The first precinctive Carabidae from Moorea, Society Islands: new Mecyclothorax spp. (Coleoptera) from the summit of Mont Tohiea

Abstract Seven species of Mecyclothorax Sharp from Moorea, Society Islands are newly described; Mecyclothorax perraulti sp. n., Mecyclothorax pahere sp. n., Mecyclothorax menemene sp. n., Mecyclothorax mahatahi sp. n., Mecyclothorax popotioaoa sp. n., Mecyclothorax mapo sp. n., and Mecyclothorax fatata sp. n. These constitute the first Mecyclothorax species described from Moorea, and the first carabid beetle species shown to be geographically restricted to that island. Each of the newly described species is most similar to a different species on the island of Tahiti, suggesting that none of the seven Moorean taxa are evolutionary end-products of autochthonous speciation within Moorea. The occurrence of precinctive Mecyclothorax species on both Moorea and Tahiti demonstrates that radiation of Mecyclothorax in the Society Islands has been facilitated by speciation events implicating both islands. Whether this speciation has been preceded by vicariance or dispersal is discussed, with the generality of a dispersal hypothesis tested using information from Society Island Nabidae (Hemiptera). Salient morphological characters for taxa in the Society and Hawaiian Islands are compared to those representing a broad survey of southwest Pacific Mecyclothorax spp. This comparison supports the independent founding of each radiation in the Societies and Hawaii from an Australian ancestral propagule, likely drawn from the ecologically general, geographically widespread Mecyclothorax punctipennis (Macleay).


Introduction
The genus Mecyclothorax Sharp is distributed throughout Australia and associated landmasses and islands including New Guinea, the Greater Sundas of Java and Borneo, Lord Howe and Norfolk Islands, and St. Paul and Amsterdam Islands of the Indian Ocean (Baehr 1998;Baehr and Lorenz 1999;Moore 1985Moore , 1992. Mecyclothorax has also diversified in New Caledonia (Jeannel 1943;Deuve 1987) and New Zealand (Liebherr and Marris 2009). But it is on two Polynesian archipelagoes that Mecyclothorax has undergone radiations that are so rich in species that these radiations dwarf the levels of diversity observed over all other areas of the generic distribution. The Hawaiian Islands house a Mecyclothorax fauna that includes 166 valid species (Liebherr 2006(Liebherr , 2007(Liebherr , 2008(Liebherr , 2009a(Liebherr , 2009b(Liebherr , 2011aLiebherr and Krushelnycky 2011), with an estimated 73 additional species in the process of description from Haleakala volcano, Maui Ialand (unpubl. data). Mecyclothorax beetles are distributed on the Hawaiian islands of Oahu, Molokai, Lanai, Maui and Hawaii Island, but not Kauai. The Society archipelago, specifically Tahiti, also supports an impressively diverse radiation, with 67 species recognized from the island of Tahiti (Perrault 1992). Why have these two Polynesian archipelagoes been home to such diverse Mecyclothorax radiations? Based on shared attributes of Tahiti and the Hawaiian Islands, Mecyclothorax have thrived in these places in association with subtropical montane rain forest that is dissected by lava flows or by low-elevation erosionally formed valleys, with the islands' orographic relief resulting in extensively subdivided habitats ranging from 1000-3000 m elevation. In Tahiti Perrault (1992) discovered that different, neighboring mountain ridges emanating from the central peak Mont Orohena mostly supported distinct species. Though some species are shared between adjacent ridges-e.g. the ridges culminating at Mont Marau and Mont Aorai-the majority of species are not so shared between different ridges (Perrault 1992: 211). These montane habitats receive anywhere from 4000-6000 mm of precipitation per year in Hawaii (Giambelluca and Schroeder 1998) and up to 8000 mm/yr in Tahiti (Mueller-Dombois and Fosberg 1998;Craig et al. 2001). That Mecyclothorax speciation is facilitated by the geological subdivision of wet to mesic montane forest habitats is strongly supported by the geographic restriction of the vast majority of species in both radiations to rain forest habitats. The high levels of diversity in these subtropical islands contrasts sharply with the low Mecyclothorax diversity resident in Australia and New Zealand (Moore et al. 1987;Liebherr and Marris 2009), where species distributions are centered on open habitats including grasslands, moorlands, riparian corridors, and dry to mesic Eucalyptus forest.
This paper extends the comparison of the Hawaiian and Tahitian Mecyclothorax radiations by describing the first collections of Mecyclothorax species from a second Society Island, Moorea. Whereas Tahiti, including the volcanoes Tahiti Nui and Tahiti Iti, or Presqu'île de Taiarapu, encompasses 1040 km 2 , with highest elevations of 2241 m and 1332 m respectively on the two constituent volcanoes, Moorea has a much smaller land area of 142 km 2 and a peak elevation of 1207 m at the summit of Mont Tohiea (Fig. 1). Rainfall is also less abundant on Moorea, though the 5000 mm/yr recorded precipitation (Craig et al. 2001) is similar to that observed in many windward areas of Hawaii that house distributions of Hawaiian Mecyclothorax. Hypotheses of sistergroup relationships for each of the new Moorean species are proposed based on morphological characteristics, with independent relationships to different Tahitian species posited for each of the Moorean species. Therefore the faunas of these two islands that are currently separated by 23 km of ocean are closely related biogeographically, though there has been sufficient isolation-as predicted by the extreme endemism previously reported on Tahiti-to have resulted in Mecyclothorax faunas on the two islands that are absolutely distinct at the species level. Characters presented by taxa across the Society Island Mecyclothorax radiation distinctively differ from those characterizing the more generalized members of the Hawaiian Mecyclothorax radiation, supporting independent colonization events from Australia and subsequent adaptive radiations for each of these lineages on their particular archipelagoes.

Taxonomic material
Type specimens of Tahitian Mecyclothorax species used in the diagnosis of the new species were borrowed from the Naturhistorisches Museum, Basel (NHMB) and the

Laboratory techniques
Specimens were killed in ethyl acetate impregnated killing jars, maintained under that atmosphere for 24 hr, and then transferred to 70% ethanol for transport to the laboratory, or preserved directly in 100% ethanol. The former specimens were pointed or platen-mounted from ethanol. The latter specimens were maintained in ethanol at -16˚ C and examined under ethanol at room temperature, or removed from ethanol and pointed when a specimen was required for a holotype or allotype specimen. Description of characters was based only on air-dried specimens viewed using a dissecting microscope under bidirectional halogen light.
Labeling is presented verbatim for holotype and associated allotype specimens. Individual lines on labels are separated by a single slash ( / ), and separate labels are indicated by a double slash ( // ). Data for the other paratypes is presented in a standardized, condensed format organized chronologically by collection. Where information is repeated in adjacent, subsequent collections, that field is removed from the data string and is to be interpolated from the previous paratype data entry. Paratypes labeled with an MBIO#### lot number were retained in 100% ethanol and returned to EMEC for ensuing DNA extraction.
Dissection, clearing and staining techniques follow Liebherr (2009a). Genitalic dissections are maintained with the specimens in polyethylene genitalia vials. Photographs were made using a Microptics® (now Visionary Digital®) macrophotographic apparatus, including strobe flash illumination conveyed via fiber-optic wands, and a transmissible light stage. Habitus photos were taken using transmitted and reflected light, with specimens mounted on microscope slides and surrounded with two internested plexiglass rings lined with translucent Mylar® film. Genitalic dissections were photographed using transmitted light.
Descriptive conventions build upon data reported by Perrault (1978aPerrault ( , 1978bPerrault ( , 1984Perrault ( , 1986Perrault ( , 1987Perrault ( , 1988Perrault ( , 1989, with both standardized body length and setal formula presented in each diagnosis. The former consists of the sum of three measurements: 1, the distance from the anteriormost labral margin to the cervical ridge, estimated when necessary by extrapolation from the lateral reaches of the ridge; 2, the median length of the pronotum; 3, the distance from the base of the scutellum, where the scutellar dorsal surface dips ventrally, to the apex of the longer elytron, measured parallel to the suture. The setal formula (e.g., 1234) is based on the number of setae on one side of the beetle, with the four numbers signifiying: 1, the number of supraorbital setae, either one or two; 2, the number setae along the lateral margin of the pronotum, either one or two; 3, the number of dorsal elytral setae in the discal portion of the third elytral interval, ranging from 0-3 in Tahitian Mecyclothorax; and 4, the number of setae at the apex of the elytra. There are either one or two supraorbital setae; if one is absent it is the anterior seta. There are either one or two pronotal setae; if one is absent, the loss occurs at the setal position near the hind angle. There may be one or two dorsal elytral setae (or three unilaterally in one instance herein), and if a seta is lost evolutionarily, it is lost from the more apical position. Plesiomorphically there are two setae near the elytral apex, an apical seta near the apex of the second elytral stria along the elytral margin, and a subapical seta located in the seventh elytral stria anywhere from dorsad the subapical sinuation to closer to the apex of stria 7. Perrault (1978a et. seq.) reported the number of setae near the elytral apex without regard to their homology. In Moorean Mecyclothorax, the apical seta is always present; the subapical may be present or absent.
Various ratios are used to characterize shapes or relative dimensions. The ocular ratio is the measurement across the outer surface of the compound eyes divided by the minimum distance across the frons between the eyes. The ocular lobe ratio is the distance from the anterior to posterior margin of the eye measured from directly above, divided by the distance from the anterior margin of the eye to the groove at the juncture of the gena and ocular lobe using the same viewpoint. Various body meaurements are presented as ratios, with the measurements including: APW, anterior pronotal width, i.e. the width between the most anteriormost pronotal margins at the front angles; MPW, maximum pronotal width; BPW, basal pronotal width measure across the hind angles; PL, median pronotal length; HuW, humeral width, or distance between the anteriormost points along the basal groove-humeral juncture, i.e., the humeral angle; MEW, maximum elytral width. In order to present infraspecific variation in body shape, a maximum of five specimens, where available, were measured to compose these ratios. All available specimens were scanned and the largest individual, the smallest individual, and representatives of both sexes were included in the sample of five. Each of these specimens was labeled with a small number label that corresponds to its entry in the character matrix (unpubl. data). This sampling produced a range of ratios, with the smallest and largest individuals often producing the most disparate ratios. The number of sampled individuals, which may range from 1 to 5, is presented as (n = X).
The configuration of the antennae-moniliform, submoniliform, filiform, or elongate filiform-is categorized using a ratio of the dimensions of the eighth antennomere; length from basal juncture with seventh antennomere to apex divided by the maximal breadth, excluding the dense setal pelage. The configuration of the metathorax is quantified by the metepisternum width/length ratio. Width is measured perpendicular to the longitudinal body axis from the lateral edge adjacent to elytral epipleuron, to the medial juncture of mesepimeron, metasternum and metepisternum. Length is measured as the length of the medial edge from mesepimeron to juncture with metepimeron.
Elytral setation is described based on the dorsal setal positions in the third elytral interval relative to overall elytral length as measure in the standardized body length. The lateral elytral setae of the ninth interval, just laterad the eighth stria, are arrayed in an anterior series starting laterad the humerus, and an apical series that terminates just anterad the subapical sinuation. The two series are presented as A + B, with variation in setal number among individuals reported in parentheses; (B -C). When a particular setal count varied, and one state was observed only rarely, the rarely observed setal number is presented alone in parentheses; i.e., B(C).
Coloration is graded relatively from flavous (i.e. yellow without melanization), to rufoflavous, and then to rufobrunneous. Colors darker than rufobrunneous may entail dominant reddish coloration, thereby leading to rufous, dark rufous, and rufopiceous, or the colors may be dominated by browns, leading from rufobrunneous to brunneous, to rufopiceous. The darkest coloration observed is piceous, or shiny coal black. These base colors may be modified by a darker cast; incomplete melanization of the surface near setae or in thicker portions of the cuticle.
Description of microsculpture follows the general terms used in Lindroth (1974). A mature male specimen was assessed when available, though no consistent differences were observed between male and female specimens; far greater differences were observable when comparing mature, sclerotized individuals versus those partially melanized and sclerotized.
Features of the male aedeagal internal sac were described based on the homology system of Maddison (1993). Terms for the ovipositors used in the female genitalic descriptions were based on those presented by Ball and Shpeley (1983) and Liebherr and Will (1998). The number of male or female individuals dissected is noted in the appropriate section of the description. Mecyclothorax Sharp was classified in the subtribe Amblytelina based on the shared presence of elongate, apically narrowed male parameres that bear setae along the ventral margin of the right paramere and at the apex of the left paramere (Liebherr 2011b, fig. 6). The degree of parameral elongation and apical narrowing varies across the taxa placed in Amblytelina by that cladistic parsimony analysis, but all members of Amblytelina differ from those assigned to subtribe Moriomorphina, wherein the male parameres are parallel-sided, broad to their broadly rounded apex, and glabrous (e.g. Liebherr 2011b, fig. 1A-B). Within Amblytelina, Amblytelus Erichson and associated genera (Baehr 2004) comprise the sister group to Mecyclothorax. Taxa across Amblytelina exhibit derived transformations of the spermathecal duct entrance into the bursa, from the plesiomorphic position near the juncture of the common oviduct and bursa, to positions remote from that juncture. For example in Amblytelus curtus (F.) the spermathecal duct joins the bursa on its ventrolateral margin, about halfway from the bursal-oviduct juncture toward the bursal apex (Liebherr 2011b, fig . 3C). In many Mecyclothorax spp., the spermathecal duct-bursal juncture has transformed to a position on the dorsal surface of the bursa, directly dorsad the bursal-oviduct juncture (as in Fig. 5). However, in females of Mecyclothorax lophoides Chaudoir of Australia the spermathecal duct enters the bursa at the plesiomorphic position at the bursal-oviduct juncture; a plesiomorphic condition also exhibited by Meonochilus bellorum Liebherr of New Zealand (Liebherr 2011b, fig. 8A). With regard to the Society Islands' Mecyclothorax fauna, the female reproductive tracts exhibit a spermathecal duct that enters the dorsal surface of the bursa copulatrix (Fig. 5), and male aedeagal internal sacs with a broadly rounded, spoon-shaped flagellar plate (Fig. 4G), both configurations shared with the Hawaiian generotype Mecyclothorax montivagus (Blackburn). These characters therefore support Perrault's (1978a, b) generic assignment of the Society Islands' fauna.
Identification key to adults of Mecyclothorax spp. known from Moorea

M. gourvesi species group
Diagnosis. Perrault (1986Perrault ( , 1988) based recognition of this species group on presence of a broadly margined pronotum, with sinuate basolateral margins and glabrous, right hind angles.
Male Genitalia. (n = 1). Aedeagal median lobe narrowed dorsoventrally in apical half, apex dorsoventrally expanded to an adze-like tip (Fig. 4A), the apical face of tip much less convex than dorsal and ventral margins; flagellar plate large, length 0.5× distance from parameral articulation to apical face of tip; right paramere extended to 0.8× distance from parameral articulation to apical face, left paramere extended nearly to tip; internal sac with broad dense ventral microtrichial field, and small, lightly spinose dorsal ostial microtrichial patch (assessed in uneverted condition).
Etymology. This species epithet honors the memory of Georges G. Perrault, the principal describer and reviser of the Tahitian Mecyclothorax fauna.
Distribution and habitat. Individuals of this species have been found from the summit at 1207 m elevation down to 1100 m. Specimens have been found microsympatrically in rotten Freycinetia stalks with M. mapo, and in association with M. mapo and M. fatata on flowering Myrsine at night, and in a pyrethrin fog sample of a mossy tree trunk. An individual of this species was the lone carabid beetle recovered from a Berlese extraction sample of leaf litter taken at 1100 m.

M. altiusculus species group
Diagnosis. All species of this group as first defined by Perrault (1986) lacked the basolateral pronotal seta, though in his grouping the pronotal shape varied from trapezoidal with a convex or straight basolateral margin, to cordate wherein the basolateral margin is sinuate anterad a nearly right hind angle. Subsequently Perrault (1988) placed three  species with basolateral pronotal setae in the group, while surmising that the species group was composed of several phylogenetic elements that might bear subdivision. The two new species placed here have the glabrous pronotal hind angles and convex basolateral pronotal margins also exhibited by M. altiusculus Britton, suggesting that these new species would remain in this group after such a subdivision.
Mecyclothorax pahere sp. n. urn:lsid:zoobank.org:act:52B4B915-D08B-4409-A118-4901879032DE http://species-id.net/wiki/Mecyclothorax_pahere Diagnosis. This species plus M. altiusculus, M. pseudaltiusculus Perrault, and M. paraltiusculus Perrault share a bisetose, trapezoidal pronotum with rounded hind angles, and a setal formula of 2122, however individuals of this new species are larger; standardized body length 7.5-7.9 mm. The pronotum is also more transverse; MPW/PL = 1.35-1.36 (n = 2). The striae are deep and distinctly punctate in their basal half, and the convex intervals are covered with dense transverse microsculpture consisting of a mixture of parallel transverse lines and transverse-mesh sculpticells 2-4× broad as long. The mesal face of the male metatibia is lined with pectinate swellings at the points of articulation of the mesal longitudinal setal series (Fig. 2B). Of the three species listed above, M. paraltiusculus is most similar, attaining a similar body size-7.0 mm-and possessing elytral microsculpture consisting of a mixture of transverse lines and transverse mesh.
Holotype male ( Etymology. The species epithet is the Tahitian word pahere, or comb in English, either the noun or verb form (Wahlroos, 2002), and being indeclinable, is to be treated as a noun in apposition. The name is indicative of the metatibial comb in the male, formed by the evaginated bosslike articulatory processes associated with with the mediolongitudinal series of tibial setae.
Distribution and habitat. The allotype female was collected in a pyrethrin fog sample from a mossy tree trunk along with one individual each of M. fatata and M. mapo. The holotype male comprised the only beetle collected in a similar situation the day earlier.
Mecyclothorax menemene sp. n. urn:lsid:zoobank.org:act:2D31BAF6-A9E3-4344-BE4D-32658F0D209E http://species-id.net/wiki/Mecyclothorax_menemene Diagnosis. This species shares an ovoid, bisetose pronotum, and indistinctly striate elytra with M. jarrigei Perrault, however the striae are even less developed in M. jarrigei, and individuals of that species at 6.2 mm are larger than the unique specimen of this new species; standardized body length 5.2 mm. This species exhibits a setal for-mula of 2121 as in M. jarrigei; setae in the new species include both discal elytral setae and the apical elytral seta positioned just laterad the apex of the second stria. The elytra are exceedingly convex in the unique holotype (Fig. 2C), perhaps due to the slightly teneral nature of the specimen leading to apical distortion of the elytra. The dense transverse microsculpture on the elytra results in a silvery metallic reflection.
Male Genitalia. (n = 1). Aedeagal median lobe evenly narrowed dorsoventrally from midlength to broadly rounded apex, the apex slightly expanded on ventral margin (Fig. 4D); flagellar plate smaller, length 0.26× distance from parameral articulation to apex (assessed in uneverted, teneral dissection); internal sac dorsal surface apparently covered with field of fine spicules, neither dorsal nor ventral ostial microtrichial patch apparent; both parameres extended apically about 0.8× distance from parameral articulation to apex. Etymology. The species epithet is the Tahitian word menemene, i.e. round or spherical in English (Wahlroos, 2002), denoting the rounded pronotum and broadly rounded, convex elytra of this species. Being indeclinable, the epithet is to be treated as a noun in apposition.

Distribution and habitat.
The unique holotype of this species was collected from a rotten log situated in a deep, wet gulch, in association with four individuals of M. mapo and one of M. popotioaoa.

M. globosus species group
Diagnosis. Species in this group are characterized by cordate pronota, the basolateral margin distinctly sinuate anterad the projected hind angle, with the lateral marginal depression very narrow and of even breadth throughout the length of the pronotum. The elytra are very convex, with the eighth interval nearly vertical at the juncture with the elytral lateral marginal depression (Perrault, 1986(Perrault, , 1988. Individuals of the included species are smaller; body lengths range 3.3-5.0 mm (Perrault, 1989). Of these, M. cupripennis shares reduced microsculpture with this species, the pronotal disc lacking any discernible sculpticells. The two species differ in elytral microsculpture. This new species is characterized by a glossy elytral integument, with only sporadic small patches of indistinct isodiametric sculpticells in transverse rows, whereas M. cupripennis is characterized by presence of a more regular, though shallow, transverse mesh on the discal elytral intervals; the sculpticells consistently visible outside the reflection of bright, direct microscope light. Body size is similar for the two species; standardized body length of the new species is 3.7 mm, that for M. cupripennis 3.5 mm (measurement made on male specimen, CUIC).
Female reproductive tract. The unique female holotype was not dissected. Etymology. The species epithet is a compounding of maha, Tahitian for the number four, and tahi, Tahitian for one (Wahlroos, 2002), indicative of the reduced setation in this species resulting in the setal formula of 1111. As tahi is indeclinable, the epithet is to be treated as a noun in apposition.
Distribution and habitat. The unique specimen was collected in pyrethrin fog sample of a mossmat in association with one specimen of M. perraulti.
Mecyclothorax popotioaoa sp. n. urn:lsid:zoobank.org:act:CD0E0395-531F-4DD2-B68C-8761EA70D2F5 http://species-id.net/wiki/Mecyclothorax_popotioaoa Diagnosis. Within the M. globosus group, this is the only species for which individuals lack dorsal elytral setae, resulting in a setal formula of 2101; one individual of the five type specimens has an asymmetrically positioned dorsal elytral seta at 0.24× length on the left elytron, however this is considered a variant condition not characterizing the species. This species is also characterized by shallow elytral striae (Fig. 3B), and much reduced microsculpture across the entire dorsum. The most similar species in the group, M. hemisphaericus Perrault, shares much reduced elytral striae, though in this species they are nearly obsolete. M. hemisphaericus also exhibits reduced setation, though presence of a single dorsal elytral seta results in a setal formula of 2111. The new species can also be distinguished from M. hemisphaericus by the more narrowly ovoid elytra, MEW/MPW = 1.43-1.48 (n = 4), versus more broadly ovoid elytra in M. hemisphaericus, MEW/MPW = 1.61 (n = 2 paratypes, MNHN). Individuals of both species are of similar size; standardized body length for this species is 3.7-4.0 mm, versus 3.5-3.9 mm for M. hemisphaericus, as determined from two examined paratypes (MNHN).
Male genitalia. (n = 1). Aedeagal median lobe evenly curved and of subequal diameter in basal half, narrowed apically to tightly rounded apex that extends little beyond apical ostial margin (Fig. 4E); internal sac lightly sclerotized, only flagellar plate visible in uneverted dissection, length of plate 0.33× distance from parameral articulation to apex; right paramere short, apex extended toward apex 0.7× distance from parameral articulation to apex, left paramere longer, extended 0.9× that distance.
Female reproductive tract. (n = 1). Bursa copulatrix very short, present as a very short lobe situated dorsad the broad common oviduct (Fig. 5C); bursal apex extended beyond evident transverse fold at base of oviduct the same distance as that from transverse fold to a line drawn between bases of basal gonocoxites; bursal surface lightly sclerotized, as membranous as surface of median oviduct; spermatheca reniform, spermathecal gland connected to base of spermatheca by a short duct; basal gonocoxite 1 with apical fringe of 3-4 setae, and 10-11 smaller setae along mesal margin, spanning ventromedial to dorsomedial surfaces of gonocoxite (Fig. 6C); apical gonocoxite 2 narrow basally with tightly rounded apex, single dorsal and lateral ensiform setae and 2 apical nematiform setae.
Etymology. The species epithet is a compounding of the Tahitian word popoti, beetle or cockroach, and oaoa, the Tahitian adjective narrow (Wahlroos, 2002), signifying the constricted pronotal base and narrow body of adult beetles of this species. As oaoa is indeclinable, the epithet is to be treated as a noun in apposition.
Distribution and habitat. Four of the five specimens recorded for the species have been collected on fern fronds, either living or dead, in all such instances associated with individuals of M. mapo. The fifth specimen was collected in association with M. mapo and the single known M. menemene by pyrethrin fogging of a mossy log complex in a deep, wet gulch.

M. viridis species group
Diagnosis. Perrault (1986) assigned four species to this group, all of which exhibit: 1, cordate pronotum with sinuate basolateral margins, and right, setose hind angles; 2, distinctly convex elytra with upraised humeral margin; and 3, presence of 1 or 2 dorsal elytral setae. He noted that the four species comprise two geographic pairs, M. Diagnosis. This species shares transverse-line elytral microsculpture and deep punctate elytral striae with M. castaneus, and individuals are of similar body size; standardized body length for this species 3.8-4.4 mm versus 3.8 mm for M. castaneus. The pronotum is of similar dimensions in the two species, with MPW/PL = 1.14-1.17 (n = 5) in this species, versus a ratio of 1.19 in M. castaneus (Perrault, 1986). The species differ in setation, with this species consistently characterized by two discal elytral setae, and therefore a setal formula of 2221, versus M. castaneus where one of the two type specimens had two discal setae on one elytron, whereas the other three elytra of the two beetles were unisetose. In addition, the pronotal base of this species is relatively broader, MPW/BPW = 1.52-1.64 (n = 5), versus a narrower base and greater ratio of 1.70 for M. castaneus.
Male genitalia. (n = 3). Aedeagal median lobe broad in basal ⅔ of length, narrowed apically to broadly rounded apex with blunt apical face, ventral portion of median lobe straight (Fig. 4F); internal sac ventrally covered with dense microspiculate field, distinct ventral or dorsal microtrichial patches absent; flagellar plate large, length 0.57× distance between parameral articulation and apex, gonopore visible on middle of dorsal surface, longitudinally radiate sclerotic ridges on inner, ventral, surface of plate; right paramere narrowly elongate (Fig. 4G), tip extended 0.85× distance from parameral articulation to apex, left paramere slightly longer, extended 0.90× such distance.
Etymology. Given that this species is most similar to M. castaneus, the common name of the Tahitian chestnut tree, Inocarpus fagifer (Parkinson) (Fabaceae)-i.e. mapo (Wahlroos, 2002)-was chosen for the species epithet. The epithet is to be treated as a noun in apposition.
Distribution and habitat. This species has been found in a variety of situations on Mont Tohiea, accounting for 68 of the 90 specimens of Mecyclothorax collected on or near the summit. Specimens have been found by sampling ferns, Angiopteris, rotten Freycinetia, Myrsine foliage and flowers, and moss-covered Wienmannia trunks and roots. In keeping with this species' numerical dominance, it has been collected in association with all other Mecyclothorax spp. known from Mont Tohiea.
Mecyclothorax fatata sp. n. urn:lsid:zoobank.org:act:14EE6095-7917-4D0F-8D25-42F7C8763199 http://species-id.net/wiki/Mecyclothorax_fatata Diagnosis. This species shares upturned pronotal margins with a visible lateral depression (Fig. 3D) and regular transverse-mesh elytral microsculpture with M. ata Perrault. Individuals of the two species are of similar body size; standardized body length 4.7-5.0 for this species versus 4.5 mm for M. ata. However this species deviates from M. ata by presence of only the anterior elytral seta resulting in a setal formula of 2211 versus 2221 for M. ata. The pronotal base is also more constricted in this species, MPW/BPW = 1.53-1.62 (n = 5) versus a ratio of 1.49 in M. ata (Perrault 1986).
Variation. Marginal setation of the apical visible ventrite in males is unstable in this species. Of the four male specimens, one individual has 4 terminal abdominal setae along the apical margin of visible ventrite 6, 2 on each side (EMEC); one has 3 apical setae, 1 on the right and 2 on the left (CUIC); and two others exhibit the usual 2 setae, 1 each side (MNHN, EMEC).
Etymology. Because this species is most similar to M. ata, the Tahitian epithet fatata, near or nearly (Wahlroos 2002) was chosen to express the similarity. The epithet is indeclinable and is to be treated as a noun in apposition.
Distribution and habitat. All six collections and eight specimens of M. fatata were made in association with the numerically dominant M. mapo. In two instances M. perraulti was also present in the sample. Five of the eight M. fatata specimens were collected from Myrsine, one from moss-covered Weinmannia, and two others from unidentified trees.

Discussion
The greatest similarity of the seven Moorean Mecyclothorax spp. to seven different Tahitian Mecyclothorax spp. (Table 1) points to independent biogeographic relationships-i.e. independent speciation events-for the seven sister-species pairs. This pattern suggests that all seven speciation events have occurred such that the descendant species occupy allopatric distributions on either side of the Moorea Channel. Isolation by this Channel leading to speciation could have conceivably involved three biogeographic phenomena: 1, vicariance between Moorea and Tahiti based on subsidence of intermediate land areas and associated oceanic incursion; 2, dispersal of an ancestral population from Tahiti to Moorea; 3, dispersal of an ancestral population from Moorea to Tahiti. The first option involving subsidence of an ancient mountain range was favored by Crampton (1917, 16;1932, 194)  Taking up the hypotheses that the biogeographic relationships of the Moorean and Tahitian taxa are due to dispersal, we can ask whether a consistent biogeographic relationhip between Mont Tohiea and an area of endemism in Tahiti is observed. If Mecyclothorax beetles dispersed between these islands, they necessarily did so over water, not in the air, as all Mecyclothorax spp. on both islands possess vestigial flight wings and a truncated metathoracic flight apparatus evolutionarily associated with winglessness (Darlington 1936). Given that restriction, dispersal would most likely have been associated with flooding events, landslides, and subsequent rafting between islands on vegetative masses broken loose from the very steep, highly eroded volcanic flanks present on both islands (Zimmerman 1948). Marau and Aorai represent the Tahitian mountain ridges most proximate to Moorea, and thus most likely to be implicated in dispersal to or from Moorea. Based on the distributional ranges of the Tahitian adelphotaxa (Table 1), there is majority support for this pattern as five of the seven species occur on either Aorai or Marau. However two of the adelphotaxa are restricted to Taiarapu, a geographically distant source or recipient.
The generality of the proposed, most prevalent dispersal pattern can be tested using information from other taxa. To date the best example of such a test involves the Moorean Nabidae, which comprise three micropterous species, all related to different micropterous species in Tahiti (Polhemus 2010 Hoch (2006) proposed that two Moorean species of Oteana Hoch (Hemiptera: Cixiidae) are adelphotaxa, suggesting that the island has had a complicated enough geographical and botanical history during its 1.52 Myr subaerial lifespan (Guillou et al. 2000) to have supported speciation within its bounds. Recent collections of Mecyclothorax beetles have all occurred near the summit of Mont Tohiea, from 1100-1207 m elevation. During his Tahitian collecting, Perrault (1986Perrault ( , 1987Perrault ( , 1988 found singleton specimens he described as three species-M. teatara Perrault, M. ferruginosus Perrault, and M. sinuatus Perrault-in habitats from 800-1000 m. During the 2006 survey that resulted in the first collections of Mecyclothorax on Mont Toheia, another undescribed Mecyclothorax was discovered in a riparian habitat at 705 m elevation on Mont Mauru in eastern Tahiti Nui (unpubl. data). Thus 700 m elevation stands as a likely lower limit for present-day distributions of Mecyclothorax beetles on Tahiti. Assuming a similar lower elevational limit for taxa in Moorea suggests that Moorea's Mt. Rotui, isolated by low-lying valleys and peaking at 899 m elevation, is the locality most likely to contain Moorean adelphotaxa to one or more of the seven species precinctive to Mont Tohiea. Other peaks on the main Mont Toheia massif-e.g. Mont Mouaroa-may also support Mecyclothorax, though this summit at 880 m elevation is connected to Mont Toheia by a ridgeline supporting montane forest, and thus would be less likely to house distinct species.
The most speciose radiations of Mecyclothorax spp. are concentrated in the Society and Hawaiian island chains. Is this shared level of extreme diversity indicative  Britton (1948) proposed the Australian M. punctipennis (his "M. ambiguus" prior to Moore's [1984] clarification) (Fig. 7A) as the species most similar to the generalized Hawaiian M. montivagus (Blackburn) (Fig. 7C) based on pronotal, elytral and aedeagal configurations (Fig. 8) (Moore et al. 1987;Baehr 2009;Liebherr and Marris 2009), the greatest morphological similarity also links Australia's M. punctipennis to the Tahitian members of the M. striatopunctatus species group (Perrault 1986) (Fig. 7B). The four M. striatopunctatus group species all display the maximal setal formula of 2222, as does M. punctipennis. Members of the M. striatopunctatus species group also exhibit an ovoid pronotum with little projected hind angles, and punctate elytral striae. Being vestigially winged, the elytra have more rounded humeri, as also observed in M. montivagus (Figs 7B,C), and the elytral striae are the deepest of the three, with lateral striae more well developed than in M. punctipennis or M. montivagus (Fig. 7). Aedeagal conformation is also very similar in males of all three species. The aedeagal lobe of the Tahitian species terminates in moderately expanded apex (Perrault 1986, figs 22, 23) similar to the aedeagal median lobe of M. punctipennis (Fig. 8A) and M. montivagus (Fig. 8B). The internal sac of the latter two taxa are also similar (data for the Tahitian species not available), sharing: 1, robust, moderately elongate shape; 2, a well-developed dorsal ostial microtrichial patch with short spicules; 3, a ventral ostial microtichial patch; 4, a broad ventral field of fine microspicules that extends to the flagellar plate; and 5, a relatively large flagellar plate. Thus it would appear that M. punctipennis also represents the generalized mainland species that is most similar to any extant Tahitian taxon, and therefore the most likely candidate to have colonized the Society Islands.
If M. punctipennis spawned the colonizing propagules that founded radiations on both Hawaiian and Society archipelagoes, were these radiations founded in a stepping-stone like manner from a source in the southwest Pacific? Such a stepping-stone pattern of colonization has been shown for the dominant Hawaiian Ohi`a, Metrosideros polymorpha (Myrtaceae) and allied species, with dispersal stemming from New Zealand and Lord Howe Island to the Marquesas and the Societies, and ultimately to the Hawaiian Islands (Wright et al. 2001;Percy et al. 2008). Elytral configuration in the two Mecyclothorax radiations argues against such a conclusion. In the Tahitian M. striatopunctatus species group, as well as most other Society Island Mecyclothorax, the eighth elytral interval just laterad the seventh stria is developed into a subcarinate margin that extends from the elytal apex to well anterad the position of the subapical elytral seta (Fig. 7E). This subcarinate ridge may be developed further in Society species, taking the form of a carinate ridge that extends anterad various lengths on the elytron (Figs 2A, B). This carinate condition has evolved several times during the evolution of Moriomorphini (Liebherr 2011b). Both the subcarinate and elongate carinate conditions were derived from the more briefly subcarinate eighth interval observable in M. punctipennis, where the distinct mesal margin of the eighth interval does not extend appreciably anterad the subapical setal depression (Fig. 7D). This brief, marked depression of the eighth interval occurs in spite of the generally shallow lateral elytral striae characteristic of this species (Fig. 7A). Several other Austral-Pacific Mecyclothorax species share the elytral configuration observed in M. punctipennis: M. ambiguus and M. lophoides (Chaudoir) of Australia; M. basepunctus Louwerens and M. lissus (Andrewes) of Java; and M. rotundicollis (White) plus M. oopteroides Liebherr and Marris (2009) from New Zealand. Therefore colonization of the Societies has been followed during diversification by evolutionary enhancement of this carina (Fig. 7E). Conversely in the Hawaiian M. montivagus, this carina is less developed than in M. punctipennis; i.e. the eighth interval is more rounded mesally, with its margin most distinct, though still broadly rounded, laterad the subapical setal depression (Fig. 7F). Thus the Hawaiian and Australian taxa share the absence of a subcarinate interval extension anterad the subapical seta. Based on the pattern of evolutionary transformation in this character, whereby the Australian condition is considered plesiomorphic (Fig. 7D), evolution has proceeded independently in two directions; 1, toward more carinate elytra in the Societies (Fig. 7E); and 2, to a less carinate, more rounded eighth interval in Hawaii (Fig. 7F). Given this pattern, it is most parsimonious to hypothesize two independent colonization events involving the same colonizing species, with subsequent independent radiations in the Tahitian and Hawaiian archipelagoes.
Independent colonization of the Societies and Hawaii by propagules derived from populations of M. punctipennis suggests several ancillary conclusions. First, both colonizing events would seem to have involved propagules derived from one of the currently most common carabid beetle species in Australia, M. punctipennis. The numerical dominance and ecological plasticity of this species, which occurs from dry Xanth- orrhea-Eucalyptus forest at sea level to open montane snow gum (Eucalyptus pauciflora Sieber) forests at over 1500 m elevation (unpubl. data), enhances the probability that were a propagule produced from this species, the colonizers would possess substantial genetic variability to allow survival as ecological pioneers across the range of habitats occupied by the beetles in Australia. This range would include the present-day leeward montane shrubland habitat of M. montivagus in Maui.
Second, if M. punctipennis were the source of colonizing propagules, colonization of both the Society and Hawaiian Islands must have occurred recently. Present evidence for the Society radiation, with species restricted to Moorea and Tahiti, would place the age of origin of the fauna at no more than 2.25 Ma were the radiation founded on Moorea, or 1.75 Ma if founded on Tahiti (Craig et al. 2001). In Hawaii, the most generalized species, M. montivagus, occurs on Haleakala, a volcano that completed the shield-building stage by 0.95 Ma (Sherrod et al. 2003). In Hawaii, there are no Mecyclothorax in Kauai, and the Oahu taxa have sister groups on Maui Nui (Liebherr 2009a;Liebherr and Krushelnycky 2011). This places Maui Nui, and most defensibly Haleakala based on extant, observable taxa, as the point of original Mecyclothorax colonization in Hawaii. Thus both island radiations have evolved over the past 1-2 Myr. Bocher (1995) found that nearly all of the 140 carabid beetle species assembled in 2 Myr old glacial deposits at Kap København, North Greenland remain extant. If the widespread, ecologically plastic M. punctipennis responded to climatic change as did the broadly distributed Holarctic coleopteran species that tracked Pleistocene climate change (Coope 1979), then a 2 Myr species duration for M. punctipennis seems a reasonable working hypothesis.
Finally, as the Tahitian and Hawaiian radiations have the same, extant sister species as their adelphotaxon, intense acceleration of the speciation rate is demonstrable in both island archipelagoes. This accelerated diversification must be directly attributable to life on these isolated volcanic islands. The subtropical Hawaiian and Society volcanic islands developed forests quickly, as orographic rainfall synergized nutrient transfer to forest plants within thousands of years of plant colonization (Kitayama et al. 1997;Vitousek et al. 1997). Evolutionary loss of flight wings in both of these island radiations enhanced speciation rate in the flightless Mecyclothorax by reducing gene flow among populations (Liebherr 1988), thereby facilitating adaptation to specific locales (Darlington 1943;Southwood 1977). Allopatric speciation among these poorly connected populations was facilitated initially by range fragmentation caused by by emplacement of new lava flows on the landscape resulting in isolated forest kipukas (Zimmerman 1948). Upon cessation of major volcanic activity, rampant erosion and valley formation isolated organisms in islands or peninsulas of montane forest habitat. The number of ecological components in these habitats was restricted by the colonization process, and those taxa present faced reduced competition for resources except among phylogenetic relatives that shared their general way of life. This restrictive competition favored subsequent specialization within what remained a more generalized way of life in the mainland source community. In the Tahitian Mecyclothorax fauna, this has resulted in species that have specialized to live predominantly in leaf litter, such as the large-bodied M. muriauxi species group (Perrault 1984) and others that are observed nearly always in arboreal microhabitats, such as species of the smallbodied M. globosus group (Perrault 1989). This ecological divergence took place without the ability to colonize remote habitat patches via winged flight. However, forest habitats anastomosed through time as lava flows became forested, allowing beetles to colonize newly available habitats terrestrially, with subsequent lava flows dissecting the landscape in new and different ways. The beetles' various ecological specializations combined with their aggregate role as one of the dominant predatory lineages, ensuring that these radiating insects persisted in the many historically colonized forest patches, thereby setting the stage for multiplicative generation of ever more allopatric species; an island variant of Noonan's (1988) continental cyclic vicariance. This multiply-layered Mecyclothorax fauna, with many sympatric species representing different sublineages of the radiation, prompted Perrault (1992) to suggest that Tahiti was colonized repeatedly by numerous waves of different Mecyclothorax colonists. However, the presence of a very similar pattern of intense levels of sympatry and extreme specialization in the much more isolated Hawaiian Islands, coupled with the extreme similarity of only the mainland M. punctipennis to any of the Hawaiian Mecyclothorax, supports evolution from a single colonization event in Hawaii. In the Society Islands, moreover, the presence of species assignable to Mecyclothorax based on appropriate generic-level characters, yet exceedingly different (e.g., Fig. 2B) from any other Mecylothorax species present anywhere beyond the Society Islands, suggests that diversification has proceeded apace within this archipelago to produce today's incredibly diverse fauna.