Tempo and Mode of Diversification in a Radiation of Endogean Ground Beetles

Received: 1 April 2017 | Revised: 23 July 2017 | Accepted: 5 September 2017 DOI: 10.1111/mec.14358

ORIGINAL ARTICLE

Speciation below ground: Tempo and mode of diversification in a radiation of endogean ground beetles

1Department of Life Sciences, Natural History Museum, London, UK Abstract 2Department of Life Sciences, Imperial Dispersal is a critical factor determining the spatial scale of speciation, which is con- College London, Ascot, UK strained by the ecological characteristics and distribution of a species’ habitat and 3Grupo de Ecologıa y Evolucion en Islas, Instituto de Productos Naturales y the intrinsic traits of species. Endogean taxa are strongly affected by the unique Agrobiologıa (IPNA-CSIC), San Cristobal de qualities of the below-ground environment and its effect on dispersal, and contrast- la Laguna, Spain ing reports indicate either high dispersal capabilities favoured by small body size 4Departamento de Zoologıa y Antropologıa Fısica, Universidad Complutense de Madrid, and mediated by passive mechanisms, or low dispersal due to restricted movement Madrid, Spain and confinement inside the soil. We studied a species-rich endogean ground beetle 5Institut de Biologia Evolutiva (CSIC- Universitat Pompeu Fabra), Barcelona, lineage, Typhlocharina, including three genera and more than 60 species, as a model Spain for the evolutionary biology of dispersal and speciation in the deep soil. A time-cali- > Correspondence brated molecular phylogeny generated from 400 individuals was used to delimit Carmelo Andujar, Department of Life candidate species, to study the accumulation of lineages through space and time by Sciences, Natural History Museum, London, – – UK. species area age relationships and to determine the geographical structure of the Email: [email protected] diversification using the relationship between phylogenetic and geographic distances

Funding information across the phylogeny. Our results indicated a small spatial scale of speciation in Natural Environment Research Council, Typhlocharina and low dispersal capacity combined with sporadic long distance, pre- Grant/Award Number: NE/M021955; European Commission, Grant/Award sumably passive dispersal events that fuelled the speciation process. Analysis of lin- Number: MSCA-IF-2015-705639; Ministerio eage growth within Typhlocharina revealed a richness plateau correlated with the de Economıa y Competitividad, Grant/Award Number: CGL2010-16944 range of distribution of lineages, suggesting a long-term species richness equilibrium mediated by density dependence through limits of habitat availability. The interplay of area- and age-dependent processes ruling the lineage diversification in Typhlo- charina may serve as a general model for the evolution of high species diversity in endogean mesofauna.

KEYWORDS Anillini, density dependence, endogean, geographic speciation, long-distance dispersal (LDD), microendemism, Typhlocharina

1 | INTRODUCTION Both the frequency and the distance of dispersal events critically influence speciation, determining its geographical scale and temporal The role of dispersal in modulating gene flow is key to the speciation dynamics (Kisel & Barraclough, 2010; Lomolino, Riddle, Whittaker, & process. Dispersal determines the balance between species cohesion Brown, 2010). For organisms with a high dispersal potential, pro- and the divergence of isolated populations (Coyne & Orr, 2004; cesses of geographic differentiation are only effective at large spatial Futuyma, 1998) and promotes the establishment of new populations scales, while differentiation is quickly erased over small distances as a source for the diversification process (Nathan & Nathan, 2014). (Finlay, 2002; Wilkinson, Koumoutsaris, Mitchell, & Bey, 2012). At

| Molecular Ecology. 2017;26:6053–6070. wileyonlinelibrary.com/journal/mec © 2017 John Wiley & Sons Ltd 6053 6054 | ANDUJAR ET AL. the opposite end of the spectrum, lineages with low dispersal capa- The soil environment is a complex and heterogeneous matrix of bilities may undergo geographic differentiation at a local scale, gen- minute and interconnected spaces, characterized by darkness, limited erating high levels of endemicity, geographical structure and positive temperature fluctuations and high relative humidity and CO2 levels species–area relationships even over short distances (Futuyma, 1998; (Russell & Appleyard, 1915). This matrix imposes major restrictions Kisel & Barraclough, 2010; Kisel, McInnes, Toomey, & Orme, 2011). to the movement of organisms, resulting in a reduced range for the If not constrained by geographic or ecological barriers, dispersal activity of an individual (Eisenbeis & Wichard, 1987). Arthropods is a stochastic process affecting individuals, which follow a trajectory adapted to live in deep soil layers (euedaphics sensu Eisenbeis & of limited movement from their site of birth (Harte, McCarthy, Tay- Wichard, 1987; endogeans sensu Giachino & Vailati, 2010) show a lor, Kinzig, & Fischer, 1999), and over time give rise to predictable general trend to body size reduction, depigmentation, shortening of macroecological patterns of decreasing similarity at greater distances extremities and loss of eyes and flight capacity (Eisenbeis & Wichard, (Baselga et al., 2013; Diniz-Filho & Bini, 2011). Such idealized homo- 1987; Gisin, 1943). Therefore, adaptations of deep soil mesofauna geneous landscapes do not exist, and the separation of habitat converge on a general decrease in dispersal potential by active patches imposes barriers to the free movement of organisms and movement, but at the same time some of these adaptations, such as modulates the divergence of populations. In addition, dispersal size reduction or the adaptation to survive underwater during flood capacity of species is constrained by the properties of the species’ events, enhance the likelihood of passive dispersal over short and habitat, given the differences in habitat availability and stability at long distances. The interplay of these active and passive dispersal ecological and geological timescales (the habitat-templet concept; events determines the probability function of dispersal (dispersal ker- Southwood, 1977; Korfiatis & Stamou, 1999). Thus, the factors shap- nel) for a species (Nathan et al., 2008) and overall will determine the ing evolutionary diversification may be similar across lineages associ- scale and dynamics of differentiation in soil lineages. ated with particular habitat types. For example, lineages of aquatic Understanding the processes that drive the distribution of soil insects differ in their dynamics of diversification due to the differ- organisms over spatial and temporal scales is essential to estimate ence in geological stability of standing or flowing water bodies (Rib- the magnitude and evolution of soil biodiversity (Ettema & Wardle, era, 2008). Furthermore, the species diversity in an area may be 2002). However, we currently lack relevant information on specia- constrained by competitive interactions that limit the number of spe- tion patterns and geographical distributions of soil lineages (Decaens,€ cies able to coexist. This may affect the dynamics of lineage growth 2010). Different studies have found high levels of cryptic diversity by reducing the rate of net diversification towards the recent, after and microendemicity in several lineages of soil arthropods (e.g., an initial diversification pulse (Alroy, 1996; Foote, 1997; Kisel et al., Andujar et al., 2015; Bennett, Hogg, Adams, & Hebert, 2016; Casale, 2011; Rabosky, 2009; Sepkoski, Bambach, Raup, & Valentine, 1981). 2009; Cicconardi, Fanciulli, & Emerson, 2013; Cicconardi, Nardi, Thus, the interaction of both dispersal constraints and habitat Emerson, Frati, & Fanciulli, 2010), pointing to a very small spatial parameters will shape the spatial and temporal dynamics of specia- scale of diversification in soil mesofauna. In contrast, other studies tion and biodiversity patterns. have proposed a high dispersal capacity mainly promoted by passive

FIGURE 1 (a) Relationship between clade age and species richness with density-dependent diversification due to geographical, ecological or other factors imposing a maximum species richness or “carrying capacity” of the clade. (a1) Individual clades (marked with different symbols) may vary in age, with square and open circle being the youngest clades. Independently of the age, when evolutionary time is enough, clades reach the equilibrium species richness which depends on its “carrying capacity” and not on clade age. The graph shows an initial richness accumulation phase where clades may vary in diversification rate with the slope of the log-diversity curve proportional to the net rate of diversification (modified from Rabosky, 2009). In this phase, there is a positive relationship between clade age and species richness (a2). As ecological niches or geographical space is filled, clades reach the richness equilibrium phase where there is no relationship between clade age and species richness and the net diversification rate will be 0 (a3). In those cases where the diversity is limited by geographical space and speciation is mainly allopatric, we additionally will expect a richness–area relationship at both accumulation and equilibrium phases and a positive area–age relationship during the accumulation phase and no correlation at the equilibrium phase. Note that if diversification is mainly mediated by other ecological factors independent of area or by vicariance without range expansion, we will expect no area–age and area– richness correlations. (b) Topological and lineage through time (LTT) expectations under a density-dependent diversification. Clades at the richness equilibrium phase can show different topologies according to the ratio between the initial speciation rate (k) and extinction rate (l), in a process that is also dependent on the age (t) and “carrying capacity” (k) of clades. One hundred simulations performed using a density- dependent model in DDD R package (Etienne et al., 2012; t = 50, K = 30, model = 1.3, condition = 3; detailed parameters in Appendix S1: Data S2). When extinction is relatively high (k = 0.7; l = 0.4), the high turnover of species along time results in a coalescence-like process that increase the probability of a long stem branch to its sister clade (b1). On the opposite, if extinction is low (k = 0.7; l = 0.01), topology is expected to show an initial diversification with long tips (b2). A randomly selected topology is shown including extant species (left) and all species (extinct lineages included; right). LTTs represent the selected topology (black line) and the 100 simulated trees (in grey). (c) Alternative predictions for the relationships between geographical (G) and phylogenetic (P) distances according to diversification mostly occurring after dispersal to adjacent areas to the species range; that is, speciation is more likely in closer areas than in more distant ones with a unidirectional filling of space (c1) or adding the possibility of distant lineages brought into close proximity secondarily (c2). Alternatively, dispersal may as probable at shorter (“adjacent”) as at longer distances potentially due to a higher passive dispersal potential (c3) ANDUJAR ET AL. | 6055

(a) (a1)

(a2) (a3)

(b) (b1) (b2)

(c) (c1) (c2) (c3) 6056 | ANDUJAR ET AL. dispersal and long-distance movements that dilute the role of geo- should be maintained through species accumulation in new areas, graphical speciation in soil mesofauna (e.g., Costello, Tiegs, & Lam- and a positive species–age correlation is maintained beyond the ini- berti, 2010; Coulson, Hodkinson, Webb, & Harrison, 2012; Nkem tial phase. Density-dependent diversification (DDD) can also be et al., 2006). apparent from the topological branching pattern of phylogenetic This study uses the subtribe Typhlocharina (Jeanne, 1973) as a trees (Etienne et al., 2012). If extinction rates represent a relatively model for the evolutionary biology of dispersal and speciation in the high fraction of the speciation rate, as overwhelmingly suggested by deep soil. Typhlocharina is a typical endogean group in the tribe Anil- the palaeontological literature (Alroy, 1996; Stanley, 1979), long stem lini (Jeannel, 1963; Coleoptera, Carabidae, Trechinae) consisting of branches are expected for clades under DDD when at richness equi- three genera: Lusotyphlus, Typhlocharis and Microcharidius, and 62 librium during enough time (Figure 1b1; see caption for details). On known species: 59 restricted to the Iberian Peninsula and 2 distributed the contrary, when extinction rates are relatively low, trees typically on both sides of the Gibraltar Strait and one endemic to Tunisia result in long tip branches (Figure 1b2). While the importance of (Perez-Gonz alez, Andujar, & Zaballos, 2017; Zaballos, Andujar, & spatial dependence effects on diversification rates has been recently Perez-Gonz alez, 2016). All species show strong morphological adapta- highlighted in some groups (Pigot, Owens, & Orme, 2010; Fritz, tions to life in deep soil layers, including reduced size (from 0.9 to Jønsson, Fjeldsa, & Rahbek, 2012; reviewed in Rabosky, 2009), the 2.9 mm), dorsoventrally flattened and parallel body shape, depigmen- role and geographical scale of these processes are still unexplored tation and anophthalmia. Many species of Typhlocharina are only for soil lineages. known from a very small area, thought to be the result of very reduced Whereas the slow movement of individuals in the soil over long active dispersal, but at the same time it has been suggested that they time periods may still lead to large geographical ranges occupied by have a high potential for passive long-range dispersal, especially a clade, the possibility of occasional passive long-distance dispersal through water-mediated transport (hydrochory; Ortuno~ & Gilgado, events (LDD; Nathan, 2005) may accelerate this process while 2011). This potential for long-range passive dispersal has been sur- favouring geographical speciation. Passive dispersal can contribute to mised in other lineages of Anillini, which have colonized distant islands the founding of new populations beyond the reach of the regular such as New Caledonia, likely through rafting (Andujar et al., 2016). movements of individuals and thus may promote speciation in the These characteristics make Typhlocharina a candidate model group to newly occupied site. If passive dispersal is more frequent at shorter study the diversification processes in endogean lineages. than longer distances, a correlation of geographic distance with lin- We first use a novel time-calibrated phylogeny of Typhlocharina eage age (phylogenetic distance) is expected (Figure 1c1). In addi- to delimit species and to explore the accumulation of lineages tion, if space is not filled uniformly, distant lineages can secondarily through space and time by species–area–age relationships (Kisel be brought into close proximity (Figure 1c2). This scenario results in et al., 2011; Figure 1a1). We consider a scenario where the endo- asymmetric plots of the phylogenetic distance vs. geographic dis- gean ancestor of all extant Typhlocharina gave rise to the modern tance (PDGD; Ribera et al., 2011), where species or populations in diversity and wide distribution across Iberia starting from a specific close geographic proximity exhibit the whole range of phylogenetic point in geography. This geographical diversification process may fol- distances, but no closely related species or populations appear at low a logistic growth model with richness constrained by area (Cour- large distances (Figure 1c2). However, this triangular distribution can tillot & Gaudemer, 1996) and regulated by the low propensity for be obscured by occasional LDD events which may result in closely dispersal in the below-ground habitat, in which case we would related populations at large geographic distances (Figure 1c3). If fre- expect an initial phase where diversity accumulates through time quent enough, they have the potential to entirely remove the geo- and space by the successive colonization of new areas, with subse- graphical structure across the phylogeny. quent isolation and speciation of the newly established populations The spatial distribution and density dependency of lineage evolu- (Figure 1a2; Kisel et al., 2011; Moen & Morlon, 2014; Rabosky, tion can be used as a framework to tease apart the roles of active 2009). After the initial phase of unconstrained growth, available and passive dispersal constraints in speciation below ground. Our space is filled and density-dependent processes may result in a rich- integrative approach to the diversification process of a typical endo- ness equilibrium phase characterized by an overall null net diversifi- gean lineage allows a better understanding of the processes ruling cation rate, such that each time a lineage goes extinct, another diversification of endogean mesofauna and show how, when digging lineage appears by speciation (Hey, 1992; Nee, Holmes, May, & Har- enough, endogean lineages become an ideal model system to study vey, 1994; Rabosky, 2009). Predictions for lineages at this second geographic speciation. stage, where richness is constrained by density-dependent processes within the available area, include a positive species richness–area 2 | MATERIAL AND METHODS relationship with greater number of species in lineages with a wider range of distribution. In addition, under this scenario, we do not 2.1 | Study group expect a significant increase in species numbers with lineage age, and no significant increase in the total area occupied by a clade with The endogean subtribe Typhlocharina (Coleoptera: Carabidae: lineage age (Figure 1a3; Rabosky, 2009). However, if there were no Trechinae) is part of the tribe Anillini, which includes about five hun- spatial constraints to total richness, a positive net diversification rate dred species of litter-, soil- and cave-dwelling beetles (Lorenz, 2005; ANDUJAR ET AL. | 6057

Sokolov, 2015) widely distributed in all continents except Antarctica. the cuticle and observed through light microscopy (Zeiss 474620- Within Anillini, Typhlocharina is the most diverse lineage, with 62 9900 microscope, Germany). Additional specimens from all 62 described species. All of them are characterized by small body size, described species, including type material for 56 of them, were lack of pigmentation, apterism, absence of eyes and a well-devel- examined, and the recommendations of the “open nomenclature” oped set of long sensorial setae in the lateral margins of the elytra (Bengston, 1988) were followed to assign identification to the holo- (umbilicate series; Jeannel, 1963). In a molecular phylogeny of the genophores (Details in Appendix S1: Data S1, with a complete list of tribe Anillini that included nine species of Typhlocharina (Andujar studied specimens in Appendix S1: Table S1). et al., 2016), the lineage was found to be monophyletic, dating back to the Paleocene (60 Mya). A detailed morphological study of 2.4 | Alignment and phylogenetic analyses Typhlocharina split the former genus Typhlocharis into three distinct genera: Lusotyphlus, Typhlocharis and Microcharidius (Perez-Gonz alez Sequences were aligned using the online version of MAFFT 6.240 et al., 2017). In most cases, the species are microendemics with very (Katoh, Misawa, Kuma, & Miyata, 2002), with the Q-INS-i algorithm restricted distributions, sometimes known from a single locality for rRNA genes and L-INS-i for protein-coding genes. The correct where, however, they can be very abundant (densities up to 4,000– translation to amino acids was checked, and all fragments were con- 3 6,000 specimens/m of soil; e.g., Perez-Gonz alez, Zaballos, & Ghan- catenated into a single data matrix in GENEIOUS 7.1.9. (Biomatters Ltd., nem, 2013). NZ). Phylogenetic analyses were conducted with maximum likelihood (ML) and Bayesian inference (BI) using the CIPRES Science Gateway 2.2 | DNA extraction and sequencing (Miller, Pfeiffer, & Schwartz, 2010). ML trees were obtained using

We sequenced 402 specimens of Typhlocharina from 116 localities, RAXML 7.2.7 (Stamatakis, 2006). Data sets were partitioned by gene including 29 of the 62 described species (Zaballos et al., 2016) and and the protein-coding genes (cox1-a and cox1-b fragments) were 70 additional populations estimated to represent 45 potentially new additionally partitioned by separating the third codon positions species. Sampled taxa include representatives from all proposed (Andujar, Serrano, & Gomez-Zurita, 2012). An independent GTR + G species groups (Ortuno~ & Gilgado, 2011; Perez-Gonz alez & Zabal- model was applied to each data partition. The best scoring ML tree los, 2013a; Zaballos & Ruiz-Tapiador, 1997; Zaballos & Wrase, was selected among 200 searches on the original alignment with dif- 1998) and cover most of the distribution range of the genus ferent randomized parsimony starting trees. Support values were (Appendix S1: Fig. S1). Two representatives of the closely related obtained with 1,000 bootstrap replicates (Felsenstein, 1985). BI was genus Geocharis were used as outgroups to root the tree. Details run in MRBAYES 3.2.3 (Huelsenbeck, Ronquist, Nielsen, & Bollback, on the studied taxa and localities are provided in Appendix S1: 2001) on the concatenated data, partitioned by gene and codon as Tables S1 and S2. before, and for each partition, the optimal substitution model was

DNA was nondestructively extracted from whole specimens selected using the Akaike information criterion (AIC) in JMODELTEST using Qiagen Dneasy and BioSprint Blood and Tissue kits (Hilden, 2.1.7 (Darriba, Taboada, Doallo, & Posada, 2012). BI consisted of Germany). Voucher specimens and DNA aliquots are kept in the IBE two independent runs, each with three hot and one cold chain, for (Barcelona), NHM (London), Universidad Complutense de Madrid 10 million generations, whereby trees were sampled every 1,000 (UCM) and the authors’ personal collections. Five DNA fragments generations. The standard deviation of split frequencies was checked were PCR-amplified and Sanger-sequenced, namely (i) the 50 end of to assess the convergence of results, as well as the mean and effec- 0 the cox1 gene (cox1-a; 658 bp); (ii) the 3 end of the cox1 gene tive sampled size (ESS) of likelihood values computed with TRACER 1.6 (cox1-b; 826 bp); (iii) the 50 end of the rrnL gene plus the complete (Rambaut, Suchard, Xie, & Drummond, 2013). The 50% majority rule trnL and the 50 end of nad1 (779 bp); (iv) a fragment of the SSU and strict consensus trees were calculated excluding 50% of the ini- nuclear ribosomal RNA gene (630 bp); and (v) a fragment of the tial trees as a conservative burn-in, ensuring that the plateau in tree

LSU nuclear ribosomal RNA gene (1,000 bp). PCRs were performed likelihood values had been reached. Trees were visualized using FIG- using PuReTaq Ready-To-Go PCR beads (GE Healthcare, UK) or Bio- TREE 1.4.2 (Rambaut, 2012), and node posterior probabilities were taq Polymerase (Bioline, London, UK), with 39 cycles using 48–52°C interpreted as support values. as the annealing temperature. The primers used for each gene fragment are given in Appendix S1: Table S3. Overall, 800 sequences of 2.5 | Dating analyses Typhlocharina were newly generated with sequence Accession nos

MF538817-MF539616 (Appendix S1: Table S1). For the Bayesian phylogenetic analyses in BEAST 1.8.1 (Drummond, Suchard, Xie, & Rambaut, 2012), the original concatenated data set was collapsed to haplotypes and the outgroups were excluded, 2.3 | Morphological study resulting in a data set of 330 terminals. We used as calibration prior A detailed morphological study of all the sequenced specimens for each gene an uniform function on the mean substitution rate (hologenophores sensu Astrin, Zhou, & Misof, 2013) was conducted. encompassing the 95% confident interval values obtained for the After DNA extraction, specimens were rinsed in lactic acid to clear same DNA fragments (cox1-a, cox1-b, rrnL, LSU and SSU genes) in 6058 | ANDUJAR ET AL.

TABLE 1 Priors on rates of evolution Initial and root age used on calibration analyses Parameter Function value Min–max values Shape Scale Values from conducted on BEAST cox1-a mean rate Uniform 0.0113 (0.0081–0.0147) n.a. n.a. Andujar et al. (2012) cox1-b mean rate Uniform 0.0145 (0.01–0.0198) n.a. n.a. Andujar et al. (2012) rrnL mean rate Uniform 0.0016 (0.001–0.00248) n.a. n.a. Andujar et al. (2012, 2016) LSU mean rate Uniform 0.0013 (0.0007–0.00242) n.a. n.a. Andujar et al. (2012, 2016) SSU mean rate Uniform 0.0003 (0.000178–0.00045) n.a. n.a. Andujar et al. (2016) Root age Gamma 60 n.a. 29.896 2.0991 Andujar et al. (2016)

the confamilial genus Carabus (Andujar, Soria-Carrasco, Serrano, & 2.6 | Species delimitation Gomez-Zurita, 2014; Andujar, Serrano, & Gomez-Zurita, 2012) and the tribe Anillini (Andujar et al., 2016; Table 1). We additionally The GMYC (Pons et al., 2006) and bPTP (Zhang, Kapli, Pavlidis, & applied a gamma prior on the root age for the most recent common Stamatakis, 2013) methods, along with a comparative morphological ancestor (crown group) of Typhlocharina. We used TREESTAT 1.6.1 study of the specimens, were applied to delimit species-level entities (Rambaut & Drummond, 2010) to recover the node age of the in Typhlocharina. GMYC was applied to the ultrametric consensus ancestor of Typhlocharina from the sample of the MCMC search of tree obtained from the dating analyses favoured by Bayes factors the BEAST analysis favoured in Andujar et al. (2016), and we used the comparisons for each of the alternative speciation models used. We

“fitdistr” option of the R package MASS to obtain the values of a used the SPLITS package (Ezard, Fujisawa, & Barraclough, 2009) in R gamma function adjusting the distribution of sampled ages, thus to conduct the single threshold (as recommended by Fujisawa, & accounting for the uncertainty associated with the age estimations Barraclough, 2013) and multiple threshold approaches. In addition, (Table 1). the bPTP analysis was conducted on the same ultrametric trees and

Analyses were conducted under the model of substitution best on the consensus trees from MRBAYES and the ML tree from RAXML. fitting to each gene and codon partition as above, and were run for bPTP was performed on the bPTP web server (J. Zhang; http://spec 50 million generations sampling one tree in every 5,000 generations. ies.h-its.org/ptp/; Accessed: 3 March 2016), excluding outgroups and In the molecular clock settings, gene partitions were uncorrelated, with 500,000 MCMC generations, Thinning = 100, burn-in = 0.2 and and two independent analyses were conducted under alternative seed = 123. Both approaches were compared with the detailed mor- molecular clock hypotheses, applying (i) an uncorrelated lognormal phological study of all sequenced specimens. The final species delim- (ULN) to all genes, or alternatively (ii) using an ULN for nuclear and itation hypothesis integrated the morphological information to refine a strict clock (SC) for mitochondrial genes. For both clock settings, the results of the GMYC and bPTP, a desirable approximation to test three independent analyses were conducted under alternative tree potential imperfections in the thresholding of molecular species priors, applying (i) a Yule speciation model (YS), (ii) a birth–death delimitation approaches (Talavera, Dinca, & Vila, 2013). model (BD) and (iii) a birth–death-incomplete sampling (BDIS) model. We repeated all the analyses with and without the inclusion of the 2.7 | Lineage age, species richness and range size constraint on the age for the ancestor of Typhlocharina, and each of these 12 alternative calibration analyses was repeated twice with Correlations between lineage age, species richness and range size different seed values (Table 2). for the main lineages of Typhlocharina were assessed to character- The alternative clock and speciation-tree settings were com- ize the patterns of diversification. The age of each lineage was pared with Bayes factors estimated with the stepping-stone and obtained as the median value of the posterior distribution of the the path-sampling algorithms in BEAST (Baele, Li, Drummond, favoured BEAST analysis, considering the stem node for each lineage.

Suchard, & Lemey, 2013) and with the SHM estimator in TRACER Species richness values included sequenced species and all remain-

1.6. Consensus trees were estimated with TREEANNOTATOR (Drum- ing species that we could unambiguously assign to these lineages mond et al., 2012) discarding the 50% initial trees as a burn-in based on the detailed morphological study (all but Microcharidius fraction, after checking ESS of likelihood, evolutionary rates and santchii;Perez-Gonz alez et al., 2017). Finally, the area of the mini- root age values, and ensuring that the tree likelihood values had mum polygon including all known localities for each lineage was reached a plateau. Posterior probabilities were considered as a used as a proxy of range size as measured using Google Earth Pro. measure of node support. We tested the magnitude and significance of the relationship ANDUJAR ET AL. | 6059

TABLE 2 Summary of results for calibration analyses with alternative clock, speciation and root age (C, constrained; F, free) priors

Root age mean (95% Root Clock Speciation Repeat (seed) Likelihood (mean/ESS) HM SS PS HPD interval) C ULN Yule_BDIS 1 (123,456)b 41,213.6/369.7 41,302.7 43,553.6 43,528.9 59.68 [51.55, 68.65] 2 (654,321) 41,201.4/605.3 41,306.2 43,575.1 43,545.5 60.99 [52.78, 69.83] C ULN Yule_BD 1 (123,456) 41,216.6/344.0 41,313.4 43,588.4 43,555.9 60.49 [52.54, 68.37] 2 (654,321)b 41,205.5/205.4 41,291.8 43,559.0 43,547.4 59.44 [51.07, 69.14] C ULN Yule 1 (123,456)b 41,240.1/361.6 41,329.9 43,777.6 43,745.1 48.98 [42.36, 56.72] 2 (654,321) 41,232.6/534.0 41,342.9 43,783.2 43,764.1 49.42 [42.26, 57.28] F ULN Yule_BD 1 (123,456)c 41,200.2/330.1 41,280.1 43,587.6 43,560.5 59.31 [49.49, 69.47] 2 (111,111) 41,214.6/208.2 41,296.7 43,597.6 43,571.4 60.4 [51.33, 71.13] F ULN Yule_BDIS 1 (123,456)b 41,203.1/268.1 41,283.9 43,563.4 43,522.6 59.06 [50.16, 68.73] 2 (111,111) 41,214.0/467.4 41,297.9 43,545.0 43,533.8 60.84 [51.3, 71.05] F ULN Yule 1 (123,456)b 41,234.1/86.3 41,321.8 43,816.7 43,776.4 47.53 [41.24, 54.36] 2 (111,111) 41,246.2/282.6 41,332.4 43,823.4 43,771.6 48.31 [40.73, 55.7] C SC_ULN Yule_BDIS 1 (123,456)a 41,374.9/392.1 41,436.1 43,620.2 43,604.9 63.22 [54.87, 71.22] 2 (654,321) 41,372.8/555.7 41,439.3 43,627.8 43,607.8 63.1 [55.78, 71.55] C SC_ULN Yule_BD 1 (123,456)a 41,371.7/331.8 41,437.3 43,672.0 43,638.1 61.45 [53.53, 69.89] 2 (654,321) 41,373.0/476.0 41,439.8 43,683.3 43,664.2 63.05 [54.63, 71.75] C SC_ULN Yule 1 (123,456)a 41,415.0/330.3 41,485.0 43,899.8 43,863.7 52.8 [46.14, 60.48] 2 (654,321) 41,429.5/548.3 41,499.9 43,885.1 43,874.7 53.77 [46.6, 61.14] F SC_ULN Yule_BD 1 (123,456) 41,380.5/43.0 41,446.7 43,671.4 43,634.7 63.05 [54.95, 72.85] 2 (111,111)a 41,374.4/433.8 41,433.5 43,641.6 43,630.1 62.9 [54.55, 72.13] F SC_ULN Yule_BDIS 1 (123,456) 41,370.5/333.5 41,454.8 43,625.2 43,599.2 63.48 [55.08, 73.8] 2 (111,111)a 41,373.4/443.0 41,444.8 43,646.0 43,608.5 61.5 [52.81, 70.44] F SC_ULN Yule 1 (123,456)a 41,430.2/449.4 41,497.9 43,889.6 43,871.3 52.54 [45.6, 60.3] 2 (111,111) 41,419.2/391.8 41,502.6 43,884.2 43,845.9 52.14 [45.07, 59.82]

Bayes factors were estimated with the SHM estimator in TRACER (HM) and the stepping-stone (SS) and path-sampling (PS) algorithms in BEAST. aFavoured of each pair with identical parameter. bSelected for delimitation and diversification analyses. cOverall favoured analyses. between species richness and lineage age, range size and lineage and the “end” age was the crown age estimated by dating analyses. age, and range size and species richness using Pearson and Spear- Models including 1, 2, 3 and 4 rate shifts were compared using chi- man correlations in R. squared tests. Analyses were repeated under alternative hypotheses on sampling coverage, using (case 1) 69% (actual % of known species included in the study; 74/107), (case 2) 50% assuming 148 species 2.8 | Diversification rates within Typhlocharina and (case 3) 25% assuming 296 species

The tempo and mode of diversification of Typhlocharina was investi- (Table 3). Analyses were performed with the packages APE (Paradis, gated using birth–death statistics on the dated ultrametric trees Claude, & Strimmer, 2004), PICANTE (Kembel et al., 2010) and TREEPAR inferred with various speciation models (YS, BD, BDIS) and the (Stadler, 2011) in the R 3.1.2 software platform. favoured ULN clock. Diversification analyses were performed with We also used BAMM 2.5 (Bayesian Analysis of Macroevolutionary trees collapsed to species (74 terminals) using ML and Bayesian Mixtures; Rabosky, 2014; Rabosky, Donnellan, Grundler, & Lovette, methods, and including models that account for diversity-dependent 2014), which uses the reversible jump Markov chain Monte Carlo diversification. approach to model dynamics of speciation, extinction and trait evolu- First, we used the TreePar approach (Stadler, 2011) with the tion on phylogenetic trees. This software takes into account hetero- “bd.shifts.optim” function to estimate discrete changes in speciation geneity of diversification rates across the tree allowing us to estimate and extinction rates considering alternative hypotheses about sam- continuous rate variation through time and among clades, and to pling completeness. TreePar analyses were run with grid = 0.1 Myr model clade-specific sampling completeness. The prior block parame- and posdiv = FALSE to allow the diversification rate to be negative ter values were chosen by BAMMTOOLS (lambda Init Prior, lambda Shift (i.e., allows for periods of declining diversity). The “start” age was 0 Prior, mu Init Prior), with a Poisson Rate Prior = 1.0. Analyses were run 6060 | ANDUJAR ET AL.

TABLE 3 Information on the eight main lineages of Typhlocharina with corresponding clades as in Figures 3 and 4

Species in this paper Sampling coverage

Range Total this New Not Clade Colour size (Km2) Age (Ma) Rich. paper Desc. sp. samp.a Case 1 Case 2 Case 3 C1A (Lusotyphlus) Pink 14,399 45.51 5 2 2 0 3 0.40 0.30 0.20 C1B (Typhlocharis) Red 147,038 45.51 32 20 8 12 12 0.63 0.47 0.31 C2A (Microcharidius clade gomezi) Yellow 41,623 45.41 7 2 2 0 5 0.29 0.21 0.14 C2B (Microcharidius clade diecki) Grey 16,204 45.41 4 1 1 0 3 0.25 0.19 0.13 C2C (Microcharidius clade quadridentata) Black 92,787 53.31 14 13 4 9 1 0.93 0.70 0.46 C3A (Microcharidius clade toletana) Green 59,502 48.33 11 11 5 6 0 1.00 0.75 0.50 C3B (Microcharidius clade monastica) Dark blue 14,550 43.8 4 3 2 1 1 0.75 0.56 0.38 C3C (Microcharidius clade belenae) Light blue 122,994 43.8 30 22 5 17 8 0.73 0.55 0.37 Total 74 29 45 33 0.69 0.50 0.25

The three last columns show the sampling probabilities used in the diversification analyses for cases 1–3 where total number of species in Typhlocharina is estimated at 104, 148 and 296, respectively. aHypothetically assigned based on morphological affinities. with four chains for 1,000,000 generations, sampling every 1,000th Typhlocharina. Linear distance matrices between the localities of the generation for a final posterior distribution estimated from 1,000 sequenced specimens were calculated using ARCGIS 9.2 (Environmen- trees. The seg length parameter was set to 0.01, and scaling, MCMC tal Systems Research Institute Inc., Redlands, CA, USA). Phylogenetic move frequencies and remaining operators were used under default distance matrices were computed from branch lengths of the ultra- settings (see Appendix S1: Data S1). Sampling probabilities were metric tree as the estimated age of divergence (Ma) between speci- assigned to the eight main lineages according to the assignation to mens using the R function cophenetic.phylo (Paradis, Claude, & these clades of described species not sampled based on morphologi- Strimmer, 2004). Bivariate plots confronting these matrices were cal traits and combined with alternative hypotheses about the total generated, and Mantel tests with 10,000 randomizations and Spear- expected number of species per clade (case 1: Sampled + man correlation were used to assess the magnitude and significance known = 100%; Case 2: Sampled + known = 75%, Sampled 50%; of the relationships (Oksanen et al., 2013). These analyses were Case 3: Sampled + known = 50%, Sampled 25%; Table 3). Conver- repeated using the phylogeny including all sequenced specimens and gence of the analyses was tested using the code library in R. Support the phylogeny pruned to species (see Section 2.6 section above) to for the presence of rate shifts was checked by Bayes factors and check for the effect of intraspecific variation. In the case of the spe- mean diversification rate and rate-through-time plots were generated cies-pruned trees, the minimum distance between localities of each using the BAMMTOOLS package of R. two species was considered as the minimum geographical distance

Finally, we used the function dd_LR of the R package DDD (Eti- between them. In parallel, for both all-specimens and species-pruned enne et al., 2012) to estimate the maximum likelihood of a diversity- phylogenies, the analyses were performed for the complete tree and dependent diversification model for the phylogenetic branching two lineage-level sections: one with three clusters and other with times of the favoured tree (BD speciation and ULN clock), conduct- eight clusters (matching the main lineages in agreement with mor- ing a bootstrap likelihood ratio test (200 bootstrap replicates) of the phology and distribution, see Section 3), in order to check the phylo- diversity-dependent model (assuming a linear dependence in specia- genetic–geographical relationships. Finally, these analyses were also tion rate with parameter K’ = diversity where speciation = 0) against performed individually for the three species which contrastingly the constant-rates birth–death model (initial values: k = 0.02, showed the largest geographical distributions (i.e., maximum dis- l = 0.01). We have conducted these analyses under the four alter- tances around 200 km between localities): Microcharidius diecki, M. native hypotheses on sampling coverage modifying missnumspec cf. elenae and M. sp.19 aff. toletana. parameter (case 0: missnumspec = 0; case 1: missnumspec = 33; case 2: missnumspec = 74; case 3: missnumspec = 222) and fitting 3 | RESULTS each of these models using two different conditionings on (i) crown age (cond = 0) and (ii) crown age and nonextinction of the phy- 3.1 | A molecular phylogeny of Typhlocharina logeny (See Etienne, Pigot, & Phillimore, 2016; Etienne et al., 2012). Bayesian and ML analyses resulted in highly congruent tree topologies (Figures 2 and 3; Appendix S1: Fig. S2), revealing three well- 2.9 | Phylogenetic vs. geographic distances supported main lineages within Typhlocharina: one for Lusoty- The correlation between phylogenetic and geographic distances was phlus + Typhlocharis (Clade C1) and two for the species of checked at multiple hierarchical levels across the phylogeny of Microcharidius, (Clades C2 and C3). The relationships among them ANDUJAR ET AL. | 6061

FIGURE 2 50% majority rule consensus tree obtained using BI in MRBAYES. Circles on nodes indicate clade support based on posterior probabilities: black, 0.95–1; dark grey, 0.90–0.94; light grey, 0.80–0.89. Images show one representative species for each of the eight main clades, scaled proportionally to their actual sizes. Maps show the distribution of the lineages [Colour figure can be viewed at wileyonlinelibrary.com]

are uncertain due to disagreement among the different analyses (MR- group, we defined eight main lineages within Typhlocharina: Lusoty-

BAYES, RAXML and BEAST with alternative clock and tree priors) and phlus (C1A), Typhlocharis (C1B) and six lineages of Microcharidius short branches. Most interspecific relationships within clades C1, C2 (three main lineages within C2 and three main clades within C3), and C3 showed posterior probability >0.95 for Bayesian and boot- identified in Figures 2 and 3, Table 3; Appendix S1: Fig. S2. strap support >80 for ML analyses, with the exception of clade BEAST analyses resulted in highly similar topologies and node monastica that lacked support for several internal nodes. In accor- ages (mean age of the root for the different analyses between 59 dance with these results and considering the morphology of the and 63 Ma; 95% HPD interval 50–71 Ma) independently of the 6062 | ANDUJAR ET AL.

FIGURE 3 Ultrametric time-calibrated tree from BEAST analyses favoured by BF comparisons (ULN clock, Birth–death speciation model, root age nonconstrained), pruned to a single terminal branch per species (74 terminal branches). Schematic tree on the left includes all terminals, and the vertical red line marks the GMYC single threshold age used as a guide to the final species hypotheses. Circles on nodes indicate clade support based on posterior probabilities: black, 0.95–1; dark grey, 0.90–0.94; light grey, 0.80–0.89. Grey bars represent the 95% HPDI of node ages. Names of clades as in Figure 3 [Colour figure can be viewed at wileyonlinelibrary.com] inclusion/exclusion of the calibration constraint on the root age, model and the constraint on the root age (Table 2, Figure 3). The the application of an SC or ULN clock to the mitochondrial genes ages for the eight main lineages within Typhlocharina are summa- or the use of a speciation BD or BDIS model (Table 2). Analyses rized in Table 3. under a Yule pure birth model resulted in the same topology but with a trend to younger ages for the root node (mean age of the 3.2 | Species delimitation in Typhlocharina root 49–54 Ma; 95% HPD interval 41–61 Ma). Comparisons based on Bayes factors favoured the use of an ULN clock on all genes The GMYC and bPTP methods were applied to the maximum clade and BD or BDIS models (Table 2). Differences in Bayes factors credibility trees obtained with the favoured ULN clock (six trees with among analyses with and without the root age constraint and alternative speciation model and with/without the root age prior). between BD or BDIS models were small. The overall best score Results for the six trees were highly consistent. bPTP and the multi- was obtained using the ULN clock on all genes, a BD speciation ple threshold approach of GMYC resulted in a higher-than-expected ANDUJAR ET AL. | 6063

TABLE 4 Summary of delineation analyses conducted with pPTP and GMYC methods on different phylogenetic trees, indicating the number of species-level entities obtained within the maximum-likelihood (ML) result, time threshold for the species delineation and confidence values when available. *** p-value < 0.001

GMYC single GMYC multiple pPTP LR test LR test Tree ML result ML result Interval null model Threshold time ML result Interval null model

BEAST (Free root age; ULN; yule) 117 78 53–99 0*** 4.201809 138 72–143 0***

BEAST (Free root age; ULN; yule_BDIS) 120 78 52–127 0*** 4.026375 130 128–131 0***

BEAST (Free root age; ULN; yule_BD) 113 79 78–94 0*** 4.155471 78 52–134 0***

BEAST (Const. root age; ULN; yule) 117 75 53–95 0*** 4.573274 84 56–139 0***

BEAST (Const. root age; ULN; yule_BDIS) 113 78 51–94 0*** 4.067902 125 70–131 0***

BEAST (Const. root age; ULN; yule_BD) 120 78 53–91 0*** 3.997224 134 54–134 0***

RAXML 122 n.a. n.a. n.a. n.a. n.a. n.a. n.a.

MRBAYES strict 126 n.a. n.a. n.a. n.a. n.a. n.a. n.a.

MRBAYES 50% majority rule 111 n.a. n.a. n.a. n.a. n.a. n.a. n.a.

FIGURE 4 Results for species–area–age relationships for the eight main clades of Typhlocharina. (a) Correlation between species richness and lineage age (richness–age relationship); (b) correlation between range size and lineage age (area–age relationship); and (c) correlation between range size and species richness (area–richness relationship) [Colour figure can be viewed at wileyonlinelibrary.com] number of species (110–130 species) for which, in many cases, the case of M. hiekei and M. wrasei matching the two already- specimens morphologically identical and collected in the same local- described morphological species, and M. cf. crespoi and M. aff. crespoi, ity (same soil sample) were split into two or more species (Table 4). and M. sp. 32 aff. carpetana and M. sp. 33. The single threshold approach of the GMYC resulted in a delimitation of 75–79 entities mostly consistent with the morphology and 3.3 | The process of diversification in Typhlocharina distribution of the studied populations (Table 4). This set was slightly modified to maximize the fit with morphology and distribution, The final hypothesis of 74 species was used to prune the maxi- resulting in a final hypothesis of 74 species. These modifications mum clade credibility trees obtained with alternative speciation included the following cases (highlighted in Appendix S1: Fig. S2 models and with/without the root age prior (see above), keeping with “*”): (i) specimens from the same locality and morphologically one representative per species (74 terminals). Across the eight main indistinguishable, split by GMYC, which were considered a single lineages of Typhlocharina, range size and species richness were entity in the final delimitation: the cases of Microcharidius scrofa, strongly positively correlated (Pearson r: .97, p value: <.001; Spear- M. cf. estrellae and M. sp. 28; (ii) specimens collected in different man r: .92, p value: <.001). No significant correlation was found localities and morphologically distinguishable recovered as one between range size and lineage age (Pearson r: .18, p value: .33; GMYC entity, which were split into two entities: the pairs M. sp. 16 Spearman r: .21, p value: .31) nor between species richness and lin- and M. sp. 17 and M. cf. farinosae and M. cf. elenae; (iii) closely eage age (Pearson r: 0.05, p value: .55; Spearman r: .33, p value: related specimens whose molecular topology and morphological .21; Figure 4). Convergence of BAMM analyses was checked using traits did not agree on reciprocal monophyly (consistent with molec- effective sample sizes of the log-likelihood estimated with the func- ular ancestral polymorphism), for which two entities were retained: tion “effectiveSize” from the R library coda and a burn-in fraction 6064 | ANDUJAR ET AL.

FIGURE 5 Relationship between geographical and phylogenetic distances by pairs of populations for the three larger lineages (C1, C2 and C3) and the eight main clades within Typhlocharina. The dotted horizontal red line on each graph marks the age threshold from GMYC analyses. Maps show the distribution of the lineages [Colour figure can be viewed at wileyonlinelibrary.com] of 20%. Effective sample sizes were in all cases higher than 500. areas. Only a few clades showed closely related species/populations

TREEPAR favoured a model with one rate shift, corresponding to a at very distant areas, which could point to rare events of long-dis- decrease in the net diversification rate around 4 Ma (note that tance passive dispersal (Figure 5; Appendix S1: Figs. S5, S6). diversification events beyond this age are not considered, as this is Lineage C1 (genus Lusotyphlus + genus Typhlocharis) showed a mostly at the intraspecific level), and with a constant diversification triangular pattern in the phylogenetic distance vs. geographic dis- from the root to that point (Appendix S1: Fig. S3). BAMM analyses tance (PDGD) plots, with no closely related populations/species at resulted in no evidence for significant shifts in the diversification large geographic distances (Pearson r = .54; p value = .009); the rate for particular lineages, but showed a trend (BAMM analyses) to same pattern was found for lineage C1B (genus Typhlocharis; r = .56; lower rates towards the present for all subclades of Typhlocharina p value = .009). Lineage C2 (genus Microcharidius part 1) showed a (Appendix S1: Fig. S4). DDD analyses for sampling coverage cases similar triangular pattern (r = .61; p value = .009), with some data 0 and 1 (missnumspec = 0 and 33, respectively) resulted in a signifi- points at the intraspecific level at large geographic distances indica- cantly better fit of the obtained phylogeny to a diversity-dependent tive of recent long-distance colonization. Within C2, lineage C2C diversification model (p value <.05) despite the relatively low power (clade quadridentata) showed a clean triangular relationship (r = .81; of the test (values from 0.15 to 0.059). For case 2 (missnum- p value = .01), whereas in the monospecific lineage C2B (clade spec = 74), results were nonsignificant, and for case 3, the analyses diecki) there was a high correlation between the different popula- were computationally untreatable due to the large size of the simu- tions of the species and their distances in the tree, but with a much lated trees (Appendix S1: Table S4). lower slope (r = .93; p value = .009; Appendix S1: Fig. S6), suggest- The phylogenetic and geographic distances at multiple levels ing recent long-distance colonization events. were strongly correlated across the phylogeny of Typhlocharina, as Lineage C3 (genus Microcharidius part 2) also showed a correla- expected if dispersal mainly leads to the colonization of adjacent tion between phylogenetic and geographic distances (r = .71; p ANDUJAR ET AL. | 6065 value = .009), although some closely related populations showed with continuous range expansion. Different factors have to be large geographic distances. The correlation was weaker in lineage responsible for the positive species–area correlation, which we pro- C3A (clade outereloi; r = .34; p value = .009), which included closely pose to be related to species interactions that prevent the further related populations at large geographic distances; this pattern was build-up of species diversity. unchanged when intraspecific relationships were not considered DDD analyses point in the same direction, resulting in a signifi- (r = .34; p value = .009; Appendix S1: Fig. S5). Similarly, the correla- cantly better fit of the obtained phylogeny to a diversity-dependent tion for lineage C3C (clade belenae) was weak (r = .28; p value = .01) diversification model when we considered all sampled and all known and even nonsignificant when intraspecific relationships were not species. Nevertheless, DDD analyses have been questioned recently considered (r = 0.11; p value = .09). Finally, the lineages C1A (genus to produce inaccurate results with a high proportion of false posi- Lusotyphlus), C2A (clade gomezi) and C3B (clade monastica) were con- tives detecting density dependence (Etienne et al., 2016), and conse- sistent with the triangular correlation (absence of closely related quently, these results should be considered cautiously. The slow- populations/species at large geographic distances), despite the low down of the rate of lineage growth towards the recent when the number of populations/species included on each of them. phylogeny is considered as whole (Appendix S1: Fig. S3) may be interpreted as an artefact that results from integrating over the diversification rates from different lineages at the equilibrium rich- 4 | DISCUSSION ness phase under a constant model (artifactual positive diversification rate; Nee et al., 1994; Rabosky, 2009) and related to 4.1 | Diversification with a reduced spatial scale for incomplete sampling (artifactual decay; Nee et al., 1994), as we have speciation observed how the inclusion of missing taxa moderates the decay in The phylogeny of Typhlocharina confirmed a strong pattern of the estimated diversification rates (Appendix S1: Fig. S3). Neverthe- microendemicity and geographic structure in all main lineages, which less, BAMM analyses should be also considered cautiously due to the together with the positive relationship between the number of spe- open discussion about the performance of the software (Moore, cies in a lineage and the area it occupies are in accordance with a Hohna,€ May, Rannala, & Huelsenbeck, 2016; Rabosky, Mitchell, & dominant role of space in constraining diversification. Chang, 2017). The three main branches of Typhlocharina (C1: Lusoty- Overall, our results indicate the existence of a limited load capac- phlus + Typhlocharis; C2: Microcharidius part 1; C3: Microcharidius ity for the richness of Typhlocharina lineages, which is related to the part 2) overlap spatially with each other through a large part of the size of the area each lineage occupies. Thus, we would expect differ- Iberian Peninsula. Each of these clades can be divided into two or ent phases during the diversification of a lineage (Figure 1a). There three morphologically distinctive subclades (Perez-Gonz alez et al., should be an initial accumulation phase where richness increases 2017) virtually without overlapping distributions (Figure 5), for a while empty areas and/or niches are filled, with the speed of species total of eight main lineages. In Lusotyphlus, Typhlocharis and within accumulation being proportional to the net diversification rate of the six clades of Microcharidius, species richness is positively corre- species formation (Rabosky, 2009). The Typhlocharina in the Iberian lated with the range size of the lineage but not with its estimated Peninsula apparently has reached the subsequent equilibrium or pla- stem age. The existence of ecological limits to species richness at teau phase, where species richness is predicted to oscillate around different scales has gained support (Ezard, Purvis, & Morlon, 2016; the load capacity, at a null net diversification rate through time (Hey, Rabosky, 2013; Rabosky & Hurlbert, 2015), although it remains a 1992; Rabosky, 2009). Under a DDD process, if extinction rates controversial topic (Harmon & Harrison, 2015; Marshall & Quental, have been high relative to speciation rates and lineages are old 2016). Positive species–area correlations have been proposed to enough, long stem branches are expected (Figure 1b). As lineages go reflect the ecological movement distributions of organisms as well as extinct, they are replaced by other more recent lineages and, thus, the in situ speciation of lineages (Kisel et al., 2011), and species rich- the shape of the phylogenetic tree frequently is characterized by a ness has been shown to be dependent on the levels of gene flow fairly recent common ancestor from which all living taxa are derived. and the size of the area, even after accounting for the synergistic This may be evident from a young crown group subtended by a long effects of other correlated factors such as habitat heterogeneity stem to its sister lineage, in analogy to the within-species processes (Kisel & Barraclough, 2010). In agreement with our results, it has of the population coalescent model (Kingman, 1982), although in a been suggested that the best evidence in favour of density-depen- phylogenetic tree this shape becomes evident only if stem lineages dent diversification (DDD) comes from the observations of no rela- are old and/or the extinction and diversification rates are relatively tionship between clade ages and clade species richness (Rabosky, high to achieve a complete species turnover within the lineage (Hey, 2009; Ricklefs, 2007), as a positive relationship between age and 1992; Nee et al., 1994; Rabosky & Hurlbert, 2015). The tree of diversity should be apparent even when clades strongly differ in Typhlocharina indeed shows some long branches leading up to the their rates of diversification if clade diversity generally increases crown groups, each of which shows a correlation between lineage through time (Rabosky, 2009). In Typhlocharina, the species diversity size and range size. Within the crown group, we observed a weak is not correlated with lineage age, and neither are lineage area and correlation between richness and crown group ages for the eight lineage age, which argues against a simple lineage growth model subclades, which suggests that rates of replacement are similar but 6066 | ANDUJAR ET AL.

(a) (b)

FIGURE 6 Hypothetical dispersal kernels for lineages with different passive dispersal potential. (2a) High passive dispersal potential; (2b) reduced passive dispersal potential. The high passive dispersal of species in (a) allows the cohesion of species over a longer geographic distances than in (b), with a lower passive dispersal [Colour figure can be viewed at wileyonlinelibrary.com] require more time in larger lineages (i.e., those lineages with more species cohesion to small spatial scales, and (ii) rare events of long- species and larger ranges), in analogy to the longer time to coales- distance dispersal that extend species ranges and lead to the estab- cence in larger populations (Appendix S1: Fig. S7). These findings are lishment of new isolated populations that become additional sources consistent with the existence of diversity-dependent diversification of diversification. limits in Typhlocharina, proportional to the size of the geographic Dispersal kernels (i.e., the statistical distribution of dispersal dis- extension of the lineage. This is expected when the reduced spatial tances) frequently show a positively skewed curve defined by a high scale of speciation is similar among the main lineages of Typhlochar- frequency of dispersive events at shorter distances and an exponen- ina, due to similar constraints in gene flow. Typhlocharina did not tial decay with distance (Nathan, 2006; Tilman & Kareiva, 1997). In exhibit a significant correlation between lineage age and species plants, the particular shape of the dispersal kernels seems to be the richness or range size, which further confirms that the diversification result of several dispersive vectors, with nonstandard vectors being of these lineages is in the plateau phase. frequently responsible for LDDs (Nathan et al., 2008). Similarly, the specificities and strength for dispersal of different soil lineages would define different dispersal kernels and a different spatial scale and 4.2 | A reduced spatial scale for speciation structure for speciation across lineages of soil mesofauna. A gener- Our results revealed a very small spatial scale of species diversifica- ally low active dispersal is likely expected in soil mesofauna as a tion in Typhlocharina, pointing to a very reduced dispersal capacity. result from the low distance of active movement of individuals; nev- The interplay of dispersion and habitat specificity has been proposed ertheless, different strategies for passive dispersal may result in con- as key parameter in determining levels of gene flow, with lower dis- trasting scenarios (Figure 6). For groups inhabiting superficial soil persal determining a finer spatial scale of gene flow and speciation, layers more exposed to dispersal vectors, higher frequency of pas- and leading to higher taxonomic diversity within a given area (Kisel sive dispersal would contribute to the species cohesion through a & Barraclough, 2010). In the case of soil mesofauna, a generally high wider space, increasing the spatial scale of speciation (Figure 6a). dispersal (mainly passive) has been proposed (e.g., Coulson et al., Adaptations to favour passive dispersal are mainly found in organ- 2012; Nkem et al., 2006; Wardle, 2002), similar to what is found for isms specialized in the soil superficial layers, such as the feather-like the microscopic components of soil communities (Green et al., 2004; wings of Ptiliidae (known as featherwing beetles) or the furcula in Ramirez et al., 2014). However, there are multiple studies, most of springtails (Ponge & Salmon, 2013) where passive dispersal seems to them applying molecular techniques, that have pointed to a much be a key living strategy (Coulson, Hodkinson, Webb, & Harrison, more constrained dispersal potential of soil mesoarthropods (Bennett 2002; Nkem et al., 2006). On the contrary, the adaptation to deep et al., 2016; Cicconardi et al., 2010, 2013). This matches the case of soil conditions could strongly constraint passive dispersal by a much Typhlocharina. Our results show a pattern of high species richness, more reduced exposition to dispersal vectors. In these cases, disper- microendemicity and geographical structure which together point to sal kernels are expected to be truncated at very short distances but (i) strong dispersal limitations that weaken gene flow and reduce still a very low chance of LDD may exist, which becomes highly ANDUJAR ET AL. | 6067 probable when considering geological timescales. Despite the passive the Gibraltar Strait about 5 Mya which led to only a few vicariance and unpredictable nature of these LDD events, they will occur more events (see Appendix S1: Fig. S8). (iii) There is a lack of evidence for likely to adjacent areas (Figure 6b), thus promoting high levels of shifts on diversification rate within the subclades of Typhlocharina microendemicity but also geographical structure in these deep soil that could be expected when a punctual event is promoting the spe- lineages (Nathan & Nathan, 2014). This dual dispersal mode could be ciation (Appendix S1: Fig. S3). And (iv), a substantial number of cases promoting contrasting processes of speciation between superficial in which species coexist syntopically (Perez-Gonz alez & Zaballos, and deep soil fauna, resulting in important differences in the distri- 2013b) involve nonsister species, and are best explained by specia- bution and diversity of their communities (Andujar et al., 2015). tion in allopatry with secondary contact due to range expansions. The relationship between geographical and phylogenetic dis- Interestingly, these cases of syntopy usually involve pairs of species tances is consistent with a process of LDD promoting colonization with large differences in size (“syntopic miniaturization”; Sokolov, of adjacent regions and dispersal constraints favouring further isola- 2013) or in morphological traits related to reproduction (e.g., type of tion and speciation. Regarding the vectors involved in the limited female genitalia, shape of aedeagus), suggesting the presence of but recurrent passive dispersal of Typhlocharina across the Iberian reinforcement. Given the coexistence among nonsister species, the Peninsula, hydrochory (i.e., the passive dispersal mechanism medi- presence of reinforcement would imply that species retain the capa- ated by waterborne transport of sediments/specimens) has been bility to interbreed through different cladogenetic events, which in hypothesized to play an important role (Ortuno~ & Gilgado, 2011). turn suggest that speciation processes occurred in allopatry, in the Hydrochory has been demonstrated as an effective mechanism absence of gene flow. transferring soil mesofauna species either at short and at long dis- If generalized across different endogean groups, the proposed pro- tances (e.g., Costello et al., 2010; Terhivuo, Lundqvist, & Saura, cess of geographical diversification with a very reduced spatial scale 2002). Here, we find that main lineages in the phylogeny are mostly for speciation could be one of the main drivers promoting and main- associated with major Iberian watersheds, wherein small-scale diver- taining a high diversity in soil communities. While these lineages sification seems to be dominant (Figure 5). In addition, in our results, would have a moderate contribution to the local diversity (alpha diver- we also found evidence for a few potential events of dispersal to sity), very high levels of species turnover are expected even at a more distant areas at both inter- and intraspecific level, mostly reduced scale, resulting in an important contribution to the overall within main watersheds and so again potentially promoted by trans- gamma diversity of soils and global ecosystems. As an example, some port of sediments/specimens by water also at larger scales. These well-sampled regions such as Corsica or Sardinia, or more restricted are the cases of Microcharidius diecki distributed in the Ebro Basin, regions such as the Maures mountains in the South of France, include M. cf. eleneae in the Guadiana Basin and M.sp19aff. toletana in the dozens of endogean endemics, whereas none or very few endemics Jucar Basin (Appendix S1: Fig. S6). are known in the above-ground communities of the same areas (Cer- These specific geographic patterns and the general “triangular” retti, Mason, Minelli, Nardi, & Whitmore, 2009; Ponel, 1993). This pat- shape of the PDGD plots suggest that LDDs, defined as those dis- tern has been observed more generally for soil beetles (Andujar et al., persal movements beyond a distance that are infrequent enough to 2015), where a moderate contribution of endogean lineages to local not contribute to the species cohesion of species, is a major mecha- diversity was described with a few endogean representatives in com- nism for speciation within Typhlocharina, in agreement with recent munities that could reach hundreds of different species. Nevertheless, studies proposing an important role of LDD in speciation (Nathan & due to the high microendemicity of these endogean lineages, species Nathan, 2014). An effect of other processes in the differentiation of turnover was very high and so the total contribution to gamma biodi- Typhlocharina, such as gradual divergence of populations over time versity and community structure notably increased. In agreement with due to the limited gene flow through distance (isolation-by-distance this pattern, our results suggest a critical role and a reduced scale of scenario), ecological differentiation, for example, due to climatic LDD in the diversification of endogean lineages and highlight the changes, or development of geological barriers (vicariance scenario), expected important contribution of these lineages to total levels and cannot be fully discarded, and probably they contributed to specia- patterns of biodiversity on Earth. tion in particular cases; however, taken together our results support a major role for speciation through colonization via LDD driving ACKNOWLEDGEMENTS diversification in Typhlocharina. We found multiple evidences for an iterative speciation process mediated by colonization and isolation of We thank Rampal Etienne for the useful recommendations about populations mostly in adjacent areas: (i) we do not identify general analyses with the DDD R package, Jose Luis Lencina for providing clines in morphological traits across geography, as expected under some specimen, Rocıo Alonso (IBE Barcelona) and Alex Aitken (NHM isolation-by-distance processes. Conversely, Typhlocharina is made London) for their support in the molecular laboratory and three up by a puzzle of genetically and morphological discrete entities with anonymous Referees for valuable comments. This study was partially small ranges of distribution. (ii) Different lineages of Typhlocharina supported by the NERC Grant NE/M021955. CA was partly founded likely colonized the Baetic region and North Africa only after the by The European Commission (MSCA-IF-2015-705639) and The Baetic region joined the Iberian Peninsula about 15 Mya, due to the Spanish “Ministerio de Economıa y Competitividad” (CGL2010- movement of the Betic-Rifean plate to the west, and the opening of 16944) during the development of this study. 6068 | ANDUJAR ET AL.

DATA ACCESSIBILITY Casale, A. (2009). Adaptive radiation in Mediterranean islands? The case of Reicheiina in Sardinia (Coleoptera, Carabidae, Scaritinae). In All the sequences here used are deposited in GenBank, and acces- S. Casellato, P. Burighel, & A. Minelli (Eds.), Life and time: The evolu- sion numbers for each sequence, with taxonomic and locality infor- tion of life and its history (pp. 75–86). Padova: Cleup. mation for the sequenced specimens, are provided in the Cerretti, P., Mason, F., Minelli, A., Nardi, G., & Whitmore, D. (Eds.) (2009). Research on the terrestrial arthropods of sardinia. Zootaxa. Appendix S1. Phylogenetic trees are deposited in Dryad: https://doi. Auckland, New Zealand: Magnolia Press. org/10.5061/dryad.bk78q. Cicconardi, F., Fanciulli, P., & Emerson, B. (2013). Collembola, the biological species concept and the underestimation of global species richness. Molecular Ecology, 2, 5382–5396. AUTHORS'S CONTRIBUTIONS Cicconardi, F., Nardi, F., Emerson, B. C., Frati, F., & Fanciulli, P. P. (2010). Deep phylogeographic divisions and long-term persistence of forest C.A. and I.R. conceived the study; C.A., P.A., S.P.G. and J.P.Z. col- invertebrates (Hexapoda: Collembola) in the North-Western Mediter- lected the specimens; C.A., P.A., A.P.V. and IR designed the analyses; ranean basin. Molecular Ecology, 19, 386–400. Costello, D. M., Tiegs, S. D., & Lamberti, G. A. (2010). Do non-native C.A., P.A. and S.P.G. did the molecular lab work and performed the earthworms in Southeast Alaska use streams as invasional corridors analyses; C.A. led the writing and all author contributed to the dis- in watersheds harvested for timber? Biological Invasions, 13, 177–187. cussion of the results and the final writing of the manuscript. Coulson, S. J., Hodkinson, I. D., Webb, N. R., & Harrison, J. A. (2002). Sur- vival of terrestrial soil-dwelling arthropods on and in seawater: Implica- tions for trans-oceanic dispersal. Functional Ecology, 16, 353–356. ORCID Coulson, S. J., Hodkinson, I. D., Webb, N. R., & Harrison, J. A. (2012). Survival of terrestrial soil-dwelling arthropods on and in seawater: Carmelo Andujar http://orcid.org/0000-0001-9759-7402 Implications for trans-oceanic dispersal. Functional Ecology, 16, 353– Alfried P. Vogler http://orcid.org/0000-0002-2462-3718 356. Courtillot, V., & Gaudemer, Y. (1996). Effects of mass extinctions on bio- Ignacio Ribera http://orcid.org/0000-0002-2791-7615 diversity. Nature, 381, 146–148. Coyne, J. A., & Orr, H. A. (2004). Speciation. Sunderland, MA: Sinauer Associates. REFERENCES Darriba, D., Taboada, G. L., Doallo, R., & Posada, D. (2012). JMODELTEST 2: More models, new heuristics and parallel computing. Nature Methods, Alroy, J. (1996). Constant extinction, constrained diversification, and 9, 772. uncoordinated stasis in North American Mammals. Palaeogeography, Decaens,€ T. (2010). Macroecological patterns in soil communities. Global Palaeoclimatology, Palaeoecology, 127, 285–311. Ecology and Biogeography, 19, 287–302. Andujar, C., Arribas, P., Ruzicka, F., Crampton-Platt, A., Timmermans, M. Diniz-Filho, J., & Bini, L. (2011). Geographical patterns in biodiversity: J. T. N., & Vogler, A. (2015). Phylogenetic community ecology of soil Towards an integration of concepts and methods from genes to spe- biodiversity using mitochondrial metagenomics. Molecular Ecology, 24, cies diversity. Brazilian Journal of Nature Conservation Research Letters 3603–3617. Natureza & Conservacßao~ , 9,1–9. Andujar, C., Faille, A., Perez-Gonz alez, S., Zaballos, J. P., Vogler, A. P., & Drummond, A. J., Suchard, M. A., Xie, D., & Rambaut, A. (2012). Bayesian Ribera, I. (2016). Gondwanian relicts and oceanic dispersal in a cos- phylogenetics with BEAUTI and the BEAST 1.7. Molecular Biology and mopolitan radiation of euedaphic ground beetles. Molecular Phyloge- Evolution, 29, 1969–1973. netics and Evolution, 99, 235–246. Eisenbeis, G., & Wichard, W. (1987). Atlas on the biology of soil arthro- Andujar, C., Serrano, J., & Gomez-Zurita, J. (2012). Winding up the pods. Berlin: Springer-Verlag. molecular clock in the genus Carabus (Coleoptera: Carabidae): Etienne, R. S., Haegeman, B., Stadler, T., Aze, T., Pearson, P. N., Purvis, Assessment of methodological decisions on rate and node age esti- A., & Phillimore, A. B. (2012). Diversity-dependence brings molecular mation. BMC Evolutionary Biology, 12, 40. phylogenies closer to agreement with the fossil record. Proceedings of Andujar, C., Soria-Carrasco, V., Serrano, J., & Gomez-Zurita, J. (2014). the Royal Society B: Biological Sciences, 279, 1300–1309. Congruence test of molecular clock calibration hypotheses based on Etienne, R. S., Pigot, A. L., & Phillimore, A. B. (2016). How reliably can Bayes factor comparisons (R Freckleton, Ed). Methods in Ecology and we infer diversity-dependent diversification from phylogenies? Meth- Evolution, 5, 226–242. ods in Ecology and Evolution, 7, 1092–1099. Astrin, J. J., Zhou, X., & Misof, B. (2013). The importance of biobanking Ettema, C. H., & Wardle, D. A. (2002). Spatial soil ecology. Trends in Ecol- in molecular taxonomy, with proposed definitions for vouchers in a ogy and Evolution, 17, 177–183. molecular context. ZooKeys, 365,67–70. Ezard, T., Fujisawa, T., & Barraclough, T. G. (2009) SPLITS: species’ limits Baele, G., Li, W. L. S., Drummond, A. J., Suchard, M. A., & Lemey, P. by threshold statistics. R package version, 1. (2013). Accurate model selection of relaxed molecular clocks in baye- Ezard, T. H. G., Purvis, A., & Morlon, H. (2016). Environmental changes sian phylogenetics. Molecular Biology and Evolution, 30, 239–243. define ecological limits to species richness and reveal the mode of Baselga, A., Fujisawa, T., Crampton-Platt, A., Bergsten, J., Foster, P. G., macroevolutionary competition. Ecology Letters, 19, 899–906. Monaghan, M. T., & Vogler, A. P. (2013). Whole-community DNA Felsenstein, J. (1985). Confidence limits on phylogenies: An approach barcoding reveals a spatio-temporal continuum of biodiversity at spe- using the bootstrap. Evolution, 39, 783–791. cies and genetic levels. Nature Communications, 4, 1892, http://org. Finlay, B. J. (2002). Global dispersal of free-living microbial eukaryote doi/10.1038/ncomms2881. species. Science, 296, 1061–1063. Bengston, P. (1988). Open nomenclature. Palaeontology, 31, 223–227. Foote, M. (1997). The evolution of morphological diversity. Annual Review Bennett, K. R., Hogg, I. D., Adams, B. J., & Hebert, P. D. N. (2016). High of Ecology and Systematics, 28, 129–152. levels of intraspecific genetic divergences revealed for Antarctic Fritz, S. A., Jønsson, K. A., Fjeldsa, J., & Rahbek, C. (2012). Diversification springtails: Evidence for small-scale isolation during Pleistocene and biogeographic patterns in four island radiations of passerine glaciation. Biological Journal of the Linnean Society, 119, 166–178. birds. Evolution, 66, 179–190. ANDUJAR ET AL. | 6069

Fujisawa, T., & Barraclough, T. G., (2013). Delimiting species using single- Moore, B. R., Hohna,€ S., May, M. R., Rannala, B., & Huelsenbeck, J. P. locus data and the generalized mixed yule coalescent approach: A (2016). Critically evaluating the theory and performance of Bayesian revised method and evaluation on simulated data sets. Systematic analysis of macroevolutionary mixtures. Proceedings of the National Biology, 62, 707–724. Academy of Sciences, 113, 201518659. Futuyma, D. J. (1998). Evolutionary biology, 3rd edn. Sunderland, MA: Sin- Nathan, R. (2005). Long-distance dispersal research: Building a network auer Associates. of yellow brick roads. Diversity and Distributions, 11, 125–130. Giachino, P., & Vailati, D. (2010). The subterranean environment. In W. Nathan, R. (2006). Long-distance dispersal of plants. Science (New York, B. A. Onlus (Ed.), Hypogean life, concepts and collecting techniques. N.Y.), 313, 786–788. Italy: WBA Handbooks, Verona. Nathan, R., & Nathan, O. (2014). Unlikely yet pivotal long dispersals. Gisin, H. (1943). Okologie€ und Lebensgemeinschaften der collembolen im Science, 344, 153–154. Schweizarischen Exkursionsgebiet Basels. Revue Suisse de Zoologie, Nathan, R., Schurr, F. M., Spiegel, O., Steinitz, O., Trakhtenbrot, A., & 50, 131–224. Tsoar, A. (2008). Mechanisms of long-distance seed dispersal. Trends Green, J. L., Holmes, A. J., Westoby, M., Oliver, I., Briscoe, D., Danger- in Ecology and Evolution, 23, 638–647. field, M., ... Beattie, A. J. (2004). Spatial scaling of microbial eukary- Nee, S., Holmes, E. C., May, R. M., & Harvey, P. H. (1994). Extinction ote diversity. Nature, 432, 747–750. rates can be estimated from molecular phylogenies. Philosophical Harmon, L. J., & Harrison, S. (2015). Species diversity is dynamic and Transactions of the Royal Society of London Series B, Biological Sciences, unbounded at local and continental scales. The American Naturalist, 344,77–82. 185, 584–593. Nkem, J. N., Wall, D. H., Virginia, R. A., Barrett, J. E., Broos, E. J., Pora- Harte, J., McCarthy, S., Taylor, K., Kinzig, A., & Fischer, M. L. (1999). Esti- zinska, D. L., & Adams, B. J. (2006). Wind dispersal of soil inverte- mating species–area relationships from plot to landscape scale using brates in the McMurdo Dry Valleys, Antarctica. Polar Biology, 29, species spatial-turnover data. Oikos, 86,45–54. 346–352. Hey, J. (1992). Using phylogenetic trees to study speciation and extinc- Oksanen, J., Blanchet, F. G., Kindt, R., Legendre, P., Minchin, P. R., tion. Evolution, 46, 640. O’Hara, R. B., ... Wagner, H. (2013). VEGAN: Community ecology pack- Huelsenbeck, J. P., Ronquist, F., Nielsen, R., & Bollback, J. P. (2001). age. R package version 2.0-10. http://cran.r-project.org/package=ve Bayesian inference of phylogeny and its impact on evolutionary biol- gan. R package ver. 2.0–8, 254. ogy. Science, 294, 2310–2314. Ortuno,~ V. M., & Gilgado, J. D. (2011). Historical perspective, new contri- Jeanne, C. (1973). Sur la classification des Bembidiides endoges de la butions and an enlightening dispersal mechanism for the endogean region euro-mediterraneenne (Col. Carabidae, Bembidiinae, Anillini). genus Typhlocharis Dieck 1869 (Coleoptera: Carabidae: Trechinae). Nouvelle Revue d’Entomologie, 3,83–102. Journal of Natural History, 45, 1233–1256. Jeannel, R. (1963). Monographie des Anillini bembidiides endoges: Coleop- Paradis, E., Claude, J., & Strimmer, K. (2004). APE: Analyses of phylogenet- tera Trechidae. Paris: Memoires du Museum national d’histoire natur- ics and evolution in R language. Bioinformatics, 20, 289–290. elle. Perez-Gonz alez, S., Andujar, C., & Zaballos, J. P. (2017). Hidden biodiver- Katoh, K., Misawa, K., Kuma, K., & Miyata, T. (2002). MAFFT: A novel sity: Total evidence phylogenetics and evolution of morphological method for rapid multiple sequence alignment based on fast Fourier traits in a highly diverse lineage of endogean ground beetles, Typhlo- transform. Nucleic Acids Research, 30, 3059–3066. charis Dieck, 1869 (Carabidae, Trechinae, Anillini). Cladistics,1–25. Kembel, S. W., Cowan, P. D., Helmus, M. R., Cornwell, W. K., Morlon, H., http://org.doi/10.1111/cla.12208 Ackerly, D. D., ... Webb, C. O. (2010). PICANTE: R tools for integrating Perez-Gonz alez, S., & Zaballos, J. P. (2013a). Tarsal tetramery and phylogenies and ecology. Bioinformatics, 26, 1463–4. extreme size reduction in Anillini (Coleoptera, Carabidae, Trechinae): Kingman, J. F. C. (1982). The coalescent. Stochastic Processes and their The case of Typhlocharis Dieck, 1869; description of three new spe- Applications, 13, 235–248. cies and definition of a new intra-generic species group. Zootaxa, Kisel, Y., & Barraclough, T. G. (2010). Speciation has a spatial scale that 3682, 249–269. depends on levels of gene flow. The American Naturalist, 175, 316– Perez-Gonz alez, S., & Zaballos, J. P. (2013b). Four new species of Typhlo- 334. charis (baetica group) (Coleoptera: Carabidae: Anillini) from south- Kisel, Y., McInnes, L., Toomey, N. H., & Orme, C. D. L. (2011). How diver- western Iberian Peninsula with notes on their biogeographical and sification rates and diversity limits combine to create large-scale spe- morphological implications. Systematic Entomology, 38, 104–122. cies-area relationships. Philosophical Transactions of the Royal Society Perez-Gonz alez, S., Zaballos, J., & Ghannem, S. (2013). Intraspecific varia- of London Series B, Biological Sciences, 366, 2514–2525. tion in Typhlocharis Dieck, 1869 (Coleoptera, Carabidae, Anillini): The Korfiatis, K., & Stamou, G. (1999). Habitat templets and the changing case of two new species of the baetica group. Zootaxa, 3710,46–60. worldview of ecology. Biology and Philosophy, 14, 375–393. Pigot, A. L., Owens, I. P. F., & Orme, C. D. L. (2010). The environmental limits Lomolino, M. V., Riddle, B. R., Whittaker, R. J., & Brown, J. H. (2010). Bio- to geographic range expansion in birds. Ecology Letters, 13,705–715. geography (4th ed.). Sunderland, MA: Sinauer Associates. Ponel, P. (1993). Coleopt eres de Massif des Maures et de la depression Lorenz, W. (2005). Systematic list of extant ground beetles of the world permienne periph erique. Faune de Provence, 14,5–23. (Insecta Coleoptera “Geadephaga”: Trachypachidae and Carabidae incl. Ponge, J. F., & Salmon, S. (2013). Spatial and taxonomic correlates of Paussinae, Cicindelinae, Rhysodinae). 2 edn. Tutzing, Germany: Wolf- species and species trait assemblages in soil invertebrate communi- gang Lorenz. ties. Pedobiologia, 56, 129–136. Marshall, C. R., & Quental, T. B. (2016). The uncertain role of diversity Pons, J., Barraclough, T., Gomez-Zurita, J., Cardoso, A., Duran, D. P., dependence in species diversification and the need to incorporate Hazell, S., ... Vogler, A. P. (2006). Sequence-based species delimita- time-varying carrying capacities. Philosophical Transactions of the tion for the DNA taxonomy of undescribed insects. Systematic Biol- Royal Society B: Biological Sciences, 371, 20150217. ogy, 55, 595–609. Miller, M, Pfeiffer, W, & Schwartz, T. (2010). Creating the CIPRES Rabosky, D. L. (2009). Ecological limits and diversification rate: Alterna- Science Gateway for inference of large phylogenetic trees. In: Pro- tive paradigms to explain the variation in species richness among ceedings of the gateway computing environments workshop (GCE). IEEE, clades and regions. Ecology Letters, 12, 735–743. New Orleans, LA, pp. 1–8. Rabosky, D. L. (2013). Diversity-dependence, ecological speciation, and Moen, D., & Morlon, H. (2014). Why does diversification slow down? the role of competition in macroevolution. Annual Review of Ecology, Trends in Ecology and Evolution, 29, 190–197. Evolution, and Systematics, 44, 481–502. 6070 | ANDUJAR ET AL.

Rabosky, D. L. (2014). Automatic detection of key innovations, rate shifts, Stamatakis, A. (2006). RAXML-VI-HPC: Maximum likelihood-based phyloge- and diversity-dependence on phylogenetic trees. PLoS ONE, 9, e89543. netic analyses with thousands of taxa and mixed models. Bioinformat- Rabosky, D. L., Donnellan, S. C., Grundler, M., & Lovette, I. J. (2014). Analy- ics (Oxford, England), 22, 2688–2690. sis and visualization of complex Macroevolutionary dynamics: An exam- Stanley, S. M. (1979). Macroevolution: Pattern and process. San Francisco, ple from Australian Scincid lizards. Systematic Biology, 63, 610–627. CA: Freeman. Rabosky, D. L., & Hurlbert, A. H. (2015). Species richness at continental Talavera, G., Dinca, V., & Vila, R. (2013). Factors affecting species delimi- scales is dominated by ecological limits. The American Naturalist, 185, tations with the GMYC model: Insights from a butterfly survey (E 572–583. Paradis, Ed). Methods in Ecology and Evolution, 4, 1101–1110. Rabosky, D. L., Mitchell, J. S., & Chang, J. (2017). Is BAMM flawed? The- Terhivuo, J., Lundqvist, E., & Saura, A. (2002). Clone diversity of Eiseniella oretical and practical concerns in the analysis of multi-rate diversifi- tetraedra (Oligochaeta: Lumbricidae) along regulated and free-flowing cation models. Systematic Biology, preprint. boreal rivers. Ecography, 25, 714–720. Rambaut, A. (2012). FIGTREE v.1.4.2 Available at: http://tree.bio.ed.ac. Tilman, D., & Kareiva, P. (1997). Spatial ecology: The role of space in popu- uk/software/figtree/. lation dynamics and interspecific interactions. Princeton: Princeton Uni- Rambaut, A., & Drummond, A. J. (2010). TREESTAT v1.6.1: tree statistic cal- versity Press. culation tool. Wardle, D. (2002). Communities and Ecosystems: Linking the aboveground Rambaut, A., Suchard, M. A., Xie, W., & Drummond, A. J. (2013). TRACER and belowground components. Princeton: Princeton University Press. v1.6. Available at: http://tree.bio.ed.ac.uk/software/tracer/. Wilkinson, D. M., Koumoutsaris, S., Mitchell, E. A. D., & Bey, I. (2012). Ramirez, K. S., Leff, J. W., Barberan, A., Bates, S. T., Betley, J., Thomas, Modelling the effect of size on the aerial dispersal of microorganisms. W., ... Fierer, N. (2014). Biogeographic patterns in below-ground Journal of Biogeography, 39,89–97. diversity in New York City’s Central Park are similar to those Zaballos, J. P., Andujar, C., & Perez-Gonz alez, S. (2016). A small carabid observed globally. Proceedings of the Royal Society B, 281, 20141988. for a huge entomologist: Typhlocharis amara, a new species dedicated Ribera, I. (2008). Habitat constraints and the generation of diversity in to Dr. Fritz Hieke. Entomologische Blatter€ und Coleoptera, 112, 465– freshwater macroinvertebrates. In J. Lancaster, & R. Briers (Eds.), 473. Aquatic Insects. Challenges to populations (pp. 289–312). Trowbridge: Zaballos, Juan P., & Ruiz-Tapiador, I. (1997). Nuevos Typhlocharis Dieck Cromwell Press. (Coleoptera, Caraboidea, Trechidae) de Espana.~ Graellsia, 52,95–106. Ribera, I., Castro, A., Dıaz, J. A., Garrido, J., Izquierdo, A., Jach,€ M. A., & Zaballos, J. P., & Wrase, D. W. (1998). Tres nuevos Typhlocharis Dieck, Valladares, L. F. (2011). The geography of speciation in narrow-range 1869 (Coleoptera, Caraboidea, Trechidae) de Navarra (Espana).~ Graell- endemics of the “Haenydra” lineage (Coleoptera, Hydraenidae, sia, 54,43–52. Hydraena). Journal of Biogeography, 38, 502–516. Zhang, J., Kapli, P., Pavlidis, P., & Stamatakis, A. (2013). A general species Ricklefs, R. E. (2007). Estimating diversification rates from phylogenetic delimitation method with applications to phylogenetic placements. information. Trends in Ecology and Evolution, 22, 601–610. Bioinformatics (Oxford, England), 29, 2869–2876. Russell, E. J., & Appleyard, A. (1915). The atmosphere of the soil: Its composition and the causes of variation. The Journal of Agricultural Science, 7,1–48. Sepkoski, J. J., Bambach, R. K., Raup, D. M., & Valentine, J. W. (1981). SUPPORTING INFORMATION Phanerozoic marine diversity and the fossil record. Nature, 293, 435– Additional Supporting Information may be found online in the sup- 437. Sokolov, I. M. (2013). A new genus and eight new species of the subtribe porting information tab for this article. Anillina (Carabidae, Trechinae, Bembidiini) from Mexico, with a cladis- tic analysis and some notes on the evolution of the genus. ZooKeys, 92,51–92. Sokolov, I. M. (2015). Review of the species of Pelodiaetodes Moore How to cite this article: Andujar C, Perez-Gonzalez S, Arribas (Coleoptera: Carabidae: Bembidiini: Anillina) of New Zealand. Zootaxa, P, Zaballos JP, Vogler AP, Ribera I. Speciation below ground: 3963, 561–582. Tempo and mode of diversification in a radiation of endogean Southwood, T. (1977). Habitat, the templet for ecological strategies? ground beetles. Mol Ecol. 2017;26:6053–6070. Journal of Animal Ecology, 46, 336–365. Stadler, T. (2011). Mammalian phylogeny reveals recent diversification https://doi.org/10.1111/mec.14358 rate shifts. Proceedings of the National Academy of Sciences of the Uni- ted States of America, 108, 6187–6192. SUPPLEMENTARY MATERIALS

Speciation below ground: Tempo and mode of diversification in a

radiation of endogean ground beetles

Carmelo Andújar, Sergio Pérez, Paula Arribas, Juan Zaballos, Alfried P. Vogler, Ignacio Ribera

Table S1. Specimens and Genbank accession number.

Table S2. Localities and collectors.

Table S3. Primers used in the study. F, forward; R, reverse.

Table S4. Initial parameters and results for the bootstrap likelihood ratio test of DDD model against the constant-rates birth-death model.

Data S1. Identification criteria and taxonomic decisions.

Data S2. Lines used in R for DDD simulations in figure 1B.

Figure S1. Distribution map of Typhlocharina.

Figure S2. Ultrametric time calibrated tree obtained in BEAST.

Figure S3. TreePar results for the ultrametric time calibrated tree obtained in BEAST

Figure S4. BAMM results for the ultrametric time calibrated tree obtained in BEAST.

Figure S5. Results for the relationship between geographical and phylogenetic distances by pares of species for the three superlineages (C1, C2 y C3) and the 8 main lineages within Typhlocharina.

Figure S6. Results for the relationship between geographical and phylogenetic distances by pares of populations within the three species with a wider range of distribution (Microcharidius diecki, M. cf elenae and M. sp 19 aff toletana).

Figure S7. Results for correlation between species richness and lineage crown age for the 8 main clades.

Figure S8. Ultrametric tree pruned to a single representative per species indicating distribution in or out from the Baetic-Rifean plate. Table S1. Specimens and Genbank accession number.

Species Voucher COD Cox1-bc-5' Cox1-bc-3' rrnL LSU SSU Lusotyphlus L. lunai BMNH-1046040 SLGU MF539269 MF539116 MF538825 MF538889 L. lunai BMNH-1046041 SLGU MF539270 MF539117 MF539000 MF538890 L. paulinoi BMNH-1424408 MIRA MF539528 L. paulinoi BMNH-1424409 MIRA MF539529 L. paulinoi BMNH-1424410 MIRA MF539530 L. paulinoi BMNH-1424411 MIRA MF539531 MF539197 MF539081 MF538961 Typhlocharis T. acutangula BMNH-1046036 SSGZ MF539265 MF539114 MF538998 MF538887 T. acutangula BMNH-1046037 SSGZ MF539266 T. acutangula BMNH-1046178 SSGZ MF539382 MF539166 MF539049 MF538938 T. armata BMNH-1046248 SALM MF539448 T. armata BMNH-1046249 SALM MF539449 MF539172 MF539057 MF538945 T. armata BMNH-1046250 SALM MF539450 T. armata CA90 SALM MF539611 MF539230 T. cf. deferreri BMNH-1041174 BENA2 MF539236 MF539105 MF538987 MF538877 T. cf. deferreri BMNH-1041970 BENA1 MF539241 T. cf. deferreri BMNH-1424413 BNMH MF539533 MF539198 MF538856 MF539082 MF538962 T. cf. deferreri BMNH-1424414 BNMH MF539534 T. cf. laurentii CA68 ALMA MF539227 MF538871 MF538979 T. martini BMNH-1046222 ALCA MF539425 KU976288 KU978713 KU978667 T. martini CA18 FPRA MF539589 HM009024 T. martini CA21 FPAR MF539591 MF539213 T. martini CA23 FPAR MF539592 MF539214 T. martini CA29 FPRA MF539597 MF539217 T. martini CA30 FPRA MF539598 KU976296 KU978722 KU978676 T. martini IBE1189 FPAR MF539615 MF539232 MF539102 MF538983 T. mixta BMNH-1046044 REVE MF539272 T. mixta BMNH-1046045 REVE MF539273 MF539119 MF539002 MF538892 T. mixta BMNH-1046046 REVE MF539274 T. mixta BMNH-1046179 REVE MF539383 T. mixta BMNH-1046180 REVE MF539384 T. mixta BMNH-1046181 REVE MF539385 T. mixta BMNH-1046182 REVE MF539386 T. mixta BMNH-1046183 REVE MF539387 T. pacensis BMNH-1046148 AJIT MF539353 MF539155 MF539038 MF538927 T. silvanoides BMNH-1046245 SNSL MF539445 KU976289 KU978623 KU978714 KU978668 T. silvanoides BMNH-1046246 SNSL MF539446 MF539056 MF538944 T. silvanoides BMNH-1046247 SNSL MF539447 T. silvanoides CA32 SNSL MF539218 T. sp. 1 aff. armata BMNH-1041166 BENA2 MF539235 MF539104 MF538817 MF538986 MF538876 T. sp. 1 aff. armata BMNH-1041891 BENA1 MF539237 T. sp. 1 aff. armata BMNH-1041946 BENA1 MF539238 T. sp. 1 aff. armata BMNH-1041962 BENA1 MF539239 T. sp. 1 aff. armata BMNH-1041963 BENA1 MF539240 T. sp. 1 aff. armata BMNH-1041971 BENA1 MF539242 T. sp. 1 aff. armata BMNH-1046051 UBRI MF539279 MF539122 MF539005 MF538895 T. sp. 1 aff. armata BMNH-1046052 UBRI MF539280 T. sp. 1 aff. armata BMNH-1046054 UBRI MF539281 T. sp. 1 aff. armata BMNH-1046187 UBRI MF539391 T. sp. 1 aff. armata BMNH-1046188 UBRI MF539392 T. sp. 1 aff. armata BMNH-1046189 UBRI MF539393 T. sp. 1 aff. armata BMNH-1046190 UBRI MF539394 T. sp. 1 aff. armata BMNH-1046310 UBRI MF539500 T. sp. 1 aff. armata BMNH-1046311 UBRI MF539501 T. sp. 2 BMNH-1046035 CARM MF539264 MF539113 MF538997 MF538886 T. sp. 3 aff. baetica BMNH-1042236 ALCU1 MF539255 MF539111 MF538993 T. sp. 3 aff. baetica BMNH-1046169 CEBO MF539374 MF539161 MF538838 MF539044 MF538933 T. sp. 3 aff. baetica BMNH-1424416 ALCU MF539535 MF539199 MF539083 MF538963 T. sp. 37 CA34 LARA MF539599 T. sp. 37 CA35 LARA MF539600 KU976298 KU978631 KU978724 KU978678 T. sp. 39 CA33 SNSL KU976297 KU978630 KU978723 KU978677 T. sp. 4 aff. baetica BMNH-1046161 SSPR MF539366 T. sp. 4 aff. baetica BMNH-1046162 SSPR MF539367 T. sp. 4 aff. baetica BMNH-1046163 SSPR MF539368 MF539159 MF538836 MF539042 MF538931 T. sp. 4 aff. baetica BMNH-1046164 SSPR MF539369 T. sp. 4 aff. baetica BMNH-1046165 SSPR MF539370 T. sp. 4 aff. baetica BMNH-1046166 GARL MF539371 T. sp. 4 aff. baetica BMNH-1046167 GARL MF539372 MF539160 MF538837 MF539043 MF538932 T. sp. 4 aff. baetica BMNH-1046168 GARL MF539373 T. sp. 4 aff. baetica BMNH-1046171 ADOH MF539375 MF539163 MF538840 MF539046 MF538935 T. sp. 40 aff. armata BMNH-1042046 APAJ MF539247 MF539107 MF538819 MF538989 MF538879 T. sp. 46 aff. laurentii BMNH-1046243 ALMA MF539444 MF539171 MF539055 T. sp. 5 BMNH-1046047 REVE MF539275 MF539120 MF539003 MF538893 T. sp. 6 BMNH-1046143 ZAOS MF539351 MF539153 MF539036 MF538925 T. sp. 6 BMNH-1046145 ZAOS MF539352 MF539154 MF539037 MF538926 T. sp. 7 BMNH-1046138 HVAR MF539346 T. sp. 7 BMNH-1046139 HVAR MF539347 MF539152 MF538834 MF539035 MF538924 T. sp. 7 BMNH-1046140 HVAR MF539348 T. sp. 7 BMNH-1046141 HVAR MF539349 T. sp. 7 BMNH-1046142 HVAR MF539350 T. sp. 7 BMNH-1046293 HVAR MF539487 MF539188 MF539073 T. sp. 7 BMNH-1046294 HVAR MF539488 T. sp. 7 BMNH-1046295 HVAR MF539489 T. sp. 8 BMNH-1046048 VBUR MF539276 T. sp. 8 BMNH-1046049 VBUR MF539277 MF539121 MF539004 MF538894 T. sp. 8 BMNH-1046050 VBUR MF539278 T. sp. 8 BMNH-1046127 GDJI MF539337 T. sp. 8 BMNH-1046128 GDJI MF539338 T. sp. 8 BMNH-1046129 GDJI MF539339 MF539149 MF539032 MF538921 T. sp. 8 BMNH-1046130 GDJI MF539340 T. sp. 8 BMNH-1046131 ALBU MF539341 T. sp. 8 BMNH-1046133 ALBU MF539342 T. sp. 8 BMNH-1046134 ALBU MF539343 T. sp. 8 BMNH-1046135 ALBU MF539344 T. sp. 8 BMNH-1046136 VLEG MF539345 MF539151 MF539034 MF538923 T. sp. 8 BMNH-1046184 VBUR MF539388 T. sp. 8 BMNH-1046185 VBUR MF539389 T. sp. 8 BMNH-1046186 VBUR MF539390 T. sp. 8 BMNH-1046285 GDJI MF539479 MF539180 MF539065 T. sp. 8 BMNH-1046286 GDJI MF539480 MF539181 MF539066 T. sp. 8 BMNH-1046287 GDJI MF539481 MF539182 MF539067 T. sp. 8 BMNH-1046288 GDJI MF539482 MF539183 MF539068 T. sp. 8 BMNH-1046289 GDJI MF539483 MF539184 MF539069 T. sp. 8 BMNH-1046290 ALBU MF539484 MF539185 MF539070 T. sp. 8 BMNH-1046291 ALBU MF539485 MF539186 MF539071 T. sp. 8 BMNH-1046292 ALBU MF539486 MF539187 MF539072 T. sp. 8 BMNH-1046308 VBUR MF539499 T. sp. 8 BMNH-1046132 ALBU MF539150 MF539033 MF538922 Microcharidius M. amara sp. n. 22 BMNH-1424384 BJOZ MF539514 M. amara sp. n. 22 BMNH-1424385 BJOZ MF539515 M. amara sp. n. 22 BMNH-1424386 BJOZ MF539516 M. amara sp. n. 22 BMNH-1424387 BJOZ MF539517 MF539192 MF538853 MF539077 MF538956 M. baeturicus BMNH-1046043 SLGU MF539271 MF539118 MF538826 MF539001 MF538891 M. bazi BMNH-1424374 ALPD MF539507 M. bazi BMNH-1424375 ALPD MF539508 M. bazi BMNH-1424376 ALPD MF539509 MF539189 MF538852 MF539074 MF538953 M. cf. crespoi BMNH-1424377 PINX1 MF539510 M. cf. crespoi BMNH-1424378 PINX1 MF539511 MF539190 MF539075 MF538954 M. cf. crespoi BMNH-1424379 PINX1 MF539512 MF539191 MF539076 MF538955 M. cf. crespoi BMNH-1424380 PINX1 MF539513 M. cf. crespoi BMNH-1424388 CFCH MF539193 MF538957 M. cf. elenae BMNH-1046038 SLGU MF539267 M. cf. elenae BMNH-1046039 SLGU MF539268 MF539115 MF538824 MF538999 MF538888 M. cf. elenae BMNH-1046056 MERI MF539283 M. cf. elenae BMNH-1046057 MERI MF539284 MF539124 MF538827 MF539007 MF538897 M. cf. elenae BMNH-1046058 MERI MF539285 M. cf. elenae BMNH-1046059 TRUJ MF539286 M. cf. elenae BMNH-1046060 TRUJ MF539287 MF539125 MF539008 MF538898 M. cf. elenae BMNH-1046061 CORN MF539288 MF539126 MF539009 MF538899 M. cf. elenae BMNH-1046062 CORN MF539289 M. cf. elenae BMNH-1046063 MERI MF539290 M. cf. elenae BMNH-1046064 MERI MF539291 M. cf. elenae BMNH-1046173 TMBU MF539377 MF539164 MF538841 MF539047 MF538936 M. cf. elenae BMNH-1046174 TMBU MF539378 M. cf. elenae BMNH-1046175 PALM MF539379 MF539165 MF538842 MF539048 MF538937 M. cf. elenae BMNH-1046191 MERI MF539395 M. cf. elenae BMNH-1046192 MERI MF539396 M. cf. elenae BMNH-1046193 TRUJ MF539397 M. cf. elenae BMNH-1046194 TRUJ MF539398 M. cf. elenae BMNH-1046195 CORN MF539399 M. cf. elenae BMNH-1046196 CORN MF539400 M. cf. elenae BMNH-1046312 MERI MF539502 M. cf. elenae BMNH-1046313 TRUJ MF539503 M. cf. elenae BMNH-1424398 AL_CV MF539521 M. cf. elenae BMNH-1424400 AL_CV MF539522 M. cf. elenae BMNH-1424401 AL_CV MF539523 M. cf. elenae BMNH-1046170 CEBO MF539162 MF538839 MF539045 MF538934 M. cf. elenae larva BMNH-1046072 MERI MF539298 M. cf. estrellae BMNH-1046257 BECE MF539455 KU976290 KU978624 KU978715 KU978669 M. cf. estrellae BMNH-1046258 BECE MF539456 M. cf. estrellae CA11 NAVA MF539586 MF539208 M. cf. estrellae CA121 BECE MF539587 M. cf. estrellae CA66 BECE MF539609 HM009031 M. cf. estrellae IBE1188 BECE MF539614 KU976329 KU978658 KU978754 KU978705 M. cf. farinosae BMNH-1046149 PREY MF539354 M. cf. farinosae BMNH-1046150 PREY MF539355 MF539156 MF539039 MF538928 M. cf. farinosae BMNH-1046151 PREY MF539356 M. cf. farinosae BMNH-1046152 PREY MF539357 M. cf. farinosae BMNH-1046153 PREY MF539358 M. cf. farinosae BMNH-1046299 PREY MF539490 M. cf. farinosae BMNH-1046300 PREY MF539491 M. cf. farinosae BMNH-1046301 PREY MF539492 M. cf. gonzaloi CA80 JA_BE MF539610 MF539099 M. cf. gonzaloi CA79 JA_BE MF539228 MF539098 MF538980 M. diecki BMNH-1046241 MNCY MF539442 MF538846 MF539054 MF538943 M. diecki BMNH-1046242 MNCY MF539443 M. diecki BMNH-1424462 POBL1 MF539570 M. diecki BMNH-1424463 POBL1 MF539571 M. diecki BMNH-1424464 POBL1 MF539572 M. diecki BMNH-1424466 POBL2 MF539573 M. diecki BMNH-1424467 POBL2 MF539574 M. diecki BMNH-1424468 POBL2 MF539575 M. diecki BMNH-1424470 POBL3 MF539576 MF539204 MF538859 MF539088 MF538968 M. diecki BMNH-1424471 POBL3 MF539577 M. diecki BMNH-1424472 POBL3 MF539578 M. diecki BMNH-1424480 SGUA2 MF539582 MF539205 MF538860 MF539089 MF538969 M. diecki CA44 MNCY MF539604 M. diecki CA41 ARTJ MF539221 MF538866 MF539094 MF538975 M. diecki CA43 MNCY MF538867 MF539095 MF538976 M. diecki CA47 SESM KU976299 KU978632 KU978725 KU978679 M. hiekei BMNH-1046076 TOMS MF539302 M. hiekei BMNH-1046078 TOMS MF539303 M. hiekei BMNH-1046079 TOMS MF539304 M. hiekei BMNH-1046114 TOMS MF539334 M. hiekei BMNH-1046077 TOMS MF539130 MF538828 MF539013 MF538903 M. josabelae CA3 SLLA MF539602 KU976295 KU978629 KU978721 KU978675 M. josabelae CA4 SLLA MF539607 MF539224 M. josabelae CA81 JA_BE2 MF538872 MF539100 MF538981 M. monasticus IBE882 MMUG MF539616 KU976333 KU978662 KU978759 KU978709 M. monasticus IBE879 MMUG MF539234 MF538875 MF538985 M. outereloi BMNH-1424394 TORT MF539518 MF539194 MF538854 MF539078 MF538958 M. outereloi BMNH-1424395 TORT MF539519 M. outereloi BMNH-1424396 TORT MF539520 M. outereloi BMNH-1424412 TORT MF539532 M. peregrinus CA40 ARTJ MF539603 MF539220 MF538865 MF539093 MF538974 M. peregrinus CA45 NAGO MF539605 MF539222 MF538868 MF539096 MF538977 M. peregrinus CA46 NAGO MF539606 MF539223 M. portilloi BMNH-1046069 GORD MF539295 M. portilloi BMNH-1046070 GORD MF539296 MF539128 MF539011 MF538901 M. portilloi BMNH-1046071 GORD MF539297 M. portilloi BMNH-1046201 GORD MF539405 M. portilloi BMNH-1046202 GORD MF539406 M. portilloi BMNH-1046203 GORD MF539407 M. scrofa BMNH-1046030 ADEH MF539260 M. scrofa BMNH-1046031 ADEH MF539261 MF538995 MF538884 M. scrofa BMNH-1046032 ADEH MF539262 M. scrofa BMNH-1046176 ADEH MF539380 M. scrofa BMNH-1046177 ADEH MF539381 M. scrofa BMNH-1046306 ADEH MF539497 M. scrofa BMNH-1046307 ADEH MF539498 M. sp. 10 BMNH-1046055 UBRI MF539282 MF539123 MF539006 MF538896 M. sp. 11 BMNH-1042245 CBNC MF539256 MF538823 MF538994 MF538883 M. sp. 11 BMNH-1042253 CBNC MF539257 M. sp. 11 BMNH-1042261 CBNC MF539258 M. sp. 12 BMNH-1424421 ALIS1 MF539540 MF539201 MF538857 MF539085 MF538965 M. sp. 12 BMNH-1424422 ALIS1 MF539541 M. sp. 12 BMNH-1424423 ALIS1 MF539542 M. sp. 13 BMNH-1424417 ALCU2 MF539536 MF539200 MF539084 MF538964 M. sp. 13 BMNH-1424418 ALCU2 MF539537 M. sp. 13 BMNH-1424419 ALCU2 MF539538 M. sp. 13 BMNH-1424420 ALCU2 MF539539 M. sp. 14 BMNH-1042212 ANBT MF539254 MF539110 MF538822 MF538992 MF538882 M. sp. 14 BMNH-1046256 ANBT MF539454 MF539174 MF538848 MF539059 MF538947 M. sp. 14 CA1 ANBT MF539590 M. sp. 14 CA25 ALCU3 MF539593 MF539216 M. sp. 14 IBE1187 ANBT MF539613 MF539231 MF538873 MF539101 MF538982 M. sp. 15 aff. gonzaloi BMNH-1046111 RTEJ MF539333 MF539147 MF539030 MF538919 M. sp. 15 aff. gonzaloi BMNH-1046231 CCANT MF539433 MF539169 MF538845 MF539052 MF538941 M. sp. 15 aff. gonzaloi BMNH-1046232 CCANT MF539434 M. sp. 15 aff. gonzaloi BMNH-1046233 CCANT MF539435 M. sp. 15 aff. gonzaloi BMNH-1046234 CCANT MF539436 M. sp. 15 aff. gonzaloi BMNH-1046235 CCANT MF539437 M. sp. 15 aff. gonzaloi BMNH-1046269 CONS MF539467 M. sp. 15 aff. gonzaloi BMNH-1046270 CONS MF539468 MF539177 MF538851 MF539062 MF538950 M. sp. 15 aff. gonzaloi BMNH-1046271 CONS MF539469 M. sp. 15 aff. gonzaloi BMNH-1046272 CONS MF539470 M. sp. 15 aff. gonzaloi BMNH-1046273 CONS MF539471 M. sp. 16 aff. toletanus BMNH-1046227 CCANT MF539429 M. sp. 16 aff. toletanus BMNH-1046228 CCANT MF539430 MF539168 MF538844 MF539051 MF538940 M. sp. 16 aff. toletanus BMNH-1046229 CCANT MF539431 M. sp. 16 aff. toletanus BMNH-1046230 CCANT MF539432 M. sp. 17 aff. toletanus BMNH-1046251 CNAJ MF539451 M. sp. 17 aff. toletanus BMNH-1046252 CNAJ MF539452 MF539173 MF538847 MF539058 MF538946 M. sp. 17 aff. toletanus BMNH-1046253 CNAJ MF539453 M. sp. 17 aff. toletanus CA14 CNAJ MF539588 MF539211 MF538862 MF539090 MF538971 M. sp. 17 aff. toletanus CA17 CNAJ MF539212 M. sp. 18 aff. toletanus BMNH-1046236 CNEG MF539438 M. sp. 18 aff. toletanus BMNH-1046237 CNEG MF539439 MF539170 MF539053 MF538942 M. sp. 18 aff. toletanus BMNH-1046238 CNEG MF539440 M. sp. 18 aff. toletanus BMNH-1046239 CNEG MF539441 M. sp. 18 aff. toletanus BMNH-1046259 FPAR MF539457 M. sp. 18 aff. toletanus BMNH-1046260 FPAR MF539458 MF539175 MF538849 MF539060 MF538948 M. sp. 18 aff. toletanus BMNH-1046261 FPAR MF539459 M. sp. 18 aff. toletanus BMNH-1046262 FPAR MF539460 M. sp. 18 aff. toletanus BMNH-1046263 FPAR MF539461 M. sp. 18 aff. toletanus CA12 CNEG MF539209 M. sp. 18 aff. toletanus CA13 CNEG MF539210 M. sp. 18 aff. toletanus CA24 CNEG MF539215 MF538863 MF539091 MF538972 M. sp. 18 aff. toletanus CA8 FPAR MF539229 M. sp. 18 aff. toletanus IBE1190 FPAR MF539233 MF538874 MF539103 MF538984 M. sp. 19 aff. toletanus BMNH-1046108 ZZAN MF539331 M. sp. 19 aff. toletanus BMNH-1046109 ZZAN MF539332 MF539146 MF539029 MF538918 M. sp. 19 aff. toletanus BMNH-1046216 ZZAN MF539419 M. sp. 19 aff. toletanus BMNH-1046217 ZZAN MF539420 M. sp. 19 aff. toletanus BMNH-1046218 ZZAN MF539421 M. sp. 19 aff. toletanus BMNH-1046219 ZZAN MF539422 M. sp. 19 aff. toletanus BMNH-1046220 ZZAN MF539423 M. sp. 19 aff. toletanus BMNH-1046221 ZZAN MF539424 M. sp. 19 aff. toletanus BMNH-1046223 CPAJ MF539426 MF539167 MF538843 MF539050 MF538939 M. sp. 19 aff. toletanus BMNH-1046224 CPAJ MF539427 M. sp. 19 aff. toletanus BMNH-1046225 CPAJ MF539428 M. sp. 19 aff. toletanus BMNH-1046264 FSCH MF539462 M. sp. 19 aff. toletanus BMNH-1046265 FSCH MF539463 MF539176 MF538850 MF539061 MF538949 M. sp. 19 aff. toletanus BMNH-1046266 FSCH MF539464 M. sp. 19 aff. toletanus BMNH-1046267 FSCH MF539465 M. sp. 19 aff. toletanus BMNH-1046268 FSCH MF539466 M. sp. 19 aff. toletanus BMNH-1424430 UNDE MF539549 M. sp. 19 aff. toletanus BMNH-1424431 TOLO MF539550 M. sp. 19 aff. toletanus BMNH-1424432 FSCH MF539551 M. sp. 19 aff. toletanus BMNH-1424433 ALCR3 MF539552 M. sp. 19 aff. toletanus BMNH-1424435 ALCR3 MF539553 M. sp. 19 aff. toletanus BMNH-1424436 ALCR3 MF539554 M. sp. 19 aff. toletanus BMNH-1424437 ALCR5 MF539555 M. sp. 19 aff. toletanus BMNH-1424438 ALCR5 MF539556 M. sp. 19 aff. toletanus BMNH-1424439 ALCR5 MF539557 M. sp. 19 aff. toletanus BMNH-1424440 ALCR5 MF539558 M. sp. 19 aff. toletanus BMNH-1424441 ALCR6 MF539559 M. sp. 19 aff. toletanus BMNH-1424442 ALCR6 MF539560 M. sp. 19 aff. toletanus BMNH-1424443 ALCR6 MF539561 M. sp. 19 aff. toletanus BMNH-1424444 ALCR1 MF539562 M. sp. 19 aff. toletanus BMNH-1424445 ALCR1 MF539563 M. sp. 19 aff. toletanus BMNH-1424446 ALCR1 MF539564 MF539203 MF539087 MF538967 M. sp. 19 aff. toletanus BMNH-1424447 ALCR2 MF539565 M. sp. 19 aff. toletanus BMNH-1424448 ALCR2 MF539566 M. sp. 19 aff. toletanus BMNH-1424449 ALCR4 MF539567 M. sp. 19 aff. toletanus BMNH-1424450 ALCR4 MF539568 M. sp. 19 aff. toletanus BMNH-1424452 ALCR4 MF539569 M. sp. 20 CA10 ESTN MF539585 MF539207 M. sp. 20 IBE1186 ESTN MF539612 KU976328 KU978657 KU978753 KU978704 M. sp. 21 BMNH-1041993 CVJE MF539243 MF539106 MF538818 MF538988 MF538878 M. sp. 21 BMNH-1042001 CVJE MF539244 M. sp. 21 BMNH-1042009 CVJE MF539245 M. sp. 21 BMNH-1042025 CVJE MF539246 M. sp. 21 BMNH-1042183 CFRA MF539248 MF539108 MF538820 MF538990 MF538880 M. sp. 21 BMNH-1042191 CFRA MF539250 M. sp. 21 BMNH-1042199 CFRA MF539252 M. sp. 21 BMNH-1046125 CVJE MF539336 MF539148 MF538833 MF539031 MF538920 M. sp. 21 CA38 ZUHE MF539601 MF539219 MF538864 MF539092 MF538973 M. sp. 23 aff. belenae BMNH-1046091 MPLA MF539314 MF539136 MF538830 MF539019 MF538908 M. sp. 23 aff. belenae BMNH-1046092 MPLA MF539315 MF539137 MF538831 MF539020 MF538909 M. sp. 24 aff. bullaquensis BMNH-1046093 GRIM MF539316 M. sp. 24 aff. bullaquensis BMNH-1046094 GRIM MF539317 MF539138 MF539021 MF538910 M. sp. 25 BMNH-1046073 ALUZ MF539299 MF539129 MF539012 MF538902 M. sp. 25 BMNH-1046074 ALUZ MF539300 M. sp. 25 BMNH-1046075 ALUZ MF539301 M. sp. 26 BMNH-1046095 ZMAY1 MF539318 MF539139 MF538832 MF539022 MF538911 M. sp. 27 BMNH-1046096 MO_CI MF539319 M. sp. 27 BMNH-1046097 MO_CI MF539320 MF539140 MF539023 MF538912 M. sp. 27 BMNH-1046211 MO_CI MF539414 M. sp. 27 BMNH-1046212 MO_CI MF539415 M. sp. 28 BMNH-1046087 ZMAY2 MF539310 M. sp. 28 BMNH-1046088 ZMAY2 MF539311 MF539134 MF539017 MF538907 M. sp. 28 BMNH-1046089 ZMAY2 MF539312 M. sp. 28 BMNH-1046208 ZMAY2 MF539412 M. sp. 28 BMNH-1046315 ZMAY2 MF539505 M. sp. 29 BMNH-1046101 PZAR MF539324 M. sp. 29 BMNH-1046102 PZAR MF539325 MF539142 MF539025 MF538914 M. sp. 30 BMNH-1046105 SCPG2 MF539328 M. sp. 30 BMNH-1046106 SCPG2 MF539329 MF539144 MF539027 MF538916 M. sp. 30 BMNH-1046107 SCPG1 MF539330 MF539145 MF539028 MF538917 M. sp. 30 BMNH-1046214 SCPG2 MF539417 M. sp. 30 BMNH-1046215 SCPG2 MF539418 M. sp. 30 BMNH-1046316 SCPG2 MF539506 M. sp. 31 BMNH-1046154 VCAB MF539359 MF539157 MF539040 MF538929 M. sp. 31 BMNH-1046155 VCAB MF539360 M. sp. 32 aff. carpetana BMNH-1046103 PZAR MF539326 M. sp. 32 aff. carpetana BMNH-1046104 PZAR MF539327 MF539143 MF539026 MF538915 M. sp. 32 aff. carpetana BMNH-1046213 PZAR MF539416 M. sp. 33 BMNH-1046098 CORI MF539321 M. sp. 33 BMNH-1046099 CORI MF539322 MF539141 MF539024 MF538913 M. sp. 33 BMNH-1046100 CORI MF539323 M. sp. 34 aff. elenae BMNH-1046156 SSPR MF539361 M. sp. 34 aff. elenae BMNH-1046157 SSPR MF539362 MF539158 MF538835 MF539041 MF538930 M. sp. 34 aff. elenae BMNH-1046158 SSPR MF539363 M. sp. 34 aff. elenae BMNH-1046159 SSPR MF539364 M. sp. 34 aff. elenae BMNH-1046160 SSPR MF539365 M. sp. 34 aff. elenae BMNH-1046172 TMBU MF539376 M. sp. 34 aff. elenae BMNH-1046302 SSPR MF539493 M. sp. 34 aff. elenae BMNH-1046303 SSPR MF539494 M. sp. 34 aff. elenae BMNH-1046305 SSPR MF539496 M. sp. 35 BMNH-1046210 GRIM MF539413 M. sp. 36 BMNH-1424424 ALIS1 MF539543 MF539202 MF538858 MF539086 MF538966 M. sp. 36 BMNH-1424425 ALIS1 MF539544 M. sp. 36 BMNH-1424426 ALIS1 MF539545 M. sp. 36 BMNH-1424427 ALIS1 MF539546 M. sp. 36 BMNH-1424428 ALIS1 MF539547 M. sp. 36 BMNH-1424429 ALIS2 MF539548 M. sp. 38 BMNH-1424402 SSUS MF539524 MF539195 MF539079 MF538959 M. sp. 41 aff. atienzai BMNH-1046080 TOMS MF539305 MF539131 MF538829 MF539014 MF538904 M. sp. 41 aff. atienzai BMNH-1046084 SALO MF539308 M. sp. 41 aff. atienzai BMNH-1046086 SALO MF539309 MF539133 MF539016 MF538906 M. sp. 41 aff. atienzai BMNH-1046090 VENC MF539313 MF539135 MF539018 M. sp. 41 aff. atienzai BMNH-1046117 SALO MF539335 M. sp. 41 aff. atienzai BMNH-1046206 SALO MF539410 M. sp. 41 aff. atienzai BMNH-1046207 SALO MF539411 M. sp. 41 aff. atienzai BMNH-1046314 SALO MF539504 M. sp. 42 aff. bazi BMNH-1046275 INVR MF539472 M. sp. 42 aff. bazi BMNH-1046276 INVR MF539473 MF539178 MF539063 MF538951 M. sp. 42 aff. bazi BMNH-1046277 INVR MF539474 M. sp. 42 aff. bazi BMNH-1046278 INVR MF539475 M. sp. 42 aff. bazi BMNH-1046304 INVR MF539495 M. sp. 42 aff. bazi CA53 INVR MF539608 MF539225 MF538869 MF539097 MF538978 M. sp. 42 aff. bazi CA55 INVR MF539226 MF538870 M. sp. 43 aff. peregrinus BMNH-1424476 SGUA1 MF539579 M. sp. 43 aff. peregrinus BMNH-1424477 SGUA1 MF539580 M. sp. 43 aff. peregrinus BMNH-1424479 SGUA2 MF539581 M. sp. 43 aff. peregrinus BMNH-1424482 SGUA3 MF539583 MF539206 MF538861 MF538970 M. sp. 43 aff. peregrinus BMNH-1424483 SGUA3 MF539584 M. sp. 44 aff. crespoi BMNH-1424404 CIMA MF539525 M. sp. 44 aff. crespoi BMNH-1424405 CIMA MF539526 MF539196 MF538855 MF539080 MF538960 M. sp. 44 aff. crespoi BMNH-1424406 CIMA MF539527 M. sp. 45 aff. scrofa BMNH-1046033 SILL MF539263 MF539112 MF538996 MF538885 M. sp. 9 BMNH-1042187 FCAL MF539249 MF539109 MF538821 MF538991 MF538881 M. sp. 9 BMNH-1042195 FCAL MF539251 M. sp. 9 BMNH-1042203 FCAL MF539253 M. sp. 9 BMNH-1042266 FCAL MF539259 M. tetramerus BMNH-1046065 MERI MF539292 M. tetramerus BMNH-1046066 MERI MF539293 MF539127 MF539010 MF538900 M. tetramerus BMNH-1046067 MERI MF539294 M. tetramerus BMNH-1046197 MERI MF539401 M. tetramerus BMNH-1046198 MERI MF539402 M. tetramerus BMNH-1046199 MERI MF539403 M. tetramerus BMNH-1046200 MERI MF539404 M. toletanus BMNH-1046279 AVIT MF539476 M. toletanus BMNH-1046280 AVIT MF539477 MF539179 MF539064 MF538952 M. toletanus BMNH-1046281 AVIT MF539478 M. toletanus CA26 AVIT MF539594 HM009025 M. toletanus CA27 AVIT MF539595 KU976294 KU978628 KU978720 KU978674 M. toletanus CA28 AVIT MF539596 HM009027 M. wrasei BMNH-1046082 MEMB MF539306 MF539132 MF539015 MF538905 M. wrasei BMNH-1046083 MEMB MF539307 M. wrasei BMNH-1046204 MEMB MF539408 M. wrasei BMNH-1046205 MEMB MF539409 Geocharis sp. CA91 SALM KU976324 KU978654 KU978749 KU978700

Geocharis sp. BMNH-1041887 AEAS MF928072 KU976285 KU978620 KU978710

Table S2. Localities and collectors

COD LOCALITY Altitude Latitude Longitude

ADEH Arroyo La Dehesa, Cumbres de San Bartolomé, Huelva, SPAIN 358 38.067 -6.767

ADOH Arroyo Dos Hermanos, Zarza-Capilla, Badajoz, SPAIN 449 38.818 -5.236

AJIT Arroyo Jituero, Azuaga, Badajoz, SPAIN 523 38.223 -5.680

AL_CV N2, Aljustrel-Castro Verde, Alentejo, PORTUGAL 175 37.779 -8.099

ALBU Río La Albuera, Almendral, Badajoz, SPAIN 309 38.625 -6.830

ALCA Alcadima, Liétor, Albacete, SPAIN 610 38.549 -2.038

ALCR1 Valdeolivas, Alcarria, Cuenca, SPAIN 940 40.511 -2.436

ALCR2 Arroyo de la Dehesa, Alcarria, Guadalajara, SPAIN 898 40.534 -2.463

ALCR3 Camino del Corral de Ratón, Alcarria, Cuenca, SPAIN 958 40.529 -2.421

ALCR4 Arroyo de la Zarzuela, Alcarria, Guadalajara, SPAIN 900 40.539 -2.407

ALCR5 Arroyo de Valdegraciano, Alcarria, Cuenca, SPAIN 830 40.489 -2.470

ALCR6 Camino del Cuento, Alcarria, Cuenca, SPAIN 890 40.503 -2.459

ALCU Valle de Alcudia, Almodovar del Campo, Ciudad Real, SPAIN 720 38.553 -4.364

ALCU1 Valle de Alcudia, Arroyo de Cerrosado, Almodovar del Campo, Ciudad Real, SPAIN 720 38.553 -4.364

ALCU2 Valle de Alcudia, Cañada del hato de Garro, Almodovar del Campo, Ciudad Real, SPAIN 735 38.539 -4.311

ALCU3 Valle de Alcudia, Ermita de Pedro Morillo, Almodóvar del Campo, Ciudad Real, SPAIN 715 38.553 -4.360

ALIS1 Arroyo de la Aliseda, Ventillas, Ciudad Real, SPAIN 720 38.487 -4.279

ALIS2 Arroyo de la Aliseda, Ventillas, Ciudad Real, SPAIN 705 38.492 -4.289

ALMA Los Lobos, Sierra Almagrera, Cuevas de Almanzora, Almeria, SPAIN 61 37.311 -1.765

ALPD Alpedrete de la Sierra, 2.2 km O, Alpedrete de la Sierra, Guadalajara, SPAIN 936 40.914 -3.433

ALPO Alportel, Alportel, Algarve, PORTUGAL 283 37.175 -7.904

ALUZ Arroyo de la Luz 10.6km NE, Arroyo de la Luz, Cáceres, SPAIN 392 39.531 -6.501

ANBT Arroyo de Navalacuenca, N-420 km 118, Brazatortas, Ciudad Real, SPAIN 778 38.518 -4.382

ANBT Arroyo de Navalacuenca, N-420 km 118, Brazatortas, Ciudad Real, SPAIN 780 38.519 -4.380

APAJ Arroyo de los Pajares, Llanos del Republicano, Villaluenga del Rosario, Cádiz, SPAIN 810 36.682 -5.357

ARTJ Artajona, Artajona, Navarra, SPAIN 470 42.626 -1.777

AVIT Arroyo de Vitoria, Villarrubia de Santiago, Toledo, SPAIN 590 40.020 -3.300

BECE Arroyo Becea, La Fuencaliente, Ciudad Real, SPAIN 627 39.178 -3.995

BENA1 La Dehesilla, Benaocaz, Cádiz, SPAIN 470 36.707 -5.457

BENA2 La Dehesilla, Benaocaz, Cádiz, SPAIN 480 36.707 -5.457

BJOZ Badajoz, 3km NO, Badajoz, Badajoz, SPAIN 218 38.913 -6.993

BNMH Las Cuevas, Benamahoma, Sierra de Grazalema, SPAIN 750 36.750 -5.427

CARM Carmona, Carmona, Sevilla, SPAIN 156 37.483 -5.617

CBNC Nava de Cabra, Cortijo de los Benítez, Cabra, Córdoba, SPAIN 995 37.486 -4.363

CCANT Cerro de las Canteras, Fuentealbilla, Albacete, SPAIN 669 39.226 -1.543

CEBO Arroyo Cebolloso, Zarza-Capilla, Badajoz, SPAIN 476 38.821 -5.280

CFCH Canafechal, CM-168, Canafechal, Algarve, PORTUGAL 51 37.194 -8.647

CFRA Cortijo de los Frailes, CP-89, Cabra, Córdoba, SPAIN 710 37.468 -4.393

CIMA Montes de Cima, 2km S, Montes de Cima, Algarve, PORTUGAL 75 37.223 -8.609

CNAJ El Cenajo, Hellín, Albacete, SPAIN 375 38.380 -1.765

CNEG Cueva Negra, Aýna, Albacete, SPAIN 707 38.545 -2.156

CONS Rambla de la Consolación, Iniesta, Cuenca, SPAIN 678 39.451 -1.503

CORI Coria 6.15km NE, Coria, Cáceres, SPAIN 299 40.019 -6.478 CORN Parque Nat. de Cornalvo, Trujillanos, Badajoz, SPAIN 333 38.983 -6.183

CORS Corsino, 1.1 Km S, Corsino, Algarve, PORTUGAL 108 37.225 -8.726

CPAJ Cañada de Pajares, Molito del Tobillo, Alpera, Albacete, SPAIN 908 39.992 -1.295

CVJE Camino Viejo a la Ermita, Cabra, Córdoba, SPAIN 970 37.481 -4.389

ESTN Río Estena, CM-4106 km 40, Helechosa de los Montes, Badajoz, SPAIN 475 39.381 -4.736

FCAL N-420 km 101, Fuencaliente, Ciudad Real, SPAIN 670 38.415 -4.304

FPAR Fuente de la Parra, Aýna, Albacete, SPAIN 874 38.571 -2.120

FPRA Fuente del Prado, Liétor, Albacete, SPAIN 760 38.560 -1.968

FSCH Fuente de Sancho, Alatoz, Albacete, SPAIN 960 39.077 -1.341

GARL Embalse de la Serena, Garlitos, Badajoz, SPAIN 398 38.840 -5.048

GDJI Río Guadajira, Aceuchal, Badajoz, SPAIN 299 38.630 -6.507

GORD El Gordo, 3km N, El Gordo, Toledo, SPAIN 316 39.883 -5.333

GRIM Grimaldo 2.3km N, Cañaveral, Cáceres, SPAIN 422 39.860 -6.350

GRZM Grazalema, Grazalema, Cádiz, SPAIN 650 36.765 -5.346

HVAR Higuera de Vargas, Higuera de Vargas, Badajoz, SPAIN 358 38.417 -6.983

INVR Barranco de San Roque, Las Inviernas, Guadalajara, SPAIN 900 40.833 -2.663

JA_BE Jalón-Bernia, Alicante, SPAIN ? 38.694 -0.051

JA_BE2 Jalón-Bernia, Alicante, SPAIN ? 38.721 -0.036

LARA Larache, MOROCCO 36 35.112 -6.143

LIET Híjar-Puente, Liétor, Albacete, SPAIN 590 38.544 -2.030

MEMB Membrío 5km NO, Membrío, Cáceres, SPAIN 380 39.549 -7.093

MERI A-5 km 334, Mérida, Badajoz, SPAIN 256 38.933 -6.283

MIRA Rio Mira, 1km S, IC1, , Alentejo, PORTUGAL 196 37.551 -8.265

MMUG EU15 Urquiola NP, Mañana-Mugarra, Bizkaia, SPAIN ? 43.145 -2.669

MNCY Santa Cruz del Moncayo, Santa Cruz del Moncayo, Zaragoza, SPAIN 705 41.868 -1.761

MO_CI Moraleja-Cilleros 5.5km NO, Moraleja, Cáceres, SPAIN 284 40.086 -6.720

MPLA Malpartida de Plasencia 2km S, Malpartida de Plasencia, Cáceres, SPAIN 393 39.967 -6.062

NAGO Nagore, Navarra, SPAIN 603 42.862 -1.379

NAVA Arroyo Castaño, Los Navalucillos, Toledo, SPAIN 750 39.610 -4.660

PABA Ribera del Guadalquivir, Pedro Abad, Córdoba, SPAIN 149 37.967 -4.483

PAGO EU01, pista a Legarrola, Pagoeta, Guipuzkoa, SPAIN ? 43.226 -2.156

PALM Palomas, Badajoz, SPAIN 325 38.699 -6.136

PINX Pincho, 1.4 km SE, Algarve, PORTUGAL 69 37.197 -8.730

PINX Pincho, 0.4 km S, Algarve, PORTUGAL 90 37.204 -8.737

POBL1 Camino de la Cadeneta, Poblet, Tarragona, SPAIN 706 41.302 1.074

POBL2 Carretera TV-7041 km18, Poblet, Tarragona, SPAIN 778 41.303 1.057

POBL3 Barranc de D'Almunt Vila, Poblet, Tarragona, SPAIN 701 41.294 1.031

PREY CM-4162, 6km N, Puerto Rey, Toledo, SPAIN 677 39.500 -5.000

PZAR Arr. Mirabella, 3.5 km S, Pozuelo de Zarzón, Cáceres, SPAIN 448 40.123 -6.439

REVE Puerto de los Reventones, Fregenal de la Sierra, Badajoz, SPAIN 500 38.250 -6.567

RTEJ Rambla del Tejar, La Pesquera, Cuenca, SPAIN 775 39.569 -1.582

SALM Cortijo de Salomón, San Roque, Cádiz, SPAIN 80 36.250 -5.380

SALO Río Salor, 7.2 km NE, Membrío, Cáceres, SPAIN 159 39.574 -6.997

SCPG1 Santa Cruz de Paniagua 3.6km E, Cáceres, SPAIN 442 40.187 -6.299

SCPG2 Santa Cruz de Paniagua 4.2km SO, Cáceres, SPAIN 566 40.164 -6.369

SESM Sesma, Navarra, SPAIN 375 42.360 -2.074 SGUA1 Sierra de Guara, Casa de Estebañón, Abizanda, Huesca, SPAIN 425 42.231 -0.232

SGUA2 Sierra de Guara, Río Calcón, Abizanda, Huesca, SPAIN 612 42.212 -0.194

SGUA3 Sierra de Guara, Río Formiga, Abizanda, Huesca, SPAIN 785 42.220 -0.164

SILL Rio Sillo, Cumbres de San Bartolomé, Huelva, SPAIN 275 38.050 -6.817

SLGU Sanlúcar de Guadiana, 1,5 km SE, Huelva, SPAIN 129 37.466 -7.454

SLGU Sanlúcar de Guadiana, 4 km SE, Huelva, SPAIN 129 37.448 -7.441

SLLA Fuente de Alcántara, Sella, Alicante, SPAIN 380 38.610 -0.256

SNSL Santuario de Nuestra Señora de la Luz, Tarifa, Cádiz, SPAIN 61 36.084 -5.621

SSGZ San Silvestre de Guzmán, 0,7 km N, Huelva, SPAIN 175 37.383 -7.350

SSPR BA-136 km 12, Sancti-Spiritus, Badajoz, SPAIN 421 38.978 -5.140

SSUS Santa Susana, 0.7 km SO, Alentejo, PORTUGAL 51 38.441 -8.395

TMBU Cortijo de Tamburrero, Valle de la Serena, Badajoz, SPAIN 408 38.685 -5.856

TOLO Arroyo de Tolonche, Alatoz, Albacete, SPAIN 900 39.113 -1.263

TOMS Regato de las Tomasas, Herrera de Alcántara, Cáceres, SPAIN 243 39.604 -7.325

TORT Tortuero, 0.4 km SO, Tortuero, Guadalajara, SPAIN 922 40.935 -3.357

TRUJ Trujillanos, 1,2 km NE, Badajoz, SPAIN 265 38.950 -6.233

UBRI Puerto Tirado, 2,7 km SE, Ubrique, Cádiz, SPAIN 527 36.650 -5.433

UNDE 3 La Unde, Ayora, Valencia, SPAIN 990 39.087 -1.218

VBUR EX-101, km 18, Valverde de Burguillos, Badajoz, SPAIN 488 38.350 -6.517

VCAB BAV-7117 4.8km SO , Valdecaballeros, Badajoz, SPAIN 401 39.206 -5.214

VENC Valdencín 3.3km S, Torrejoncillo, Cáceres, SPAIN 310 39.887 -6.398

VLEG Valverde de Leganés, Badajoz, SPAIN 377 38.659 -6.963

ZAOS Arroyo Zaos, Oliva de la Frontera, Badajoz, SPAIN 347 38.300 -6.905

ZMAY1 Zarza la Mayor 4.5 km NE, Cáceres, SPAIN 289 39.363 -7.274

ZMAY2 Zarza la Mayor 4km S, Cáceres, SPAIN 326 39.846 -6.841

ZUHE El Navazuelo, Zuheros, Córdoba, SPAIN 1025 37.480 -4.337

ZZAN CUV-7031, 5.7km NE , Zafra de Záncara, Cuenca, SPAIN 911 39.905 -2.493

AEAS Arroyo del Espino-Arroyo del Sieso, El Bosque, Cádiz 320 36.776 -5.503

Table S3. Primers used in the study. F, forward; R, reverse.

Type DNA Gene Primer S Primer sequence (5’- 3’) Described in: Mitochondrial Cox1-a lco1490 F GGTCAACAAATCATAAAGATATTGG (Folmer et al., protein coding 1994) hco2198 R TAAACTTCAGGGTGACCAAAAAATCA (Folmer et al., 1994) Mitochondrial Cox1-b Jerry F CAACATTTATTTTGATTTTTTGG (Simons et al., protein coding (M202) 1994) Pat (M70) R TCCA(A)TGCACTAATCTGCCATATTA (Simons et al., 1994) Mitochondrial rrnL 16SaR F CGCCTGTTTAWCAAAAACAT (Simons et al., ribosomal (M14) 1994) 16s- R GGTCCCTTACGAATTTGAATATATCCT (Simons et al., ND1a 1994) (M223) Nuclear LSU LS58F F GGGAGGAAAAGAAACTAAC (Ober, 2002) ribosomal (D1) LS998R R GCATAGTTCACCATCTTTC (Ober, 2002) (D3) Nuclear SSU 5' F GACAACCTGGTTGATCCTGCCAGT (Shull et al., 2001) ribosomal b5.0 R TAACCGCAACAACTTTAAT (Shull et al., 2001)

Nuclear SSU 18Sai F CCTGAGAAACGGCTACCACATC (Wray et al., 1993) ribosomal 18Sbi R GAGTCTCGTTCGTTATCGGA (Wray et al., 1993)

Table S4. Initial parameters and results for the bootstrap likelihood ratio test (200 bootstrap replicates; endmc = 200) of the diversity-dependent model (assuming a linear dependence in speciation rate with parameter K' = diversity where speciation = 0; ddmodel = 1.3) against the constant-rates birth-death model.

Sampling coverage Conditioning initparsoptDD initparsoptCR pValue Power of test Case 0: missnumspec = 0 Cond = 0 lambda 0.02, mu 0.01, k’ 200 lambda 0.02, mu 0.01 0.010 0.11 Case 0: missnumspec = 0 Cond = 1 lambda 0.02, mu 0.01, k’ 200 lambda 0.02, mu 0.01 0.015 0.15 Case 1: missnumspec = 33 Cond = 0 lambda 0.02, mu 0.01, k’ 200 lambda 0.02, mu 0.01 0.045 0.09 Case 1: missnumspec = 33 Cond = 1 lambda 0.02, mu 0.01, k’ 200 lambda 0.02, mu 0.01 0.010 0.06 Case 2: missnumspec = 74 Cond = 0 lambda 0.02, mu 0.01, k’ 200 lambda 0.02, mu 0.01 0.070 0.11 Case 2: missnumspec = 74 Cond = 1 lambda 0.02, mu 0.01, k’ 200 lambda 0.02, mu 0.01 0.090 0.14 Case 3: missnumspec = 222 Cond = 0 lambda 0.02, mu 0.01, k’ 200 lambda 0.02, mu 0.01 N.F N.F Case 3: missnumspec = 222 Cond = 1 lambda 0.02, mu 0.01, k’ 200 lambda 0.02, mu 0.01 N.F N.F initparsoptDD: Initial parameters for optimizing Density-Dependent model; initparsoptCR: Initial parameters for optimizing Constant Rate model; N.F. Analyses not finished after 2 weeks of computing, presumably untreatable due to size of simulations.

Data S1. Identification criteria and taxonomic decisions

1) The specimen clearly belongs to a known, described species. This is, the specimen completely match the morphological information of such species and shows all its diagnostic traits. In this case, it is identified with the name of the species. Ej. BMNH1046040-Typhlocharis lunai Serrano & Aguiar, 2006.

2) The specimen probably belongs to a known species, but it cannot be verified with the available data. This occurs, for example, when a specimen match the main diagnostic traits of a species but also some variability from the type material and the identification is not satisfactory; or when a specimen is damaged or incomplete and certain key characters are missing. In this case, it is labeled with the abbreviation “cf.” (latin confer – compare) before the species name. Ej. BMNH1041970-Typhlocharis cf. deferreri Zaballos & Pérez- González, 2011.

3) The specimen has obvious affinities with a known species, but possibly belongs to a different, closely related species. This occurs when the specimen shows morphological differences beyond the usual levels of intraspecific variation (Pérez-González et al., 2013) but still shares key diagnostic characters with a known species. In this case, it is labeled with the abbreviation “aff.” (latin affinis – related) before the species name. In this case, it has been numbered and treated as potential new species. Ej. BMNH1424404-Microcharidius sp. 42 aff. crespoi Serrano & Aguiar, 2008.

4) The specimen clearly belongs to a new species. Generally, this occurs when the specimen shows autapomorphic traits that allows unambiguous differentiation from any other species. In this case, it is labeled with the abbreviation “sp.” (species) and an identification number. Ej. BMNH1424387 Microcharidius sp. 20.

Data S2. Lines used in R for DDD simulations in figure 1B.

# sim_dd_07s_04e_30k_50y

SimTES_dd_07s_04e_30k_50y <- vector("list", 2) #create an empty list of 2 elements class(SimTES_dd_07s_04e_30k_50y) <- "multiPhylo" #make this list a multiPhylo object

SimTAS_dd_07s_04e_30k_50y <- vector("list", 2) #create an empty list of 2 elements class(SimTAS_dd_07s_04e_30k_50y) <- "multiPhylo" #make this list a multiPhylo object for (i in 1:100) { dd_KI_sim(c(0.0001,0.0001,2, 0.7,0.4,30, 49), 50, ddmodel = 1.3)->sim_dd_07s_04e_30k_50y SimTES_dd_07s_04e_30k_50y[[i]]<-sim_dd_07s_04e_30k_50y$tes SimTAS_dd_07s_04e_30k_50y[[i]]<-sim_dd_07s_04e_30k_50y$tas } par(mar=c(2,2,2,2),pty="m",mfrow=c(2,3)) plot(sim_dd_07s_04e_30k_50y$tes) plot(sim_dd_07s_04e_30k_50y$tas) ltt.plot(sim_dd_07s_04e_30k_50y$tes) ltt.plot(sim_dd_07s_04e_30k_50y$tas) mltt.plot(SimTES_dd_07s_04e_30k_50y, legend = FALSE) mltt.plot(SimTAS_dd_07s_04e_30k_50y, legend = FALSE)

#sim_dd_07s_001e_30k_50y

SimTES_dd_07s_001e_30k_50y <- vector("list", 2) #create an empty list of 2 elements class(SimTES_dd_07s_001e_30k_50y) <- "multiPhylo" #make this list a multiPhylo object

SimTAS_dd_07s_001e_30k_50y <- vector("list", 2) #create an empty list of 2 elements class(SimTAS_dd_07s_001e_30k_50y) <- "multiPhylo" #make this list a multiPhylo object for (i in 1:100) { dd_KI_sim(c(0.0001,0.0001,2, 0.7,0.01,30, 49), 50, ddmodel = 1.3)->sim_dd_07s_001e_30k_50y SimTES_dd_07s_001e_30k_50y[[i]]<-sim_dd_07s_001e_30k_50y$tes SimTAS_dd_07s_001e_30k_50y[[i]]<-sim_dd_07s_001e_30k_50y$tas } par(mar=c(2,2,2,2),pty="m",mfrow=c(2,2)) plot(sim_dd_07s_001e_30k_50y$tes) plot(sim_dd_07s_001e_30k_50y$tas) ltt.plot(sim_dd_07s_001e_30k_50y$tes) ltt.plot(sim_dd_07s_001e_30k_50y$tas) mltt.plot(SimTES_dd_07s_001e_30k_50y, legend = FALSE) mltt.plot(SimTAS_dd_07s_001e_30k_50y, legend = FALSE)

SUPPORTING INFORMATION Figure S1. Distribution map of Typhlocharina. Red: localities sampled in this study; grey: Typhlocharis localities obtained from the literature and unpublished data, including those for the 33 non sampled species.

Figure S2. Ultrametric time calibrated tree obtained in BEAST with parameter favored by BF comparisons (ULN clock, yule-birth-death speciation model, root age non constrained). The vertical red line indicate the GMYC single threshold, used as a guide to obtain species hypothesis, where the '*' indicates those cases with disagreement between GMYC result and morphology. Names of clades as in Fig. 3 of the main text. Numbers on nodes indicate posterior probabilities, and grey bars represent the 95% HPDI for the estimated node age.

figure S2. Continued.

Figure S2. Continued.

Figure S3. TreePar results for the ultrametric time calibrated tree obtained in BEAST with parameter favored by BF comparisons (ULN clock, yule-birth-death speciation model, root age non constrained) and pruned to species. A) sampling coverage case 1 (sampling represents 69% of known species; i.e. sampled plus other known species represent the total existing species within the genus). B) sampling coverage case 2 (sampling represents 50% of existing species). C sampling coverage case 3 (sampling represents 25% of existing species). In all cases a model with one rate shift corresponding to the interphase from the interspecific to the intraspecific level was obtained, what is interpreted as no shifts detected by TreePar during the speciation phase of the group. Blue line: net diversification (speciation-extinction); Red line: turnover=extinction/speciation.

Figure S4. BAMM results for the ultrametric time calibrated tree obtained in BEAST with parameter favored by BF comparisons (ULN clock, yule-birth-death speciation model, root age non constrained) and pruned to species. A) sampling coverage case 1 (sampling represents 69% of known species; i.e. sampled plus other known species represent the total existing species within the genus). B) sampling coverage case 2 (sampling represents 50% of existing species). C sampling coverage case 3 (sampling represents 25% of existing species). Details on the interpretation of graphs can be obtained in the BAMM tutorials (Mitchell & Rabosky, 2016).

Figure S5. Results for the relationship between geographical and phylogenetic distances by pares of species for the three superlineages (C1, C2 y C3) and the 8 main lineages within Typhlocharina. The dotted horizontal red line on each graph represents the single threshold GMYC result. Maps show the distribution of the lineages.

Figure S7. Results for correlation between species richness and lineage crown age for the 8 main clades in the phylogeny of Typhlocharina which show geographical and morphological consistence.

Figure S8. Ultrametric tree pruned to a single representative per species indicating, roughly, distribution in or out from the Baetic-Rifean plate.

References

Folmer, O., Black, M., Hoeh, W., Lutz, R., Vrijenhoek, R., 1994. DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Mol. Mar. Biol. Biotechnol 3, 294–299. Mitchell, J.S., Rabosky, D.L., 2016. Bayesian model selection with Bayesian Analysis of Macroevolutionary Mixtures: effects of the model prior on the inferred number of diversification shifts. Methods Ecol. Evol. doi:10.1111/2041-210X.12626 Ober, K. a., 2002. Phylogenetic relationships of the carabid subfamily Harpalinae (Coleoptera) based on molecular sequence data. Mol. Phylogenet. Evol. 24, 228–248. doi:10.1016/S1055-7903(02)00251-8 Shull, V.L., Vogler, a P., Baker, M.D., Maddison, D.R., Hammond, P.M., 2001. Sequence alignment of 18S ribosomal RNA and the basal relationships of Adephagan beetles: evidence for monophyly of aquatic families and the placement of Trachypachidae. Syst. Biol. 50, 945–969. doi:10.1080/106351501753462894 Simons, 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. Ann. Entomol. Soc. Am. doi:10.1080/17470210902990829 Wray, C.G., Lee, J.J., Desalle, R., 1993. Extraction and enzymatic characterization of foraminiferal DNA. Micropaleontology 39, 69–73.