, and maintains that position for approximately 3.5 minutes, during which time the wasp was not visibly agitated but was unable to extricate itself. The beetle appeared to remain healthy and active in the subsequent hours, seemingly unaffected by the stinging attempt. These actions are also shown in Video 4: https://vimeo.com/688212175.