Ripiphorus caboverdianus, Batelka, Jan & Straka, Jakub, 2011

Batelka, Jan & Straka, Jakub, 2011, Ripiphorus caboverdianus sp. nov. — the first ripiphorid record from the Macaronesian volcanic islands (Coleoptera: Ripiphoridae: Ripiphorinae), Zootaxa 2792, pp. 51-62 : 52-60

Ripiphorus caboverdianus sp. nov.

( Figs. 3–12 )

Type specimens. Holotype (1 ɗ): dried specimen mounted on a white label, ( NMPC) labeled: CAPE VERDE Isl. / BOAVISTA—rock N of / Sal Rei, 20.X.2009 / J. Batelka & J. Straka lgt. [printed label]”. Paratypes: 13 ɗɗ, 19 ΨΨ [partly dry-mounted, partly preserved in 96% ethyl alcohol], ( NMPC, JBCP, JSCP), same data as the holotype.

Diagnosis. Small species (4.0–5.0 mm), characterized by milky white elytra (partly translucent in most specimens) and hyaline hind wings. Abdomen with translucent membranous cuticle between the dorsal and ventral surface of ventrites 1–2. Female hind tibia slightly curved, 2× wider at the apex than at its base, metatarsomere 1 slender, parallel-sided, 5× as long as wide.

Description. Head short, compressed anterioposteriorly, vertex raised to a distinct blunt protuberance, frons flat, without median excavation, labrum broad, shallowly emarginate. Eyes placed laterally, entire, without incision. Antennae inserted above eyes, weakly sclerotized, light brown with fuscous tips. Maxillary palpi light brown, last segment slightly longer than preceding one, parallel, not enlarged at apex. Body surface and all extremities densely covered by short whitish setae. Elytra abbreviated, scale-like, rounded at apex. Hind wings longer than abdomen, unfolded. Sexual dimorphism developed as follows:

Male. Body black, legs light brown. Abdomen parallel, black. Pygidium rounded. Antennae 11-segmented, antennomeres 1 and 2 poorly visible, shortened and almost hidden in antennal cavity, antennomeres 3–10 strongly biflabellate, antennomere 11 uniflabellate. Hind wings always hyaline, PC+C+ScA+ScP+RA and radial cell almost without pigmentation. Hind tibia narrow, 3.0× as wide at apex as at base, inner edge of end of hind tibia narrow, without emargination. Metatarsomere 1 slender, parallel-sided. Tarsal claws serrate, teeth thin and hardly distinguishable. Aedeagus long and slender, simple, without apical hooks, slightly curved apically. Parameres concrescented, expanded on one side and emarginated at apex.

Female. Body black, metasternum in some specimens orange, legs black. Abdomen bulky, widest in its posterior third (character not visible in dried specimens), mostly orange, sometimes with black stripes or spots dorsally and laterally. Pygidium protruded, longer than broad to support ovipositor. Antennae 10-segmented, antennomeres 1 and 2 poorly visible, shortened and almost hidden in antennal cavity, antennomeres 3–10 uniflabellate, light brown, occasionally fuscous at tip. Hind wings hyaline, with lightly fuscous PC+C+ScA+ScP+RA and radial cell.

Hind tibia 6.8× as long as wide at apex, length ratio of tarsus (claws excluded) to tibia equal to 4:5. Tarsal claws with six shorter inner teeth and one distinct long apical tooth each.

Etymology. The species-name is patronymic and refers to the Cape Verde archipelago.

Relationships. R. caboverdianus sp. nov. is similar to R. arabiafelix Batelka, 2009 from Yemen and Oman in having milky white elytra but differs as follows. In the female of R. caboverdianus sp. nov., the hind tibia is 2× as wide at the apex as at the base and metatarsomere 1 is 5× as long as wide; in the female of R. arabiafelix , the hind tibia is 3× wider at the apex than at the base and metatarsomere 1 is 4× as long as wide. The teeth on the tarsal claws are less numerous in the female of R. caboverdianus sp. nov. Ripiphorus arabiafelix further differs from the newly described species by having PC+C+ScA+ScP+RA and the radial cell dark fuscous in both sexes. The Afrotropical species of Ripiphorus were reviewed in Batelka (2009a). Ripiphorus caboverdianus sp. nov. and R. arabiafelix are the only described Old World species with white elytra (apomorphy); in addition, at least one undescribed species from southern Africa has similarly whitish elytra (J. Batelka, unpublished data). All three taxa are probably not closely related, and they elytral coloration need not to be a synapomorphy. More likely it represents a homoplastic adaptive response to arid conditions or correlate with the color of preferred flowers.

Collecting conditions. Beetles and bees were sampled by sweeping and individual collecting on Aerva javanica [hereafter referred to as Aerva ] at the southern slope of a small rock formation in the northern part of Sal Rei town (16°11’N, 22°54’W; Figs. 1–2 ). The area was about 100× 50 m in size, cca 20–40 m a. s. l. The vegetation had a steppe character and included solitary bushes of the invasive species Prosopis juliflora (Sw.) DC. Dominant herbs other than Aerva included one species of the Brassicaceae , several species of grasses and one species of the Nyctaginaceae (probably Boerhavia or Commicarpus ). The weather was sunny, almost calm, with sporadic clouds. Heavy rain fell in the evening on 18 October 2009 in Sal Rei, and heavy rains occurred all over the archipelago in the second half of September.

Behavior of Ripiphorus imagines. Ripiphorus females were observed ovipositing into apical parts of Aerva inflorescences among flower-buds that had not yet come into flower. We assume that the free-living first instar larvae hatch when these buds come into flower and are visited by females of the host bee(s). Ripiphorus females were very agile when searching for suitable places for oviposition on the blossoms, but remained stationary for several tens of seconds when ovipositing. They extend the ovipositor underneath towards the head; their wings remain always unfolded ( Figs. 3–8 ).

Ripiphorus males were observed flying around flowering Aerva plants only when they were visited by ovipositing females. As no beetles in copula were observed, the female had probably been fertilized earlier and the observed behavior of the males probably was of little consequence for their fitness. Males were not observed to visit flowers for more than a few seconds, or to feed elsewhere.

Supposed host. Together with the ripiphorid, 46 specimens of one species of the genus Halictus Latreille, 1804 (subgenus Seladonia Robertson, 1918 ) were captured on the flowers of Aerva . We identified this probable host as Halictus lucidipennis Smith, 1853 . Báez et al. (2005) recorded two Halictus species, H. varipes Morawitz, 1876 and H. lucidipennis , from the CVA. While Ebmer (1988) considered both species closely related and possibly subspecies of one taxon, the two species names were already synonymized by Sakagami & Ebmer (1987). The synonymy is still accepted ( Pauly 2008) and the nomenclature given by Báez et al. (2005) is thus outdated. Halictus lucidipennis is widely distributed from the CVA through Africa and Asia to Gobi and Thailand ( Ebmer 1988, Sakagami & Ebmer 1987). We also collected H. lucidipennis on the islands of Fogo and Santiago in the CVA. According to Báez et al. (2005) the species should occur on all large islands of the archipelago except Maio and Brava (probably a sampling artefact).

No other Apoidea were collected on the flowers of Aerva at the locality, and only a few specimens of other bees ( Amegilla , Ceylalictus , Megachile and Thyreus ) were captured there. None of them could be a possible host of Ripiphorus caboverdianus sp. nov. Amegila, Ceylalictus and Megachile should be excluded because of their inappropriate size and Thyreus because of its parasitic development (probably on Amegila). However, Megachile probably visits Aerva because the only specimen we caught carried one first instar larva of Ripiphorus (see Note). No nesting sites of Halictus were found nearby.

Name of the visited plant. Sánchez-Pínto et al. (2005) reported from the CVA Aerva persica (Burm.f.) Merr [= A. javanica Juss. , = A. tomentosa Forssk. ]. The same name, A. persica (Burn. [sic!]) Merill. [sic!], was also used for the species widespread in the Central Sahara by Ozenda (1977). However, Jongbloed (2003) reported this plant from the mountainous region of the U.A.E. and Aluka (2008) recorded it from Mauritania across Africa into Arabia, India and Madagascar under the name A. javanica (Burm.f.) Juss. ex Schult [= A. tomentosa ]. For the purpose of this paper we follow Jongbloed (2003) and Aluka (2008) without trying to delve into this nomenclatorical and taxonomical problem.

Description of the free-living first instar larva ( Fig. 12 ).

Length 0.5–0.7 mm, width 0.15–0.17 mm.

Body elongate, subcylindrical, fusiform, widest through metathorax, abdomen attenuated apically; color yellowish brown, mandibles and areas bearing stemmata darker.

Head subtriangular, distinctly longer than wide (length:width ratio equal to 7:6), rounded in front, without median longitudinal dorsal line; posterior margin with strong, transverse, slightly emarginate thickening, terminating laterally below dorsal ocelli, collum distinct. Surface with mesh of minute reticulations.

Epicranial and frontal sutures lacking.

Stemmata positioned laterally on the head, behind antennal foramina; two stemmata turned dorsally, three latero-ventrally, the other laterally; all subequal in size.

Antennae slender, three-segmented, including terminal seta slightly shorter than length of head. First segment short, one third as long as second segment, wider than second segment, directed laterally at a right angle to plane of body. Second segment asymmetrical in dorsal view, with anterior margin evenly but slightly curved and posterior margin slightly curved but obliquely narrowed approximately over apical half; posterior surface evenly expanded from base to apex, with circular sensory pit at apical two-fifths and one microseta midway between sensory pit and apex; apex ventrally with elongate and slender sensory appendix slithly more than 0.4× as long as second segment. Third segment slender, nearly straight, almost 1.5× times as long as second one. Terminal seta about 1.5× as long as third segment.

Antennal foramina large, well defined, placed laterally, contiguous with distinct, weakly sclerotized subgenal line extending from below ocelli to mandibular foramen.

Clypeus and labrum fused into nasale.

Mandibles moving inward in antero-ventral plane, not extending beyond margin of head when open, apices overlapping when closed. Base greatly enlarged for muscle attachment, apex slender, falciform, strongly incurved, inner margin without teeth or tooth-like ridges. Condyle large, medially placed.

Maxillae long, slender, weakly sclerotized, posteriorly attenuated, separated by about three times their own width. Surface bearing a conspicuous long seta at anterior one-third. Cardo not visible, probably reduced or fused with stipes. Palpi three segmented, directed laterally, four-fifths as long as second and third antennal segments together; first two segments asymmetrical in ventral view, second segment slightly longer than third, together about three-fourths as long as third segment; second segment with a ventral apical seta; third segment with a large oblique sensory pore on anterior surface slightly before middle.

Mentum and submentum very weakly sclerotized.

Labium indistinctly separated from mentum, simple, weakly sclerotized. Palpi vestigial, apparently represented by two unconfined papillae placed at line connecting posterior margins of insertions of maxillary palpi.

Thorax about 0.3× as long as body, about twice as long as head. Nota well sclerotized, surface with numerous sensory pits and two pairs of setae, nota of meso- and metathorax incompletely fused with pleura, mesothoracic spiracle on line of partial fusion. Each sternum consisting of one well-developed plate bearing well-developed seta on each side near lateral margin; setae placed anteriorly on prosternum and at middle of lateral margin on meso- and metasternum.

Legs of all pairs subequal in length. Coxae broader than long, hind coxae widely separated, their distance almost twice the length of coxa, posteriolateral angle of all coxae distinctly concave dorsally for partial reception of base of femora; dorsal surface of all coxae with one microseta and two sensory pits, ventral surface with setae and several sensory pits; all setae on anterior coxae more strongly developed but shorter than those on middle and hind coxae. Trochanters subtriangular with numerous sensillae and long seta, slightly more than half as long as femora. Femora slightly longer than and more than twice as wide as tibio-tarsi, increasing in length but decreasing in width from fore to hind femora. Surface of each femur with a series of very fine and feebly elevated lines, anterior margin with apical seta, posterior margin with four setae, ventral surface with longitudinal row of three setae and one microseta near anterior subbasal margin. Tibio-tarsi with three setae on anterior margin, two setae on posterior margin, and one sensilla beyond middle on outer face; claws present, single, slightly curved, about as long as width of apex of tibio-tarsus, at least partly surrounded at base by integumental thickening; pretarsus highly developed, obovate, flattened, membranous with fine and short longitudinal thickenings.

Abdomen cylindrical, gradually and evenly attenuated from base to apex, segment 9 about one-third as wide as segment 1, segments 1-9 wider than long; segment 10 nearly two times as long as wide with all sclerites fused into a cone; spiracles of segment 8 subequal to those of mesothorax, remaining segments without spiracles; sternites 1- 9 entire, each with anterior pair of microsetae, sternites 1-7 with posterior row of six macrosetae, sternites 8 and 9 with four apicolateral macrosetae, those of tergite/sternite 8 with the outer pair twice as long as inner pair. Apicolateral macrosetae of segment 9 about as long as segment 10.

Differential diagnosis. The larva is similar to the free-living first instar larva of Ripiphorus smithi Linsley et McSwain, 1950 , the only other Ripiphorus first instar larva described in detail to date. However, the body of the new species is smaller, the head is slightly longer, the terminal seta of the antenna shorter (only about 1.5× as long as antennal segment 3) and the apicolateral macrosetae are 1.5× as long as in R. smithi . First instar larva descriptions of R. subdipterus Bosc, 1792 ( Chobaut 1906) and R. solidaginis (Pierce, 1902) ( Pierce 1904) do not allow precise comparison.

Note. First instar larvae were observed individually on flowers of Aerva and were found attached to the bodies of adult Ripiphorus ( Figs. 5, 7 and 8 ), Halictus bees, and in one case to the carpenter bee Megachile sp. ( Table 1 ). They moved very fast and changed their position on the inflorescence, probably in response to vibrations caused by Halictus bees and females of Ripiphorus . They were attached by the mandibles, mostly on the ventral side of the vector’s body and usually in the membranous parts between the head and the thorax or between the thorax and the abdomen, rarely on legs (one individual) and wings (one individual) (cf. Linsley et al. 1952: 300 and Batra 1965: 377). Usually one larva was found on a single vector; two larvae were found together in four cases. Linsley et al. (1952: 300) reported up to 11 larvae and Batra (1965: 377) up to four larvae together on a single vector. Although Ripiphorus adults were not reported as vectors of first instar larvae by Linsley et al. (1952) or Batra (1965), Pierce (1904: 25) found some larvae attached to adults and speculated about the role of Ripiphorus adults as a principal vector. However, a high number of larvae on the females of Ripiphorus caboverdianus sp. nov. is more likely a side effect of their permanent presence on the flowers. The number and specificity of the vectors apparently depends on local conditions. Moreover, attached larvae might be easily overlooked given their tiny size. Collecting and transportation of specimens (see Material and methods) also negatively affected the preservation of larvae on their vectors ( Table 1 ).

Host associations of Ripiphorus . About 70 species of the genus Ripiphorus have been described so far ( Batelka 2009a). Hosts of Ripiphorus have been reported for several, mostly North American species ( Table 2 ). They belong to four genera of the family Halictidae ( Augochlora , Dieunomia , Halictus and Lasioglossum ) and to one genus of the family Apidae (Diadasia) . Supposed host association with Halictus observed on Boavista thus fits well the known host preferences.

Presence of Ripiphorus in the CVA. The Macaronesian beetle fauna has been catalogued in the Azores ( Borges et al. 2005), Madeira and related islands ( Borges et al. 2008), the Salvagens Islands ( Erber & Wheater 1987), the Canary Islands ( Machado & Oromí 2000) and the CVA (Oromí et al. 2005). No Ripiphoridae have been reported from these islands and archipelagoes so far.

No evidence, even indirect, allows us to consider Ripiphorus caboverdianus sp. nov. endemic to the CVA. While old studies presumed that the CVA are of an Early Cretaceous volcanic origin ( Assunçao et al. 1968) and supposed the existence of a long-lived hot spot for some 120 Mya ( Morgan 1981 fide Bernoulli et al. 2007), current research shows that the CVA arose probably in the Early or Middle Miocene (about 20–18 Mya) due to no signs of pre-Miocene subaerial volcanic activity ( Bernoulli et al. 2007). The eastern islands of Boavista, Maio and Sal belong to the oldest ones within the archipelago. All three islands probably had a much larger area in the recent geological past, and Boavista and Maio might have been connected during sea level fluctuation in the Pliocene ( Jesus et al. 2002). However, the fauna and flora of the archipelago is characteristic by extremely low proportion of endemic taxa in comparison with the similarly old Canary Islands. Only 16 endemic genera of insects occur in the CVA ( Arechavaleta et al. 2005), of which only six belong to the Coleoptera (Oromí et al. 2005) . In comparison, the Canary Islands (maximum age of 22 Mya) harbor at least 42 endemic beetle genera ( Machado & Oromí 2000). To put these numbers in further perspective, the Galápagos Islands (maximum age of 4 Mya) harbor only seven endemic genera (Peck 2006). This may indicate that the biota of the CVA is younger than the islands. Brochmann et al. (1997) supposed that the vascular flora in the CVA is not older than a few hundred thousand years because of the absence of palaeoendemic taxa and a low level of endemism, with only one endemic angiosperm genus present ( Arechavaleta et al. 2005). This is much smaller number than the 19 endemic genera in the Canary Islands ( Stuessy & Crawford 1998). The flora and fauna of the CVA might have been affected by mass extinction(s) and biodiversity bottlenecks in the past, for example after long periods of drought. Ripiphorus caboverdianus sp. nov. may thus represent recent colonization by a species occurring somewhere on the coast of the West Africa, especially given that both the visited plant and the host bee are widespread in Africa and Asia.

R. fasciatus (Say) Lasioglossum (Dialictus) pruinosum Robert- Melander & Brues (1903) USA

son

Conservation remarks. The type locality of R. caboverdianus sp. nov. is under strong pressure for further development: an apartment resort has been built at the western edge ( Fig. 1 ), a new hotel resort is under construction at the south-east edge ( Fig. 2 ), and a large stone quarry has been opened at the eastern edge of the rock. Moreover, the locality will soon be overgrown by the invasive bush Prosopis juliflora , as many young shoots were already observed in open areas at the locality. This will probably exterminate Aerva together with the host bee. If conservation of localities with a higher plant or insect diversity is prioritized locally over species-poor habitats similar to the type locality of R. caboverdianus sp. nov., this species might be endangered or exterminated. Examination of Halictus collected at several other CVA localities and habitats (Boavista, Sal Rei, sand dunes; Fogo, Sao Felipe, Acacia steppe; and Santiago, Tarafal , edge of secondary forest) did not reveal any Ripiphorus first instar larvae. This may indicate possible specific requirements of R. caboverdianus sp. nov. on the abundance and simultaneous occurrence of the visited plant and the host bee when a higher vector diversity at the locality may decrease the rate of infestation by first instar larvae.

Dispersal abilities of the Ripiphoridae . Volcanic oceanic islands without previous connection to mainland are ideal to study long-range dispersal. Natural colonization of oceanic islands by insects falls into the following four categories: (1) passive aerial dispersal by wind currents, (2) rafting on organic matter, (3) phoresy on the surface of rafting or flying animals and (4) active aerial dispersal of winged species. Almost no data are available on the dispersal abilities of the Ripiphoridae , and their biogeography is inadequately known. Only the following records have been published from volcanic oceanic islands so far:

(A) Ripidius pectinicornis Thunberg, 1806 (Ripidiinae) is reported from Oahu isl. (Hawaii) ( Falin 2001). Most specimens came at light around the Honolulu International Airport. The species parasitizes the cockroach Blattella germanica , which is not native to Hawaii and is spreading worldwide by ships and cargo ( Falin 2001). Ripidius pectinicornis has undoubtedly been introduced to Hawaii by recent trade activities.

(B) Two ripiphorid genera and species have been reported from Santa Cruz isl. (Galápagos): Trigonodera lineata (Champion, 1891) (Pelecotominae) and Ancholaemus acuminatus Fairmaire, 1904 (Micholaeminae) ( Causton et al. 2006, Peck 2006). The occurrence of both continental species in only one island in the archipelago and lack of records before 1964 support the opinion that both species were introduced to the island in timber imported from South America for the building of the Darwin Station ( Causton et al. 2006, Peck 2006).

(C) Schilder (1923) described the subspecies Macrosiagon benschi insularum Schilder, 1923 (Ripiphorinae) from Anjouan isl. ( Comoros), which lies some 400 km west of northern Madagascar. One of us (JB) examined the type specimen of this subspecies and found it similar to the Madagascan nominate subspecies Macrosiagon benschi benschi Alluaud, 1902 . Moreover, there are another two closely related taxa described and known only from Madagascar: Macrosiagon sodalis (Waterhouse, 1883) , and Macrosiagon benschi var. [sic!] s eyrigi Pic, 1952. Examination of the types and additional specimens of these taxa revealed that they form a unique species-group (possibly consisting of only one species) characterized by a flattened spiny process at the apex of antennomere 1. This apomorphic character does not occur in any other species of the genus Macrosiagon Hentz, 1830 and the oceanic dispersal of the ancestor of M. benschi insularum from Madagascar to the Comoros (the other direction is less probable) seems well justified (J. Batelka, unpublished data). The host (vector) of this Macrosiagon species group is unknown. Members of the genus Macrosiagon parasitize larvae of wasps and bees (families Apidae , Crabronidae , Halictidae , Pompilidae , Scoliidae , Sphecidae , Tiphiidae and Vespidae (Batelka & Hoehn 2007)) , which are strong fliers with good dispersal abilities. Moreover, Macrosiagon females lay eggs in clusters, even hundreds of eggs per female ( Jarvis 1922, Clausen 1940), which increases the chance that the infested bee or wasp will carry several first instar larvae ( Jarvis 1922). Phoresy of the free-living first instar larvae in Macrosiagon is similar to Ripiphorus ( Jarvis 1922, Clausen 1940).

Figures

FIGURES 3 – 8. Females of Ripiphorus caboverdianus sp. nov. lying eggs into the tips of inflorescences of Aerva javanica. Arrows in Figs. 5, 7 and 8 indicate freely moving first instar larva. Not to scale.

FIGURES 9 – 12. 9 — Female of Ripiphorus caboverdianus sp. nov., dorsal view, 4.0 mm; 10 — Female hind tibia and tarsomeres, 2.5 mm; 11 — Male (holotype), dorsal view, 4.5 mm; 12 a—first instar larva, dorsal view, 0.66 mm, 12 b—first instar larva, ventral view, 0.6 mm.

FIGURES 1 – 2. Type locality of Ripiphorus caboverdianus sp. nov. 1 — view towards west; 2 — view towards south-east (town of Sal Rei with a lagoon and newly built hotel complex close to the rock).

Tables

TABLE 1. Numbers of the Ripiphorus first instar larvae attached to Ripiphorus adults and to bees. * = larvae were found freely in a vial with 10 Ripiphorus and 25 Halictus specimens preserved in ethyl alcohol.

Vector R. caboverdianus sp. nov. (female) R. caboverdianus sp. nov. (male) H. lucidipennis (female) H. lucidipennis (male)	One larva 4 2 2 -	Two larvae 4 - - 1
Megachile sp. unknown*	1 16	- -
Discussion

TABLE 2. A review of host records for Ripiphorus.

Ripiphorus species	Host bee	Primary reference(s) with the potential association	Country of record
R. schwarzi (LeConte)	Augochlora (Augochlora) pura (Say)	LeConte (1880)	USA
R.? stylopides (Newman)	Augochlora (Augochlora) pura (Say)	Barber (1939)	USA
Ripiphorus sp.	Augochlora (Oxystoglossella) nominata Michener	Eickwort & Eickwort (1972)	Costa Rica
R. solidaginis Pierce	Dieunomia (Epinomia) triangulifera Vachal	Pierce (1904), Wcislo et al. (1994)	USA
R. epinomiae Linsley et McSwain	Dieunomia (Epinomia) nevadensis (Cresson)	Linsley & McSwain (1950)	USA
R. nomiae Rivnay	Dieunomia (Epinomia) nevadensis (Cresson)	Linsley & McSwain (1950); as Nomia pattoni Cresson	USA
Ripiphorus sp.	Dieunomia (Epinomia) nevadensis (Cresson)	LeConte (1880)	USA