OR I G I N A L A R T I C L E

Substratum karstificability, dispersal and genetic structure in astrictly subterranean beetle

Valeria Rizzo1 | David S�anchez-Fern�andez1,2 | Roc�ıo Alonso1 | Josep Pastor3 |

Ignacio Ribera1

1Institute of Evolutionary Biology (CSIC-

Universitat Pompeu Fabra), Barcelona,

Spain

2Instituto de Ciencias Ambientales,

Universidad de Castilla-La Mancha, Toledo,

Spain

3Museu de Ci�encies Naturals (Zoologia),

Barcelona, Spain

Correspondence

Ignacio Ribera, Institute of Evolutionary

Biology, Barcelona, Spain.

Email: ignacio.ribera@ibe.upf-csic.es

Funding information

Universidad de Castilla-La Mancha;

European Social Fund; AEI/FEDER, Grant/

Award Number: CGL2010-15755, CGL2016-

76705-P

Editor: Isabel Sanmart�ın

Abstract

Aim: The deep subterranean environment is an ideal system to test the effect of

physical constraints on the ecology and evolution of species, as it is very homoge-

neous and with simple communities. We studied the effect of substratum karstifica-

bility in the dispersal of the strictly subterranean Troglocharinus ferreri (Reitter)

(Coleoptera, Leiodidae) by comparing the genetic diversity and structure of popula-

tions in limestone (more soluble) and dolostone (less soluble) in the same karstic

system.

Location: Troglocharinus ferreri is only known from c. 100 vertical shafts in an area

of <500 km2 SW of Barcelona (Spain).

Methods: We sequenced mitochondrial and nuclear markers of a representative

sample to identify main lineages within T. ferreri and estimate their temporal origin,

and used mitochondrial data of 129 specimens from 41 caves to reconstruct their

demographic history and estimate dispersal among caves.

Results: Troglocharinus ferreri diverged from its sister in the Early Pliocene, with an

initial divergence of the sampled populations in the Early Pleistocene. The best

demographic model was a constant population size with a fast population increase

in the middle Pleistocene. The ancestral population was likely in limestone, with a

probability of transition from limestone to dolostone triple to that from dolostone

to limestone, suggesting a higher permeability of limestone to the transit of individ-

uals. Populations in dolostone caves had lower gene flow between them and a

stronger isolation by distance, although the low genetic variability for the studied

markers and the lower abundance of dolostone caves decreased the statistical

power of the analyses.

Main conclusions: Our results point to the physical characteristic of the substratum

as a determinant of dispersal and gene flow, potentially conditioning the long-term

evolution of subterranean biodiversity.

K E YWORD S

gene flow, isolation, substratum permeability, subterranean environment, Troglocharinus,

troglomorphism

DOI: 10.1111/jbi.13074

Journal of Biogeography. 2017;44:2527–2538. wileyonlinelibrary.com/journal/jbi © 2017 John Wiley & Sons Ltd | 2527

1 | INTRODUCTION

It has long been recognized that some physical properties of the

habitats may constraint the characteristics of the organisms living in

them. Some of these constraints may affect dispersal and gene flow

between populations, thus potentially also acting at longer, evolu-

tionary time-scales. This general principle has been articulated in dif-

ferent theoretical frameworks (see e.g. Grime, 1977 for plants, or

the “habitat templet” of Southwood, 1977 for terrestrial inverte-

brates), but it has been difficult to precisely identify which habitat

characteristics determine these constraints, especially those promot-

ing dispersal, divergence and speciation. Some of the best studied

examples include long-term habitat stability, especially in aquatic sys-

tems, both marine (mediated through the type of larval development,

e.g. Emlet, 1995) and freshwater (Ribera & Vogler, 2000; see Ribera,

2008 for a review). In terrestrial systems, likely due to their higher

complexity, it has been more difficult to identify habitat characteris-

tics linking processes affecting individuals or populations to those

affecting metapopulations or species, and ultimately whole lineages

(but see Papadopoulou, Anastasiou, Keskin & Vogler, 2009 for an

example with Coleoptera).

A system that potentially may simplify this complexity is the

deep subterranean environment. In the deep parts of caves and the

associated network of voids and fissures the environmental condi-

tions of the habitat are extremely constant and homogeneous, with

a permanent darkness and nearly constant humidity and temperature

through the year, which is approximately equal to the average

annual temperature of the surface (Culver & Pipan, 2009; Poulson &

White, 1969; Racovitza, 1907). Subterranean organisms have a very

limited range of options to take advantage of local climatic hetero-

geneities, and their general lack of mobility reduces the possibility of

migration when conditions become unfavourable. In addition, the

general scarcity of resources in the deep subterranean environment

imposes stringent requirements on the species, resulting in simple

communities showing low local diversity (Culver, 1976; Poulson &

Culver, 1969).

The subterranean environment is a discontinuous medium with a

highly variable degree of connectivity, and populations of troglobitic

species are in general more fragmented than populations of species

living on the surface (Crouau-Roy, 1989; Culver & Pipan, 2009). Due

to the difficulty to disperse, gene flow among geographically close

populations can be very restricted, resulting in a strong micro-ende-

mism (Faille, Bourdeau, Bell�es & Fresneda, 2015; Faille, T€anzler &

Toussaint, 2015; Juberthie, Delay & Bouillon, 1980). These condi-

tions offer a particularly favourable situation for exploring the rela-

tionship between environmental factors and the genetic structure of

populations, as due to the geographical proximity and the generally

homogeneous environment the number of confounding variables

should be greatly reduced.

One of the main factors potentially constraining dispersal within

the subterranean environment is the karstificability of the substra-

tum, which is highly dependent on its lithology and geochemical

composition. The role of geological barriers (i.e. non-karstificable

strata) has been traditionally recognized as a factor shaping the dis-

tribution and evolution of the subterranean fauna (e.g. Bell�es, 1973;

Bell�es & Mart�ınez, 1980), and the degree of fragmentation of the

karst has recently being related with ongoing speciation processes in

some Pyrenean subterranean species (Faille, Bourdeau, Bell�es &

Fresneda, 2015; Faille, T€anzler & Toussaint, 2015). These studies,

however, did not establish a direct link between the geology of the

substratum and the genetic structure of the studied species, and the

factors potentially affecting dispersal could not be identified with

precision. Ideally, the role of karstificability in determining dispersal

and genetic structure should be investigated with populations of the

same species in different geological substrata, but otherwise with

similar general characteristics and with the same biogeographical

history.

Here we study a system that matches these conditions, the

troglobitic species Troglocharinus ferreri (Reitter) (Coleoptera, Leioidi-

dae, Leptodirini) in the subterranean environment of the Garraf mas-

sif, in the vicinity of Barcelona (NE of the Iberian peninsula). In a

previous study we found that this strictly subterranean genus

expanded its range from the central Pyrenees to the coastal area of

Catalonia in the early Pliocene, where it subsequently diversified

during the late Pliocene and the Pleistocene (Rizzo, Comas, Fadrique,

Fresneda & Ribera, 2013). Within the coastal clade of the genus

Troglocharinus, T. ferreri is distributed in the Garraf massif, isolated

by Pleistocene sedimentary basins of three rivers (Llobregat in the

north-east, Anoia in the north and Foix in the south-west, Figure 1).

Preliminary results suggested a high level of genetic diversity within

T. ferreri, consistent with the complex geological structure of the pla-

teau of the Garraf massif (Rizzo et al., 2013), with one mitochondrial

lineage mostly associated with a more homogenous limestone of

Cretaceous origin and its sister (including a recognized subspecies, T.

ferreri pallaresi Bell�es) inhabiting also some areas with Jurassic and

Triassic dolostone. Dolomite (CaMg(CO3)2) has a similar structure to

calcite but with approximately half of its Ca replaced by Mg. The

dissolution rate of dolomite is considerably lower than that of calcite

(Goldscheider & Drew, 2007), and in a recent study Appelo and

Postma (2005) estimated that under the same chemical and physical

conditions of the water it takes more than 100 times longer to reach

95% saturation for dolomite than for calcite. In addition, erosion in

dolomitic karsts results in fine sand that fills sinkholes and caves,

partially blocking the ground water circulation and further reducing

the development of a subterranean medium (Gilli, 2015; Renault,

1970). This lower karstificability should result in an environment less

permeable for the subterranean fauna, with impeded dispersal and

reduced contact between populations in different areas of the

massif.

Using a comprehensive sample of T. ferreri across the different

geological layers found through its entire geographical range we

investigate the relationship between the karstificability of the sub-

stratum and the genetic diversity and structure of the populations

living on it, which may be an indication of differences in the rate of

2528 | RIZZO ET AL.

dispersal between them. Since we compare individuals in a reduced

geographical space, originating from the same colonization event and

with a shared evolutionary history and the same ecology, biology

and physiology, it is expected that differences among them are

mainly the result of the geology of the substratum in which they are

found. Our specific objectives are (1) determine the phylogeographi-

cal structure and demographic history of T. ferreri within its entire

distributional range; (2) reconstruct the direction and frequency of

the transitions between different types of substratum; and (3) com-

pare the genetic structure of populations within geological layers

with different degrees of karstificability.

2 | MATERIALS AND METHODS

2.1 | Geological and taxonomic background

The Garraf Massif is a calcareous horst with an area of around

500 km2 in the north-east of the Iberian Peninsula, 30 km south-

west of the city of Barcelona. It is part of the Catalonian littoral

cordillera, composed mainly of Jurassic and Cretaceous limestone

and dolostone and isolated by Quaternary sedimentary layers from

the rest of the coastal area (Daura, Sanz, Forno, Asensio & Juli�a,

2014; Figures 1 and 2). The central part of the massif is domi-

nated by Cretaceous limestone rocks (Moreno, 2007), where most

of the karst formations have developed, while the Jurassic dolo-

mite strata in the high plains present less significant karst devel-

opment and fewer dolines (Rubinat, 2004). Outcrops of

Cretaceous marls and marly limestone are also found in a few

areas (Moreno, 2007), while the Triassic carbonates comprise

strata that only crop out in the north-east of the Garraf Massif

rim (Figure 2). There are more than three hundred shafts docu-

mented in the central Garraf Massif, which have been explored

since the late 19th century (Cardona i Oliv�an, 1990), in which the

fauna is well-documented due to the long biospeleological tradi-

tion of the city of Barcelona (e.g. Bell�es, 1973; Espa~nol, 1961;

Lagar, 1954; Zariquieyi, 1917).

Troglocharius ferreri currently includes three recognized sub-

species, T. ferreri ferreri, T. ferreri pallaresi and T. ferreri abadi Lagar,

all restricted to the Garraf Massif (Fresneda & Salgado, 2017; Sal-

gado, Blas & Fresneda, 2008). Four other subspecies were described

based mostly on small differences in the shape of the pronotum and

the antennae, but are now considered to be synonyms of T. ferreri

ferreri due to the presence of multiple populations with intermediate

morphologies (Salgado et al., 2008; Appendix S1a). The two recog-

nized subspecies were maintained in part due to the geological isola-

tion of their populations (Salgado et al., 2008). Troglocharinus ferreri

ferreri is known from 123 cavities distributed through the Garraf

Massif (Fresneda & Salgado, 2017; Appendix S2a), although this

number may certainly increase with additional explorations. The

other two subspecies have, however, very restricted distributions: T.

ferreri abadi is only known from three cavities in close proximity situ-

ated in the north-west of the Massif, and T. ferreri pallaresi from four

cavities in the north-east, also in close proximity (Fresneda & Sal-

gado, 2017).

Castelldefels

Sitges

F IGURE 1 Distribution of the coastal clade of the genus Troglocharinus, with the main tectonic and geological substrata (modified fromRizzo et al., 2013). Black squares, T. ferreri; circles, other species in the coastal clade. Colour codes: yellow, Tertiary and Quaternarysedimentary basins; green, folded Mesozoic cover; pink and purple: Hercynian substratum; grey: Quaternary filling of Miocene and Quaternaryfractures. Lines represent tectonic features in standard geological notation. Note the isolation of the distribution area of T. ferreri byQuaternary sediments

RIZZO ET AL. | 2529

2.2 | Taxon sampling and DNA sequencing

We sequenced a total of 126 specimens of T. ferreri ferreri from 40

shafts distributed through its entire known distribution area, and

three T. ferreri pallaresi from the type locality (Avenc de Montmany)

(Figure 2; Appendix S1a). Caves were chosen with the aim to include

the whole range of geographical, geological and taxonomic diversity,

including the type locality of the species (Avenc d’en Roca) and

those of the synonymized subspecies, of which we obtained two

(Appendix S1a). Despite several attempts we could not obtain speci-

mens of T. ferreri abadi for study, and we are not aware of recent

records (the last collected specimens to our knowledge being from

Cova Miserachs in 1987, J. Comas personal communication, 2015).

The three shafts in which the subspecies is known are very dry,

and the specimens, always in reduced numbers, have only being

found in the deepest, more humid areas (Salgado et al., 2008).

Specimens were collected in caves with the use of baits laid

24 hr in advance or through manual searches. We sequenced five

specimens of each cave whenever available (Appendix S1a). Virtually

all cavities in the Garraf Massif are vertical shafts with no or very

limited horizontal development, and in most cases beetles were col-

lected at the base of the shaft, so the coordinates of the entrance

are a good approximation to the actual location of the samples. To

increase the statistical power of some analyses of molecular diversity

we pooled specimens from caves known to be part of the same sys-

tem, connected through the subterranean medium and without any

apparent geological discontinuity between them (Figure 2;

Appendix S1a).

DNA extractions of single specimens were non-destructive, using

commercial kits (mostly DNeasy Tissue Kit; Qiagen, Hilden, Ger-

many) following the manufacturer’s instructions. Vouchers and DNA

samples are kept in Institute of Evolutionary Biology, Barcelona

(IBE). We amplified fragments of five mitochondrial and four nuclear

genes in eight amplification reactions: (1) 50 end (the “barcode”,

cox1-50) and (2) 30 end of the cytochrome c oxidase subunit (cox1-

30); (3) an internal fragment of the cytochrome oxidase b (cob); (4) 50

end of the large ribosomal unit plus the Leucine transfer plus the 30

end of NADH dehydrogenase subunit 1 (rrnL+trnL+nad1); (5) 50

end of the small ribosomal unit, 18S rRNA (SSU); (6) an internal frag-

ment of the large ribosomal unit, 28S rRNA (LSU); and one fragment

each of (7) the histone H3 (H3) and (8) wingless (Wg) (see

Appendix S2b for the primers used, and Appendix S1a for the speci-

mens sequenced for each of the fragments). Owing to the generally

low levels of variability found in preliminary analyses, only a limited

sample of specimens was sequenced for the nuclear markers

(Appendix S1a). New sequences (285) were deposited in the EMBL

F IGURE 2 Simplified geological map with the location of the studied caves (circles). Blue, caves in clade 1; yellow, caves in clade 2A(limestone lineage); red, caves in clade 2B (dolostone lineage); black, Troglocharinus ferreri pallaresi (upper right corner). Caves with mixedcolours had specimens of different clades. Geological substratum: yellow, dolostone; green, limestone; grey and pink, different non-karstificablematerials (modified from ICC, 2010)

2530 | RIZZO ET AL.

database under accession numbers LT797170–LT797446,

LT799421–LT799428) (Appendix S1a).

2.3 | Phylogenetic and phylogeographical analyses

We aligned the gene fragments with MAFFT 6 (http://mafft.cbrc.jp/

alignment/ server). Protein-coding genes were aligned using the

FFT-NS-1 algorithm, and ribosomal genes with the Q-INS-i algo-

rithm, which considers the secondary structure (Katoh, Asimenos &

Toh, 2009).

To reconstruct the general phylogenetic relationships and time of

divergence among the main lineages within T. ferreri we built a data

matrix with all mitochondrial and nuclear markers, using as out-

groups a selection of 26 specimens from both the coastal and Pyre-

nean clades of Troglocharinus following the topology obtained in

Rizzo et al. (2013) (Appendix S1a). We reconstructed phylogenetic

relationships with RAXML 7 (Stamatakis, Hoover & Rougemont,

2008) using GTR+I+G as an evolutionary model and a partition by

gene, separating the two cox1 fragments (due to the unequal sam-

pling) and pooling the two mitochondrial ribosomal genes (rrnL+trnL).

The optimum topology was that of the best likelihood amongst 100

replicates, and node support was estimated with 500 bootstrap repli-

cates using the CAT approximation (Stamatakis et al., 2008). We

analysed the nuclear sequence data separately using the same

methods.

To obtain an estimation of the divergence time we analysed the

matrix in BEAST 1.8 (Drummond, Suchard, Xie & Rambaut, 2012),

using as priors the rates obtained by Cieslak, Fresneda and Ribera

(2014) for the same group of organisms and genes, calibrated using

the tectonic separation of Sardinia from mainland Europe c. 33 Ma

(Schettino & Turco, 2006). We used an uncorrelated relaxed clock

with normally distributed prior average rates of 0.015 substitutions/

site/MY for mitochondrial protein coding genes (cox1-5, cox1-3, cob,

nad1), 0.004 for nuclear ribosomal (LSU, SSU), and 0.006 for mito-

chondrial ribosomal genes (rrnL+trnL), all of them with a standard

deviation of 0.001. For the nuclear protein coding genes (H3, Wg)

we used default flat priors. We constrained the monophyly of the

coastal and Pyrenean clades, respectively, following the topology

obtained in Rizzo et al. (2013) and ran two independent analyses for

200 million (M) generations, logged every 5000, with 20 M (10%) as

burn-in fractions. Preliminary analyses with the complete matrix

using either a Yule speciation (YL) or a Birth-Death with incomplete

sampling model (BD, Stadler, 2009) failed to converge, so we used a

sample of 27 specimens of T. ferreri with complete sequence data,

including all main lineages as identified in the RAXML analysis and a

wide representation of different caves and types of substratum. We

used a GTR+I+G and HYK evolutionary models for each of the two

mitochondrial and nuclear partitions, respectively, and TRACER 1.6

(Drummond et al., 2012) to assess convergence, to measure the

effective sample size of each parameter and to check that the used

burn-in fraction was sufficient. We run two different sets of analyses

using YL or BD models, comparing their likelihoods using 100 repli-

cas of the Akaike’s information criterion for MCMC samples (AICM)

statistic (known to perform better than harmonic mean estimators,

Baele, Li, Drummond, Suchard & Lemey, 2013) in TRACER 1.6.

2.4 | Demographic history

To estimate the demographic history of T. ferreri we used the mito-

chondrial cox1-5 fragment only, for which the sampling was most

complete and presented sufficient variation, with a GTR+I+G evolu-

tionary model and an a priori mean rate of 0.015 substitutions/site/

MY with a standard deviation of 0.001. We run two analyses in

BEAST 1.8, one with a strict clock and a second with a relaxed lognor-

mal, and comparing their likelihoods using 100 replicas of the AICM

statistic. Analyses were run for 100 M generations and convergence

and burn-in fraction were assessed as in previous BEAST analyses.

We compared four demographic models implemented in BEAST

1.8: constant population size (CT), population expansion (ES), logistic

growth (LG) and exponential growth (EL). We also constructed a

Bayesian skyline plot (BSP, Drummond, Rambaut, Shapiro & Pybus,

2005) with the results of a separate analysis. The likelihood of the

trees built with the different models were compared with 100 repli-

cas of the AICM statistic.

2.5 | Estimation of transition rates betweendifferent geological substrata

To estimate the transition rates between different geological sub-

strates we assigned each of the caves to different categories accord-

ing to the dissolvability of the substratum in which they are found,

as estimated with detailed lithological maps (ICC, 2010;

Appendix S1a). We divided the caves in two categories: (1) Creta-

ceous limestone and (2) Jurassic and Triassic dolostone. Due to the

lower number of dolostone caves we pooled all dolostone caves in a

single category. We did an ancestral trait reconstruction in BEAST 1.8,

with the geological type of the cave as a discrete character and a

reduced cox1-5 matrix randomly selecting one specimen per cave

when multiple haplotypes in the same cave were monophyletic

(Appendix S1a). We used a single partition with a simple HKY model

with estimated nucleotide frequencies, as preliminary analyses

showed convergence problems when using more complex models.

We set up an asymmetric discrete phylogeographical (CTMC) model

(Lemey, Rambaut, Drummond & Suchard, 2009), using as priors a

uniform distribution between 0 and 1 for the frequencies, and a

Gamma with shape 1 and scale 1 for the rates. We also imple-

mented the best clock and demographic model as estimated above

and uniform root frequencies for the trait. The analyses were run for

100 M generations, and convergence was assessed as in the previ-

ous analyses.

2.6 | Genetic versus geological or geographicaldistances

To estimate the role of the lithology in the genetic isolation

among T. ferreri populations we compared the relationship

RIZZO ET AL. | 2531

between phylogenetic and geographical distances in populations in

dolostone and limestone. Given the heterogeneity of the spatial

distribution of dolostone and limestone in the Garraf massif, in

addition to the simple geographical distance between caves we

also used a weighted measure taking into account the permeability

to the subterranean fauna of the different geological substrata in

between them.

Phylogenetic distances were obtained in MESQUITE 3.8 (http://

mesquiteproject.org) from the ultrametric tree obtained with the

cox1-5 sequence, using the best demographic and clock models as

estimated previously. Geographical distances were measured from

the cave entrance and specimens from the same cave were con-

sidered to have a geographical distance of zero. These distances

were expressed in decimal degrees and calculated in ARCGIS 9.3

(ESRI, Redlands, CA, USA). For the geological data we used the

“Mapa de grups litol�ogics de Catalunya LITO250M_v1 (2010)”

downloaded from the Institut Cartogr�afic i Geol�ogic de Catalunya

(www.igc.cat), converted into a raster file at a resolution of 0.001°

(c. 90 m). We pooled all dolostone caves (Jurassic and Triassic) in

a single category, and added two non-karstificable layers (with no

known caves), recognizing a total of four layers: (1) Cretaceous

limestone; (2) Jurassic and Triassic dolostone; (3) sedimentary

rocks with no or few carbonated constituents (mainly marly lime-

stone); (4) metamorphic or sedimentary siliceous rocks (marls,

slate) (Figure 2; Appendixes S1a, S2c). We assigned a value of

geological resistance to each one of these categories as a proxy

of the suitability for the dispersal of the subterranean fauna, with

lower values for cells with higher geologic permeability (i.e. low

resistance). We tested the sensitivity of our results to different

values of resistance by trying five different combinations, always

maintaining the ranking of substratum resistance (1 < 2 < 3 < 4)

(Appendix S2c).

We used the software CIRCUITSCAPE 2.2 (McRae, 2006; McRae,

Dickson, Keitt & Shah, 2008) to estimate the pairwise connectivity

(or current flow) between the cells with observed presences (i.e.

caves with fauna) according to the five resistance maps (Appendixes

S2c,d). CIRCUITSCAPE uses circuit theory to model connectivity in

heterogeneous landscapes. Landscapes are represented as conduc-

tive surfaces, with low resistances assigned to landscape features

types that are most permeable to movement or best promote gene

flow, and high resistances assigned to movement barriers. We

selected the pairwise option so that connectivity, or current flow,

was calculated between all pairs of observed presence cells, and

these pairwise current maps were then overlapped (and averaged) to

obtain a single cumulative current (or connectivity) map

(Appendix S2d).

The values of these five cumulative connectivity maps were then

correlated (Pearson’s correlation coefficient) with genetic distances

using a Mantel test with 1,000 randomizations, to estimate whether

the phylogenetic distance between populations in caves in different

substratum was best explained by geographical distance alone or by

a compound measure of geological isolation and geographical

distance.

2.7 | Genetic diversity and population structure

We compared population genetic statistics between caves from dif-

ferent geological substrata, to test if their different karstificability

resulted in significant differences in dispersal among populations that

could be reflected in their current genetic characteristics. Caves

were classified as being on limestone or dolostone as described

above (Appendix S1a). To increase the number of specimens per

population, when caves were in the same substratum, in close prox-

imity and within the same geological unit (i.e. without any geological

or lithological discontinuity) we pooled individuals within the same

clade following the topology of the previous phylogeographical anal-

yses (Appendix S1a). We collapsed the cox1-5 haplotypes of each

population in Mesquite v3 and estimated Fst values between caves

in ARLEQUIN v3.5 (Excoffier, Laval & Schneider, 2005). We then com-

pared the slope of the regression between geographical distance and

Fst. If due to the lower permeability of the substratum nearby caves

in dolostone are more isolated between them than caves in lime-

stone, the Fst value should increase at a higher rate with geographi-

cal distance, resulting in a stepper slope. For each cave we identified

all caves within a radius of 5 km using ARCGIS 9.3 and plotted the

geographical distance (km) against the Fst. We separated three types

of caves pairs considering the type of substratum, limestone-lime-

stone, limestone-dolostone and dolostone-dolostone, and estimated

the slope of the linear regression for each of them.

3 | RESULTS

3.1 | Phylogenetic and phylogeographical analyses

The complete final data set (Appendix S1a) included 5,173 aligned

nucleotides, with no length variation except for some regions in the

LSU gene and a variation in one or two Asparagine residues in a

repetitive region in the gene Wg. The topology of the coastal clade

of Troglocharinus was very similar to that obtained in Rizzo et al.

(2013), with T. ferreri monophyletic and sister to the rest of the spe-

cies with very good support. Within T. ferreri two well-supported

clades were recovered (clades 1 and 2), although with only moderate

support for most of the internal nodes in each of them

(Appendix S3a). The nuclear sequence only recovered a mono-

phyletic T. ferreri, although with low support, and with no support

for internal relationships (Appendix S3b).

In the calibrated analysis in BEAST with a representative sample of

T. ferreri the best tree model was BD (AICM YL = 28,674.8,

BD = 28,625.4), although with a poorer convergence. The topology

and age estimations of both analyses were, however, almost identi-

cal, with changes in some poorly supported terminal nodes

(Appendix S3c,d). The separation between the coastal and Pyrenean

clades was estimated to have occurred at the end of the Miocene-

early Pliocene (5 � 1 Ma), with the separation between T. ferreri and

the rest of species in close proximity (4.5 � 1 Ma). The sampled T.

ferreri had a last common ancestor in the transition Pliocene-Pleisto-

cene (2.6 � 0.6 Ma), and most of the variation within the different

2532 | RIZZO ET AL.

lineages of T. ferreri had a Middle to Late Pleistocene origin

(Appendix S3c).

3.2 | Demographic history

The analysis of the cox1-5 data (129 specimens, Appendix S1a)

assuming a strict clock had better AICM than when assuming a log-

normal clock (AICM strict = 4072.6; relaxed = 4084.2). We thus

implemented a strict clock in all subsequent analyses using the cox1-

5 data.

The best demographic model was constant population size (CT)

(Table 1). The BSP (Figure 3) suggested no substantial changes for

most of the history of the species, with a rapid population growth

starting at c. 250–300 Ka and a pronounced shift at <100 Ka. In all

subsequent analyses we implemented a CT model.

The Bayesian tree obtained with a strict clock and a CT coales-

cent model had good support for the two main clades within T. fer-

reri (Figure 4), but only moderate or poor support for most of the

lineages. Most haplotypes found in single caves were either exclu-

sive or only found in other caves in close geographical proximity.

Only one cave had individuals with mitochondrial haplotypes in the

two main clades, the Avenc San Cristofol, and only three had individ-

uals with haplotypes in different well supported lineages within each

of the clades (Cuneta, Sogre and Bufi) (Figures 2 and 4). All four

caves with individuals of mixed origins were in limestone

(Appendix S1a). Of the 13 caves in dolostone, none had individuals

with haplotypes in different main lineages, although due to the

higher number of limestone caves (28) differences were not signifi-

cant as measured with a Fisher’s exact test in a 2 9 2 contingency

table.

3.3 | Estimation of transition rates betweengeological substrata

The reduced matrix selecting one specimen per cave when all sam-

pled specimens were monophyletic had 82 haplotypes

(Appendix S1a). The most likely lithology of the ancestral population

was reconstructed as limestone, with a probability of 0.74 over 0.24

dolostone (Figure 4). Similar probabilities were reconstructed for

clade 2 (0.6 vs. 0.4), while clade 1 was clearly reconstructed as hav-

ing originated from limestone caves (0.98, Figure 4). Within clade 1

there were some transitions to Jurassic or Triassic dolostones, but all

very recent and with only one potential reversal to limestone. This

reversal affected two caves (Roca and Rigol) at c. 1.3 km from each

other but in different geological substratum (limestone and Jurassic

dolostone, respectively), which populations had specimens with

mixed haplotypes (Figure 4). Within clade 2 there was one geneti-

cally and geologically isolated lineage sister to the rest, T. ferreri pal-

laresi in Montmany cave, in Triassic dolostones (Figure 4). The

remaining specimens were grouped in two poorly supported clades,

one entirely on limestone caves (clade 2A) and the other recon-

structed as having originated in Jurassic dolostone with only 56%

probability (clade 2B). Within the later there was only one reversal

to limestone, affecting two caves (Cristofol and Bufi) also in close

proximity (c. 0.6 km) and surrounded by dolostone. In both caves

there were also specimens with haplotypes belonging to other clades

in limestone caves (Figure 4), suggesting that they have a fauna of

mixed origin.

There were large differences in the geographical setting of the

different clades. Clade 1 (limestone, with only some caves in Jurassic

dolostone) occupies a large geographical area within the Garraf mas-

sif (mean geographical distance between caves of 7.0 km); while the

two lineages within clade 2 (2A in limestone and 2B in Jurassic dolo-

stone) both occupy a restricted area with a mean geographical dis-

tance between caves in each of them of 1.2 km (Figure 2).

The estimated transition rate from limestone to dolostone was c.

three times that of dolostone to limestone (1.46 vs. 0.51), although

with largely overlapping marginal probability distributions (Figure 4).

3.4 | Genetic versus geological or geographicaldistances

The correlation between geographical and phylogenetic distance was

strongly significant both for the whole T. ferreri and for the two

clades separately, with or without the inclusion of T. ferreri pallaresi

(Montmany cave) (Table 2). None of the schemes with different val-

ues for the geological resistance could improve the correlation

obtained with the simple geographical distance (Table 2), although

the best (also with highly significant correlations) was the one with

the highest resistance values for sedimentary and metamorphic

rocks, which do not have void interstices (Table 2, Appendix S2c).

The phylogenetic distances within the two main clades were very

similar but, as noted above, clade 1 covers a much wider area (Fig-

ure 2), which resulted in significantly different slopes of the regres-

sion lines between genetic and geographical distances (Table 2).

Thus, in clade 2, with both dolostone and limestone caves, the same

phylogenetic distances were attained at much shorter distances than

in clade 1, mostly with limestone caves.

3.5 | Genetic structure in populations in limestoneversus dolostone

The relationship between geographical distance and Fst for caves

closer than 5 km was non-significant for pairs of caves in limestone

TABLE 1 Value of Akaike’s information criterion for MCMCsamples (AICM) of the four tested demographic models. CT,constant population size; ES, population expansion; LG, logisticgrowth; EL, exponential growth. Lower values of AICM indicate abetter adjustment to the model

Model AICM SD DAICM

CT 4008.9 0.08 –

ES 4013.3 0.14 �4.4

LG 4046.7 0.21 �37.8

EL 4012.6 0.10 �3.7

RIZZO ET AL. | 2533

(N = 188, y = 0.38 + 0.04x; p = .37; r2 = .02). This relationship was

significant for pairs of caves in different substratum (limestone-

dolostone; N = 146, y = 0.39 + 0.07x; p = <.01; r2 = .06), and espe-

cially for pairs of caves in dolostone, with a higher slope (N = 22,

y = 0.28 + 0.14x; p < .001; r2 = .50).

4 | DISCUSSION

Despite the reduced extension of its geographical range, T. ferreri

displays a strong phylogeographical structure, as could be expected

from a strictly subterranean species with very limited dispersal capa-

bilities (Kane, Culver & Jones, 1992). This phylogeographical varia-

tion, together with the general homogeneity of the subterranean

environment and the lack of climatic differences due to its limited

geographical range, render subterranean species ideal systems to

study the effect of physical constraints imposed by the habitat (see

e.g. Faille, Bourdeau, Bell�es & Fresneda, 2015; Faille, T€anzler &

Toussaint, 2015 for examples with a different group of Coleoptera).

According to our reconstruction, and in agreement with Rizzo

et al. (2013), the ancestor of the coastal clade of the genus

Troglocharinus colonized the area at the end of the Pliocene-early

Pleistocene, likely in a window of favourable climatic conditions

before the onset of the mediterranean climate c. 3.2 Ma (Suc, 1984).

This range expansion from the ancestral area in the Pyrenees to the

coast may have involved surface displacements, as there is no conti-

nuity in the subterranean medium between the Pyrenean and coastal

areas (Rizzo et al., 2013). The resistance of the species of Troglochar-

inus to relatively high temperatures (up to 20–23°C, Rizzo, S�anchez-

Fern�andez, Fresneda, Cieslak & Ribera, 2015) would allow surface

displacements in periods with low seasonality and high precipitation,

such as the Pliocene-Pleistocene transition (Jim�enez-Moreno, Fau-

quette & Suc, 2010; Suc & Cravatte, 1982). The subsequent forma-

tion of sedimentary basins in the rivers surrounding the Garraf

massif likely contributed to the isolation of the populations that

would become the current T. ferreri (Rizzo et al., 2013).

The demographic reconstruction suggested a constant population

size for most of the history of the species but with a fast population

expansion starting at c. 0.3 Ma. This expansion is in agreement with

the available data of the development of the subterranean environ-

ment in the Garraf massif (Daura et al., 2014). According to these

authors, the caves in the Garraf plateau have an endogenous origin,

and the subterranean network of cavities and fissures did not devel-

oped completely, opening to the surface, before the middle-upper

Pleistocene. The accessibility to an extensive subterranean system in

the massif could have allowed the population expansion of T. ferreri.

It must be noted that the ancestor of the coastal clade of Troglochar-

inus (and thus of T. ferreri) must have had all the morphological and

physiological modifications currently associated with the subter-

ranean life (Cieslak et al., 2014; Rizzo et al., 2013), so it could be

expected that the subterranean environment was occupied when-

ever it become accessible.

We found a highly significant linear correlation between geo-

graphical and phylogenetic distance in all main lineages within T. fer-

reri, despite the reduced geographical area. This correlation is

habitually interpreted as isolation by distance (Slatkin, 1993), without

dispersal barriers or corridors that could, respectively, increase or

decrease isolation with independence of geographical distance.

Within the Garraf Massif the subterranean environment is not con-

tinuous, as there are geological discontinuities and non-calcareous

layers which in principle should act as barriers to dispersal through

the underground. This was apparent in the case of the subspecies T.

ferreri pallaresi, isolated from the rest of the populations by a layer

of non-karstificable Triassic Keuper marl (ICC [Institut Cartogr�afic de

Catalunya], 2010). However, and contrary with our expectations, the

attempt to weight geographical distances with a priori values for the

permeability to dispersal of the different substrata did not result in

any improvement in the correlation with phylogenetic distance,

Time (Ma)

0 0.5 1 1.5 20

1.E0

1.E1

1.E2

F IGURE 3 Bayesian skyline plot of theanalyses of the cox1-5 matrix of 129specimens of Troglocharinus ferreri,assuming a strict clock with a rate of 0.015substitutions/site/MY. Thin lines, 95%confidence intervals; horizontal axis, time(MY); vertical axis, effective population size(NeT)

2534 | RIZZO ET AL.

RA1106 Geolegs

VR80 Infern

PB1 Sogre *

RA1160 Marçal

VR133 Troneda

VR108 Esquerra

PB15 Cristofol **

VR44 Llinia

RA1105 Pompeu Fabra

VR130 Troneda

RA1100 Roc

RA1097 Rigol 2

VR81 Infern

PB14 Cristofol **

VR41 Puigmolto

VR33 Verdaguer

PB8 Trons

RA1193 Pasant

VR58 Marcel

VR53 Benjamin Digò

VR42 Puigmolto

RA1107 Geolegs

PB6 Trons

VR70 Verdiell

VR118 Comes

VR83 Pleta

VR3 Montmany

VR62 Cuneta *

VR59 Marcel_1.938_41.351

VR73 Vermell

PB12 Cristofol **

RA748 Alzines Gran

RA1098 Rigol 2

RA1093 Rigol

RA1092 Roca

PB3 Pere

VR90 Bufi *

VR105 Gori

VR91 Bufi *

VR112 Banderes

RA1159 Pepi

VR48 Morgan i Comas

VR27 Corral Nou

VR126 Svabroc

RA1192 Pasant

VR117 Esteles

RA1111 Trons

VR68 VerdiellAF172 Bardissa

VR104 GoriRA1194 Gori

VR55 Cabeza

VR115 Banderes

RA1108 Sadurnì

PB11 Sogre *

VR50 Benjamin Digò

PB22 Roca

RA1109 Sadurni

VR111 Esquerra

VR46 Llinia

PB21 Pompeu Fabra

VR49 Benjamin Digò

VR82 Infern

VR119 Comes

RA1156 Pere

AI1065 Roc

VR128 Svabroc

PB2 Pere

VR92 Bufi *

VR47 Llinia

RA1099 Roc

VR52 Benjamin Digò

PB5 Serrano

VR26 Corral Nou

PB7 Trons

VR31 Guerrillers

VR76 Vermell_1.939_41.347

VR45 Llinia

RA1163 Marçal

VR69Verdiell

RA1104 Pompeu Fabra

VR60 Cuneta *

1

[0.97,0.03]

[0.02,0.98]

[0.03,0.97]

[0.04,0.96]

[0.26,0.74]

[0,1]

1

[0,1]

[0.02,0.98]

[0.23,0.77]

[0,1]

[0.56,0.44]

[0.02,0.98]

1

[0.11,0.89]

1

[0.01,0.99]

[0.03,0.97]

[1,0]

1

1

1

[0,1]

[0.02,0.98]

[0,1]

[0.01,0.99]

[0.03,0.97]

1

[0.02,0.98]

[0,1]

[0,1]

1

1

[0.01,0.99]

1

[0.23,0.77]

1

1

[1,0]

[1,0]

[1,0]

1

[0.06,0.94]

[0.01,0.99]

1

[0,1]

1

[0.01,0.99]

[0.12,0.88]

[0,1]

1

[0,1]

1

[0,1]

[0.07,0.93]

1

1

1

[0.01,0.99]

1

1

[0.02,0.98]

[0.59,0.41]

1

1

1

1

[0.36,0.64]

[0.06,0.94]

[0,1]

[0,1]

[0.53,0.47]

[1,0]

[0.43,0.57]

[0,1]

[0,1]

[0.92,0.08]

11

[0.02,0.98]

Clade 2

Clade 2B

Clade 2A

Clade 1

Troglocharinus ferreri

6420

0.5

1

1.5

2

No. Transitions/Ma

Den

sity

F IGURE 4 Reconstruction of theancestral lithology of the substratum inBEAST, using a simplified cox1-5 matrix anda strict clock with a rate of 0.015substitutions/site/MY and a constantpopulation size. Numbers in nodes,probabilities of respectively dolostone (red)and limestone (blue) (only when >0). Withstars, specimens from caves with fauna ofmixed origin (two stars, different mainclades; one, different subclades). SeeFigure 2 for the location of these caves.Insert: Marginal probability distribution ofthe reconstructed transition rates fromlimestone to dolostone (grey) anddolostone to limestone (blue). Habitusphotograph by A. Messeger

RIZZO ET AL. | 2535

although the best result was obtained with the highest resistance

values for non-karstificable materials (sedimentary and metamorphic

rocks). It must be considered that, although highly significant, this

correlation was relatively small, with low slopes and a high disper-

sion except when the geographically and genetically very isolated

population of T. ferreri pallaresi was included in the analyses. The

low variance explained by distance alone suggests a more complex

scenario, with possibly a number of uncontrolled factors determining

dispersal acting at small scales, although it may just be that our

exploration of the parametric space was insufficient.

Other than the possible existence of what could be considered

strong barriers, preventing all movement through the subterranean

habitat, the different composition of the substratum did also influ-

ence the dispersal capabilities and in consequence the genetic struc-

ture of the studied populations. The Garraf Massif is broadly divided

into two bands of different composition perpendicular to the coast,

limestone in the south and dolostone in the north (Figure 2). Our

results suggest that the initial colonization was in limestone, more

soluble and thus likely to have developed a subterranean network of

fissures earlier than the dolostone. The initial colonization of lime-

stone may have biased the estimation of the transition probabilities

(Davis, Midford & Maddison, 2013), which had also wide marginal

distributions likely due to the low number of dolostone caves. But in

any case, our results indicate a much higher rate of transitions from

limestone to dolostone than vice-versa, suggesting that limestone is

more permeable to the movements of the subterranean fauna. This

higher permeability agrees with the fact that in most dolostone

caves only closely related specimens were found, in many cases

forming a monophyletic lineage exclusive of a cave or a group of

caves in close geographical proximity, although differences with

limestone caves were not statistically significant. It is interesting to

note that the limestone caves with fauna of mixed origin were all in

the same area, and in close proximity of the caves in dolostone

inhabited by populations of clade 2 (see the location of the caves in

Figure 2). There is the possibility that other factors favoured the

transit of the subterranean fauna in this area, including some early

transitions from limestone to dolostone in this clade. The high num-

ber of tectonic faults (ICC [Institut Cartogr�afic de Catalunya], 2010)

may be one of these factors.

The general relationship between phylogenetic and geographical

distances was also indicative of differences between the permeabil-

ity of dolostone and limestone to the movement of the subterranean

fauna. Thus, the clade with the highest proportion of dolostone

caves (clade 2) had a steeper slope than the clade mostly on lime-

stone caves (clade 1), suggesting a lower permeability to the dis-

placement of the fauna. A similar result was obtained for the degree

of genetic isolation (as measured with the Fst) between the popula-

tion of a cave and that of its closest neighbours. For caves in dolo-

stone Fst values increased much faster with distance than when

caves were in limestone, clearly indicating a stronger isolation likely

due to the increased difficulty of displacement through the

substratum.

There are few systems in which it is possible to trace the effect

of a physical constraint from the individual to the macroevolution of

the lineage. In this sense, the highly fragmented subterranean habitat

offers an excellent opportunity, with the additional advantage of its

abiotic and biotic simplicity and stability. Knowing the factors leading

to the origin of the current diversity can greatly help to inform man-

agement and conservation decisions, but the lack of data of both

genetic diversity and population viability in this unique system ham-

pers our understanding of the potential effect of environmental

changes (Rizzo et al., 2015; S�anchez-Fern�andez et al., 2016).

ACKNOWLEDGEMENTS

We thank Aliga speleo Group of Barcelona, Jordi Comas, Javier Fres-

neda and Enric Lleopart for help and support in the field; Pau Balart,

Ana Izquierdo and Arnaud Faille for laboratory work, and Jordi

Comas for multiple comments and suggestions on the distribution

and taxonomy of T. ferreri. We also thank the editor (Isabel San-

mart�ın) and three referees for the useful comments to earlier ver-

sions of the manuscript. D.S.-F. was supported by a “Juan de la

Cierva” postdoctoral contract from the Spanish Ministry of Economy

and Competitiveness and another post-doctoral contract funded by

Universidad de Castilla-La Mancha and the European Social Fund

(ESF). This work was partly funded by projects CGL2010-15755 and

CGL2016-76705-P (AEI/FEDER, UE) to I.R.

REFERENCES

Appelo, C. A. J., & Postma, D. (2005). Geochemistry, groundwater and pol-

lution. Boca Raton, FL: CRC Press.

Baele, G., Li, W. L. S., Drummond, A. J., Suchard, M. A., & Lemey, P.

(2013). Accurate model selection of relaxed molecular clocks in Baye-

sian phylogenetics. Molecular Biology and Evolution, 30, 239–243.

TABLE 2 Linear regression between phylogenetic distances,Euclidean geographical distances (in decimal degrees) and distancesweighted by resistance schemes A, B & E (values for C&D wereintermediate between B&E, data not shown). See Appendixes S2c,dfor the resistance values applied to the different geologicalsubstrata; and Appendix S1a and Figure 4 for the composition ofthe clades (clade “2[A+B]” refers to Clade 2 with the exclusion ofT. ferreri pallaresi)

Distance Clade Intercept SE (i) Slope SE (s) r2 p

Geographical 1 0.22 0.004 1.78 0.051 .27 <.01

2 0.20 0.003 3.57 0.059 .77 <.01

2(A+B) 0.18 0.004 4.91 0.233 .32 <.01

A 1 0.32 0.003 0.001 0.001 .00 n.s.

2 0.17 0.005 0.02 0.001 .58 n.s.

2(A+B) 0.29 0.005 0.01 0.001 .11 n.s.

B 1 0.29 0.004 0.02 0.002 .02 n.s.

2 0.14 0.005 0.08 0.002 .60 n.s.

2(A+B) 0.20 0.005 0.04 0.003 .16 n.s.

E 1 0.21 0.006 0.11 0.005 .11 <.01

2 0.07 0.006 0.24 0.005 .65 <.01

2(A+B) 0.15 0.006 0.13 0.007 .27 <.01

2536 | RIZZO ET AL.

Bell�es, X. (1973). Un nuevo Bathysciinae del Macizo de Garraf. Mis-

cel�anea Zool�ogica, 3, 45–49.

Bell�es, X., & Mart�ınez, A. (1980). La geolog�ıa y la especiaci�on de los

Bathysciinae en la regi�on del Pened�es (Catalunya, Espa~na). M�emoires

de Biosp�eologie, 7, 221–223.

Cardona i Oliv�an, F. (1990). Grans cavitats de Catalunya, segon volum: El sis-

tema mediterrani i la Depressi�o Central. Barcelona: Espeleo Club de Gracia.

Cieslak, A., Fresneda, J., & Ribera, I. (2014). Life-history specialization

was not an evolutionary dead-end in Pyrenean cave beetles. Proceed-

ings of the Royal Society B: Biological Sciences, 281, 20132978.

Crouau-Roy, B. (1989). Population studies on an endemic troglobitic bee-

tle: Geographical patterns of genetic variation, gene flow and genetic

structure compared with morphometric data. Genetics, 121, 571–582.

Culver, D. C. (1976). The evolution of aquatic cave communities. The

American Naturalist, 110, 945–957.

Culver, D. C., & Pipan, T. (2009). The biology of caves and other subter-

ranean habitats. Oxford: Oxford University Press.

Daura, J., Sanz, M., Forno, J. J., Asensio, A., & Juli�a, R. (2014). Karst evo-

lution of the Garraf Massif (Barcelona, Spain): Doline formation,

chronology and archaeo-palaeontological archives. Journal of Cave

and Karst Studies, 76, 69–87.

Davis, M. P., Midford, P. E., & Maddison, W. (2013). Exploring power and

parameter estimation of the BiSSE method for analyzing species

diversification. BMC Evolutionary Biology, 13, 38.

Drummond, A. J., Rambaut, A., Shapiro, B., & Pybus, O. G. (2005). Baye-

sian coalescent inference of past population dynamics from molecular

sequences. Molecular Biology and Evolution, 22, 1185–1192.

Drummond, A. J., Suchard, M. A., Xie, D., & Rambaut, A. (2012). Bayesian

phylogenetics with BEAUti and the BEAST 1.7. Molecular Biology and

Evolution, 29, 1969–1973.

Emlet, R. B. (1995). Developmental mode and species geographic range

in regular sea urchins (Echinodermata: Echinoidea). Evolution, 49,

476–489.

Espa~nol, F. (1961). Fauna cavern�ıcola de la provincia de Barcelona 1:

Invertebrados. Cat�alogo Espelol�ogico de la Provincia de Barcelona, 1,

29–46.

Excoffier, L., Laval, G., & Schneider, S. (2005). Arlequin (version 3.0): An

integrated software package for population genetics data analysis.

Evolutionary Bioinformatics Online, 1, 47.

Faille, A., Bourdeau, C., Bell�es, X., & Fresneda, J. (2015). Allopatric specia-

tion illustrated: The hypogean genus Geotrechus Jeannel, 1919

(Coleoptera: Carabidae: Trechini), with description of four new spe-

cies. Arthropod Systematics and Phylogeny, 73, 439–455.

Faille, A., T€anzler, R., & Toussaint, E. F. A. (2015). On the way to specia-

tion: Shedding light on the karstic phylogeography of the micro-ende-

mic cave beetle Aphaenops cerberus in the Pyrenees. Journal of

Heredity, 106, 692–699.

Fresneda, J., & Salgado, J. M. (2017). Cat�alogo de los Cole�opteros Leiodi-

dae Cholevinae Kirby, 1837 de la pen�ınsula Ib�erica e islas Baleares.

Monografies del Museu de Ci�encies Naturals, 7, 1–308.

Gilli, E. (2015). Karstology. Karsts, caves and springs. Boca Raton, FL: CRC

Press.

Goldscheider, N., & Drew, D. (Eds.) (2007). Methods in Karst hydrogeology:

IAH: international contributions to hydrogeology. Boca Raton, FL: CRC

Press.

Grime, J. P. (1977). Evidence for the existence of three primary strategies

in plants and its relevance to ecological and evolutionary theory. The

American Naturalist, 111, 1169–1194.

ICC [Institut Cartogr�afic de Catalunya]. (2010). Mapa de grups litol�ogics de

Catalunya 1:250.000 LITO250M_v1. Barcelona: Institut Cartogr�afic i

Geol�ogic de Catalunya.

Jim�enez-Moreno, G., Fauquette, S., & Suc, J.-P. (2010). Miocene to Plio-

cene vegetation reconstruction and climate estimates in the Iberian

Peninsula from pollen data. Review of Palaeobotany and Palynology,

162, 403–415.

Juberthie, C., Delay, B., & Bouillon, M. (1980). Sur I’existence d’un milieu

souterrain superficiel en zone non calcaire. Comptes-rendus de

l’Acad�emie des Sciences de Paris, 290(D), 49–52.

Kane, T. C., Culver, D. C., & Jones, R. T. (1992). Genetic structure of

morphologically differentiated populations of the amphipod Gam-

marus minus. Evolution, 46, 272–278.

Katoh, K., Asimenos, G., & Toh, H. (2009). Multiple alignment of DNA

sequences with MAFFT. Methods in Molecular Biology, 537, 39–64.

Lagar, A. (1954). Los Bathysciinae de la provincia de Barcelona. Speleon,

5, 247–259.

Lemey, P., Rambaut, A., Drummond, A. J., & Suchard, M. A. (2009). Baye-

sian phylogeography finds its roots. PLoS Computational Biology, 5(9),

e1000520.

McRae, B. H. (2006). Isolation by resistance. Evolution, 60, 1551–1561.

McRae, B. H., Dickson, B. G., Keitt, T. H., & Shah, V. B. (2008). Using cir-

cuit theory to model connectivity in ecology and conservation. Ecol-

ogy, 10, 2712–2724.

Moreno, J. A. (2007). Bioestratigraf�ıa del Aptiense del macizo del Garraf

(NE de la Pen�ınsula Ib�erica). Geogaceta, 41, 131–134.

Papadopoulou, A., Anastasiou, I., Keskin, B., & Vogler, A. P. (2009). Com-

parative phylogeography of tenebrionid beetles in the Aegean archi-

pelago: The effect of dispersal ability and habitat preference.

Molecular Ecology, 18, 2503–2517.

Poulson, T. L., & Culver, D. C. (1969). Diversity in terrestrial cave com-

munities. Ecology, 50, 153–158.

Poulson, T. L., & White, W. B. (1969). The cave environment. Science,

165, 971–981.

Racovitza, E. G. (1907). Essai sur les probl�emes biosp�eologiques. Archives

de Zoologie Exp�erimentale et G�en�erale. 4�eme s�erie, VI, 371–488.

Renault, P. H. (1970). La formation des cavernes, Que Sais-je?. Paris: PUF.

Ribera, I. (2008). Chapter 15: Habitat constraints and the generation of

diversity in freshwater macroinvertebrates. In J. Lancaster & R. A.

Briers (Eds.), Aquatic insects: Challenges to populations (pp. 289–311).

Wallingford, UK: CAB International.

Ribera, I., & Vogler, A. P. (2000). Habitat type as a determinant of

species range sizes: The example of lotic–lentic differences in

aquatic Coleoptera. Biological Journal of the Linnean Society, 71,

33–52.

Rizzo, V., Comas, J., Fadrique, F., Fresneda, J., & Ribera, I. (2013). Early

Pliocene range expansion of a clade of subterranean Pyrenean bee-

tles. Journal of Biogeography, 40, 1861–1873.

Rizzo, V., S�anchez-Fern�andez, D., Fresneda, J., Cieslak, A., & Ribera, I.

(2015). Lack of evolutionary adjustment to ambient temperature in

highly specialized cave beetles. BMC Evolutionary Biology, 15, 10.

Rubinat, F. (2004). Cat�aleg Espeleol�ogic del Mass�ıs de l’Ordal: Barcelona.

Barcelona: Centre Excursionista �Aliga.

Salgado, J. M., Blas, M., & Fresneda, J. (2008). Fauna Ib�erica. Vol. 31:

Coleoptera: Cholevidae. Madrid: CSIC.

S�anchez-Fern�andez, D., Rizzo, V., Cieslak, A., Faille, A., Fresneda, J., &

Ribera, I. (2016). Thermal niche estimators and the capability of poor

dispersal species to cope with climate change. Scientific Reports, 6,

23381.

Schettino, A., & Turco, E. (2006). Plate kinematics of the Western

Mediterranean region during the Oligocene and Early Miocene. Geo-

physical Journal International, 166, 1398–1423.

Slatkin, M. (1993). Isolation by distance in equilibrium and non-equili-

brium populations. Evolution, 47, 264–279.

Southwood, T. R. E. (1977). Habitat, the templet for ecological strategies?

Journal of Animal Ecology, 46, 337–365.

Stadler, T. (2009). On incomplete sampling under birth-death models and

connections to the sampling-based coalescent. Journal of Theoretical

Biology, 261, 58–66.

Stamatakis, A., Hoover, P., & Rougemont, J. (2008). A rapid bootstrap

algorithm for the RAxML web servers. Systematic Biology, 57, 758–

771.

RIZZO ET AL. | 2537

Suc, J. P. (1984). Origin and evolution of the Mediterranean vegetation

and climate in Europe. Nature, 307, 429–432.

Suc, J. P., & Cravatte, J. (1982). �Etude palynologique du Plioc�ene de Cat-

alogne (nord-est de l’Espagne). Paleobiologie Continentale, 13, 1–31.

Zariquieyi, R. (1917). Sobre el g�enere Troglocharinus (Ins. Col.). Treballs de

la Instituci�o Catalana d’Hist�oria Natural, 3, 283–294.

BIOSKETCH

Valeria Rizzo is a PhD student in the Institute of Evolutionary

Biology in Barcelona. This paper is part of his thesis dissertation,

focussed on the evolution of a lineage of subterranean beetles

(genus Troglocharinus).

Author contributions: V.R. and I.R. conceived the work; V.R., J.P.

and I.R. led the specimen and data collection; V.R. and R.A.

obtained the molecular data; V.R., D.S-F. and I.R. analysed the

data; V.R. and I.R. led the writing and all authors contributed to

the discussion of results and writing.

SUPPORTING INFORMATION

Additional Supporting Information may be found online in the sup-

porting information tab for this article.

How to cite this article: Rizzo V, S�anchez-Fern�andez D,

Alonso R, Pastor J, Ribera I. Substratum karstificability,

dispersal and genetic structure in a strictly subterranean

beetle. J Biogeogr. 2017;44:2527–2538. https://doi.org/

10.1111/jbi.13074

2538 | RIZZO ET AL.

1

Substratum karstificability, dispersal and genetic structure in a strictly

subterranean beetle

Valeria Rizzo1, David Sánchez-Fernández1,2, Rocío Alonso1, Josep Pastor3 and Ignacio

Ribera1*

1Institute of Evolutionary Biology (CSIC-Universitat Pompeu Fabra), Barcelona, Spain

2 Instituto de Ciencias Ambientales. Universidad de Castilla-La Mancha. Campus

Tecnológico de la Fábrica de Armas, Toledo, Spain

3Museu de Ciències Naturals (Zoologia), Barcelona, Spain

*Correspondence: I. Ribera, Institute of Evolutionary Biology, Passeig Maritim de la

Barceloneta, 37-49, 08003, Barcelona, Spain.

E-mail: ignacio.ribera@ibe.upf-csic.es

SUPPORTING INFORMATION

Appendix S1 Additional materials:

(a) List of the studied material, with accession numbers of the sequences used and the

lithology and geographic coordinates of the caves.

Appendix S2 Additional methods:

(a) Location of all known caves in the Garraf massif (yellow circles), including those in

which the presence of Troglocharinus ferreri has been recorded (red circles).

(b) Primers used in the study.

(c) Categorisation of the different geological substrata according to the geological map

of the Garraf Massif (ICC 2010), and resistance values given to the main lythological

types in the correlation between phylogenetic distance and weighted geographical

distances.

2

(d) Resistance maps, applying the resistance values in Appendix S2c (Figs A to E) to

the different general lithological categories. Blue circles, caves with specimens in clade

1; red circles, caves with specimens in clade 2 (see Figs 4,5).

Appendix S3 Additional results:

(a) Best tree obtained in RAxML using all five mitochondrial and four nuclear markers.

Numbers in nodes, fast bootstrap support values (when >50%). See text and Appendix

S1a for details on the specimens included and the methods.

(b) Best tree obtained in RAxML using only the four nuclear markers for a reduced

representation of T. ferreri and outgroups. Numbers in nodes, fast bootstrap support

values (when >50%). Se text and Appendix S1a for details on the specimens included

and the methods.

(c) Majority rule consensus tree obtained in BEAST using all five mitochondrial and

four nuclear markers and a reduced representation of T. ferreri, using as tree priors a

Yule speciation. Numbers in nodes, posterior probabilities (when >0.5, in red) and

estimated age with 95% confidence interval in Million years (Ma) for selected nodes. Se

text and Appendix S1a for details on the specimens included and the methods.

(d) Majority rule consensus tree obtained in BEAST using all five mitochondrial and

four nuclear markers and a reduced representation of T. ferreri, using as tree priors a

birth death with incomplete sampling. Numbers in nodes, posterior probabilities (when

>0.5). Se text and Appendix S1a for details on the specimens included and the methods.

Substratumkarstificability,dispersalandgeneticstructureinastrictlysubterraneanbeetleValeriaRizzo,DavidSánchez-Fernández,RocioAlonso,JosepPastorandIgnacioRibera

AppendixS1(a)List of the studied material, with accession numbers of the sequences used and the lithology and geographic coordinates of the caves.Typelocalitiesinbold.SubspeciesinsquarebracketsarecurrentlyconsideredjuniorsynonymsofT.ferreriferreri.Inbold,newsequences(285).Analyses:(1)calibrationanalyses;(2)demographic&geologicalreconstruction

AnalysesNo Genus sp spp ex.voucher cave (1) (2) lythotype collector year cavegroupsforFst UTM bar cox1 16S nad1 cyb 18S 28S H3 WG

1 Troglocharinus ferreri [codinai] IBE-VR73 AvencVermell x Triassicdolostone V.Rizzo 2011 411430 4578066 LT797202 LT797341 LT7973732 Troglocharinus ferreri [codinai] IBE-VR74 AvencVermell Triassicdolostone V.Rizzo 2011 411430 4578066 LT7972163 Troglocharinus ferreri [codinai] IBE-VR75 AvencVermell Triassicdolostone V.Rizzo 2011 411430 4578066 LT7972174 Troglocharinus ferreri [codinai] IBE-VR76 AvencVermell x x Triassicdolostone V.Rizzo 2011 411430 4578066 LT797189 LT797335 LT7973785 Troglocharinus ferreri [codinai] IBE-VR77 AvencVermell Triassicdolostone V.Rizzo 2011 411430 4578066 LT7972186 Troglocharinus ferreri [codinai] IBE-VR81 AvencVermell x Triassicdolostone V.Rizzo 2011 411430 4578066 LT7972197 Troglocharinus ferreri [fonti] IBE-PB10 AvencdeSantRoc Jurassicdolostone J.Fresneda 2006 Cabeza+SanRoc+Puigmoltò 409725 4574299 LT797273 LT7973968 Troglocharinus ferreri [fonti] IBE-PB9 AvencdeSantRoc Jurassicdolostone J.Fresneda 2006 Cabeza+SanRoc+Puigmoltò 409725 4574299 LT7972779 Troglocharinus ferreri [fonti] IBE-RA1099 AvencdeSantRoc x Jurassicdolostone J.Pastor 2013 Cabeza+SanRoc+Puigmoltò 409725 4574299 LT79727810 Troglocharinus ferreri [fonti] IBE-RA1100 AvencdeSantRoc x Jurassicdolostone J.Pastor 2013 Cabeza+SanRoc+Puigmoltò 409725 4574299 LT79717711 Troglocharinus ferreri [fonti] MNCN-AI1065 AvencdeSantRoc x x Jurassicdolostone J.Fresneda 2006 Cabeza+SanRoc+Puigmoltò 409725 4574299 LT797272 HF912496 HF912612 HF912547 HF912547 HF912582 HF912594 LT797353 LT79739312 Troglocharinus ferreri ferreri IBE-VR49 AvencBenjaminDigò x calcite V.Rizzo 2010 409213 4571473 LT79729113 Troglocharinus ferreri ferreri IBE-VR50 AvencBenjaminDigò x calcite V.Rizzo 2010 409213 4571473 LT797185 LT79737014 Troglocharinus ferreri ferreri IBE-VR51 AvencBenjaminDigò calcite V.Rizzo 2010 409213 4571473 LT797285 LT797355 LT79737615 Troglocharinus ferreri ferreri IBE-VR52 AvencBenjaminDigò x calcite V.Rizzo 2010 409213 4571473 LT79729216 Troglocharinus ferreri ferreri IBE-VR53 AvencBenjaminDigò x calcite V.Rizzo 2010 409213 4571473 LT797286 LT797356 LT79737917 Troglocharinus ferreri ferreri IBE-RA1194 Avencd'enGori x calcite J.Pastor 2011 404677 4578670 LT79719318 Troglocharinus ferreri ferreri IBE-VR103 Avencd'enGori calcite J.Pastor 2011 404677 4578670 LT79719419 Troglocharinus ferreri ferreri IBE-VR104 Avencd'enGori x x calcite J.Pastor 2011 404677 4578670 LT797190 LT797336 LT79739820 Troglocharinus ferreri ferreri IBE-VR105 Avencd'enGori x calcite J.Pastor 2011 404677 4578670 LT79728021 Troglocharinus ferreri ferreri IBE-VR106 Avencd'enGori calcite J.Pastor 2011 404677 4578670 LT797279 LT797366 LT79742222 Troglocharinus ferreri ferreri IBE-VR68 Avencd'enJordiVerdiell x calcite V.Rizzo 2011 Verdiell+Bardissa+Esteles 404980 4582126 LT79723723 Troglocharinus ferreri ferreri IBE-VR69 Avencd'enJordiVerdiell x calcite V.Rizzo 2011 Verdiell+Bardissa+Esteles 404980 4582126 LT797231 LT797372 LT79741024 Troglocharinus ferreri ferreri IBE-VR70 Avencd'enJordiVerdiell x calcite V.Rizzo 2011 Verdiell+Bardissa+Esteles 404980 4582126 LT79723825 Troglocharinus ferreri ferreri IBE-VR71 Avencd'enJordiVerdiell x calcite V.Rizzo 2011 Verdiell+Bardissa+Esteles 404980 4582126 LT797232 LT797343 LT79738126 Troglocharinus ferreri ferreri IBE-VR72 Avencd'enJordiVerdiell calcite V.Rizzo 2011 Verdiell+Bardissa+Esteles 404980 4582126 LT79723927 Troglocharinus ferreri ferreri IBE-RA1192 Avencd'enPasant x calcite V.Rizzo 2011 408654 4571082 LT79729428 Troglocharinus ferreri ferreri IBE-RA1193 Avencd'enPasant x calcite V.Rizzo 2011 408654 4571082 LT79718029 Troglocharinus ferreri ferreri IBE-VR95 Avencd'enPasant calcite V.Rizzo 2011 408654 4571082 LT79729530 Troglocharinus ferreri ferreri IBE-VR97 Avencd'enPasant calcite V.Rizzo 2011 408654 4571082 LT797293 LT797358 LT79739531 Troglocharinus ferreri ferreri IBE-PB2 Avencd'enPere x x calcite J.Pastor 2012 408932 4572805 LT79720732 Troglocharinus ferreri ferreri IBE-PB3 Avencd'enPere x calcite J.Pastor 2012 408932 4572805 LT79720833 Troglocharinus ferreri ferreri IBE-RA1156 Avencd'enPere x x calcite J.Pastor 2012 408932 4572805 LT79721134 Troglocharinus ferreri ferreri IBE-PB22 Avencd'enRoca x x calcite J.Pastor 2013 406911 4584181 LT79722135 Troglocharinus ferreri ferreri IBE-RA1091 Avencd'enRoca calcite J.Pastor 2013 406911 4584181 LT797224 LT79942736 Troglocharinus ferreri ferreri IBE-RA1092 Avencd'enRoca x calcite J.Pastor 2013 406911 4584181 LT79727637 Troglocharinus ferreri ferreri IBE-RA1108 AvencdeCanSadurní x calcite J.Pastor 2013 408593 4578273 LT797198 LT797321 LT79738038 Troglocharinus ferreri ferreri IBE-RA1109 AvencdeCanSadurní x x calcite J.Pastor 2013 408593 4578273 LT797199 LT797338 LT79740039 Troglocharinus ferreri ferreri IBE-VR107 Avencdel'Esquerrà calcite J.Pastor 2011 Esquerrà+Banderes 4045826 4578039 LT79725540 Troglocharinus ferreri ferreri IBE-VR108 Avencdel'Esquerrà x calcite J.Pastor 2011 Esquerrà+Banderes 4045826 4578039 LT79725641 Troglocharinus ferreri ferreri IBE-VR110 Avencdel'Esquerrà x calcite J.Pastor 2011 Esquerrà+Banderes 4045826 4578039 LT797242 LT797345 LT79741842 Troglocharinus ferreri ferreri IBE-VR111 Avencdel'Esquerrà x calcite J.Pastor 2011 Esquerrà+Banderes 4045826 4578039 LT79725743 Troglocharinus ferreri ferreri IBE-AF172 AvencdelaBardissa x calcite J.Comas 2009 Verdiell+Bardissa+Esteles 405009 4583047 LT797229 HF912457 HF912610 LT797318 LT79739744 Troglocharinus ferreri ferreri IBE-VR60 AvencdelaCuneta x calcite V.Rizzo 2010 406617 4579692 LT797259 LT79741545 Troglocharinus ferreri ferreri IBE-VR61 AvencdelaCuneta x calcite V.Rizzo 2010 406617 4579692 LT797230 LT797312 LT797328 LT79737746 Troglocharinus ferreri ferreri IBE-VR62 AvencdelaCuneta x calcite V.Rizzo 2010 406617 4579692 LT797236 LT79942447 Troglocharinus ferreri ferreri IBE-VR43 AvencdelaLínia calcite V.Rizzo 2010 404260 4581712 LT79723448 Troglocharinus ferreri ferreri IBE-VR44 AvencdelaLínia x calcite V.Rizzo 2010 404260 4581712 LT797233 LT79731949 Troglocharinus ferreri ferreri IBE-VR45 AvencdelaLínia x calcite V.Rizzo 2010 404260 4581712 LT79721350 Troglocharinus ferreri ferreri IBE-VR46 AvencdelaLínia x calcite V.Rizzo 2010 404260 4581712 LT797206 LT797414

51 Troglocharinus ferreri ferreri IBE-VR47 AvencdelaLínia x calcite V.Rizzo 2010 404260 4581712 LT79723552 Troglocharinus ferreri ferreri IBE-RA1158 AvencdelaPepi x Jurassicdolostone J.Comas 2012 Guerrillers+Pepi 410412 4571645 LT79717253 Troglocharinus ferreri ferreri IBE-RA1159 AvencdelaPepi x Jurassicdolostone J.Comas 2012 Guerrillers+Pepi 410412 4571645 LT79717054 Troglocharinus ferreri ferreri IBE-VR124 AvencdelaPepi x Jurassicdolostone V.Rizzo&J.Comas 2011 Guerrillers+Pepi 410412 4571645 LT797171 LT79733055 Troglocharinus ferreri ferreri IBE-VR83 AvencdelaPleta x x calcite V.Rizzo 2011 Morgan+Pleta 409018 4569729 LT797248 LT797327 LT79737556 Troglocharinus ferreri ferreri IBE-VR85 AvencdelaPleta calcite V.Rizzo 2011 Morgan+Pleta 409018 4569729 LT79725257 Troglocharinus ferreri ferreri IBE-VR86 AvencdelaPleta calcite V.Rizzo 2011 Morgan+Pleta 409018 4569729 LT79725358 Troglocharinus ferreri ferreri IBE-VR87 AvencdelaPleta calcite V.Rizzo 2011 Morgan+Pleta 409018 4569729 LT79725459 Troglocharinus ferreri ferreri IBE-VR130 AvencdelaTroneda x Jurassicdolostone V.Rizzo 2011 4098983 45735143 LT79727160 Troglocharinus ferreri ferreri IBE-VR131 AvencdelaTroneda Jurassicdolostone V.Rizzo 2011 4098983 45735143 LT79717861 Troglocharinus ferreri ferreri IBE-VR132 AvencdelaTroneda Jurassicdolostone V.Rizzo 2011 4098983 45735143 LT797269 LT797352 LT79738562 Troglocharinus ferreri ferreri IBE-VR133 AvencdelaTroneda x Jurassicdolostone V.Rizzo 2011 4098983 45735143 LT797176 LT797333 LT79739063 Troglocharinus ferreri ferreri IBE-RA1112 AvencdelesAlzines calcite J.Pastor 2013 Alzines+AlzinesGran 409001 4579126 LT797262 LT79737464 Troglocharinus ferreri ferreri IBE-RA1113 AvencdelesAlzines calcite J.Pastor 2013 Alzines+AlzinesGran 409001 4579126 LT79726065 Troglocharinus ferreri ferreri IBE-RA748 AvencdelesAlzinesGran x x calcite J.Pastor 2012 Alzines+AlzinesGran 408864 4578726 LT797261 LN849293 LT797349 LT79742166 Troglocharinus ferreri ferreri IBE-VR112 AvencdelesBanderes x calcite V.Rizzo&J.Comas 2011 Esquerrà+Banderes 404079 4575189 LT79725867 Troglocharinus ferreri ferreri IBE-VR113 AvencdelesBanderes calcite V.Rizzo&J.Comas 2011 Esquerrà+Banderes 404079 4575189 LT79718868 Troglocharinus ferreri ferreri IBE-VR114 AvencdelesBanderes calcite V.Rizzo&J.Comas 2011 Esquerrà+Banderes 404079 4575189 LT797250 LT79734869 Troglocharinus ferreri ferreri IBE-VR115 AvencdelesBanderes x calcite V.Rizzo&J.Comas 2011 Esquerrà+Banderes 404079 4575189 LT797249 LT797347 LT79738270 Troglocharinus ferreri ferreri IBE-PB12 AvencdeSanCristofol x calcite J.Pastor 2012 4087109 45716595 LT79726471 Troglocharinus ferreri ferreri IBE-PB13 AvencdeSanCristofol calcite J.Pastor 2012 4087109 45716595 LT797265 LT79942672 Troglocharinus ferreri ferreri IBE-PB14 AvencdeSanCristofol x x calcite J.Pastor 2012 4087109 45716595 LT79726673 Troglocharinus ferreri ferreri IBE-PB15 AvencdeSanCristofol x x calcite J.Pastor 2012 4087109 45716595 LT79719174 Troglocharinus ferreri ferreri IBE-RA1014 AvencdeSanMarçal calcite J.Pastor 2012 405355 4578350 LT79731775 Troglocharinus ferreri ferreri IBE-RA1160 AvencdeSanMarçal x calcite J.Pastor 2012 405355 4578350 LT79719276 Troglocharinus ferreri ferreri IBE-RA1162 AvencdeSanMarçal calcite J.Pastor 2012 405355 4578350 LT79721277 Troglocharinus ferreri ferreri IBE-RA1163 AvencdeSanMarçal x calcite J.Pastor 2012 405355 4578350 LT797205 LT79741378 Troglocharinus ferreri ferreri IBE-VR90 AvencdelBufi x calcite J.Pastor 2011 409241 4572260 LT79718479 Troglocharinus ferreri ferreri IBE-VR91 AvencdelBufi x calcite J.Pastor 2011 409241 4572260 LT79729980 Troglocharinus ferreri ferreri IBE-VR92 AvencdelBufi x x calcite J.Pastor 2011 409241 4572260 LT797263 LT79742081 Troglocharinus ferreri ferreri IBE-VR26 AvencdelCorralNou x x calcite V.Rizzo 2010 CorralNou+SerranoArbonés 404164 4573240 LT797246 HF912477 HF912631 HF912564 HF912530 HF912588 HF912588 LT797346 LT79740482 Troglocharinus ferreri ferreri IBE-VR27 AvencdelCorralNou x calcite V.Rizzo 2010 CorralNou+SerranoArbonés 404164 4573240 LT797245 LT797307 LT79732983 Troglocharinus ferreri ferreri IBE-VR29 AvencdelCorralNou calcite V.Rizzo 2010 CorralNou+SerranoArbonés 404164 4573240 LT79736284 Troglocharinus ferreri ferreri IBE-VR118 AvencdelPladelesComes x calcite V.Rizzo&J.Comas 2011 408588 4583479 LT79722885 Troglocharinus ferreri ferreri IBE-VR119 AvencdelPladelesComes x x calcite V.Rizzo&J.Comas 2011 408588 4583479 LT797275 LT797354 LT79740586 Troglocharinus ferreri ferreri IBE-VR120 AvencdelPladelesComes calcite V.Rizzo&J.Comas 2011 408588 4583479 LT797241 LT797344 LT79740387 Troglocharinus ferreri ferreri IBE-VR121 AvencdelPladelesComes calcite V.Rizzo&J.Comas 2011 408588 4583479 LT797227 LT797364 LT79742388 Troglocharinus ferreri ferreri IBE-VR41 AvencdelPuigmoltò x Jurassicdolostone V.Rizzo&J.Pastor 2010 Cabeza+SanRoc+Puigmoltò 409884 4574286 LT797267 LT797350 LT79739189 Troglocharinus ferreri ferreri IBE-VR42 AvencdelPuigmoltò x Jurassicdolostone V.Rizzo&J.Pastor 2010 Cabeza+SanRoc+Puigmoltò 409884 4574286 LT797175 LT797332 LT79738990 Troglocharinus ferreri ferreri IBE-PB1 AvencdelSogre x calcite J.Pastor 2012 4025889 45756006 LT79728191 Troglocharinus ferreri ferreri IBE-PB11 AvencdelSogre x calcite J.Pastor 2012 4025889 45756006 LT79725192 Troglocharinus ferreri ferreri IBE-RA1106 AvencdelsGeolegs x Jurassicdolostone J.Pastor 2013 4061138 45812030 LT797204 LT79741293 Troglocharinus ferreri ferreri IBE-RA1107 AvencdelsGeolegs x x Jurassicdolostone J.Pastor 2013 4061138 45812030 LT797196 LT797320 LT79738394 Troglocharinus ferreri ferreri IBE-VR30 AvencdelsGuerrillers x Jurassicdolostone V.Rizzo 2010 Guerrillers+Pepi 412114 4572180 LT797173 HF912481 HF912633 HF912566 HF912532 HF912589 LT797331 LT79738895 Troglocharinus ferreri ferreri IBE-VR31 AvencdelsGuerrillers x Jurassicdolostone V.Rizzo 2010 Guerrillers+Pepi 412114 4572180 LT797174 HF91248296 Troglocharinus ferreri ferreri IBE-PB6 AvencdelsTrons x calcite J.Pastor 2013 409177 4571473 LT79718697 Troglocharinus ferreri ferreri IBE-PB7 AvencdelsTrons x calcite J.Pastor 2013 409177 4571473 LT79728898 Troglocharinus ferreri ferreri IBE-PB8 AvencdelsTrons x calcite J.Pastor 2013 409177 4571473 LT79728999 Troglocharinus ferreri ferreri IBE-RA1110 AvencdelsTrons calcite J.Pastor 2013 409177 4571473 LT797304100 Troglocharinus ferreri ferreri IBE-RA1111 AvencdelsTrons x calcite J.Pastor 2013 409177 4571473 LT797290101 Troglocharinus ferreri ferreri IBE-VR117 AvencEsteles x calcite V.Rizzo&J.Comas 2011 Verdiell+Bardissa+Esteles 402417 4581000 LT797203 LT797371102 Troglocharinus ferreri ferreri IBE-VR126 AvencHannaSvacbroc x calcite J.Pastor 2011 409295 4571435 LT797305103 Troglocharinus ferreri ferreri IBE-VR128 AvencHannaSvacbroc x x calcite J.Pastor 2011 409295 4571435 LT797287 LT797357 LT797394104 Troglocharinus ferreri ferreri IBE-VR78 AvencInfern calcite V.Rizzo 2011 409241 4572260 LT797181105 Troglocharinus ferreri ferreri IBE-VR79 AvencInfern calcite V.Rizzo 2011 409241 4572260 LT797179 LT797334 LT797419106 Troglocharinus ferreri ferreri IBE-VR80 AvencInfern x calcite V.Rizzo 2011 409241 4572260 LT797182107 Troglocharinus ferreri ferreri IBE-VR82 AvencInfern x calcite V.Rizzo 2011 409241 4572260 LT797183108 Troglocharinus ferreri ferreri IBE-VR54 AvencJoanCabeza x Jurassicdolostone V.Rizzo 2010 Cabeza+SanRoc+Puigmoltò 409466 4574236 LT797268 LT797351 LT797392109 Troglocharinus ferreri ferreri IBE-VR55 AvencJoanCabeza x Jurassicdolostone V.Rizzo 2010 Cabeza+SanRoc+Puigmoltò 409466 4574236 LT797270110 Troglocharinus ferreri ferreri IBE-VR56 AvencMarcel Triassicdolostone V.Rizzo 2010 411126 4578561 LT797200 LT797339 LT797401111 Troglocharinus ferreri ferreri IBE-VR57 AvencMarcel Triassicdolostone V.Rizzo 2010 411126 4578561 LT797201 LT797340 LT797402

112 Troglocharinus ferreri ferreri IBE-VR58 AvencMarcel x Triassicdolostone V.Rizzo 2010 411126 4578561 LT797214113 Troglocharinus ferreri ferreri IBE-VR59 AvencMarcel x Triassicdolostone V.Rizzo 2010 411126 4578561 LT797215114 Troglocharinus ferreri ferreri IBE-VR48 AvencMorganiComas x calcite V.Rizzo 2010 Morgan+Pleta 409006 4569728 LT797247 LT797326 LT797384115 Troglocharinus ferreri ferreri IBE-PB20 AvencPompeuFabra calcite J.Pastor 2013 409741 4579499 LT797209116 Troglocharinus ferreri ferreri IBE-PB21 AvencPompeuFabra x calcite J.Pastor 2013 409741 4579499 LT797240117 Troglocharinus ferreri ferreri IBE-RA1104 AvencPompeuFabra x calcite J.Pastor 2013 409741 4579499 LT797210118 Troglocharinus ferreri ferreri IBE-RA1105 AvencPompeuFabra x x calcite J.Pastor 2012 409741 4579499 LT797197 LT797337 LT797399119 Troglocharinus ferreri ferreri IBE-PB4 AvencSerrano_Arbonés calcite J.Pastor 2012 CorralNou+SerranoArbonés 405608 4573168 LT797243120 Troglocharinus ferreri ferreri IBE-PB5 AvencSerrano_Arbonés x calcite J.Pastor 2012 CorralNou+SerranoArbonés 405608 4573168 LT797244121 Troglocharinus ferreri ferreri IBE-PB23 CovadeCanRigol1 Jurassicdolostone J.Pastor 2013 Rigol1+Rigol2 4079740 45844133 LT797222122 Troglocharinus ferreri ferreri IBE-RA1093 CovadeCanRigol1 x Jurassicdolostone J.Pastor 2013 Rigol1+Rigol2 4079740 45844133 LT797225123 Troglocharinus ferreri ferreri IBE-RA1094 CovadeCanRigol1 Jurassicdolostone J.Pastor 2013 Rigol1+Rigol2 4079740 45844133 LT797226124 Troglocharinus ferreri ferreri IBE-PB24 CovadeCanRigol2 Jurassicdolostone J.Pastor 2013 Rigol1+Rigol2 4080340 45954133 LT797223125 Troglocharinus ferreri ferreri IBE-RA1097 CovadeCanRigol2 x Jurassicdolostone J.Pastor 2013 Rigol1+Rigol2 4080340 45954133 LT797220 LT797342 LT797417126 Troglocharinus ferreri ferreri IBE-RA1098 CovadeCanRigol2 x x Jurassicdolostone J.Pastor 2013 Rigol1+Rigol2 4080340 45954133 LT797195 LT797411127 Troglocharinus ferreri ferreri IBE-VR33 CovadeCanVerdaguer x Jurassicdolostone V.Rizzo 2010 409056 4583183 LT797297 HF912484 LT797359 LT797406128 Troglocharinus ferreri ferreri IBE-VR34 CovadeCanVerdaguer Jurassicdolostone V.Rizzo 2010 409056 4583183 LT797297 LT797363 LT797409129 Troglocharinus ferreri ferreri IBE-VR35 CovadeCanVerdaguer Jurassicdolostone V.Rizzo 2010 409056 4583183 LT797298 HF912486 LT797360 LT797407130 Troglocharinus ferreri pallaresi IBE-RA1095 AvencdeCanMontmany x Triassicdolostone J.Pastor 2013 413413 4585701 LT797187 LT799428 LT797387131 Troglocharinus ferreri pallaresi IBE-RA1096 AvencdeCanMontmany x Triassicdolostone J.Pastor 2013 413413 4585701 LT797274132 Troglocharinus ferreri pallaresi IBE-VR3 AvencdeCanMontmany x Triassicdolostone J.Comas 2009 413413 4585701 HG915334 HF912480 HF912632 HF912565 HF912531 HF912601 LT797361133 Troglocharinus elongatus elongatus IBE-VR1 AvencdelaPlanaAncosa x outgroup(Coast) J.Comas 2009 LT797301 HF912462 HF912619 HF912553 HF912516 LT797437134 Troglocharinus elongatus mateui IBE-VR4 CovadelGarrofet outgroup(Coast) J.Comas 2009 LT797314135 Troglocharinus elongatus ollai IBE-VR17 CovadelaOlla x outgroup(Coast) V.Rizzo&J.Comas 2010 LT799422 HF912470 HF912627 HF912561 HF912524136 Troglocharinus elongatus portai IBE-VR24 MinesdePontons1 outgroup(Coast) V.Rizzo&J.Comas 2010 LT797302137 Troglocharinus elongatus pyniareti IBE-VR38 AvencdelsPinyarets x outgroup(Coast) J.M.Victoria 2010 LT797284 HF912489 HF912636 HF912569 HF912535 LT797442138 Troglocharinus elongatus IBE-AF115 CovadelaSensada x outgroup(Coast) J.Comas&LluísAuroux 2009 HF912454 HF912606 HF912543 HF912504 HF912577 LT797438139 Troglocharinus elongatus IBE-VR7 CovadelPas outgroup(Coast) V.Rizzo&J.Comas 2009 LT797315140 Troglocharinus elongatus IBE-RA1176 MinasdePontons x outgroup(Coast) P.Balart 2014 LT797303 LT797308141 Troglocharinus espanoli IBE-VR23 CovadelTraça x outgroup(Coast) F.Fadrique 2010 LT797300 HF912475 HF912630 HF912563 HF912529 LT797443 LT797369142 Troglocharinus fonti IBE-AF181 AvencPladelafornesa x outgroup(Pyrenees) C.Bourdeau&A.Faille 2009 HF912458 LT797429 LT797434 LT797445143 Troglocharinus fonti NHM-IRC26 Covad'Ormini x outgroup(Pyrenees) J.Fresneda 2001 GU356902 GU356801 GU356848 HF912585 GU356993144 Troglocharinus fonti schuettei IBE-VR37 CovadelaFontMentidora x outgroup(Pyrenees) J.Fresneda,I.Ribera&V.Rizzo 2010 HF912488 HF912534145 Troglocharinus hustachei IBE-VR36 ForatdelGel x outgroup(Pyrenees) A.Meseguer 2010 HF912487 HF912635 HF912568 HF912533 LT797435 LT797367146 Troglocharinus impellitierii NHM-IRC22 CovaPalomera outgroup(Pyrenees) J.Fresneda 2001 HF912501 HF912617 HF912514 HF912599147 Troglocharinus jacasi IBE-VR2 CovadelesRondes outgroup(Coast) J.M.Victoria 2009 HF912629 HF912526 LT797323148 Troglocharinus jacasi IBE-VR22 CovadelesRondes outgroup(Coast) V.Rizzo&J.Comas 2010 LT799423 HF912474 HF912528149 Troglocharinus kiesenwetteri andresi IBE-VR13 AvencdelClast outgroup(Coast) V.Rizzo&J.Comas 2010 LT799421 HF912466 HF912623 HF912557 HF912520150 Troglocharinus kiesenwetteri sanllorensi IBE-RA899 AvencQuimSolves x outgroup(Coast) J.Pastor 2013 LT797311 LT797425 LT797313 LT797446151 Troglocharinus kiesenwetteri sanllorensi IBE-VR11 CovesdeMura x outgroup(Coast) V.Rizzo 2010 LT797306 HF912464 HF912621 HF912555 HF912518 HF912587152 Troglocharinus olerdolai IBE-VR102 AvencdeOlerdola x outgroup(Coast) V.Rizzo&J.Comas 2011 LT797316153 Troglocharinus orcinus acevedoi IBE-VR20 Covad'enCartanyà x outgroup(Coast) F.Fadrique 2010 HF912472 LT797428 HF912527 LT797433 LT797436 LT797322154 Troglocharinus orcinus acevedoi IBE-VR10 CovadeCanMasiet x outgroup(Coast) V.Rizzo&F.Fadrique 2010 LT797283 HF912463 HF912620 HF912554 HF912517 HF912586155 Troglocharinus orcinus orcinus IBE-VR9 CovadelaFebrò x outgroup(Coast) V.Rizzo 2009 HG915332 HF912495 HF912642 HF912575 HF912540 HF912590 LT797444 LT797325156 Troglocharinus orcinus orcinus NHM-IRC42 CovadelaFebrò x outgroup(Coast) F.Fadrique 2002 HF912502 HF912618 HF912552 HF912515 HF912600157 Troglocharinus pallisei IBE-RA1190 CovadelaRibadelTalc x outgroup(Coast) J.Pallisé 2014 LT797282 LT797309 LT797426 LT797431 LT797440158 Troglocharinus pallisei IBE-RA1191 CovadelaRibadelTalc x outgroup(Pyrenees) J.Fresneda&V.Rizzo 2013 LT797296 LT797310 LT797427 LT797432 LT797441159 Troglocharinus patracoi IBE-VR16 AvencdelPetrecò x outgroup(Coast) V.Rizzo&J.Comas 2010 HF912469 HF912626 HF912560 HF912523 LT797324160 Troglocharinus quadricollis MNCN-AI578 GralleradeCastellet x outgroup(Pyrenees) J.Fresneda 2003 HF912497 HF912613 HF912548 HF912510 HF912595161 Troglocharinus senenti IBE-VR6 QuerantGrandePaús x outgroup(Pyrenees) J.Fresneda 2002 HF912493 HF912640 HF912573 HF912538 HF912603 LT797368 LT797416162 Troglocharinus senenti MNCN-AI585 QuerantGrandePaús x outgroup(Pyrenees) J.Fresneda 2002 HF912498 HF912614 HF912549 HF912511 HF912596163 Troglocharinus subilsi IBE-RA1189 GralleradeCambrils x outgroup(Pyrenees) J.Fresneda&V.Rizzo 2013 LT797424 LT799425 LT797430 LT797439

3

Appendix S2 Additional methods:

(a) Location of all known caves in the Garraf massif (yellow circles), including those in

which the presence of Troglocharinus ferreri has been recorded (red circles).

4

(b) Primers used in the study

Gene Name Sense Sequence Reference cox1 Jerry (M202) F CAACATTTATTTTGATTTTTTGG 5

Pat (M70) R TCCA(A)TGCACTAATCTGCCATATTA 5

Chy F T(A/T)GTAGCCCA(T/C)TTTCATTA(T/C)GT 3

Tom R AC(A/G)TAATGAAA(A/G)TGGGCTAC(T/A)A 3

Tom-2 R A(A/G)GGGAATCATTGAATAAA(A/T)CC 3

cob CB3 F GAGGAGCAACTGTAATTACTAA 1 CB4 R AAAAGAAA(AG)TATCATTCAGGTTGAAT 1 rrnL-nad1 16saR (M14) F CGCCTGTTTA(A/T)CAAAAACAT 5

16Sa R ATGTTTTTGTTAAACAGGCG 5

16Sb R CCGGTCTGAACTCAGATCATGT 5

16SAlf1 R GCATCACAAAAAGGCTGAGG 6

ND1A (M223) R GGTCCCTTACGAATTTGAATATATCCT 5

SSU 5' F GACAACCTGGTTGATCCTGCCAGT 4

b5.0 R TAACCGCAACAACTTTAAT 4

LSU Ka F ACACGGACCAAGGAGTCTAGCATG 3

Kb R CGTCCTGCTGTCTTAAGTTAC 3

H3 H3aF F ATGGCTCGTACCAAGCAGAC(AG)CGC 2 H3aR R ATATCCTT(AG)GGCAT(AG)AT(AG)GTGAC 2 Wg Wg550F F ATGCGTCAGGARTGYAARTGYCAYGGYATGTC 7 WgAbRZ R CACTTNACYTCRCARCACCARTG 7 References:

1. Barraclough, T.G., Hogan, J.E. & Vogler, A.P. (1999) Testing whether

ecological factors promote cladogenesis in a group of tiger beetles (Coleoptera:

Cicindelidae). Proceedings of the Royal Society, Series B, Biological Sciences,

266, 1061–1067.

2. Colgan, D. J., McLauchlan, A., Wilson, G. D. F., Livingston, S. P., Edgecombe,

G. D., Macaranas, J., Cassis, G. & Gray, M. R. (1998) Histone H3 and U2

snRNA DNA sequences and arthropod molecular evolution. Australian Journal

of Zoology, 46(5), 419-437.

3. Ribera, I., Fresneda, J., Bucur, R., Izquierdo, A., Vogler, A.P., Salgado, J.M. &

Cieslak, A. (2010) Ancient origin of a western Mediterranean radiation of

subterranean beetles. BMC Evolutionary Biology, 10, 29.

4. 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. Systematic Biology, 50, 945–969.

5. Simon, C., Frati, F., Beckenbach, A., Crespi, B., Liu, H. & Flook, P. (1994)

5

Evolution, weighting, and phylogenetic utility of mitochondrial gene-sequences

and a compilation of conserved polymerase chain-reaction primers. Annals of

the Entomological Society of America, 87, 651–701.

6. Vogler, A.P., DeSalle, R., Assmann, T., Knisley, C.B. & Schultz, T.D. (1993)

Molecular population genetics of the endangered tiger beetle Cicindela dorsalis

(Coleoptera, Cicindelidae). Annals of the Entomological Society of America, 86,

142–152.

7. Wild, A. & Maddison, D.R. (2008) Evaluating nuclear-protein coding genes for

phylogenetic utility in beetles. Molecular Phylogenetics and Evolution, 48, 877-

891.

AppendixS2(c)CategorisationofthedifferentgeologicalsubstrataaccordingtothegeologicalmapoftheGarrafMassif(ICC2010),andresistancevaluesgiventothemainlythologicaltypesinthecorrelationbetweenphylogeneticdistanceandweightedgeographicaldistances.

EPIGRAF Original description Lythologic category 1 2 3 4 5CAlc Calcáries bioconstruídes i calcarenites. Albiá. Biogenic calcite 1 1 1 1 1CKa Calcarenites, calcáries bioclástiques, margues. Aptiá superior-Cenomaniá. calcite with sandstone, bioclastic calcite, margue 1 1 1 1 1

CVBcd Calcáries amb intercalacions dolomítiques. Valanginiá-Barremiá. calcite with dolomita intercalations 1 1 1 1 1NMcl Calcáries biomicritiques. Tortoniá. biomicritic calcite 1 1 1 1 1CAld Dolomies. Aptiá superior - Albiá. dolomite 1 2 5 20 200CAm Margues i margocalcáries. Aptiá. margue-calcite 1 2 5 20 200

Dc Calcosquists i calcáries argiloses. Devoniá mitjá. sandstone and calcite-schist 1 2 5 20 200Jd Dolomies i calcáries. Jurássic - Cretaci inferior. dolostone and calcite 1 2 5 20 200

NMgx Guixos. Tortoniá. gypsum 1 2 5 20 200PPEc Calcáries i dolomies. Ilerdiá. calcite and dolomite 1 2 5 20 200Tm1 Calcáries micrítiques i dolomies. Triásic mitjá-superior. micritic calcite and dolomite 1 2 5 20 200Tm2 Gresos i argiles. Fácies Muschelkalk mitjá. Triásic mitjá. gypsum and sandstone 1 2 5 20 200Tm3 Dolomies i calcáries. Fácies Muschelkalk superior. Triásic mitjá-superior. dolomite and calcite 1 2 5 20 200

ÃOrp Pissarres micacítiques i pissarres sorrenques. Cambroordoviciá o Ordoviciá. metamorphic rocks with no or few carbonated constituents 1 3 10 60 600ÃOrq Quarsites i pissarres quarsitoses. Cambroordoviciá o Ordoviciá. metamorphic rocks with no or few carbonated constituents 1 3 10 60 600Capc Pissarres, calcáries i calcosquists. Carbonífer. metamorphic rocks with no or few carbonated constituents 1 3 10 60 600Capl Lidites i pissarres silíciques. Carbonífer. metamorphic rocks with no or few carbonated constituents 1 3 10 60 600

Fd Dics de possible diábasi. Carbonífer-Permiá. metamorphic rocks with no or few carbonated constituents 1 3 10 60 600Ggd Granodiorites i granits alcalins. Carbonífer-Permiá. metamorphic rocks with no or few carbonated constituents 1 3 10 60 600NMc Conglomerats massius cimentats. Burdigaliá inferior. sedimentary rocks with no or few carbonated constituents 1 3 10 60 600

NMca Conglomerats amb matriu argilosa sense cimentar. Aragoniá superior - Vallesiá. sedimentary rocks with no or few carbonated constituents 1 3 10 60 600NMcv Conglomerats amb matriu argilosa. Aquitaniá-Burdigaliá. sedimentary rocks with no or few carbonated constituents 1 3 10 60 600NMe Calcarenites esculloses, biomicrites i biorudites. Serraval-liá-Tortoniá. sedimentary rocks with no or few carbonated constituents 1 3 10 60 600NMg Gresos de gra gros localment amb conglomerats. Serraval-liá-Tortoniá. sedimentary rocks with no or few carbonated constituents 1 3 10 60 600

NMm Margues amb intercalacions de calcáries. Tortoniá superior. sedimentary rocks with no or few carbonated constituents 1 3 10 60 600NMs Sorres argiloses de gra mitjá. Serraval-liá-Tortoniá. sedimentary rocks with no or few carbonated constituents 1 3 10 60 600NPa Argiles molt plástiques amb restes de fauna. Pliocé. sedimentary rocks with no or few carbonated constituents 1 3 10 60 600NPc Conglomerats amb matriu sorrenca poc cimentats. Pliocé. sedimentary rocks with no or few carbonated constituents 1 3 10 60 600

PEma Margues i argiles alternant amb calcáries. Cuisiá-Luteciá. sedimentary rocks with no or few carbonated constituents 1 3 10 60 600PPa Argiles vermelles i gresos. dipósits de bauxita. Paleocé. sedimentary rocks with no or few carbonated constituents 1 3 10 60 600SDc Calcáries noduloses i pissarres sericítiques. Siluriá - Devoniá inferior. sedimentary rocks with no or few carbonated constituents 1 3 10 60 600

Tk Margues i calcáries margoses. Fácies Keuper. Triásic superior. sedimentary rocks with no or few carbonated constituents 1 3 10 60 600Ll_Qpa Plana al-luvial i/o deltaica del Llobregat. Holocé. metamorphic or sedimentary siliceous rocks 1 4 15 80 800Ll_Qt1 Terrassa del Llobregat i afluents. Es troba 2 m sobre el nivell del riu. Holocé. metamorphic or sedimentary siliceous rocks 1 4 15 80 800Ll_Qt2 Terrassa del Llobregat i afluents. Es troba 3 m sobre el nivell del riu. Plistocé terminal - Holocé basal. metamorphic or sedimentary siliceous rocks 1 4 15 80 800Ll_Qt3 Terrassa del Llobregat i afluents. Es troba entre 15 i 20 m sobre el nivell del riu. Plistocé superior. metamorphic or sedimentary siliceous rocks 1 4 15 80 800Ll_Qt4 Terrassa del Llobregat. Es troba 40 i 50 m sobre el nivell del riu. Plistocé. metamorphic or sedimentary siliceous rocks 1 4 15 80 800

MAR Mar sea 1 4 15 80 800

NMag Argiles, gresos i conglomerats. Serraval-liá-Vallesiá. metamorphic or sedimentary siliceous rocks 1 4 15 80 800NMal Argiles fossilíferes i llims. Burdigaliá - Serraval-liá inferior. metamorphic or sedimentary siliceous rocks 1 4 15 80 800NMas Argiles blaves molt plástiques i sorres. Serraval-liá-Tortoniá. metamorphic or sedimentary siliceous rocks 1 4 15 80 800

NPs Sorres i argiles sorrenques. Pliocé. metamorphic or sedimentary siliceous rocks 1 4 15 80 800Q Sediments recents de fons de valls, rieres i peu de mont. Holocé. metamorphic or sedimentary siliceous rocks 1 4 15 80 800

Qbcn Plana al-luvial del Pla de Barcelona. Plistocé. metamorphic or sedimentary siliceous rocks 1 4 15 80 800Qc Crostes de calitx. Plistocé. metamorphic or sedimentary siliceous rocks 1 4 15 80 800Qd Cordons de dunes litorals. Holocé. metamorphic or sedimentary siliceous rocks 1 4 15 80 800Qg Peu de mont). Plistocé. metamorphic or sedimentary siliceous rocks 1 4 15 80 800Qp Sediments de platja. Holocé superior. metamorphic or sedimentary siliceous rocks 1 4 15 80 800

Qpa Plana al-luvial. Graves, sorres i lutites. Holocé superior. metamorphic or sedimentary siliceous rocks 1 4 15 80 800Qpa2-3 Plana al-luvial correlacionable amb les terrasses fluvials 2 i 3. Plistocé. metamorphic or sedimentary siliceous rocks 1 4 15 80 800

Qpc Argiles i sorres palustres. Plistocé. metamorphic or sedimentary siliceous rocks 1 4 15 80 800Qr Dipòsits dels llits actuals de les rieres i dels torrents. Holocé. metamorphic or sedimentary siliceous rocks 1 4 15 80 800

Qt0-1 Llit actual, plana d'inundació ordinária i terrassa mÚs baixa (0-2 m). Holocé recent. metamorphic or sedimentary siliceous rocks 1 4 15 80 800Qt1 Terrassa fluvial. Graves, sorres i lutites. Holocé. metamorphic or sedimentary siliceous rocks 1 4 15 80 800Qt2 Terrassa fluvial. Graves, sorres i lutites. Plistocé terminal - Holocé basal. metamorphic or sedimentary siliceous rocks 1 4 15 80 800

Qt2-3 Terrassa fluvial que engloba les terrasses 2 i 3. Plistocé. metamorphic or sedimentary siliceous rocks 1 4 15 80 800Qt3 Terrassa fluvial. Graves, sorres i lutites. Plistocé superior. metamorphic or sedimentary siliceous rocks 1 4 15 80 800Qv Ventalls al-luvials antics. Plistocé. metamorphic or sedimentary siliceous rocks 1 4 15 80 800

Qva1-3 Ventalls i plana al-luvial de la Riera de les Arenes. Plistocé. metamorphic or sedimentary siliceous rocks 1 4 15 80 800Sf Pissarres ampelítiques, fil-lites i sericites. Siluriá. metamorphic or sedimentary siliceous rocks 1 4 15 80 800

Tbc Conglomerats silícics basals. Fácies Buntsandstein. Triásic inferior. metamorphic or sedimentary siliceous rocks 1 4 15 80 800Tbg Alternanéa de gresos silícics i argiles. Fácies Buntsandstein. Triásic inferior. metamorphic or sedimentary siliceous rocks 1 4 15 80 800

Troglocharinus ferreri, family Leiodidae)

Valeria Rizzo, David Sánchez-Fernández, Rocio Alonso, Josep Pastor and Ignacio Ribera

Fig. S2

A

0 - 24

24 - 78

78 - 148

148 - 343

343 - 1,329

T. ferreri

(Troglocharinus ferreri, family Leiodidae)

Valeria Rizzo, David Sánchez-Fernández, Rocio Alonso, Josep Pastor and Ignacio Ribera

Fig. S2

High connnectivity

Low connectivity

Substratum karsti�cability, dispersal and genetic structure in a strictly subterranean beetleValeria Rizzo, David Sánchez-Fernández, Rocio Alonso, Josep Pastor and Ignacio Ribera

Appendix S2 (d)

High connnectivity

Low connectivity

C

Low connectivity

High connectivity

D

Low connectivity

High connectivity

E

0 - 9

9 - 30

30 - 76

76 - 314

314 - 1,286High connectivity

Low connectivity

6

Appendix S3 Additional results:

(a) Best tree obtained in RAxML using all five mitochondrial and four nuclear markers.

Numbers in nodes, fast bootstrap support values (when >50%). See text and Appendix

S1a for details on the specimens included and the methods.

0.04

ferreri_RA1100_Roc

ferreri_VR54_Cabeza

kiesenwetteri_L_VR11

ferreri_VR112_Banderes

ferreri_RA1194_Gori

ferreri_VR60_Cuneta

ferreri_VR80_Infern

ferreri_VR113_Banderes

ferreri_PB1_Sogre

orcinus_VR10

ferreri_VR26_CorralNouferreri_VR29_CorralNou

ferreri_RA1163_Marcal

ferreri_VR57_Marcel

ferreri_VR58_Marcel

ferreri_VR68_Verdiell

ferreri_PB10_Roc

ferreri_RA1105_PompeuFabra

ferreri_VR51_Benjamin

ferreri_RA1106_Geolegs

ferreri_RA1095_Montmany

ferreri_VR106_Gori

ferreri_VR46_Llinia

ferreri_VR77_Vermell

ferreri_VR115_Banderes

ferreri_VR47_Llinia

espanoli_VR23

ferreri_RA1160_Marcal

ferreri_VR128_Svabroc

elongatus_RA1176

ferreri_VR131_Troneda

ferreri_PB22_Roca

ferreri_VR119_Comes

ferreri_VR70_Verdiell

ferreri_VR91_Bufi

ferreri_VR97_Pasan

ferreri_VR78_Infern

ferreri_PB3_Pere

ferreri_VR75_Vermell

ferreri_RA1192_Pasan

ferreri_VR111_Esquerra

elongatus_VR7

ferreri_VR45_Llinia

ferreri_VR104_Gori

ferreri_VR85_Pleta

ferreri_RA1097_Rigollferreri_VR33_Verdaguer

ferreri_PB12_Cristofol

ferreri_VR81_Infern

fonti_VR37

ferreri_VR69_Verdiell

ferreri_PB8_Trons

ferreri_VR44_Llinia

ferreri_VR76_Vermell

ferreri_VR133_Troneda

ferreri_RA1111_Trons

ferreri_PB20_PompeuFabra

ferreri_RA1093_Rigol

fonti_IRC26

ferreri_RA1107_Geolegs

ferreri_VR56_Marcel

ferreri_RA1096_Montmany

ferreri_VR124_Pepi

quadricollis_AI578

ferreri_PB15_Cristofol

ferreri_PB7_Trons

ferreri_VR30_Guerrillers

ferreri_PB13_Cristofol

ferreri_RA1158_Pepi

kiesenwetteri_L_RA899

ferreri_VR121_Comes

ferreri_RA1193_Pasan

ferreri_VR50_Benjamin

elongatus_VR4

ferreri_PB24_Rigoll

ferreri_VR108_Esquerra

ferreri_VR87_Pleta

elongatus_VR38

ferreri_RA1092_Roca

ferreri_RA1156_Pere

ferreri_RA1108_Sadurni

fonti_AF181

ferreri_VR35_Verdaguer

ferreri_VR52_Benjamin

hustachei_VR36

ferreri_VR74_Vermell

ferreri_VR82_Infern

ferreri_RA1094_Rigol

ferreri_VR114_Banderes

ferreri_VR120_Comes

elongatus_VR24

ferreri_RA1112_Alzines

ferreri_PB23_Rigol

ferreri_RA1104_PompeuFabra

ferreri_VR132_Troneda

ferreri_VR90_Bufi

ferreri_RA1099_Roc

pallisei_RA1190

ferreri_VR43_Llinia

ferreri_RA1109_Sadurni

ferreri_VR3_Montmany

ferreri_AF172_Bardissa

ferreri_PB6_Trons

ferreri_PB21_PompeuFabra

ferreri_VR73_Vermell

ferreri_PB2_Pere

ferreri_VR59_Marcel

ferreri_VR42_Puigmolto

senenti_VR6

orcinus_VR9

ferreri_PB14_Cristofol

ferreri_VR110_Esquerra

ferreri_RA748_AlzinesGran

ferreri_VR49_Benjamin

senenti_AI585

orcinus_VR20

ferreri_VR79_Infern

ferreri_RA1110_Trons

ferreri_VR53_Benjamin

ferreri_VR130_Troneda

ferreri_VR31_Guerrillers

impellitieri_IRC22

ferreri_VR86_Pleta

ferreri_VR41_Puigmolto

ferreri_AI1065_Roc

ferreri_VR83_Pleta

ferreri_VR72_Verdiell

ferreri_PB4_Serrano

ferreri_VR62_Cuneta

ferreri_VR103_Gori

ferreri_VR126_Svabroc

ferreri_PB5_Serrano

pallisei_RA1191

ferreri_PB11_Sogre

ferreri_VR92_Bufi

patra_VR16

ferreri_VR55_Cabeza

ferreri_VR27_CorralNou

ferreri_VR95_Pasan

ferreri_VR34_Verdaguer

ferreri_VR117_Esteles

ferreri_VR48_Morgan

orcinus_IRC42

ferreri_PB9_Roc

ferreri_RA1162_Marcal

elongatus_VR1

ferreri_RA1113_Alzines

ferreri_RA1098_Rigoll

ferreri_RA1091_Roca

ferreri_VR118_Comes

ferreri_RA1159_Pepi

ferreri_VR107_Esquerra

ferreri_VR61_Cuneta

ferreri_VR105_Gori

ferreri_VR71_Verdiell

61

63

64

86

49

59

71

100

100

98

94

51

60

92

71

76

7972

77

88

95

97

82

67

90

57

99

99

51

59

66

79

100

56

66

75

60

51

69

81

65

60

88

73

69

100

100

93

99

73

63

100

95

76

80

92

53

62

100

62

78

60

100

10068

100

61

97

63

98

94

58

94

60

75

64

56

Troglocharinus ferreri

clade 1

clade 2

T. ferreri pallaresi

7

(b) Best tree obtained in RAxML using only the four nuclear markers for a reduced

representation of T. ferreri and outgroups. Numbers in nodes, fast bootstrap support

values (when >50%). Se text and Appendix S1a for details on the specimens included

and the methods.

0.0060

ferreri_VR48_Morgan

ferreri_RA1108_Sadurni

ferreri_RA1096_Montmany

ferreri_VR61_Cunetaferreri_RA1097_Rigoll

ferreri_VR117_Esteles

ferreri_VR114_Banderes

ferreri_VR132_Troneda

ferreri_VR51_Benjamin

ferreri_VR124_Pepi

ferreri_VR44_Llinia

ferreri_VR3_Montmany

ferreri_VR133_Troneda

ferreri_VR120_Comes

ferreri_VR42_Puigmolto

ferreri_VR34_Verdaguer

ferreri_VR128_Svabroc

ferreri_VR69_Verdiell

ferreri_RA1014_Marcal

patra_VR16

ferreri_VR56_Marcel

ferreri_VR30_Guerrillers

ferreri_VR54_Cabeza

ferreri_AF172_Bardissa

ferreri_VR57_Marcel

elongatus_VR4

ferreri_RA1098_Rigoll

ferreri_VR27_CorralNou

ferreri_VR73_Vermell

ferreri_VR35_Verdaguer

ferreri_RA1107_Geolegs

ferreri_PB10_Roc

ferreri_RA1105_PompeuFabraferreri_VR50_Benjamin

espanoli_VR23

olerd_VR102

orcinus_VR9

ferreri_VR92_Bufi

orcinus_VR20

ferreri_VR115_Banderes

ferreri_VR104_Gori

ferreri_VR26_CorralNou

ferreri_RA1112_Alzines

senenti_VR6

jacasi_VR2

ferreri_VR33_Verdaguer

ferreri_VR106_Gori

ferreri_VR46_Llinia

ferreri_VR83_Pleta

ferreri_VR110_Esquerra

elongatus_VR7

ferreri_VR97_Pasan

ferreri_RA748_AlzinesGranferreri_VR71_Verdiell

ferreri_RA1106_Geolegs

ferreri_VR76_Vermell

ferreri_VR79_Infern

ferreri_VR41_Puigmolto

hustachei_VR36

ferreri_VR119_Comes

ferreri_AI1065_Roc

ferreri_VR29_CorralNou

ferreri_VR60_Cuneta

ferreri_RA1109_Sadurni

ferreri_RA1095_Montmany

ferreri_VR53_Benjamin

ferreri_RA1163_Marcal

ferreri_VR121_Comes

6987

58

83

87

50

58

91

71

8

(c) Majority rule consensus tree obtained in BEAST using all five mitochondrial and

four nuclear markers and a reduced representation of T. ferreri, using as tree priors a

Yule speciation. Numbers in nodes, posterior probabilities (when >0.5, in red) and

estimated age with 95% confidence interval in Million years (Ma) for selected nodes. Se

text and Appendix S1a for details on the specimens included and the methods.

9

00.511.522.533.5

pallisei_RA1191_L_1.195_41.320

ferreri_RA1107_Geolegs_1.878_41.377

orcinus_VR20_L_1.162_41.317

ferreri_RA1109_Sadurni_1.892_41.351

ferreri_VR26_CorralNou_1.854_41.303

quadricollis_AI578_L_0.824_42.254

elongatus_RA1176_L_1.531_41.409elongatus_VR7_L_1.722_41.589

ferreri_VR3_Montmany_1.964_41.418

patra_VR16_L_1.862_41.562

fonti_VR37_L_1.090_42.243

elongatus_VR38_L_1.488_41.354

senenti_AI585_L_1.025_42.019

ferreri_VR119_Comes_1.907_41.396

orcinus_IRC42_L_1.016_41.266

orcinus_VR10_L_1.129_41.299

ferreri_VR128_Svabroc_1.916_41.287

kiesenwetteri_VR11_L_1.976_41.687

ferreri_RA1105_PompeuFabra_1.920_41.359

espanoli_VR23_L_1.307_41.394

fonti_AF181_L_1.233_42.273

olerd_VR102_L_1.709_41.302

ferreri_VR54_Cabeza_1.917_41.313

ferreri_VR124_Pepi_1.928_41.289

ferreri_RA1013_Serrano_1.874_41.304

ferreri_VR110_Esquerra_1.839_41.323

ferreri_RA1156_Pere_1.908_41.299

ferreri_RA1014_Marcal_1.867_41.349ferreri_VR83_Pleta_1.913_41.272

ferreri_AI1065_Roc_1.920_41.313

elongatus_VR1_L_1.489_41.447

elongatus_AF115_L_1.549_41.519

subilsi_RA1189_L_1.195_41.320

jacasi_VR2_L_1.554_41.471

ferreri_PB15_Cristofol_1.910_41.295

senenti_VR6_L_1.025_42.019

fonti_IRC26_L_1.225_42.207

ferreri_VR76_Vermell_1.939_41.347

ferreri_PB14_Cristofol_1.910_41.295

kiesenwetteri_RA899_L_2.017_41.657

ferreri_RA1095_Montmany_1.964_41.418

ferreri_VR30_Guerrillers_1.950_41.296

orcinus_VR9_L_1.016_41.266

impellitieri_IRC22_L_1.176_42.286

ferreri_VR92_Bufi_1.917_41.295

ferreri_RA748_AlzinesGran_1.912_41.355

pallisei_RA1190_L_1.195_41.320

ferreri_VR61_Cuneta_1.882_41.361ferreri_VR71_Verdiell_1.861_41.384

hustachei_VR36_L_0.953_42.031

ferreri_RA1158_Pepi_1.928_41.289

ferreri_VR104_Gori_1.860_41.352

ferreri_RA1098_Rigoll_1.900_41.406

0.26

0.77

0.6

0.34

0.98

0.53

0.96

1

0.97

0.72

0.58

0.6

1

0.72

0.58

1

1

1

0.9

0.75

1

0.36

0.95

0.97

0.92

0.87

0.91

0.87

0.75

0.88

1

1

1

1

0.67

1 0.99

0.96

0.36

0.82

0.36

1

1

0.480.34

1

0.91

0.91

0.69

0.88

1

0.96

(d) Majority rule consensus tree obtained in BEAST using all five mitochondrial and

four nuclear markers and a reduced representation of T. ferreri, using as tree priors a

birth death with incomplete sampling. Numbers in nodes, posterior probabilities (when

>0.5). Se text and Appendix S1a for details on the specimens included and the methods.