Molecular Ecology (2003) doi: 10.1046/j.1365-294X.2003.01914.x ' 2003 Blackwell Publishing Ltd Blackwell Publishing Ltd. Phylogeography of the harlequin beetle-riding pseudoscorpion and the rise of the Isthmus of Panam? J . A . ZEH, * D . W. ZEH * and M. M. BONILLA ? * Department of Biology and Program in Ecology, Evolution & Conservation Biology, University of Nevada, Reno, NV 89557, USA; ? Department of Biochemistry, University of Nevada, Reno, USA Abstract Molecular and geological evidence indicates that the emergence of the Isthmus of Panam? influenced the historical biogeography of the Neotropics in a complex, staggered manner dating back at least 9 Myr BP . To assess the influence of Isthmus formation on the biogeog- raphy of the harlequin beetle-riding pseudoscorpion, Cordylochernes scorpioides , we ana- lysed mitochondrial COI sequence data from 71 individuals from 13 locations in Panam? and northern South America. Parsimony and likelihood-based phylogenies identified deep divergence between South American and Panamanian clades. In contrast to low hap- lotype diversity in South America, the Panamanian Cordylochernes clade is comprised of three highly divergent lineages: one clade consisting predominantly of individuals from central Panam? (PAN A), and two sister clades (PAN B1 and PAN B2) of western Panama- nian pseudoscorpions. Breeding experiments demonstrated a strictly maternal mode of inheritance, indicating that our analyses were not confounded by nuclear-mitochondrial pseudogenes. Haplotype diversity is striking in western Atlantic Panam?, where all three Panamanian clades can occur in a single host tree. This sympatry points to the existence of a cryptic species hybrid zone in western Panam?, a conclusion supported by interclade crosses and coalescence-based migration rates. Molecular clock estimates yield a divergence time of ? ? ?? 3 Myr between the central and western Panamanian clades. Taken together, these results are consistent with a recent model in which a transitory proto-Isthmus enabled an early wave of colonization out of South America at the close of the Miocene, followed by sea level rise, inundation of the terrestrial corridor and then a second wave of colonization that occurred when the Isthmus was completed ? ? ?? 3 Myr bp . Keywords : COI gene, Cordylochernes scorpioides , harlequin beetle, mitochondrial DNA, Neotropics, phylogeny Received 7 April 2003; revision received 3 June 2003; accepted 3 June 2003 Introduction In joining North America to the island continent of South America, the rise of the Isthmus of Panam? was a geological event with profound consequences for the biogeography of the Neotropics (Simpson 1980). On land, the uplifting of the Isthmus enabled the ?Great American Interchange?: a two-way migration and intermingling of terrestrial lineages that had been separated since the Cretaceous (Marshall et al . 1979; Stehli & Webb 1985; Pindell & Barrett 1990). In the ocean, the closure of the Central American Seaway severed gene flow and brought about evolutionary divergence between populations of marine species in the eastern Pacific and tropical western Atlantic Oceans. The resulting transis- thmian sister-species pairs (?geminate species?, Jordan 1908) have been used extensively to infer rates of molecular evolu- tion in a wide range of marine taxa, including snapping shrimp (Knowlton et al . 1993; Knowlton & Weigt 1998), echinoderms (Lessios 1998) and fishes (Bermingham et al . 1997). Studies of multiple transisthmian species pairs are providing growing evidence that cessation of gene flow did not occur simultaneously across pairs of sister species but was a staggered process spanning several million years (Bermingham & Lessios 1993; Knowlton et al . 1993; Knowlton & Weigt 1998; Marko 2002). These findings are consistent with palaeogeographical reconstructions that indicate a gradual formation of the Isthmus from its emergence as an Correspondence: Jeanne A. Zeh. Fax: +1 775 784 1302; E-mail: jaz@unr.edu 2 J . A . Z E H , D . W . Z E H and M . M . B O N I L L A ' 2003 Blackwell Publishing Ltd, Molecular Ecology , 10.1046/j.1365-294X.2003.01914.x island chain in the middle to late Miocene (15?6 Myr bp ) to its completion as a terrestrial corridor ? 3 Myr bp (Coates & Obando 1996; Coates et al . 2003). The marine fossil record suggests equally gradual uplift-associated extinctions extending back more than 10 Myr (Jackson et al . 1993; Collins 1996; Vermeij 2001). Nonetheless, it is still com- monly assumed that the final seaway closure date of 3 Myr bp provides a reliable basis for molecular clock calibrations (e.g. Baldwin et al . 1998; Metz et al . 1998; Stillman & Reeb 2001). Obviously, this assumption can yield greatly overestimated rates of nucleotide sequence divergence in marine geminate species whose isolation predated completion of the Isthmus (Knowlton et al . 1993; Knowlton & Weigt 1998; Marko 2002). For terrestrial species, there is also ample evidence that the emergence of the Isthmus influenced the biogeography of the region in a complex, staggered manner dating back at least 9 Myr (Stehli & Webb 1985). The mammalian fossil record indicates a limited, but significant, exchange of taxa between North and South America during the early stages of Isthmus formation. Giant sloths of South American origin first appear in North America between 9.5 and 9.0 Myr bp , whereas procyonids (the racoon family) of North American origin are present in South America by 6.0 Myr bp (Marshall et al . 1979). This movement of terrestrial taxa between the two continents long before final completion of the land bridge has generally been attributed to ?island- hopping? dispersal along the emerging island chain (Simpson 1950). However, the island-hopping model for coloniza- tion that predated completion of the Isthmus provides a questionable explanation for the occurrence in extreme western Panam? of mitochondrial lineages of freshwater fishes estimated to have diverged from their putative South American source populations ? 5 Myr bp (Bermingham & Martin 1998). Marine conditions present a dispersal barrier to primary freshwater fishes (Myers 1938). As an altern- ative to the island-hopping model, Bermingham & Martin (1998) therefore proposed that formation of the Isthmus involved the emergence of a short-lived terrestrial corridor at the close of the Miocene (4?7 Myr bp ). Compared with what is known for marine species and terrestrial vertebrates, the impact of the rise of the Isthmus on the historical biogeography of terrestrial invertebrates is less well understood. Here, we present the results of a phylogeographical study of the harlequin beetle-riding pseudoscorpion, Cordylochernes scorpioides. This pseudo- scorpion has proved to be an excellent model system for molecular, laboratory and field investigations of intra- population genetic incompatibility (Zeh 1997; Newcomer et al . 1999; Zeh & Zeh 2001), sexual selection (Zeh et al . 1997, 1998) and speciation (Zeh & Zeh 1994, 2000; Wilcox et al . 1997). This study significantly extends the findings of a previous investigation based on a limited geographical sampling of four populations (Wilcox et al . 1997). Phylo- genetic analysis of mitochondrial DNA (mtDNA) sequence data from 71 individuals sampled from 13 locations in Panam? and northern South America reveals a pattern of deep divergence between C. scorpioides clades and a geo- graphical distribution of mtDNA haplotypes consistent with the history of colonization, extinction, divergence and re-colonization hypothesized by the Bermingham/Martin (B/M) model of the formation of the Isthmus of Panam?. Materials and methods Collections Pseudoscorpions for this study were collected from locations in north central Trinidad and eastern French Guiana and from 11 sites in Panam? (Table 1; Fig. 1). The Isthmus of Panam? is transected by a central cordillera that reaches an Table 1 Locality data of 71 Cordylochernes scorpioides individuals sequenced for the phylogeographical component of the study. WPP, western Pacific Panam?; WAP, western Atlantic Panam?; CP, central Panam? Collection site Code Latitude/longitude Popn No. of specimens Ojo de Agua, Chiriqu? Province, Panam? ODA 08? 44? N/82? 44? W WPP 10 Bocas del Toro Province, Panam? BDT 09? 00? N/82? 16? W WAP 14 La Fortuna, Chiriqu? Province, Panam? FORT 08? 43? N/82? 14? W WAP 3 Tol?, Chiriqu? Province, Panam? TOLE 08? 14? N/81? 42? W CP 1 Cerro Campana, Panam? Province, Panam? CC 08? 44? N/79? 58? W CP 4 Barro Colorado Island, Colon Province, Panam? BCI 09? 09? N/79? 50? W CP 3 Fort Sherman, Colon Province, Panam? FS 09? 17? N/79? 56? W CP 5 Bohio Peninsula, Colon Province, Panam? BOHIO 09? 12? N/79? 50? W CP 2 Camino Madden, Panam? Province, Panam? MAD 09? 05? N/79? 37? W CP 9 R?o Platanares, Panam? Province, Panam? PLAT 09? 13? N/79? 00? W CP 1 El Llano Cart?, Panam? Province, Panam? LC 09? 18? N/78? 58? W CP 8 Blanchisseuse Road, Trinidad TRIN 10? 45? N/61? 19? W TRIN 6 Kaw Mountains, French Guiana FG 04? 32? N/52? 04? W FG 5 C O R D Y L O C H E R N E S P H Y L O G E O G R A P H Y 3 ' 2003 Blackwell Publishing Ltd, Molecular Ecology , 10.1046/j.1365-294X.2003.01914.x elevation of 3000 m in western Panam? and decreases to 200 m in central Panam?. Our collections were made over a region spanning 418 km from central to western Panam? where sites were located on either the Atlantic (northern) or the Pacific (southern) versant of the cordillera. Pseudoscor- pions were collected from decaying trees ( Brosimum spp. or Ficus spp. in Panam? and Trinidad; Parahancornia fasciculata in French Guiana) or were removed from harlequin beetles captured on newly dead or dying trees. Specimens were maintained alive before freezing at - 80 ? C. DNA extraction, polymerase chain reaction and sequencing Total genomic DNA was extracted from each individual using the 2 ? CTAB protocol described previously (Zeh et al . 1992a). An ? 450 bp fragment of the mitochondrial cytochrome oxidase subunit 1 ( COI ) gene, corresponding to the region between 1751 and 2191 bp in the Drosophila yakuba mtDNA sequence (Simon et al . 1994), was amplified using a degenerate version of the primer C1-J-1751 (alias Ron) (5 ? -GGAKCACCTGATATAGCATTYCC-3 ? ) and the primer C1-N-2191 (alias Nancy) (5 ? -CCCGGTAARATT- AAAATATAAACTTC-3 ? ) (Simon et al . 1994). The 25 m L polymerase chain reaction (PCR) mix contained ? 10 ng of genomic DNA, 50 m m KCl, 10 m m Tris?HCl (pH 9.0), 0.1% Triton X-100, 2 m m MgCl 2 , 0.1 m m dNTPs and 0.5 units Titanium Taq DNA polymerase (Clontech). PCR amplification conditions involved an initial 2 min melting step at 94 ? C, followed by 32 iterations of the following cycle: 94 ? C for 60 s, 52 ? C for 90 s and 72 ? C for 105 s, with a final 7- min extension at 72 ? C. PCR templates were prepared for sequencing by electrophoresing 12 m L of the reaction through a 1% agarose gel (Gibco-BRL) stained with ethidium bromide. Amplification products were excised from gels and purified using Promega Wizard minicolumns. BigDye sequencing reactions (6 m L) containing ? 100 ng of purified PCR product and 3.2 pmoles of the C1-J-1751 primer were analysed using an ABI Prizm 3730 automated sequencer, according to the manufacturer?s protocols (PE Applied Biosystems). Inheritance of mtDNA haplotype To test for the possible presence of nuclear-mitochondrial pseudogenes (Numts) which can confound phylogenetic analysis (Bensasson et al. 2001), we examined the pattern of sequence inheritance (maternal or biparental) in several sets of offspring derived from matings between males and females of differing mtDNA haplotype. These offspring sequences were not included in the phylogenetic and gene flow analyses described below. Phylogenetic reconstruction Both morphological and molecular evidence indicates that Lustrochernes is the genus most closely related to Cordylochernes (Muchmore 1974; Wilcox et al. 1997), and L. consocius was therefore used as the outgroup in phylogenetic analyses in order to root the C. scorpioides trees. Maximum parsimony (MP) and maximum likelihood (ML) searches were carried out on aligned sequences, as implemented in paup Version 4.0b8 (Swofford 2002), and all minimal trees were saved. Hierarchical likelihood ratio tests were performed using modeltest Version 3.06 (Posada & Crandall 1998) in order to determine the best-fit ML substitution model available from the set of 56 models implemented Fig. 1 Collection sites and geographical distribution of Cordylochernes mtDNA clades. Numbers within clade symbols indicate the number of individuals collected at the site belonging to that clade (see Fig. 2 for clade composition). Collection site codes, ODA: Ojo de Agua, BDT: Bocas del Toro, FORT: La Fortuna, TOLE: Tol?, CC: Cerro Campana, FS: Fort Sherman, BCI: Barro Colorado Island, BOH: Bohio Peninsula; MAD: Camino Madden, PLAT: Rio Platanares, LC: El Llano Cart? Road, TRIN: Trinidad, FG: French Guiana (see Table 1). 4 J . A . Z E H , D . W . Z E H and M . M . B O N I L L A ' 2003 Blackwell Publishing Ltd, Molecular Ecology, 10.1046/j.1365-294X.2003.01914.x in the program. For both tree estimation procedures, sup- port for each clade was estimated by analysing 500 boot- strap replicate datasets (Felsenstein 1985). Pairwise genetic distances between haplotypes were calculated using both Kimura?s 2-parameter model (K2P) and maximum likelihood methods. Finally, sequences were translated in mcclade Version 4.05 (Maddison & Maddison 2002) using the invertebrate mitochondrial code for analysis of amino acid substitutions. Migration rate estimation We used a maximum likelihood approach based on coale- scence theory (Beerli & Felsenstein 2001) to estimate patterns of gene flow between C. scorpioides populations. This method assumes neither symmetric migration rates nor equal subpopulation sizes and allows for the incorporation of both coalescence within subpopulations and migration events that switch lineages from one population to another (Beerli & Felsenstein 2001). Data were analysed using migrate (Beerli 2002), a program that employs a Metropolis-Hastings Markov chain Monte Carlo algorithm (MH) to estimate ML effective population sizes (Q = 2Nfm ) and migration rates (2Nfm), where Nf is the female effective population size, m is the mutation rate per generation per site, and 2Nfm is the number of migrant females per generation. In our migrate simulations, we used an empirically determined Ti/Tv (calculated from the ML tree using the ?state changes and stasis? function in mcclade), empirical base frequencies and starting parameters based on FST estimates calculated using the program (Beerli 2002). The search strategy for obtaining the ML estimates involved 10 short chains with 10 000 sampled trees and 3 long chains with 100 000 sampled trees. To obtain more precise estimates, a second analysis was performed with parameter estimates from the first run as starting values. Results Haplotypes Examination of the mtDNA COI sequences of the 71 Cordylochernes scorpioides field-collected individuals identified 24 different haplotypes (GenBank Accession nos AY332244 to AY332267), with K2P pairwise distances ranging from 0.23 to 15.44%. Distances estimated using maximum likelihood were appreciably higher, particularly for more divergent sequences, and ranged from 0.23 to 32.61% (Table 2). The ML genetic distances were calculated, based on modeltest selection of the TrN + I substitution model (Akaike Information Criterion; Posada & Crandall 1998), with a rate matrix of A fi C = 1.00; A fi G = 12.58; A fi T = 1.00; C fi G = 1.00; C fi T = 14.07; G fi T = 1.00 and a proportion of invariable sites (I) equal to 0.73. Within French Guiana (FG), Trinidad (TRIN) and central Panam?, sequence variation was quite limited. A single haplotype was found in the six specimens from Trinidad, whereas the five individuals from French Guiana yielded two haplotypes that differed by a single nucleotide. Similarly, although 9 distinct haplotypes were identified among the 33 C. scorpioides collected from central Panam?, they differed by a maximum of only 3 nucleotides. By contrast, in western Panam?, mtDNA sequence variation was extremely high. Among the 17 individuals sampled from this region, we identified Table 2 Pairwise genetic distances between major clades of Cordylochernes scorpioides. Distances were calculated from 439 nucleotide sites and corrected for multiple hits using either Kimura?s 2-parameter model (K2P) or a maximum likelihood (TrN + I) substitution model (see Results). Approximate divergence times (Myr) are based on a divergence rate of 2.3% per Ma calculated for arthropod mtDNA (Brower 1994) Comparison Kimura?s 2-parameter distances Maximum likelihood distances (TrN + I) Mean distance (%) Range (%) Mean divergence (Myr) Range (Myr) Mean distance (%) Range (%) Mean divergence (Myr) Range (Myr) South America vs. Panam? 13.26 11.72?15.44 5.77 5.10?6.71 25.03 22.22?32.61 10.88 9.66?14.18 Panam? A vs. Panam? B 7.15 6.22?7.76 3.11 2.70?3.38 9.82 8.21?10.76 4.27 3.57?4.68 Panam? B1 vs. Panam? B2 3.99 3.27?4.48 1.74 1.42?1.95 4.78 3.81?5.49 2.08 1.66?2.38 South America vs. Panam? A 12.62 11.72?13.45 5.49 5.10?5.85 23.57 22.22?25.46 10.25 9.66?11.07 South America vs. Panam? B 14.31 13.67?15.44 6.22 5.94?6.71 27.41 24.34?32.61 11.92 10.58?14.18 South America vs. Panam? B1 14.59 13.67?15.44 6.34 5.94?6.71 29.61 27.37?32.61 12.87 11.90?14.18 South America vs. Panam? B2 14.02 13.72?14.36 6.10 5.96?6.24 25.22 24.34?26.39 10.96 10.58?11.47 Panam? A vs. Panam? B 7.15 6.22?7.76 3.11 2.70?3.38 9.82 8.21?10.76 4.27 3.57?4.68 Panam? A vs. Panam? B1 6.99 6.22?7.49 3.04 2.70?3.26 9.67 8.21?10.76 4.20 3.57?4.68 Panam? A vs. Panam? B2 27.31 6.46?7.76 3.18 2.81?3.38 9.97 8.49?10.65 4.33 3.69?4.63 Panam? B1 vs. Panam? B2 3.99 3.27?4.48 1.74 1.42?1.95 4.78 3.81?5.49 2.08 1.66?2.38 C O R D Y L O C H E R N E S P H Y L O G E O G R A P H Y 5 ' 2003 Blackwell Publishing Ltd, Molecular Ecology, 10.1046/j.1365-294X.2003.01914.x 16 different haplotypes distinguishable by up to 33 nucleo- tide substitutions (Fig. 2). Of these, 12 haplotypes were unique to western Panam? and 4 were almost identical to haplotypes found in individuals sampled from central Panam?. Phylogenetic analysis We used unambiguously aligned sequence from Lustrochernes consocius (Wilcox et al. 1997) to root phylogenetic trees of the 24 different C. scorpioides haplotypes. Unweighted MP analysis identified a South American clade (SA = FG + TRIN; Fig. 2) as the strongly supported monophyletic sister taxon of a more diverged and polymorphic Panamanian clade (bootstrap = 100%; tree length = 162; CI = 0.815). The South American (SA) and Panamanian (PAN) pseudoscorpions are separated by mean K2P and ML distances of 13.26 and 25.03%, respectively (Table 2). Within the Panamanian clade, C. scorpioides haplotypes are further divided into two main groups, a clade consisting largely of individuals from central Panam? (PAN A) and a clade composed entirely of individuals from western Panam? (PAN B). Mean K2P and ML distances between the PAN A and PAN B clades are 7.15 and 9.82%. The PAN B clade is further subdivided into two well-differentiated, monophy- letic lineages, PAN B1 and PAN B2, separated by mean K2P and ML distances of 4.00 and 4.78%. These two clades correspond loosely to western Atlantic Panam? (WAP) and western Pacific Panam? (WPP), respectively (Fig. 2). Support for the three clades within Panam? was strong, with bootstrap values ranging from 91 to 100%. Rooting of trees on L. consocius proved problematic for ML analyses based on the two substitution models chosen by modeltest. The first model, selected by the Akaike Information Criterion, was TrN + I with a rate matrix of A fi C = 1.00; A fi G = 12.69; A fi T = 1.00; C fi G = 1.00; C fi T = 11.16; G fi T = 1.00 and a proportion of invariable sites equal to 0.67 (?LnL = 1352.67). The second model, selected by the hierarchical likelihood ratio test, was K81uf + G with a substitution rate matrix of A fi C = 1.00; A fi G = 13.36; A fi T = 1.20; C fi G = 1.20; C fi T = 13.36; G fi T = 1.00 and a gamma parameter equal to 0.13 (?LnL = 1353.84). The two models recovered the same poorly resolved and weakly supported tree (data not shown) in which the South American C. scorpioides clade is placed between the PAN B1 and PAN A clades. The ML tree was appreciably longer (17 steps) and had a higher level of homoplasy than the MP trees (tree length = 179; CI = 0.737). Because the L. consocius and the C. scorpioides haplotypes exhibit high levels of sequence divergence at third positions (mean uncorrected pairwise distance equal to 45.4%), we investigated the possibility that the ML tree Fig. 2 Cordylochernes phylogenetic relation- ships based on 24 unique mtDNA COI haplotypes of pseudoscorpions sampled from French Guiana, Trinidad and Panam?. The phylogram is one of 28 most parsi- monious trees and is topologically equivalent to the maximum likelihood tree (TrN + I substitution model) of the Cordylochernes clade (see Results). Numbers above branches indicate number of nucleotide substi- tutions (but note that terminal branches of length 1 are unlabelled). Values below branches represent bootstrap estimates (> 80%). Haplotypes are identified by collection site, with numbers in parentheses indicating number of individuals from a particular site that exhibit the haplotype. For collection site codes, see Fig. 1. The Cordylochernes tree was rooted using sequence from Lustrochernes consocius (Wilcox et al. 1997). 6 J . A . Z E H , D . W . Z E H and M . M . B O N I L L A ' 2003 Blackwell Publishing Ltd, Molecular Ecology, 10.1046/j.1365-294X.2003.01914.x was misleading as a consequence of substitutional satura- tion at third-positions. Restriction of phylogenetic analysis to first and second positions recovered a single MP (tree length = 21; CI = 0.952) and a single ML tree (?LnL = 512.67; HKY substitution model with Ti/Tv = 4.43) that were identical in topology and that placed the TRIN and FG haplotypes as outgroup taxa to the PAN A and PAN B clades. When L. consocius was excluded from analysis and trees were rooted on the SA clade, MP and ML analyses carried out on all three codon positions produced concordant results. Of 439 total characters, 81 were polymorphic and 75 were parsimony informative, with an average nucleotide composition of 24.01% A, 17.65% C, 16.39% G and 41.94% T. A branch and bound parsimony analysis produced 28 MP trees that differed only slightly in topo- logy (tree length = 111; CI = 0.820). Likelihood analysis, based on the TrN + I substitution model described above (see ?Haplotypes?), produced a most probable tree (?lnL = 1153.52; Fig. 2) that was almost identical to the C. scorpioides topology obtained from MP analysis of the Lustrochernes- rooted tree shown in Fig. 2. To test for rate constancy of nucleotide substitutions in C. scorpioides, we carried out a log-likelihood ratio test using paup. This analysis indicated that evolutionary substitutions in C. scorpioides mtDNA occur in an approximately clock-like fashion (?ln = 1162.50 clock enforced tree vs. 1153.52 nonclock tree, c 2 = 17.96, d.f. = 22, P > 0.10). Inheritance of mtDNA haplotype The translated sequences exhibited high amino acid sequence homology with mtDNA COI proteins from other arthropods (data not shown) and no missense or stop codons were found. The pattern of inheritance of mtDNA haplotype was examined in 12 crosses between males and females of differing mtDNA haplotype, as follows: (i) PAN A maternal haplotype ? PAN B1 paternal haplotype (N = 8); (ii) PAN B1 maternal haplotype ? PAN A paternal haplotype (N = 3); and (iii) PAN A maternal haplotype ? PAN B2 paternal haplotype (N = 1). In all the 24 offspring assayed (1 male and 1 female offspring per cross), DNA sequencing demon- strated maternal inheritance of the mtDNA COI sequence. Taken together, these data provide clear evidence that our sequence analyses were not confounded by nuclear- mitochondrial pseudogenes. Gene flow estimation To estimate gene flow, C. scorpioides sequences were initially assigned to one of three populations representing major geographical regions: (i) South America (SA, comprising Trinidad and French Guiana), (ii) central Panam? (CP), and (iii) western Panam? (WP). This analysis was carried out using a three-island, full migration model, in which there were no a priori restrictions on gene flow. Pairwise estimates indicate moderate, unidirectional gene flow from central to western Panam? (MLE, 2Nfm = 1.507; see Table 3). Migration between the two Panamanian populations and South America was estimated to be either nonexistent or extremely low. Based on calculation of the 95% profile confidence intervals, as implemented in migrate (Table 3), only the South America to central Panam? migration rate was estimated to be significantly greater than zero (MLE, 2Nfm = 0.0467). Given these initial results, CP was specified as the source population in a second migrate simulation was run to estimate gene flow between the Panamanian popula- tions. For this analysis, WP was subdivided into the two geographical areas separated by the central cordillera: western Atlantic Panam? (WAP) and western Pacific Panam? (WPP). Gene flow from central to western Panam? was highest on the Pacific slope of the cordillera (CP fi WPP, 2Nfm = 1.2613; CP fi WAP, 2Nfm = 0.6090). Within western Panam?, gene flow across the cordillera was unidirectional Table 3 Maximum likelihood estimates of gene flow between Cordylochernes scorpioides populations from South America (SA), central Panam? (CP) and western Panam? (WP). The analysis was carried out using an unrestricted migration matrix model with variable subpopulation size. The ML estimate (bold) and the 95% profile confidence intervals (in parentheses, below) are shown for population sizes (Q = 2Nfm ) and number of immigrant females per generation (2Nfm), where Nf is the female effective population size and m is the mutation rate per generation per site Population i Q South America fi i Central Panam? fi i Western Panam? fi i South America 0.0059 ? 0.0000 0.0000 (0.0035?0.0115) (0.0000?0.1430) (0.0000?0.2718) Central Panam? 0.0037 0.0467 ? 0.0000 (0.0026?0.0094) (0.0351?0.2048) (0.0000?0.0895) Western Panam? 0.0220 0.0000 1.5070 ? (0.0144?0.0433) (0.0000?0.5626) (0.5487?3.1605) C O R D Y L O C H E R N E S P H Y L O G E O G R A P H Y 7 ' 2003 Blackwell Publishing Ltd, Molecular Ecology, 10.1046/j.1365-294X.2003.01914.x from the Pacific to the Atlantic side (WPP fi WAP, 2Nfm = 0.9770; WAP fi WPP, 2Nfm = 0.0000 (Table 4). Discussion Classical taxonomy based on morphological characters classifies many plants and animals as single species with pan-neotropical distributions. For example, the fig tree, Ficus insipida (Croat 1978), the boas, Boa constrictor, Corallus enydris and Epicrates cenchria (Henderson & Hedges 1995), the macaws, Ara macao (Ridgely 1981) and Ara militaris (Wege & Long 1995), and the paca, Agouti paca (Eisenberg 1989) are each classified as a single species, ranging through- out the Neotropics from Central America to southern Brazil or northern Argentina. However, the validity of such morphologically based species designations is being called into question, as a growing number of molecular phylogeo- graphical studies demonstrates pronounced genetic differen- tiation over relatively short distances in several putative species of neotropical birds (Seutin et al. 1993), freshwater fishes (Bermingham & Martin 1998; Perdices et al. 2002), frogs (Ryan et al. 1996) mammals (da Silva & Patton 1998) and snakes (Henderson & Hedges 1995). Similarly, this phylogeographical study of Cordylochernes scorpioides suggests that the species status of many terres- trial arthropods with apparently pan-neotropical distri- butions may also need to be reassessed. The harlequin beetle-riding pseudoscorpion was described by Beier (1948) as a single species, ranging from Costa Rica to southern Brazil, on the basis of his morphological examination of hundreds of specimens from several countries in South and Central America. However, the results of our mtDNA COI sequencing study demonstrate extensive genetic differentiation between C. scorpioides populations from Panam? and northern South America, with a maximum ML nucleotide divergence of almost 33%. This extreme level of divergence is associated with complete postzygotic incompatibility between C. scorpioides individuals from central Panam? and both French Guiana (Zeh & Zeh 1994) and Trinidad (JA Zeh, unpublished data). Not surprisingly, our migrate simulations suggest essentially no gene flow between the South American and Isthmian populations. Populations of C. scorpioides from South and Central America also differ dramatically in levels of haplotype diversity. Whereas sequence variation both within and between populations in Trinidad and French Guiana is quite limited, the Panamanian isthmus supports three polymorphic and highly divergent clades (PAN A, PAN B1 and PAN B2), with mtDNA haplotypes that differ by up to 11% in ML genetic distance. In the Bocas del Toro province of western Atlantic Panam?, the level of within-population sequence polymorphism is particularly striking, with rep- resentatives of all three Panamanian clades collected from a single decaying Ficus tree. Likewise, haplotypes from both the PAN B1 and PAN B2 clades were found in a group of three WAP pseudoscorpions dispersing together on a harlequin beetle. Only in the central region of Panam? was sequence diversity limited, with all individuals clustering into the relatively uniform PAN A clade. The sympatry of the three highly divergent mtDNA haplotype clades in western Panam? points to the existence of a cryptic species boundary in this region. Consistent with this inter- pretation are preliminary results from ongoing laboratory crosses that indicate partial breakdown in postzygotic reproductive compatibility between individuals from the PAN A, PAN B1 and PAN B2 clades (JA Zeh, unpublished data). In addition, the pattern of gene flow that emerges from our coalescent analysis of migration rates indicates moderate levels of gene flow from central to western Panam? along both sides of the central cordillera. Within western Panam?, gene flow appears to be unidirectional from the Pacific to the Atlantic versant of the cordillera. Presumably, western Panam? represents the southeastern limit of the PAN B clade, and sampling of C. scorpoides populations from Costa Rica is now required to determine the geographical distribution of this mtDNA lineage. Table 4 Maximum likelihood estimates of gene flow between Cordylochernes scorpioides populations from central Panam? (CP), western Atlantic Panam? (WAP) and western Pacific Panam? (WPP). Based on the results shown in Table 3, the analysis was performed with central Panam? specified as the source population. Abbreviations as in Table 3 Population i Q Central Panam? fi i Western Atlantic Panam? fi i Western Pacific Panam? fi i Central Panam? 0.0076 ? ? ? (0.0053?0.0164) Western Atlantic Panam? 0.0069 0.6090 ? 0.9770 (0.0033?0.0169) (0.4318?2.3606) (0.1821?3.5262) Western Pacific Panam? 0.0149 1.2613 0.0000 ? (0.0059?0.0558) (0.3626?5.3773) (0.0000?0.0006) 8 J . A . Z E H , D . W . Z E H and M . M . B O N I L L A ' 2003 Blackwell Publishing Ltd, Molecular Ecology, 10.1046/j.1365-294X.2003.01914.x Our analysis of mtDNA evolutionary rate constancy failed to reject a molecular clock for Cordylochernes, indicat- ing that genetic distances can be used to estimate approx- imate divergence times between C. scorpioides clades. We therefore used a molecular clock approach to provide a temporal framework for assessing the diversification of C. scorpioides mitochondrial lineages within the context of the major geological events that shaped the formation of the Isthmus of Panam?. The absence of appropriate data for direct calibration of a Cordylochernes mtDNA clock necessitated reliance on external rate calibration. We adopted the calibration of 2.3% pairwise divergence per million years, estimated by Brower (1994) on the basis of uncorrected or K2P mtDNA nucleotide distances in a vari- ety of recently diverged arthropods. When applied to the K2P distances for C. scorpioides, this rate resulted in an esti- mate of divergence time between populations from South America and Panam? that corresponds approximately to the Miocene/Pliocene boundary (mean clade divergence of 5.3 Myr; Table 2). Within the Panamanian clades, esti- mated divergence times were 2.9 Myr between central and western Panam? (PANA and PANB), and 1.6 Myr between western Atlantic Panam? and western Pacific Panam? (PAN B1 and PAN B2). The 2.3% calibration yielded higher divergence times when applied to the ML nucle- otide distances, particularly for the more diverged clades: 10.0 Myr between the clades from South America and Pan- am?; 3.9 Myr between central and western Panamanian clades, and 1.9 Myr between the two western Panamanian clades. These ML-based estimates, although still providing western and central Panam? divergence times consistent with the B/M model, may also shed light on the deep divergence between C. scorpioides from Panam? and South America east of the Andes mountains. The 10 Myr, ML- based age of the split between South American and Pana- manian C. scorpioides lineages corresponds approximately to the timing of final uplifting of the Andes, calculated from geological data to have occurred 8.5?8 Ma (Lundberg et al. 1998). However, it is important to note that the diver- gence times based on ML genetic distances are highly tentative, given the absence of an appropriate maximum likelihood calibrated clock for terrestrial arthropods. The biogeographical distribution of C. scorpioides mtDNA lineages detected in this study yields a phylogenetic signal of the geological events associated with the emergence of the Isthmus of Panam?. Our analysis reveals a deep split between western and central Panamanian clades that pro- vides compelling evidence for an early colonization event followed by isolation and re-colonization of the Isthmus. The estimated divergence time between the two main Pan- amanian clades is largely consistent with the emergence of a short-lived terrestrial corridor at the close of the Miocene (5?7 Myr bp), as proposed by Bermingham & Martin (1998). This would have enabled a first wave of colonization out of South America, across the corridor and into what is now Costa Rica. Subsequent inundation of this nascent Isthmus during the Pliocene sea level rise (Haq et al. 1987) would have resulted in mass extinction of isthmian terrestrial and freshwater taxa. According to the B/M model, colonists surviving in Costa Rica?s high-elevation Talamanca region became genetically isolated from source populations in northwestern South America. Interestingly, the most highly diverged of the C. scorpioides mtDNA haplotypes in this study (the PAN B clade) were collected from western Panama, near the eastern edge of the Talamanca range. The B/M model then posits a second, geographically more lim- ited wave of colonization that reached western Panam?. This coincided with final completion of the Isthmus ? 3 Myr bp (Coates & Obando 1996), after convergence of the South American and Caribbean plates resulted in uplift- ing of eastern Panam? and northwestern Colombia. We propose that a second wave of colonization out of north- western South America is responsible for the current dis- tribution of the C. scorpioides PAN A clade that extends across much of the Isthmus as far as western Panam?. The absence of phylogeographical structure within central Panam? suggests that the late-Pliocene colonization of the Isthmus by C. scorpioides occurred with a rate of spread even more rapid than that documented for freshwater fishes (Bermingham & Martin 1998), presumably as a con- sequence of the dispersal ability of the pseudoscorpion?s harlequin-beetle host (see Zeh et al. 2003). Because the Bocas del Toro region of western Panama harbours the greatest diversity of mtDNA haplotypes, it might be argued that this region represents a ?centre of dispersal? and that the posited second wave of colonization of the Isthmus took place in a west-to-east direction, that is, with an origin in Costa Rica rather than in South America. However, this hypothesis is not consistent with the results obtained from our migrate analysis. Unlike traditional FST-based methods for estimating migration rates, the coalescent-based, maximum-likelihood approach in migrate is capable of evaluating not only the magnitude, but also the direction of gene flow. Results from the migrate simu- lations identified central Panam? as the source of the PAN A (central Panam?) clade individuals in western Panam? and not vice versa. Moreover, the western origin hypo- thesis requires additional, ad hoc explanations to account for the absence of the PAN B1 and B2 clades in central Panam?. It therefore seems to provide a less parsimonious explanation for the geographical distribution of C. scorpioides haplotypes in Panam? than the South American origin hypothesis. Nonetheless, determining the mtDNA haplotypes of indi- viduals from Costa Rica and northwestern Colombia is clearly required to resolve this issue. Intriguingly, a shared history defined by the geological development of the Isthmus has led to very different genetic consequences in the case of the pseudoscorpion?s C O R D Y L O C H E R N E S P H Y L O G E O G R A P H Y 9 ' 2003 Blackwell Publishing Ltd, Molecular Ecology, 10.1046/j.1365-294X.2003.01914.x obligate dispersal agent, the giant harlequin beetle, Acrocinus longimanus (Beier 1948; Zeh & Zeh 1992). Throughout its range, the pan-neotropical A. longimanus co-occurs with C. scorpioides and is a pioneer species in the diverse com- munity of saproxylic invertebrates that utilizes decaying trees in the families Moraceae and Apocynaceae. C. scorpioides gains access to these rich but patchily distributed and ephemeral habitats by hitchhiking under the elytra of the beetle (Zeh & Zeh 1992). Because of this obligate associ- ation, pseudoscorpion colonization is restricted to the brief period when newly fallen or dying trees attract A. longimanus males and females for mating and oviposi- tion (Zeh et al. 1992b). The C. scorpioides populations in dead trees then remain marooned for two or three generations until the harlequin beetle larvae complete development and the pseudoscorpions can climb on board newly eclosed adult beetles to disperse en masse. Because C. scorpioides? capacity for colonization is strictly dictated by dispersal of the harlequin beetle, it was anti- cipated that the pattern and depth of cladogenesis in the pseudoscorpion and the beetle would yield clear phylo- genetic signals of shared biogeographical history. This proved not to be the case. A recent mtDNA COI phylogenetic study of A. longimanus (Zeh et al. 2003) produced a phylo- geny of the harlequin beetle topologically similar to that of its hitchhiking pseudoscorpion. However, there is no apparent congruence between timing of branching events in the harlequin beetle phylogeny and either the temporal pattern of phylogenesis in the pseudoscorpion or the B/M model of dispersal and vicariance. Even with adjustment for an approximate twofold difference in generation time, the pseudoscorpions still exhibit estimated divergence times that are an order of magnitude greater than those in the beetles (Zeh et al. 2003). Such a lack of concordance could be explained by a pseudoscorpion/harlequin beetle relationship that evolved only after the first postulated colonization of Central America from South America 5?7 Ma. However, this hypothesis seems unlikely and is not parsimonious, involving, as it would, multiple, independent origins of the complex suite of behaviours associated with dispersal on harlequin beetles (Zeh & Zeh 1992). An alternative hypothesis invokes difference between C. scorpioides and A. longimanus in their mode of reproduction (Zeh et al. 2003). Whereas the pseudoscorpion is viviparous (Zeh 1997), the harlequin beetle is oviparous. It has been argued elsewhere that, because viviparity involves an intimate relationship between the developing embryo and its mother, postzygotic isolation should evolve more rapidly in live-bearing species than in oviparous ones (Zeh & Zeh 2000, 2001). Consequently, the threshold above which genetic divergence translates into genetic incompatibility and reproductive isolation may be significantly lower in the viviparous pseudoscorpion than in the egg-laying beetle. In the context of the reproductive mode hypothesis, the ? 3 Myr hiatus between the postulated early wave of colonization across the Panamanian terrestrial corridor and the second wave appears to have resulted in genetic divergence sufficient to create barriers to gene flow in C. scorpioides but not in A. longimanus. We hypothesize that, in the harlequin beetle, the phylogenetic signal of col- onization and vicariance associated with the formation of the Isthmus of Panam? has been obscured, though not fully erased, by historical and contemporary gene flow. Intrinsic biological characteristics may thus have played a significant role in determining the extent to which the extrinsic geological events associated with the rise of the Isthmus of Panam? have shaped phylogeographical pat- terns in the pseudoscorpion and its harlequin beetle host. Acknowledgements We are especially grateful for the logistical support provided by the Smithsonian Tropical Research Institute and Ratibor Hartman in Panam? and Chris Starr and Ronnie Hernandez in Trinidad. We also thank La Autoridad Nacional del Ambiente (A.N.A.M) and the Forestry Division of the Republic of Trinidad and Tobago for permission to collect in Panam? and Trinidad, respectively. This research was funded by grants from the National Geographic Society (grant 5333-94) and the US National Science Foundation (MCB-0085335, DEB-0115555, IBN-0115986). MMB was supported by an NSF Research Experiences for Undergraduates award. Finally, for efficient and prompt processing of our sequencing samples, we thank Joan Rowe and Craig Osborne of The Nevada Genomics Center (NGC). The NGC is supported by NSF EPSCoR and NIH BRIN (P20 RR16464) grants. References Baldwin JD, Bass AL, Bowen BW, Clark WH (1998) Molecular phylogeny and biogeography of the marine shrimp Penaeus. Molecular Phylogenetics and Evolution, 10, 399?407. Beerli P (2002) MIGRATE: documentation and program, part of lamarc, Version 1.5. Revised 7 August 2002. Available at: http:// evolution.genetics.washington.edu/lamarc.html [Downloaded: 11 January 2003]. Beerli P, Felsenstein J (2001) Maximum likelihood estimation of a migration matrix and effective population sizes in n subpopula- tions by using a coalescent approach. Proceedings of the National Academy of Sciences of the USA, 98, 4563?4568. Beier M (1948) Phoresie und Phagophilie bei Pseudoscorpionen. ?esterreichische Zoologische Zeitschrift, 1, 441?497. Bensasson D, Zhang DX, Hartl DL, Hewitt GM (2001) Mitochon- drial pseudogenes: evolution?s misplaced witnesses. Trends in Ecology and Evolution, 16, 314?321. Bermingham E, Lessios HA (1993) Rate variation of proteins and mitochondrial DNA evolution as revealed by sea urchins separated by the Isthmus of Panam?. Proceedings of the National Academy of Sciences of the USA, 90, 2734?2738. Bermingham E, Martin AP (1998) Comparative mtDNA phylo- geography of neotropical freshwater fishes: testing shared history to infer the evolutionary landscape of lower Central America. Molecular Ecology, 7, 499?517. 10 J . A . Z E H , D . W . Z E H and M . M . B O N I L L A ' 2003 Blackwell Publishing Ltd, Molecular Ecology, 10.1046/j.1365-294X.2003.01914.x Bermingham E, McCafferty S, Martin A (1997) Fish biogeography and molecular clocks: perspectives from the Panamanian Isthmus. In: Molecular Systematics of Fishes (eds Kocher T, Stepien C), pp. 113?128. Academic Press, San Diego. Brower AVZ (1994) Rapid morphological radiation and conver- gence among races of the butterfly Heliconius erato inferred from patterns of mitochondrial DNA evolution. Proceedings of the National Academy of Sciences of the USA, 91, 6491?6495. Coates AG, Aubry M-P, Berggren WA, Collins LS, Kunk M (2003) Early Neogene history of the Central American arc from Bocas del Toro, western Panam?. Geological Society of America Bulletin, 115, 271?287. Coates AG, Obando JA (1996) The geologic evolution of the Central America Isthmus. In: Evolution and Environment in Tropical America (eds Jackson JBC, Budd AF, Coates AG), pp. 21? 56. University of Chicago Press, Chicago. Collins LS (1996) Environmental changes in Caribbean shallow waters relative to the closing tropical American seaway. In: Evolution and Environment in Tropical America (eds Jackson JBC, Budd AF, Coates AG), pp. 130?167. Chicago University Press, Chicago. Croat TB (1978) Flora of Barro Colorado Island. Stanford University Press, Stanford, CA. Eisenberg JF (1989) Mammals of the Neotropics, Vol. I. The Northern Neotropics. University of Chicago Press, Chicago. Felsenstein J (1985) Confidence limits on phylogenies: an approach using the bootstrap. Evolution, 39, 783?791. Haq BU, Hardenbol J, Vail PR (1987) Chronology of fluctuating sea levels since the Triassic. Science, 235, 1156?1167. Henderson RW, Hedges SB (1995) Origin of West Indian popu- lations of the geographically widespread boa Corallus enydris inferred from mitochondrial DNA sequences. Molecular Phylo- genetics and Evolution, 4, 88?92. Jackson JBC, Jung P, Coates AG, Collins LS (1993) Diversity and extinction of tropical American mollusks and emergence of the Isthmus of Panam?. Science, 260, 1624?1626. Jordan DS (1908) The law of geminate species. American Naturalist, 42, 73?80. Knowlton N, Weigt LA (1998) New dates and new rates for diver- gence across the Isthmus of Panam?. Proceedings of the Royal Society of London, Series B, 265, 2257?2263. Knowlton N, Weigt LA, Solorzano LA, Mills DK, Bermingham E (1993) Divergence in proteins, mitochondrial DNA, and repro- ductive compatibility across the Isthmus of Panam?. Science, 260, 1629?1632. Lessios HA (1998) The first stage of speciation as seen in organ- isms separated by the Isthmus of Panam?. In: Endless Forms (eds Howard DJ, Berlocher SH), pp. 186?201. Oxford University Press, Oxford. Lundberg JG, Marshall LG, Guerrero J, Horton B, Malabarba MCSL, Wesselingh F (1998) The stage for Neotropical diversi- fication: a history of tropical South American rivers. In: Phylogeny and Classification of Neotropical Fishes (eds Malabarba LR, Rreis RE, Vari RP, Lucena ZM, Lucena CAS), pp. 13?48. Edipucrs, Porto Alegre, Brazil. Maddison DR, Maddison WP (2002) MACCLADE 4: Analysis of Phylo- geny and Character Evolution, Version 4.05. Sinauer Associates, Sunderland, MA. Marko PB (2002) Fossil calibration of molecular clocks and the divergence times of geminate species pairs separated by the Isthmus of Panam?. Molecular Biology and Evolution, 19, 2005? 2021. Marshall LG, Butler RF, Drake RE, Curtis GH, Tedsford RH (1979) Calibration of the Great American Interchange. Science, 204, 272?279. Metz EC, Gomez-Guitierrez G, Vacquier VD (1998) Mitochondrial DNA and bindin gene sequence evolution among allopatric species of the sea urchin genus Arbacia. Molecular Biology and Evolution, 15, 185?195. Muchmore WB (1974) Clarification of the genera Hesperochernes and Dinocheirus (Pseudoscorpionida, Chernetidae). Journal of Arachnology, 2, 25?36. Myers GS (1938) Fresh-water fishes and West Indian zoogeo- graphy. Smithsonian Report, 1937, 766?773. Newcomer SD, Zeh JA, Zeh DW (1999) Genetic benefits enhance the reproductive success of polyandrous females. Proceedings of the National Academy of Sciences of the USA, 96, 10236?10241. Perdices A, Bermingham E, Montilla A, Doadrio I (2002) Evolu- tionary history of the genus Rhamdia (Teleostei: Pimelodidae) in Central America. Molecular Phylogenetics and Evolution, 25, 172? 189. Pindell JL, Barrett SF (1990) Geological evolution of the Caribbean region: a plate-tectonic perspective. In: The Caribbean Region, the Decade of North American Geology, Vol. H (eds Dengo G, Case JE), pp. 405?432. Geological Society of America, Boulder, CO. Posada D, Crandall KA (1998) modeltest: testing the model of DNA substitution. Bioinfomatics, 14, 817?818. Ridgely RS (1981) A Guide to the Birds of Panam?. Princeton Univer- sity Press, Princeton, NJ. Ryan MJ, Rand AS, Weigt LA (1996) Allozyme and advertisement call variation in the tungara frog, Physalaemus pustulosus. Evolu- tion, 50, 2435?2453. Seutin G, Brawn J, Ricklefs R, Bermingham E (1993) Genetic diver- gence among populations of a tropical passerine, the streaked saltator (Saltator albicollis). Auk, 110, 117?126. da Silva MNF, Patton JL (1998) Molecular phylogeography and the evolution and conservation of Amazonian mammals. Mole- cular Ecology, 7, 475?486. Simon C, Frati F, Beckenbach A, Crespi B, Liu H, Flook P (1994) Evolution, weighting and phylogenetic utility of mitochondrial gene sequences and a compilation of conserved polymerase chain reaction primers. Annals of the Entomological Society of America, 87, 651?701. Simpson GG (1950) History of the fauna of Latin America. American Scientist, 38, 261?389. Simpson GG (1980) Splendid Isolation, the Curious History of South American Mammals. Yale University Press, New Haven, CT. Stehli FG, Webb SD (1985) The Great American Biotic Interchange. Plenum Press, New York. Stillman JH, Reeb CA (2001) Molecular phylogeny of eastern Pacific porcelain crabs, genera Petrolisthes and Pachycheles, based on the mtDNA 16S rDNA sequence: phylogeographic and systematic implications. Molecular Phylogenetics and Evolution, 19, 236?245. Swofford DL (2002) PAUP*: Phylogenetic Analysis Using Parsimony (* and Other Methods), Version 4.0b 10. Sinauer Associates, Sun- derland, MA. Vermeij GJ (2001) Distribution, history, and taxonomy of the Thais clade (Gastropoda: Muricidae) in the Neogene of tropical America. Journal of Paleontology, 75, 697?705. Wege DC, Long AJ (1995) Key Areas for Threatened Birds in the Neo- tropics. Bird Life International, Cambridge. Wilcox TP, Hugg L, Zeh JA, Zeh DW (1997) Mitochondrial DNA sequencing reveals extreme genetic differentiation in a cryptic C O R D Y L O C H E R N E S P H Y L O G E O G R A P H Y 11 ' 2003 Blackwell Publishing Ltd, Molecular Ecology, 10.1046/j.1365-294X.2003.01914.x species complex of neotropical pseudoscorpions. Molecular Phy- logenetics and Evolution, 7, 208?216. Zeh JA (1997) Polyandry and enhanced reproductive success in the harlequin beetle-riding pseudoscorpion. Behavioral Ecology and Sociobiology, 40, 111?118. Zeh JA, Newcomer SD, Zeh DW (1998) Polyandrous females dis- criminate against previous mates. Proceedings of the National Academy of Sciences of the USA, 95, 13732?13736. Zeh DW, Zeh JA (1992) Dispersal-generated sexual selection in a beetle-riding pseudoscorpion. Behavioral Ecology and Sociobio- logy, 30, 135?142. Zeh DW, Zeh JA (1994) When morphology misleads: interpopula- tion uniformity in sexual selection masks genetic divergence in harlequin beetle-riding pseudoscorpion populations. Evolution, 48, 1168?1182. Zeh DW, Zeh JA (2000) Reproductive mode and speciation: the viviparity-driven conflict hypothesis. Bioessays, 22, 938?946. Zeh JA, Zeh DW (2001) Reproductive mode and the genetic bene- fits of polyandry. Animal Behaviour, 61, 1051?1063. Zeh DW, Zeh JA, Bermingham E (1997) Polyandrous, sperm- storing females: carriers of male genotypes through episodes of adverse selection. Proceedings of the Royal Society of London, Series B, 264, 119?125. Zeh DW, Zeh JA, Bonilla MM (2003) Phylogeography of the giant harlequin beetle (Acrocinus longimanus). Journal of Biogeography, 30, 1?7. Zeh DW, Zeh JA, Coffroth M-A, Bermingham E (1992a) Popula- tion-specific DNA fingerprints in a neotropical pseudoscorpion (Cordylochernes scorpioides). Heredity, 69, 201?208. Zeh DW, Zeh JA, Tavakilian G (1992b) Sexual selection and sexual dimorphism in the harlequin beetle Acrocinus longimanus. Bio- tropica, 24, 86?96. David and Jeanne Zeh are faculty in the Department of Biology and the Program in Ecology, Evolution and Conservation Biology at the University of Nevada, Reno. Their research interests lie at the interface between behavioural ecology and molecular genetics, with a focus on the role of genomic conflicts in sexual selection and speciation. They are currently investigating the potential importance of gametic interactions and postcopulatory processes in the evolution of cryptic species of Neotropical arthropods. Melvin Bonilla is an undergraduate student, majoring in biochemistry and sociology at the University of Nevada, Reno. He aims to pursue graduate studies in molecular biology.