Specimens were taken on loan for study from the following institutions: Cornell University Insect Collection (CUIC), Ithaca, NY, US; Essig Museum of Entomology (EMEC), University of California, Berkeley, CA, US; Field Museum of Natural History (FMNH), Chicago, IL, US; Lincoln University Entomology Research Collection (LUNZ), Lincoln, NZ; Museum of Comparative Zoology (MCZ), Harvard University, Cambridge, MA, US; The Natural History Museum (NHML), London, UK; New Zealand arthropod Collection, Mt. Albert, NZ (NZAC); Tasmanian Museum and Art Gallery (TMAG), Hobart, AU; Zoological Museum, University of Copenhagen (ZMUC), Copenhagen, DK; Zoologisches Staatssammlung (ZSM), München, Bavaria, DE.

Specimens were examined using a Wild M5 microscope with halogen ring-light illumination at 6–100× magnification. Genitalic dissections were made after specimens were relaxed in hot deionized water containing a few drops of Kodak Photo-Flo® detergent, with the dissections accomplished using modified minutens and watchmakers’ forceps. The dissected structures were cleared in 10% cold KOH overnight, then neutralized in dilute acetic acid, and subsequently held in glycerin for viewing and ultimate storage in polyethylene genitalia vials mounted on the specimen pin. Female reproductive tract structures were stained in Kodak Chlorazol Black® suspended in methyl cellosolve after the acetic acid neutralizing step, and then transferred after 15 minutes’ staining time to glycerin for viewing. These structures were likewise stored in genitalia vials. Specimen localities were recorded using the area definitions of Crosby et al. (1976).

Body proportions and shapes were quantified using mensural characters determined using an ocular reticle. These include eye size as the ocular ratio MHW / mFW, with MHW being maximal head width across the compound eyes, and mFW is minimum frons width between the compound eyes. Pronotal shape is represented by several ratios—APW / BPW, MPW / BPW, and MPW / PL—where APW is apical pronotal width measured at the front angles, MPW is maximal pronotal width, BPW is basal pronotal width measured between the basal angles, and PL is pronotal length measured along the midline. Elytral shape is quantified using HuW / MEW, and MEW / EL; that is HuW = width across the humeral angles, MEW = maximal elytral width, and EL = elytral length measured from the base of the scutellum to the apex of the left elytron. Standardized body size is quantified as the sum of HL + PL + EL, where HL is the midline distance between the anterior margin of the labrum and the ridge demarking the juncture between the vertex and cervix, and PL and EL are as defined above. Migadopines also exhibit variously laterally expanded pro- and mesotarsomeres, that development quantified as the ratio of the maximal lateral breadth of tarsomere 2, divided by the median tarsomere length; w / l. Terminology used for describing male genitalia was based on Lindroth (1957), with Liebherr and Will (1998) followed for interpretation of the female gonocoxae and bursa copulatrix.

Cladistic analysis was performed under the maximum parsimony criterion based on a data matrix (Suppl. material 1) comprising 74 characters for 42 taxa (Table 1), this matrix an amended expansion of the 57 character × 29 taxon matrix of Roig-Juñent (2004). The additional taxa principally include those from New Zealand and Australia phylogenetically allied with Amarotypus edwardsii. The twenty characters added to the matrix of Roig-Juñent are each indicated below by an asterisk following the character number, while character 43 of Roig-Juñent was deleted as it is autapomorphic. Characters numbered 56 and 57 of Roig-Juñent (2004) were also deleted: 56 because it incorrectly coded Amarotypus edwardsii to lack the setose sensillum of the apical gonocoxite, and 57 because a helminthoid sclerite as defined by Liebherr and Will (1998) is not present in migadopine taxa. The resultant matrix was originally edited and entered using Winclada ver. 1.00.08 (Nixon 2002) and subsequently imported into Mesquite ver. 3.81 (Maddison and Maddison 2023b). Using Mesquite, the matrix was submitted to TNT (Goloboff and Morales 2023) with the Zephyr ver. 3.31 (Maddison and Maddison 2023a) package for both parsimony tree and jackknife resampling searches. Tree search was done using Mesquite’s default commands for TNT modified to as follows: “Hold” increased to 1000000, “replic” set to 1000, and tbr used “nofillonly”. For jackknife analysis Mesquite’s default commands for TNT were used and 1000 replications were done. Tree statistics- CI, RI, and length- are given as reported by Mesquite.

The newly amended matrix was also analyzed using Bayesian inference. Bayesian analysis was run under a symmetric model in which the frequency of each state is equal and all characters informative using MrBayes ver. 3.2.7, (Ronquist et al. 2012). To ensure that an average standard deviation of split frequencies (ASDSF) below 0.01 was achieved (Ronquist et al. 2009), and that likelihood scores and all parameter values reached a stable plateau, an initial analysis was run using the command “stoprule = yes” with “stopval = 0.009998.” Tracer (Rambaut et al. 2018) was used to examine trace files resulting from this run and the effective sample size (ESS) of the parameters used to assess convergence and stationarity. The trees in a burn-in period of 25% of the generations were excluded. While the ASDSF went below 0.01 in only 1.2 million generation for the initial analysis, the ESS values for a few parameters were still below 200, the rule of thumb threshold (Nascimento et al. 2017). The stop value command was removed and a second analysis of 6 million generations for four runs of eight chains was conducted. This analysis reached an ASDSF of 0.0087 and all ESS values viewed in Tracer were > 630. The majority-rule consensus tree of post-burn-in trees was calculated to determine Bayesian posterior probabilities (PP) of clades.

Character states for the 75 characters are listed below. Character number is based on Roig-Juñent (2004), that sequence modified only by additions or deletions. Characters newly used in this analysis are indicated by an asterisk following the character number. Multistate characters are considered non-additive; i.e., the states are unordered.

Character 1: Seta of mandibular scrobe; absent (0), present (1).

Character 2*: Maxillary stipes setation; 1 (0); 2 at base (1); 3, 2 basally, 1 medially (2); 4, 3 basally, 1 medially (3); 5 along length (4); 9 along length (5).

Character 3: Galea of maxillary palps; 2 articles (0), 1 article (1).

Character 4*: Number of setae on mentum; 2 straddling midline near mentum tooth (0), 4, 2 near midline, 2 basolaterally (1), many bordering margins of mentum (2).

Character 5*: Number of setae on submentum; 2 (0), 4 (1), 6 (2), 8 or more (3).

Character 6: Submentum; separated from mentum (0), fused to mentum, at least in central region (1).

Character 7: Mentum tooth; simple, angulate (0), bilobed or blunt with median concavity (1), absent, no forward expansion medially (2).

Character 8: Paraglossae; long (0), distinct (1), undifferentiated (2).

Character 9: Setae of glossal sclerite; four (0), two (1), one (2).

Character 10: Setae of paraglossae; absent (0), present (1).

Character 11: Labial and maxillary palps; elongate (0) short and wide, subrounded (1).

Character 12: Antennae; short, reaching base of elytron (0), long, reaching the basal third of elytra (1), very long, reaching the middle third of elytra (2).

Character 13: First four antennal segments; pubescent from apex of fourth segment (0), glabrous (1).

Character 14: Supraorbital setae; 2 each side (0), only 1 each side (posterior) (1), only 1 each side (anterior) (2), absent (3).

Character 15*: Posterior supraorbital seta; between eyes, anteriad their hind margin (0), at or posteriad hind margin of eyes (1).

Character 16: Neck; present (0), absent (1).

Character 17: Depression bordering hind margin of eye; absent (0), present (1).

Character 18*: Protibial antennal cleaner; isochaetous, longitudinally sulcate, posterior spur subapical (0), anisochaetous, transverse, posterior spur displaced proximally (1).

Character 19: Fourth male protarsomere apex; truncate (0), bilobed, with both lobes equal (1), bilobed, with anterior lobe more developed than posterior (2).

Character 20: Fourth male mesotarsomere; truncate (0), bilobed with anterior lobe more developed than posterior (1).

Character 21: Fourth male metatarsomere; truncate (0), bilobed, with anterior lobe more developed than posterior (1).

Character 22: Male protarsomeres; normally dilated (0), more distinctly dilated (1).

Character 23: Male protarsomeres; 1–3 with adhesive setae (0), 1–4 or 2–4 with adhesive setae (1).

Character 24: Male mesotarsomeres; 1–4 dilated (0), 1–3 dilated (1), not dilated (2).

Character 25: Male mesotarsomeres; without adhesive setae (0), with adhesive setae (1).

Character 26*: Unguitractor plate of fifth tarsomere; with short projection (0), with elongate projection nearly 1/5 length of ungues (1; Fig. 1).

Character 27: Female protarsomeres; not dilated (0), dilated (1).

Character 28: Female protarsomeres; with only 2 rows ventral setae (0), with expansive field of ventral setae (1).

Character 29: Female mesotarsomeres; 1–4 not dilated (0), dilated (1).

Character 30: Female mesotarsomeres; with only 2 rows ventral setae (0), with expansive fields of ventral setae (1).

Character 31*: Protibial apex; moderately expanded (0); robust overall, very broad apically (1).

Character 32*: Anterior pronotal margin; lined with microsetae across breadth (0), microsetae present medially but absent near front angles (1), microsetae absent medially but present near front angles (2)

Character 33*: Posterior pronotal margin; lined with microsetae across breadth (0), microsetae present medially, absent near basal angles (1).

Character 34: Lateral border of pronotum; broad (0), narrow (1).

Character 35: Anterior and posterior breadths of pronotum; of equal or subequal breadth (0), posterior breadth markedly greater than anterior (1).

Character 36: Anterior marginal bead of pronotum; not marked, at least incomplete medially (0), marked, distinct across breadth (1).

Character 37: Form of pronotum; wider before middle with base constricted (0), sides subparallel (1), sides diverging toward the back, width maximal at basal margin (2).

Character 38: Basal pronotal setae; present (0), absent (1).

Character 39: Lateral pronotal setae; present (0), absent (1).

Character 40: Prosternal process; not extended beyond procoxae except as vertical carina (0), extended posteriad procoxae w/o contacting mesosterum (1), dorsally extended posteriad procoxae, overlapping mesosternum (2).

Character 41: Elytral humeri; rounded (0), angulate to right (1).

Character 42: Basal border of elytra; absent (0), present (1).

Character 43: Elytral striae 1–9; absent, reduced to be untraceable on surface (0), discontinuous (1), fine, continuous (2).

Character 44: Punctation of elytral striae 1–9; absent (0), fine (1), distinct (2).

Character 45: Parascutellar stria; short striole (0), extended to apex of elytra (1).

The nomenclature for elytral striae follows the interpretation of homology for the parascutellar striole and stria 1 given by Will (2020). When impressed, the parascutellar striole is directly adjacent to the scutellum on the elytra and typically continuous with the basal marginal border of the elytra. In many Migadopini, the parascutellar striole has the appearance of a stria that is nearly the length of the elytra and joined to stria 1 in the apical third. In such cases the first two apparent intervals have a homologous correspondence to the branches of the 2^nd anal vein of the insect wing.

Character 46*: Parascutellar seta; present (0), absent (1).

Character 47: Bases of elytral striae 1 and 2; fused in a common trunk basally (0), not fused, no common trunk (1)

Character 48: Parascutellar stria and elytral stria 1; fused apically (0), not fused (1).

Character 49: Striae 5 and 6; separated at base (0), joined basally (1).

Character 50*: Stem of fused striae 5 and 6; short (0), long (1).

Character 51: Ninth elytral stria; not crimped on the base (0), bent inwards at the base and attached to the eighth (1).

Character 52*: Elytral marginal setae in interval 9; 6–7 (0); 10–14 (1); 16–22 (2).

Character 53: Elytral coloration; uniform, concolorous on disc (0), with subapical patch of reddish color (1).

Character 54: Elytral marginal coloration; margin concolorous to evenly paler apically (0), margin with irregularly undulated pale areas apically (1).

Character 55*: Metathoracic flight wings; present (0), brachypterous, apex extended more than half elytral length (1), reduced, beetles apterous (2).

Character 56: Male aedeagal median lobe base; open ventrally (0), closed ventrally (1).

Character 57: Male median lobe basal carina; absent (0), present (1).

Character 58: Dorsoventral width of median lobe; thin (0), broad (1).

Character 59*: Ostium position at male median lobe apex; on right (ventral) side of lobe (0), on left (dorsal) side of lobe (1); on medial surface of lobe, dorsal when everted (2).

Character 60*: Ostium opening; broad opening on dorsal aedeagal surface (0), constricted, a small opening on aedeagal surface (1).

Character 61: Ventral region of male median lobe apex; straight, narrow (0), expanded euventrally (1).

Character 62: Male left paramere; glabrous (0), with a few setae (1), with many setae (2).

Character 63: Male left paramere; elongate (0), shape more conchoid (1).

Character 64: Male left paramere apex; sclerotized (0), membranous (1).

Character 65 Male right paramere; with 2 apical setae (0), with a row of ventral setae (1), with 2 rows of ventral setae (2).

Character 66: Sclerites X and Y of male aedeagal internal sac; present (0), absent (1).

Character 67: Ramus associated with female gonocoxite 1; absent (0), present (1).

Character 68: Female gonocoxite IX; 2-segmented (0), unsegmented, basal and apical gonocoxite fused (1).

Character 69: Female basal gonocoxite 1 (or basal portion of fused gonocoxites); apparently glabrous, without elongate setae across ventral surface (0), ventral surface covered with setae (1); with apical border of microtrichia (2), with apical border of palmate microtrichia (3).

Character 70*: Female apical gonocoxite (or apical portion of fused gonocoxites); without medioventral ensiform macrotrichia (0), with 4 mediodorsal ensiform macrotrichia (1).

Character 71*: Ligular sclerite of female bursa copulatrix; present (0), absent (1). This sclerotized structure lies on the ventral surface of the common oviduct, and is not associated with the bursal wall.

Character 72*: Ligular sclerite configuration; elongate (0); longitudinal ridges (1); sclerotized plate (2); cockscomb configuration (3). This character coded only for taxa exhibiting state 1, character 71.

Character 73*: Bursa copulatrix tubular diverticulum; absent (0), present (1). This diverticulum is elongate without an apical expansion, and it joins the bursa copulatrix distad the bursal juncture with the common oviduct. It is not interpreted as a primary spermatheca as it does not expand at the distal end, lacks any taenidial coils often observed in a primary spermatheca, and lacks a gland (Liebherr and Will 1998, fig. 3).

Character 74*: Bursa copulatrix basal diverticulum; absent (0), present (1).

The phylogenetic hypothesis of migadopine taxa was used to establish a context within which historical biogeographic events could be interpreted by converting the taxon cladogram to a taxon-area cladogram wherein all terminals were represented by areas of endemism. The areas of endemism were defined broadly so that migadopine taxa were geographically restricted relative to the scale of the adopted areas of endemism, and thus all terminals could be coded as single areas. For purposes of this analysis, Tasmania was combined with mainland Australia into a single area of endemism (AU), and both South and North Islands of New Zealand were combined into a single area (NZ). The Auckland and Antipodes Islands represent the Campbell Plateau (CP), considered for this analysis to be an area of endemism distinct from New Zealand. South America is considered as three different areas, one located in the northern high Andes of Ecuador, above 4000 m altitude (ESA), a second, in southern South America including mainly subantarctic moorland and Nothofagus forest and Patagonian steppes (SSA), and a third Neotropical region (NEO) comprising tropical lowlands that house one ultimate outgroup taxa, Cicindis horni Bruch. The two cicindine genera Cicindis and Archaecicindis were included in the phylogenetic analysis to enhance character polarizations, but the Cicindini–Archaecicindis johnbeckeri Bänninger and C. horni–represent the Inabresian zoogeographic pattern (Jeannel 1942; Kavanaugh and Erwin 1991) that spans eastern tropical South America to the Persian Gulf east of Africa. The Inabresian Region was vicariated by the opening of the Atlantic Ocean during the Cretaceous, 119–105 Ma (McLoughlin 2001). This pattern is not analyzed further in the biogeographic analysis as the Holarctic outgroups Loricerinae and Elaphrinae, i.e. Elaphrus and Blethisa, are considered the successive, closest outgroups to Migadopinae.

Nodal optimizations on the taxon-area cladogram (Fig. 8) were calculated by RASP (Yu et al. 2015, 2019) using 1,000,000 cycles of the Bayesian MCMC BBM algorithm. Nodal optimizations were considered unambiguous (unitary) when the maximal probability of any particular optimization exceeded 0.95. When the maximal probability for any particular optimization was less than 0.95, values for all optimizations exceeding 0.01 are presented for interpretation.

Parsimony analysis resulted in 16 shortest trees of 276 step-length; CI of 38, RI of 75. Five nodes within the ingroup Migadopinae collapse in the 283-strep strict consensus tree: 1, four Pseudomigadops spp. are collapsed into a polytomy.; 2, Calyptogonia atra and Nebriosoma fallax and the sister taxa Decogmus chalybeus and Dendromigadops gloriosus are unresolved adelphotaxa to the clade subtended by Monolobus spp.; and 4 and 5, three species each of Loxomerus and Stichonotus comprise tritomies. The equally parsimonious trees also differ in proposed relationships of three outgroups—Blethisa multipunctata, Elaphrus clairvillei, and Loricera foveata—with the consensus tree collapsing the nodes subtending these taxa. Respective monophyly for the four focal genera—Amarotypus, Amaroxenus, Migadopiella, and Amarophilus—is supported.

Bayesian analysis resulted in a majority-rule consensus tree consistent with relationships of the parsimony analysis, though the Bayesian consensus tree is less resolved (Suppl. material 2). As parsimony analysis elucidates characters informing the various relationships, and produces a more resolved consensus cladogram, subsequent analysis and discussion is restricted to the parsimony-based hypothesis (Fig. 5)

The results of the cladistic analysis mirror those of Roig-Juñent (2004) in the placement of Amarotypus— plus affiliated taxa from Tasmania and the South Island of New Zealand recognized as Amarotypini by Johns (2010) and Larochelle and Larivière (2022)—as a subset of taxa nested within Migadopini. Were Amarotypini retained as a valid taxon, this result would render Migadopini paraphyletic. Therefore Amarotypini Erwin (1985) is synonymized under Migadopini. Erwin’s (1985) proposal for Amarotypini was based on the presence in Amarotypus of an elongate unguitractor plate at the apex of the fifth tarsomere (character 21; Fig. 1). Though this synapomorphy is unique and distinctive, its occurrence within the heterobathmy of characters evolving during diversification of Migadopinae restricts its phylogenetic significance to the definition of a subsidiary clade within the tribe Migadopini we name the “Amarotypus clade.”

Monophyly for the family-group names, Migadopinae, Aquilicini, and Migadopini is supported based on the following sets of characters.

Migadopinae. The subfamily is diagnosed by: 1, antennomeres 1–4 glabrous except for apical setae (character 13); 2, presence of a single supraorbital seta in the posterior setal position each side of head (character 14); 3, head without a posterior constriction, or neck (character 16, reversed to presence of a neck in Monolobus); 4, lateral and basal pronotal setae absent (characters 37, 38); 5, male aedeagal internal sac sclerites X and Y (as observed in outgroups and Broscini; e.g. Liebherr et al. 2011, fig. 3b) absent (character 66).

Within Migadopinae, Aquilex diabolica Moret is placed as the sister group to taxa constituting Tribe Migadopini, and thus the status of Aquilicini Moret, 2005 is corroborated. Aquilex is excluded from its adelphotaxon Migadopini, as it lacks the elongate parascutellar stria extended beyond elytral mid-length of Migadopini (Moret 2005, fig. 1; character 45). Aquilex also shares elongate labial paraglossae (character 8) with the outgroups, whereas members of the Migadopini exhibit shortened paraglossae, most extremely when the paraglossae are not well differentiated from the glossa. The male protarsomeres of Aquilex have the basal three tarsomeres bearing squamose setae on the ventral surface as observed in the outgroups, whereas Migadopini have either the four basal tarsomeres with such setae, or tarsomeres 2–4 so clothed (character 23).

Cladistic analysis supports taxonomic recognition of the four genera comprising the Amarotypus clade—Amarotypus, Amarophilus, Migadopiella plus Amaroxenus (Fig. 5)—via three synapomorphic state changes: 1, presence of four glossal setae (character 9); 2, the uniquely synapomorphous presence of an elongate unguitractor plate of the fifth tarsomere (character 26, Fig. 1); and 3, distinctly punctate elytral striae (character 44), though such punctation is polymorphically present or absent in Amarophilus murchisonorum (Suppl. material 1).

Amarotypus stands as sister group to the other three genera, with its monophyly defined by highly derived female gonocoxal morphology, including large macrosetae spanning the juncture between the basal and apical gonocoxite (Fig. 3A), and the presence of four large ensiform mediodorsal setae on the apical gonocoxite (character 70; Fig. 3A). The submentum of Amarotypus also exhibits four setae versus the presence of six setae in the other three genera (Larochelle and Larivière 2022) plus Stichonotus.

Monophyly of the clade comprising the other three genera—Amaroxenus, Amarophilus, and Migadopiella—is unambiguously supported by: 1, posterior pronotal margin without microsetae across breadth (character 33); 2, pronotum with subparallel lateral margins (character 35; though Migadopiella octoguttata shares a basally broader pronotum with species of Amarotypus); 3, female gonocoxite unipartite (character 68; though no female is known for M. octoguttata).

Both Amaroxenus spp. exhibit a cordate, basally constricted pronotum (character 37, Fig. 2B, C), though such a cordate pronotum also occurs in Migadopiella convexipennis. Tarsal characters parsimoniously overrule pronotal shape in this instance, as Migadopiella + Amarophilus synapomorphously possess a symmetrically bilobed protarsomere 4 (character 19), and a truncate metatarsomere 4 (character 21). Moreover, males of Migadopiella spp. synapomorphously exhibit truncate mesotarsomeres 4 (character 20). Migadopiella monophyly is also supported by the male aedeagal ostium opening on the left side of the median lobe (character 59; Baehr 2009, fig. 3).

And to conclude, Amarophilus monophyly is defined by: 1, the presence of foreshortened antennae that reach the base of the elytra (character 12); and 2, at least partially discontinuous elytral striae (character 43), though this condition is also observed in Amarotypus edwardsii. Amarophilus spp., in this analysis, are also distinguished by broadly dilated male protarsomeres (character 22, Larochelle and Larivière 2022, fig. 34). That said, the degree of protarsomere dilation in this clade is evolutionarily plastic, with both Amarotypus murchisonorum (Larochelle and Larivière 2022, fig. 26) and Amaroxenus embersoni (Fig. 2B) of this data set also possessing very broad male protarsomeres.

Subfamily Migadopinae [as Migadopidae] Chaudoir, 1861: 510 (type genus Migadops C. O. Waterhouse, 1842: 136)

Tribe Migadopini Chaudoir, 1861

Monolobina Jeannel, 1938: 13 (type genus Mololobus Solier, 1849: 189; synonymy Roig-Juñent, 2004: 12)

Amarotypini Erwin, 1985: 468 (type genus Amarotypus Bates, 1872: 50; New Synonymy)

Tribe Aquilicini Moret, 2005: 30 (type genus Aquilex Moret, 1989: 246; New Rank)

In order to phylogenetically test the monophyly of the migadopine genera proposed by Baehr (2009) and Larochelle and Larivière (2022), we describe below two species of Amaroxenus that complement the species pairs representing Amarotypus, Amarophilus, and Migadopiella.

Cladistic relationships for Migadopinae versus the closest outgroups included in the phylogenetic analysis (Fig. 5)—represented by Loricera foveata, Blethisa multipunctata, and Elaphrus clairvillei—define a biogeographic hypothesis (Fig. 8) that supports initial diversification of Migadopinae in South America (Fig. 8, node 72). This node ancestral to the Migadopinae is optimized at probability 0.93 to the Holarctic (A), interpreted as a South American origin of the Migadopinae with its sister group in Laurasia.

Within South America, Aquilex is placed as the adelphotaxon to all other Migadopinae (Fig. 8, node 71), supporting recognition of the two sister tribes Aquilicini and Migadopini. Node 71 is ambiguously optimized (Table 2) as Holarctic (A, 0.29), equatorial montane South America (B, 0.39), or southern South America (C, 0.22).

Subsequently, diversification of the migadopine genera, Rhytidognathus, Antarctonomus, Pseudomigadops, Migadops, Migadopidius, and Lissopterus took place within South America (Fig. 8, subtending nodes 70, 69). The RASP analysis then posits that occupation of Australia by the grade of taxa including Calyptogonia, Nebriosoma, Decogmus, and Dendromigadops (Fig. 5) incorporated an ancestral area (Fig. 8, node 68) that can be variously optimized as South America (C, 0.56), Australia (D, 0.23), or the union of South America and Australia (C + D, 0.21) (Table 2).

Such geographic adjacency is reiterated by optimization of node 67 circumscribing the Antipodean areas, Australia, Campbell Plateau, and New Zealand, with ancestral area states including South America (C, 0.74), Australia (D, 0.12), or the union of the two (C + D, 0.10). The Campbell Plateau was isolated (node 66) prior to isolation of Tasmania and New Zealand (node 65), with optimization probabilities of an ancestral area calculated as either Australia (D, 0.61), New Zealand (0.24), or the union of those (D + E, 0.07) (Table 2).

The taxon-area cladogram posits an origin of the Campbell Plateau fauna as sister group to the Tasmanian Stichonotus clade and the amarotypine clade rooted in New Zealand. Node 66 subtending this area relationship is variously optimized to Australia (D, 0.29), New Zealand (E, 0.29), or the Campbell Plateau (F, 0.26) (Table 2), suggesting non-hierarchical area relationships of these terranes.

Subsequent optimization of node 65 supports isolation of Australia from New Zealand, with either Australia (D, 0.61), New Zealand (E, 0.24) or a both areas as ancestral (C + D, 0.7) (Table 2). Ancestral area optimizations for both nodes 66 and 65 strongly suggest an independent history for the Campbell Plateau versus New Zealand west of the Alpine Fault; i.e. North Island and western South Island. The New Zealand clade rooted with Amarotypus sister to the Stichonotus lineage of Tasmania and southeast Australia stands as sister to the Campbell Plateau radiation of Calathosoma, Loxomerus, and Taenarthrus.

The most subordinate area relationship connecting New Zealand and Australia is defined by placement Migadopiella spp. within the New Zealandian Amarotypus, Amaroxenus, and Amarophilus (node 62, fig. 8). RASP optimizes this node to either New Zealand at probability 0.86, or a combined New Zealand plus Australia at a minority probability of 0.11.

Darlington (1965) proposed that flight-capable migadopine taxa such as the South American Antarctonomus plus the Queensland, Australian Dendromigadops represent the ancestral stock of the group, with such winged taxa dispersing between South America and Australia to establish a southern disjunct distribution. Although all three immediate outgroup representatives–Blethisa multipunctata, Elaphrus clairvillei, and Loricera foveata (Fig. 5)–are characterized by macroptery, only the phylogenetically subordinate ingroup taxa Antarctonomus complanatus, Decogmus chalybeus, and Dendromigadops gloriosus exhibit fully developed flight wings. Analogously, Amarotypus edwardsii exhibits brachypterous flight-wings—i.e. reduced in length by approximately half—while being placed within a clade otherwise characterized by totally reduced flight wings; aptery. These several instances may be explained by the demonstrably frequent evolution of flightlessness across Carabidae among taxa occupying montane and islandic habitats (Darlington 1943; Kavanaugh 1985), with that repeated evolution violating the principal of parsimony (Trueman et al. 2004). Conversely, re-evolution of fully developed flight wings from apterous ancestors has been argued for Phasmatodea (Whiting et al. 2003; Forni et al. 2022). Regardless, the earliest diverging lineages within Migadopinae are not characterized by macroptery, and so Darlington’s hypothesis of ancestral dispersal by migadopines across open southern oceans is not corroborated.

Darlington’s (1965) proposal that ancestral colonizing migadopine taxa were flight capable conforms to an interpretation that dispersal to novel terranes occurs via winged beetles. Yet the trans-Tasman colonization of Tasmania by Migadopiella occurred from among a set of wingless taxa. Also, among the Campbell Plateau migadopine taxa, four species—Loxomerus nebrioides, L. katote Johns, L. huttoni, and Calathosoma rubromarginatum (Johns 2010)—occupy the Auckland Islands, with that archipelago comprising Miocene shield volcanoes overlying Cretaceous granite (Denison and Coombs 1977), yet the Pleistocene-aged Antipodes Islands (Scott et al. 2013) house the similarly flightless Loxomerus brevis. The occurrence of L. brevis on the Antipodes is thus best explained by overwater dispersal, with such an instance again occurring within a brachypterous clade.

Based on phylogenetic analysis of extant taxa, we predict future discovery of fossil taxa representing tribe Migadopini from Antarctica. We already have the first example of an Antarctic carabid beetle corroborating a trans-Antarctic biogeographic relationship through the discovery of the fossil, Antarctotrechus balli Ashworth and Erwin (2016). In this instance, Antarctotrechus is a member of a clade within the tribe Trechini comprising Trechisibus Motschulsky of South America, the Falkland Islands and South Georgia, and Tasmanorites Jeannel of Tasmania, Australia. The fossil Antarctotrechus is dated 20–14 Ma, i.e., Early to Mid-Miocene, well after the Oligocene opening of the Southern Ocean, suggesting that it was part of a fauna already evolving in isolation from related taxa on opposite sides of the southern world. It was deposited in materials consistent with mixed forest and tundra vegetation, including Nothofagus (southern beech) and Ranunculus (buttercup), indicating a riparian habitat. The present phylogenetic hypothesis incorporating austral Migadopini (Fig. 5) lays out characters that may be evaluated should a fossil Antarctic migadopine become available.

Recently, a much older Cretaceous amber fossil from Myanmar—Cretomigadops bidentatus (Liu et al. 2023a)—has been described as a member of Migadopinae based on a first instar larva encased within Burmese Kachin amber dated to 99 Ma. The fossil exhibits synapomorphies characterizing Carabidae, though placement as Migadopinae is based solely on the presence of two retinacular teeth on the mandible; a larger tooth in a plesiomorphic position near mid-length on the mandible, and a second smaller tooth more basad along the medial mandibular margin. Such a second mandibular tooth is reported for third instar larvae of Loxomerus brevis and L. nebrioides (Johns 1974), as well as larvae of Omophron Latreille, Tribe Omophronini (Thompson 1979). The Cretaceous Cretomigadops larva differs from Loxomerus larvae in: 1, the second, basal mandibular tooth being much larger relative to the distal tooth; 2, the structure of the antennal sensorium; 2, the very long legs; 3, the slender and very elongate urogomphi; and 4, paired ungual claws of equal length. The occurrence of Cretamigadops in Kachin amber necessitates additional assumptions concerning the historical biogeography based on extant taxa. Confirmation of a trans-Tethyan migadopine distribution can be corroborated through discovery of fossilized adult Migadopinae in Kachin Amber. Of course, more extensive taxonomic representation of known migadopine larval stages among extant taxa would also allow confirmation that the basal, larval retinacular tooth serves as a synapomorphy for Migadopinae. Given that the phylogenetic nexus between the Laurasian outgroups and Gondwanan Migadopinae occurred across what is now the Neotropics (Crowson 1980; McLoughlin 2001), the phylogenetic position of Burmese Migadopinae is predicted to be sister group to Aquilicini + Migadopini. The above-presented hypothesis for Gondwanan vicariance and trans-Tasman dispersal would remain unaffected, with the Aquilicini + Migadopini hypothesized to have evolved on a fragmenting Gondwana.