Biological Journal of the Linnean Society, 2018, 123, 712–727. With 4 figures.

Phylogeography of the Chilean red cricket Cratomelus

armatus (Orthoptera: Anostostomatidae) reveals high
cryptic diversity in central Chile
FERMÍN M. ALFARO1,2†, CARLOS P. MUÑOZ-RAMÍREZ3,4†, ÁLVARO ZÚÑIGA-REINOSO5,

Downloaded from https://academic.oup.com/biolinnean/article-abstract/123/4/712/4925800 by guest on 23 July 2019

STEVEN A. TREWICK6 and MARCO A. MÉNDEZ5*
1
Laboratorio de Entomología Ecológica, Departamento de Biología, Facultad de Ciencias, Universidad
de La Serena, Casilla 554, La Serena, Chile
2
Instituto de Investigación Multidisciplinar en Ciencia y Tecnología, Universidad de La Serena, La
Serena, Chile
3
Facultad de Ciencias, Universidad Católica de la Santísima Concepción, Concepción, Chile
4
Centro de Investigación en Biodiversidad y Ambientes Sustentables (CIBAS), Universidad Católica de
la Santísima Concepción, Concepción, Chile
5
Laboratorio de Genética y Evolución, Facultad de Ciencias, Departamento de Ciencias Ecológicas,
Universidad de Chile, Santiago, Chile
6
Ecology Group, Massey University, Private Bag 11-222, Palmerston North, New Zealand

Received 28 November 2017; revised 2 February 2018; accepted for publication 2 February 2018

We analysed the phylogeographical history of the red cricket Cratomelus armatus (Orthoptera: Anostostomatidae)
from central and southern Chile using 248 mitochondrial DNA COI sequences. Phylogenetic analyses revealed mul-
tiple lineages that were highly structured geographically. The two main lineages (north and south) were parapatric,
with a contact zone at the latitude of Concepción (~36.6°S), and have an estimated divergence time of 2 Mya. Deep
divergence and a species delineation analysis suggest that these lineages should be considered as different species.
The north lineage exhibited four well-supported subclades whose divergence times occurred during the Largest
Patagonian Glaciation between 0.84 and 1.1 Mya. Signals of demographic expansion in southern areas indicate
a more recent history for the south lineage (southern Chile). A positive correlation between latitude and genetic
distances between populations suggests postglacial colonization of southern areas. Bayesian estimations of popula-
tion size over time placed a bottleneck at ~150 kya. Our results support a role for glaciations in shaping contrasting
patterns of genetic diversification in C. armatus. More intensive past glaciations may have promoted diversification
in central Chile, whereas subsequent glaciations, with stronger impacts in southern areas, could have constrained
diversification in southern Chile. We discuss the taxonomic implications of our findings and hypothesize a contrast-
ing role for glaciation on patterns of genetic diversification in central and southern Chile.

ADDITIONAL KEYWORDS:  cryptic diversity – glaciations – high genetic diversity – mitochondrial DNA –
multiple refugia.

INTRODUCTION et al., 2011). Phylogeographical research has revealed

that isolation and postglacial expansion from refugia
Pleistocene climatic oscillations had important conse-
are among the most important processes explaining
quences for species distributions, genetic diversity and
genetic patterns in the Northern Hemisphere (Avise,
the evolutionary history of taxa, particularly in tem-
2000; Hewitt, 2004). In contrast to the vast phylogeo-
perate regions (Avise, 2000; Hewitt, 2000, 2004; Sérsic
graphical research conducted in temperate regions of
North America and Europe, phylogeographical studies
*Corresponding author. E-mail: mmendez@uchile.cl on taxa from temperate regions of South America have
†
These authors contributed equally to this work. accumulated at a much lower pace (Beheregaray, 2008;

© 2018 The Linnean Society of London, Biological Journal of the Linnean Society, 2018, 123, 712–727 712
 PHYLOGEOGRAPHY OF CRATOMELUS ARMATUS  713

Turchetto-Zolet et al., 2013), limiting our understand- patterns of genetic diversity across extensive glaciated
ing of the processes driving biodiversity patterns in and unglaciated areas of southern South America for
this region. Predictably, Pleistocene glaciations and the first time in an insect species.
the Andean orogenesis have played important roles Cratomelus armatus is endemic to central and
in shaping biodiversity in southern South America southern Chile (from 34 to 45°S), living in environ-
(Ruzzante et al., 2006; Sérsic et al., 2011). However, ments from the coast to the Andes Range (Elgueta,
inferred evolutionary histories of taxa in this region Cammouseight & Carbonnel, 1999; Alfaro et al., 2015).
appear more complex than previously assumed, with This distribution encompasses the biogeographical
variation in phylogeographical patterns indicating subregions of Central Chile and Sub-Antartica and the
mixed responses to environmental change (Sérsic biogeographical provinces of Santiago, Maule and the

Downloaded from https://academic.oup.com/biolinnean/article-abstract/123/4/712/4925800 by guest on 23 July 2019

et al., 2011; Victoriano et al., 2012). Valdivian Forest (Morrone, 2006). This wide distribu-
A general pattern found in several phylogeographi- tion includes areas of southern Chile that were glaci-
cal studies conducted in central and southern Chile ated and not glaciated during the Pleistocene, making
is a decrease of genetic diversity towards higher lati- this species a suitable model to study the impact of
tudes (e.g. Rodríguez-Serrano, Cancino & Palma, 2006; glacial cycles. Cratomelus armatus are important in
Himes, Gallardo & Kenagy, 2008; Victoriano et al., forest trophic webs as predators of a range of small
2008). This pattern has supported the idea of glacial invertebrates and as prey for small to large carnivores
refugia in northern areas with more stable climatic (Jaksić et al., 1990; Jimenez et al., 1990). The species
conditions that allowed the persistence of taxa during is flightless and has been documented as highly terri-
the Last Glacial Maximum (LGM) and recolonization torial, with cannibalistic behaviour and fossorial hab-
and expansion into southern areas after glaciations its (Angulo, 2001) that may suggest high site fidelity
(Palma et al., 2005; Lessa, D’Elía & Pardiñas, 2010). and perhaps constrained vagility. The extensive geo-
Although this pattern has been observed in several graphical distribution across various environmental
taxa (see Sérsic et al., 2011), some studies have found conditions, some of which are considered old and cli-
patterns that are in contrast to this general trend and matically stable, and the potentially low to moderate
seem to reveal a more complex scenario. For instance, dispersal capability predict high phylogeographical
studies analysing patterns of genetic diversity in fish structure.
(Unmack et al., 2009), lizard (Vera-Escalona et al., Here, we used mitochondrial DNA (hereafter
2012) and freshwater crustacean (Xu et al., 2009) mtDNA) to investigate the phylogeographical his-
species have shown weak or no relationship between tory of C. armatus and to analyse patterns of lineage
genetic diversity and glacial impact. Indeed, some of diversification and genetic diversity across its entire
these studies (Xu et al., 2009) have also found high geographical distribution. We test the predictions of
genetic diversity in areas that were putatively covered high genetic diversity and structure in populations
by the ice sheet, suggesting a more complex scenario, occupying areas that are considered to have been eco-
with potential cryptic refugia within the hypothesized logically stable through the Pleistocene (northern and
boundaries of the ice sheet. west-central; Fig. 1) and low genetic diversity in areas
These contrasting patterns only emphasize that that were glaciated, as a consequence of postglacial
it is still too soon to make generalizations, and more colonization. As unglaciated areas are considered more
research needs to be conducted to gain a better under- stable, we also predict high demographic stability, and
standing of the phylogeographical complexity of the therefore genetic patterns consistent with constant
region. In addition, the strong taxonomic bias in phy- population size over time. Glaciated areas (south-
logeographical studies conducted in the region, which ern areas) are expected to show signals of population
have focused on terrestrial vertebrates and plants expansion and recent bottlenecks.
(Beheregaray, 2008; Sérsic et al., 2011; Turchetto-Zolet
et al., 2013), has limited a more general understand-
ing of the historical processes impacting diversity in MATERIAL AND METHODS
southern Chile. In particular, the most diverse group
of animals on Earth, the insects, has so far received lit- Taxon sampling
tle attention in temperate areas of South America (but We examined a total of 248 individuals of C. arma-
see Zúñiga-Reinoso et al., 2016). tus from 46 localities in Chile (Table 1, Fig. 1). We
In this study, we examined the phylogeography also included in the phylogenetic analyses as out-
of a large cricket that is widely distributed in cen- groups five additional individuals of the congeneric
tral and southern Chile, the Chilean red cricket or species Cratomelus integer, which is parapatric with
‘grillo rojo’ (Cratomelus armatus Blanchard, 1851, C. armatus to the north (Table 1, Fig. 1). The sampled
Anostostomatidae) with the aim of investigating area encompasses most of the known geographical

© 2018 The Linnean Society of London, Biological Journal of the Linnean Society, 2018, 123, 712–727
714  F. M. ALFARO ET AL.

Downloaded from https://academic.oup.com/biolinnean/article-abstract/123/4/712/4925800 by guest on 23 July 2019

Figure 1.  Sampling localities of Cratomelus armatus used in the study. The numbers represent the identification of locali-
ties (see Table 1).

distribution of the species, which extends over 1200 Laboratorio de Entomología Ecológica, Universidad de
km, from the Cardenal Caro Province in the north La Serena, Chile (LEULS) and the Phoenix group col-
(O’Higgins, 34°S) to the southern Palena Province (Los lection, Massey University, New Zealand.
Lagos, 42°S; Fig. 1). For species identification, we fol-
lowed Gorochov (1999). The collected individuals were
fixed and stored in 99% ethanol and deposited in the DNA extraction, amplification and sequencing
repository of the Laboratorio de Genética y Evolución, DNA was extracted from muscle tissue using a
Universidad de Chile, Chile (GEVOL), the collection of modified salt extraction method (Jowett, 1986;

© 2018 The Linnean Society of London, Biological Journal of the Linnean Society, 2018, 123, 712–727
 PHYLOGEOGRAPHY OF CRATOMELUS ARMATUS  715

Table 1.  Sampling locations and genetic diversity estimated for the populations of Cratomelus armatus

Site no. Locality Latitude Longitude N π Hd

Cratomelus integer
1 Parque Nacional La Campana −32.9167 −71.1500 1 – –
2 Ritoque −32.8254 −71.4790 1 – –
3 Leyda −33.6017 −71.4799 3 – –
Cratomelus armatus
4 Llico −34.7463 −72.0650 6 0.01987 0.8
5 Iloca −34.9576 −72.1844 4 0.00261 0.8333

Downloaded from https://academic.oup.com/biolinnean/article-abstract/123/4/712/4925800 by guest on 23 July 2019

6 Putú −35.2116 −72.2856 7 0.0056 0.9524
7 Constitución −35.3665 −72.3867 4 0.00436 1
8 Reserva Nacional Federico Albert −35.7316 −72.5366 6 0.0027 0.8
9 Tregualemu −35.9767 −72.7519 12 0.00147 0.7879
10 Buchupureo −36.0795 −72.8038 3 0.01307 0.6667
11 Cobquecura −36.1396 −72.7541 6 0.00044 0.3333
12 Pingueral −36.5288 −72.9240 1 – –
13 Coihueco −36.6399 −71.8163 10 0.02891 0.9778
14 Talcahuano −36.7304 −73.1145 5 0.03791 0.8
15 Coyanco −36.7461 −72.5202 3 0.00959 1
16 Hualpén −36.7965 −73.1541 3 0 0
17 Laguna Pineda −36.8279 −72.9261 12 0.00705 0.9697
18 Coronel −37.0097 −73.1310 2 0 0
19 Parque Nacional Nahuelbuta −37.8164 −72.9429 10 0.0075 0.7111
20 Vegas Blancas −37.8223 −72.7387 6 0.00166 0.8667
21 Cerro Ñielol −38.7271 −72.5921 7 0.00952 0.9048
22 Teodoro Schmidt −38.9995 −73.0908 5 0.00523 0.4
23 Vilcún −38.8672 −71.8966 1 – –
24 Huerquehue −39.1386 −71.6664 1 – –
25 Lican Ray −39.4964 −72.1526 4 0.00588 0.5
26 Malihue (Fundo St. Olga) −39.7470 −72.6735 5 0.00471 0.9
27 Neltume −39.8511 −71.9256 10 0.00206 0.7778
28 Reserva Nacional Mocho Choshuenco −39.9138 −72.1461 7 0.00735 0.9524
29 Parque Oncol −39.7085 −73.3316 1 – –
30 Valdivia −39.8138 −73.1935 5 0.00549 0.9
31 La Unión −40.1740 −73.4699 8 0.00616 0.9643
32 Parque Nacional Alerce Costero −40.2119 −73.3905 6 0.0034 0.9333
33 San Juan de la Costa −40.2616 −73.3394 7 0.00411 0.7143
34 Parque Nacional Puyehue −40.7262 −72.2829 13 0.00446 0.8846
35 Puerto Octay −40.9282 −72.8718 6 0.00331 0.7333
36 Cascadas (Cascadas bridge) −41.0850 −72.6347 10 0.00726 0.8667
37 Volcán Osorno −41.1746 −72.5152 5 0.00549 0.9
38 Punta Larga −41.1874 −72.9826 4 0.00741 1
39 Parque Katalapi −41.5200 −72.7506 3 0.00087 0.6667
40 Chiloe (near Parque Nacional Chiloé) −42.4172 −74.0786 6 0.00401 0.8667
41 Chiloé (Puntra) −42.1192 −73.8111 1 – –
42 Chiloé (Cucao) −42.6398 −74.1078 3 0.00436 0.6667
43 Chiloé (Ancud) −41.8718 −73.8233 1 – –
44 Chiloé (north point of Huillinco Lake) −42.6432 −74.0115 9 0.00806 0.8889
45 Chiloé (Yaldad) −43.0927 −73.7293 6 0.00784 0.8
46 Caleta Leptepú −42.5445 −72.5219 4 0.00065 0.5
47 Chaitén −42.9131 −72.7097 4 0.00261 0.5
48 El Amarillo −42.9987 −72.4431 6 0.00131 0.6

Hd, haplotype diversity; N, sample size; π, nucleotide diversity.

© 2018 The Linnean Society of London, Biological Journal of the Linnean Society, 2018, 123, 712–727
716  F. M. ALFARO ET AL.

Sunnucks & Hales, 1996). We amplified par- summarized using the maximum clade credibility cri-
tial sequences of the mtDNA gene cytochrome teria in TreeAnnotator v2.4.4 (distributed as part of
oxidase I (COI) using primers C1-J-2183 BEAST), discarding the first 20% of the trees as burn-
(5′-CAACATTTATTTTGATTTTTTGG-3′) and TL2- in. The summarized tree was visualized and edited for
N-3014 (5′-AATTCCGCACATTGCCTAATCATTA-3′) illustration purposes using Figtree v1.4.2 (Rambaut,
(Simon et al., 1994). Previous studies with related 2008). Complementarily, to examine haplotype rela-
crickets have shown that this marker provides good tionships and the frequency distribution across space
resolution for phylogeographical analysis (e.g. Trewick we constructed a haplotype network using the median
& Morgan-Richards, 2004; Pratt, Morgan-Richards & joining algorithm (Bandelt, Forster & Röhl, 1999)
Trewick, 2008; Brettschneider et al., 2009; Chappell, implemented in PopArt v1.7.1 (Leigh & Bryant, 2015).

Downloaded from https://academic.oup.com/biolinnean/article-abstract/123/4/712/4925800 by guest on 23 July 2019

Trewick & Morgan-Richards, 2012). The reaction mix- Following the results from the phylogenetic analy-
ture included 3 mM MgCl2, 0.2 mM dNTPs, 0.2 µM each ses, we conducted a barcode gap analysis to look for
primer, 1 U Taq polymerase (Invitrogen, Carlsbad, CA, evidence of species-level differences between major
USA) and 50–100 ng total DNA. The thermal protocol clades using the automatic procedure ABGD described
for the PCR was 94 °C for 5 min, followed by 36 cycles of by Puillandre et al. (2012). This analysis uses the bar-
94 °C for 45 s, 45–50 °C for 30 s and 72 °C for 60 s, with code gap, which is the gap observed when divergence
a final extension at 72 °C for 2 min. The PCR products among individuals of the same species is smaller than
were visualized in agarose gels and sequenced using divergence among individuals from different species,
the same primers. DNA sequences were edited and to automatically find groups that might correspond to
aligned in BIOEDIT v7.0.5.2 (Hall, 1999) and trans- different potential species. A range of prior intraspe-
lated into amino acids in Mega 6 to check for codon cific divergence values from 0.001 to 0.1 was assayed
stops and frame shifts that could indicate alignment (in ten steps), applying a relative gap width (X) of 1.5
errors and the potential presence of nuclear copies and using three options for genetic distance (JC69,
of the mitochondrial gene (Numts; Song, Moulton & K80 and simple distance). This analysis was performed
Whiting, 2014). Levels of substitution saturation were through the Web server of ABGD (http://wwwabi.snv.
analysed with Xia’s test (Xia et al., 2003) in DAMBE, jussieu.fr/public/abgd/abgdweb.html).
version 5.1.5 (Xia & Xie, 2001). All sequences are avail-
able in GenBank (MG202165–MG202417).
Analyses of genetic diversity and demography
To estimate genetic diversity across the distribution of
Time-calibrated genealogy and species C. armatus, we estimated several diversity statistics,
delimitation analysis namely haplotype diversity (Hd), nucleotide diversity
A dated genealogical reconstruction was estimated (π) and genetic distance among geographically close
by Bayesian inference with the program BEAST populations for all sampling localities, using the pro-
version 2.4.4 (Bouckaert et al., 2014). We included gram DnaSP v5.10.01 (Librado & Rozas, 2009). Given
all the sequences of C. armatus and the sequences that phylogenetic analyses revealed multiple lineages,
of the congeneric species C. integer as an outgroup demographic analyses were conducted separately for
(MG202413–MG202417). The HKY + G evolutionary each of the two main lineages that were supported
model was selected as the best-fit model using the pro- as distinct by the ABGD analysis (see the results sec-
gram jModelTest version 0.1.1 (Posada, 2008) using tion below and Fig. 2). This approach was preferable
the Bayesian information criterion. The strict-clock to compare the north and south areas of the species
model was selected as the molecular clock prior, using distribution, impacted in a different manner by glacia-
a substitution rate of 0.017 substitutions per site per tions. Signals of demographic expansion and/or stabil-
million years, following recently published rates for ity were evaluated in several ways. We first estimated
insects (Papadopoulou, Anastasiou & Vogler, 2010) Fu’s Fs (Fu, 1997) and Tajima’s D statistics, evaluating
and orthopterans (Allegrucci, Trucci & Sbordoni, 2011; their significance using 1000 simulations in Arlequin
Kaya & Çiplak, 2016). v3.5.1.3 (Excoffier & Lischer, 2010). Significantly nega-
We conducted two independent analyses to check tive values for these statistics imply recent population
for consistency in the results. Each analysis was run expansion, whereas values not significantly different
for 50 million generations, sampling trees every 10 from zero are interpreted as evidence of stable popula-
000 generations. The program Tracer v. 1.6 (Rambaut tion size. We used Bayesian skyline plot (BSP) analyses
& Drummond, 2009) was used to visualize the traces of implemented in the program BEAST v1.8.0 (Drummond
the Markov chain Monte Carlo runs and to check that et al., 2012) to investigate population changes over
the effective sample size of model parameters were time. This coalescent approach uses genealogical
> 200 (indication of convergence). Sampled trees were information to estimate important population genetic

© 2018 The Linnean Society of London, Biological Journal of the Linnean Society, 2018, 123, 712–727
 PHYLOGEOGRAPHY OF CRATOMELUS ARMATUS  717

Downloaded from https://academic.oup.com/biolinnean/article-abstract/123/4/712/4925800 by guest on 23 July 2019

Figure 2.  Bayesian phylogenetic tree for Cratomelus armatus based on mitochondrial DNA. Numbers on nodes are
Bayesian posterior probabilities, and the colour of branches represents the different lineages (see Fig. 3).

© 2018 The Linnean Society of London, Biological Journal of the Linnean Society, 2018, 123, 712–727
718  F. M. ALFARO ET AL.

parameters to reconstruct the effective population size presented low saturation as indicated by Xia’s test
(Ne) over time using a Bayesian statistical approach. [index of substitutional saturation (Iss) < critical
Priors and settings were kept as in the dated genea- Iss (Iss.c); P < 0.0001], which supported their suitabil-
logical analyses, but using the coalescent Bayesian ity for phylogenetic inference. The nucleotide frequen-
skyline option as the tree prior. We ran 100 million cies were 31.16% (A), 37.87% (T/U), 16.09% (C) and
iterations, sampling every 10 000 iterations to gener- 14.88% (G); the overall transition/transversion ratio
ate 100 000 parameter estimates for two independent (R) was 3.69, a value that is typical of mtDNA data
runs. Likelihood values of demographic plots for each sets (nuclear data sets are usually below R = 2; Nei
genetic group were visualized with the program Tracer & Kumar, 2000). This transition/transversion ratio,
v1.5.0 (Rambaut & Drummond, 2009). along with the absence of internal codon stops and

Downloaded from https://academic.oup.com/biolinnean/article-abstract/123/4/712/4925800 by guest on 23 July 2019

Complementarily, a Mantel test (Mantel, 1967) was frame shifts in the sequence alignment, suggests the
also performed to compare patterns of isolation by dis- absence of nuclear copies of mtDNA in our data set.
tance between the north and south lineages. Impacted
areas that are not under equilibrium will show weak
or no correlation between geographical distance and Genealogical reconstruction and estimation
genetic structure, whereas areas that have been more of divergence times
stable will show a stronger correlation. The Mantel test The Bayesian topology supported the existence of two
was conducted in Alleles In Space (AIS; Miller, 2005). highly divergent lineages (north and south clades;
Fig. 2), with an estimated divergence of 2.06 Mya [95%
high posterior density (HPD): 1.54–2.5 Mya], that come
Testing the impact of glaciations on genetic into contact at the latitude of Concepción (~36.4°S).
diversity and structure Indeed, two sampling sites, Coihueco (13) and
We conducted various tests to assess the impact of gla- Talcahuano (14) (see Table 1), harboured individuals of
ciation on genetic diversity. We performed correlation both lineages. The north lineage, which encompasses
tests between measures of genetic diversity and lati- samples collected from Llico (34°S) to Coronel (37°S),
tude. The general model suggests that the extent of gla- was the most diverse, with four lineages showing high
ciation increased with latitude, predicting a decrease in geographical structure (lineages A–D; Fig. 2). The geo-
genetic diversity from north to south. We first looked graphical limits of these lineages broadly coincide with
for correlation between genetic diversity and latitude three major river systems, the Mataquito, Maule and
with all localities, as the distribution of populations Itata rivers. Although relationships between subclades
strongly follows a latitudinal gradient. However, the A–D were not well resolved, geographical structure
presence of distinct lineages in our phylogenetic analy- and high branch support for these clades support their
ses suggested that these should be treated separately, recognition as differentiated groups. The divergence
so we then focused the analysis on populations of the times for these lineages ranged from 1.08 to 0.49 Mya
southernmost clade (south of 37°S), which presented (see Table 2), with an estimated age for the most recent
different degrees of exposure to glacial coverage. common ancestor (MRCA) of 1.08 Mya (95% HPD:
Three measures of genetic diversity were also used 0.84–1.36 Mya). In contrast to the north lineage, the
to assess for geographical correlation, namely nucleo- south lineage showed shallow phylogenetic relation-
tide diversity (π), haplotype diversity and the average ships and little geographical structure, despite being
number of pairwise differences between populations.
Analysis using pairwise genetic distances and latitude
used only pairwise distances of one decimal degree or Table 2.  Divergence time estimates with 95% HPD con-
less (~111 km) to represent local diversity. We used the fidence intervals for clades based on mitochondrial DNA
midpoint between the compared localities to obtain a COI of Cratomelus armatus
single latitude value for pairwise comparisons. Finally,
to visualize the variation of genetic distances across Clade MRCA (Mya) 95% HPD
the distribution of the species better, we performed a
Clade A 0.14 0.05–0.24
genetic landscape shape interpolation analysis with
Clade B 0.39 0.25–0.54
the program AIS (Miller, 2005).
Clade C 0.49 0.32–0.66
Clade D 0.69 0.46–0.91
Clades BC 0.84 0.61–1.04
RESULTS Clade A(BC) 0.93 0.69–1.15
North (ABCD) 1.08 0.84–1.36
We obtained a 765 bp alignment of mtDNA COI gene South 0.53 0.38–0.67
sequences comprising 167 variable sites and a total of North–South 2.06 1.54–2.5
117 haplotypes from 248 individuals. The sequences

© 2018 The Linnean Society of London, Biological Journal of the Linnean Society, 2018, 123, 712–727
 PHYLOGEOGRAPHY OF CRATOMELUS ARMATUS  719

Downloaded from https://academic.oup.com/biolinnean/article-abstract/123/4/712/4925800 by guest on 23 July 2019

Figure 3.  Median joining network of mitochondrial DNA of Cratomelus armatus from central and southern Chile, showing
geographical distribution of haplogroups. The haplotypes are represented by circles whose sizes are proportional to their
frequencies. Colours represent the different geographical groups where haplotypes are present. The light blue shape shows
the extent of the ice cover during the Last Glacial Maximum (Clapperton, 1993).

distributed across a larger geographical area. The esti- The haplotype network was consistent with the
mated age of the southern lineage was ~50% younger genealogical reconstruction and showed a complex
than the age of the northern lineage, with a time to the structure with five main haplogroups separated by
MRCA of 0.53 Mya (95% HPD: 0.38–0.67 Mya). a large number of nucleotide substitutions (Fig. 3).

© 2018 The Linnean Society of London, Biological Journal of the Linnean Society, 2018, 123, 712–727
720  F. M. ALFARO ET AL.

Cratomelus armatus with haplogroup B haplotypes with a clear, wide barcode gap located in the distance
came from localities north of the Biobio River, whereas range 0.05–0.06 (Supporting Information, Fig. S1A).
those with group D haplotypes were associated with Furthermore, the method detected two stable candi-
localities south of the Maule River (between the Maule date species regardless of the distance option chosen,
and Itata rivers). Haplogroup C was associated with with estimated prior maximum divergence of intraspe-
localities south of the Mataquito River, whereas hap- cific diversity (P) as large as 6%. These putative spe-
logroup A was associated with localities north to the cies corresponded to the two main lineages found in the
mouth of the Mataquito River. Individuals collected in Bayesian genealogical analyses (north and south).
localities south of the Biobio River (37–43°S) belonged
to the south lineage. The four haplogroups north of the
Patterns of genetic diversity and demographic

Downloaded from https://academic.oup.com/biolinnean/article-abstract/123/4/712/4925800 by guest on 23 July 2019

Biobio River were represented by haplotypes at similar
frequencies that did not form star-like topologies. In analyses
contrast, the south lineage included a few haplotypes Considering all the sampling sites, both nucleotide
at high frequency, each with numerous rare derived diversity (π) and haplotype diversity (Hd) showed no
haplotypes differing by one to a few nucleotide substi- correlation with latitude (Supporting Information,
tutions. This star-like topology is often associated with Fig. S2A, B), against expectations of a major glacial
recent demographic expansions. impact on genetic diversity towards high latitudes.
Genetic diversity (nucleotide and haplotype diver-
sity) was relatively similar across the distribution of
Genetic differentiation and species C. armatus, except for three localities where nucleotide
delimitation analysis diversity was markedly high (i.e. outliers; Supporting
Genetic distances between the north and south clades Information, Fig. S3): Llico, Coihueco and Talcahuano.
and between subclades within the north clade were The high nucleotide diversity in these three localities
fairly high (Table 3), revealing high cryptic diversity can be explained by the presence of haplotypes from
within the nominal species C. armatus, especially different lineages (see Fig. 3).
in its northern range (from the Biobio River to the Regardless of the inclusion or exclusion of these loca-
north). The average (uncorrected) number of pairwise tion samples, correlation analyses showed no significant
differences between the north and the south lineages relationship between latitude and genetic diversity,
was 52.34 (p-distance = 0.068). The average number either for π (R = −0.311, P = 0.12 including admixed
of pairwise differences between clades A, B, C and D localities, and R = 0.2, P = 0.34 excluding admixed
(subclades within the north clade) ranged from 22.7 localities) or for Hd (R = −0.024, P = 0.91 including
(p-distance = 0.03) to 31.5 (p-distance = 0.041), with admixed localities, and R = −0.04, P = 0.85 excluding
the smallest difference being between clades A and C admixed localities; see Supporting Information, Fig.
and the largest between B and D (Table 3). S2). In contrast, genetic distance among geographically
The automatic barcode gap species delimitation ana- close populations did show a tendency to decrease with
lysis (ABGD) using all the available distance options latitude, as can be seen in the landscape genetic analy-
(JC69, K80 and simple distance) produced consistent sis (Fig. 4). This pattern is explained, at least in part,
results (see Supporting Information, Fig. S1 for full by the richer lineage diversity and phylogeographical
results). We therefore refer only to the results obtained structure present in the north and central portions of
with the K80 evolutionary model. The results showed the species distribution, clearly evident in the geneal-
a multimodal pairwise genetic distance distribution ogy and network (Figs 2, 3).
Given that genealogy and species delineation anal-
yses suggested two lineages of species status within
Table 3.  Average number of pairwise differences (uncor-
the nominal C. armatus distribution, it makes sense
rected) between haplogroups based on mitochondrial DNA
COI of Cratomelus armatus (see Fig. 3) 
to treat these lineages separately for further analyses.
For the case of patterns of genetic diversity vs. lati-
Haplogroup A C D B South tude, only the south lineage was evaluated because
it is distributed in an area where glaciations were
A – 22.71 28.20 29.01 54.24 more relevant during the most recent glacial periods.
C 0.03 – 24.89 26.51 51.65 Neither nucleotide diversity (R = −0.041, P = 0.87) nor
D 0.037 0.032 – 31.50 52.91 haplotype diversity (R = −0.42, P = 0.073) decreased
B 0.038 0.035 0.041 – 50.60 with latitude for the south lineage (Supporting
South 0.071 0.068 0.069 0.066 – Information, Fig. S2), in contrast to our expectations.
Genetic distance among populations, however, did
Below diagonal, genetic divergence between clades (p-distance). show a significant north–south decrease (R = −0.57,

© 2018 The Linnean Society of London, Biological Journal of the Linnean Society, 2018, 123, 712–727
 PHYLOGEOGRAPHY OF CRATOMELUS ARMATUS  721

Downloaded from https://academic.oup.com/biolinnean/article-abstract/123/4/712/4925800 by guest on 23 July 2019

Figure 4.  Genetic landscape shape interpolation analyses for the distribution of Cratomelus armatus based on mtDNA of
all samples (A), north lineage samples (B) and south lineage samples (C). The x- and y-axes correspond to the coordinates of
the geographical locations of the samples examined in this study.

P = 0.000), indicating a more homogeneous (less struc- latitudes in the interpopulation genetic diversity land-
tured) distribution of genetic diversity in southern scape (shown in Fig. 4C). Strikingly, this visualization
areas (Supporting Information, Fig. S4). The land- also showed a moderate increase in among-population
scape genetic analysis confirmed this latitudinal genetic distances in Chiloé, one of the southernmost
pattern by showing deeper valleys towards higher areas included in this study.

© 2018 The Linnean Society of London, Biological Journal of the Linnean Society, 2018, 123, 712–727
722  F. M. ALFARO ET AL.

Demographic analyses indicated different demo- Allegrucci et al. (2013), analysing several weta spe-
graphic histories for the north and south lineages. cies from New Zealand, showed that genetic distances
For the north lineage, both neutrality tests, Fs and within species ranged from 0.3 to 1.7%, whereas gen-
Tajima’s D, showed no significant departures from etic distances between species ranged from 1.9 to 9.3%.
zero, indicating stable population size and long-term Trewick (2008), who analysed genetic diversity within
population history (Table 4). The Mantel test also the Sigaus species complex, found genetic distances to
showed a high and significant correlation between geo- range from 3 to 8.3% between major haplogroups, sug-
graphical and genetic distances (R = 0.72, P < 0.001) gesting the need for further taxonomic work. Likewise,
for the north lineage. In contrast, for the south lineage within the South African species Libanasidus vitta-
the Fs and Tajima’s D tests showed negative and sig- tus, genetic divergence of 3.3% between clades was

Downloaded from https://academic.oup.com/biolinnean/article-abstract/123/4/712/4925800 by guest on 23 July 2019

nificant departures from zero, suggesting past bottle- reported by Brettschneider et al. (2009), and a further
necks and a more recent population history (Table 4). examination of morphological data found several dif-
Consistently, the Mantel test for this group showed a ferences that suggested new species/lineages. Several
weak (although significant) correlation between gen- studies have found that divergences in the range of
etic distance and geographical distance (R = 0.17, 2–3% for insects are consistent with traditional taxo-
P < 0.001). nomic limits (Hajibabaei et al., 2006; Lefébure et al.,
The Bayesian skyline plot analyses showed con- 2006). Our results, which show between-clade genetic
sistent patterns with the above tests. For the north distances ranging from 3 to 6.9%, indicate a long his-
clade, the analysis inferred constant population size tory of isolation between major clades and suggest the
over time (Supporting Information, Fig. S5A), whereas possibility of a complex of cryptic species within what
for the south clade it inferred a demographic expan- is currently known as C. armatus. Whether all these
sion over time (Supporting Information, Fig. S5B) that divergent lineages have evolved reproductive isola-
started ~100–150 kya and coincided with the OIS 6 tion, and therefore become different species, is a ques-
glacial event, the most extensive in the past half tion that requires further investigation. Nevertheless,
million years (Mortyn et al., 2003). Altogether, these the deep divergence between samples from the north
results strongly suggest a more recent demographic and the south of the distribution of C. armatus (a
history for the southern lineage, consistent with a sce- p-distance of 0.07) indicates a long history of isola-
nario of lower stability in southern Chile. tion between these lineages and strongly suggests
distinct species. These lineages are parapatric at the
latitude of Concepcion, with some localities harbour-
ing individuals from both lineages (Talcahuano and
DISCUSSION
Coihueco). It is likely that the Biobio River (part of
Deep divergence supports cryptic species the largest river network in Chile and whose mouth
The deep divergence between the north and south discharges near Concepcion) had historically contrib-
lineages and the ABGD analysis revealed that these uted to reduce gene flow between the north and south
lineages most probably represent two different spe- lineages. However, the fact that these lineages are
cies. In other regions harbouring a richer diversity in contact in some localities suggests that migration
of Anostostomatid crickets, such as South Africa and across the river is not fully interrupted, and therefore
Australasia, the analysis of molecular data (COI gene flow must be restricted by some other reproduct-
gene) has proved useful to gain a better understand- ive barriers apart from geographical ones (e.g. eco-
ing of interspecific vs. intraspecific variation (e.g. logical, behavioural, genetics). It has been recorded
Trewick & Morgan-Richards, 2004; Pratt et al., 2008; that C. armatus displays strong territoriality, with
Brettschneider et al., 2009; Chappell et al., 2012). aggressive behaviour that commonly results in canni-
balism in controlled experiments (Angulo, 2001). This
type of behaviour has been associated with geograph-
ical segregation and parapatric distributions in some
Table 4.  Results of the neutrality test for the two princi- taxa (e.g. Nevo, Naftali & Guttman, 1975; Arif, Adams
pal lineages (north and south) of Cratomelus armatus & Wicknick, 2007; Jankowski, Robinson & Levey,
2010), but further studies are required to test whether
Lineage N Tajima’s D P-value Fu’s Fs P-value
this factor plays a role in maintaining the parapatry
North 61 0.38 0.72 −3.4 0.18
between the north and south lineages of C. armatus.
South 187 −1.94 0.00* −24.6 0.00* Nevertheless, molecular evidence is strong in sug-
gesting that these lineages should be considered two
N is the number of individuals.
different species; therefore, we recommend detailed
*Statistically significant at the 0.05 level. morphological descriptions to describe them formally.

© 2018 The Linnean Society of London, Biological Journal of the Linnean Society, 2018, 123, 712–727
 PHYLOGEOGRAPHY OF CRATOMELUS ARMATUS  723

The northern lineage, with its higher diversity of maintained by the Pleistocene glacial and interglacial
lineages, may well represent more cryptic species, periods (Trewick, Wallis & Morgan-Richards, 2000).
because divergences between subclades are > 3%. Some of our results, however, did not follow the
However, detailed ecological and morphological stud- expected pattern of a stronger impact at higher lati-
ies, as well as additional molecular markers, would be tudes. Genetic diversity (nucleotide and haplotype
necessary to test this hypothesis. diversity estimated for each sampling site) did not
Regardless of our current ability to identify different decrease with latitude. Similar results have previously
morphological species, our data suggest the existence been reported for freshwater fish (Unmack et al., 2009;
of an elevated cryptic diversity within the C. armatus Muñoz-Ramírez et al., 2014). These mixed patterns
complex that warrants and encourages more research could suggest that C. armatus has recovered rapidly

Downloaded from https://academic.oup.com/biolinnean/article-abstract/123/4/712/4925800 by guest on 23 July 2019

into the processes that have promoted this diversity. from the inferred bottlenecks or that its populations
Understanding the taxonomic limits of this cryptic have not been impacted by the most recent glaciations;
diversity is important to make correct sense of bio- at least, not to a sufficient degree to decrease genetic
logical processes. For instance, morphological and diversity severely. Perhaps only older (previous to the
behavioural studies conducted on C. armatus from LGM) and more intense glaciations have had a detect-
Concepcion (Angulo, 2001) suggested that this spe- able impact on the demography of C. armatus, provid-
cies has a large variance in morphology and behaviour. ing enough time for populations to recover genetic
However, based on our study two species can be found diversity to the present observed levels. Indeed, our
in this area; therefore, these previous analyses might BSP analysis (Supporting Information, Fig. S5) esti-
inadvertently be reporting variation of two different mates the start of the demographic expansion for the
species instead of one, overestimating intraspecific southern lineage at ~150 kya, coinciding with the end
variation. Additional data and the incorporation of of one of the most extensive glaciations in the past
the two other Cratomelus species will strengthen our 400 kyr (the OIS 6 event; Mortyn et al., 2003) and not
understanding of species limits within the genus and during the LGM (~23 kya) as has been often found for
help to test evolutionary processes underlying the ori- other taxa (e.g. Ruzzante et al., 2006; Vera-Escalona
gin of this diversity. et al., 2012). Cratomelus armatus may have persisted
in large numbers during less severe glaciation events
(insect population sizes are often several orders of
The impact of glaciations on C. armatus magnitude larger than those of vertebrates), favoured
Although still limited in number, phylogeographical perhaps by its fossorial habits (Angulo, 2001), water-
studies generally agree with the notion of a severe efficient physiology (Nespolo, Artacho & Castañeda,
impact of glacial cycles on species’ intraspecific diver- 2007) and generalist feeding behaviour (Angulo, 2001).
sity in southern South America (e.g. Victoriano et al., Our results, along with other reported cases of demo-
2008; Sérsic et al., 2011). Therefore, molecular pat- graphic expansion pre-dating the LGM (e.g. Zemlak
terns reflecting a gradient of increasing impact from et al., 2010; Vianna et al., 2011), support a scenario
North to South are to be expected. in which the LGM was not equally detrimental for
Our results supported this general view. Lower lin- all taxa.
eage diversity and signals of population expansion in Although our demographic results demonstrate
the southern range of C. armatus suggest a stronger a stronger impact of glaciations in southern areas
impact of glaciations in southern Chile. In contrast, of Chile, they do not necessarily contradict a role of
high levels of lineage diversity and signals of greater glaciations in central Chile. It has been recently sug-
demographic stability in the northern range of the spe- gested that glaciations may have promoted diversi-
cies suggest a lower impact of glaciations in central fication in several taxa distributed along mountain
Chile. These results are consistent with previous stud- ranges (Wallis et al., 2016). These authors suggest that
ies, supporting the hypothesis that this area has been latitudinal phylogeographical breaks in several south-
a general refugium for several taxa (e.g. Rodríguez- ern Andes taxa are evidence of isolation in multiple
Serrano et al., 2006; Victoriano et al., 2008; Azpilicueta, glacial refugia resulting from transverse glacial bar-
Marchelli & Gallo, 2009; Palma et al., 2012). Likewise, riers. This hypothesis is consistent with our finding of
related orthopterans at similar latitudes in New several lineages in the northern range (central Chile)
Zealand show higher diversity in the north compared of C. armatus. Although central Chile has often been
with the south of their ranges (Morgan-Richards, assumed to be less impacted by glaciations compared
Trewick & Wallis, 2001; Bulgarella et al., 2014), and with southern Chile, some glaciations, particularly the
the phylogeographical structure of some species, such most severe ones, could have contributed to the diver-
as the alpine scree weta Deinacrida connectens, has sification of C. armatus in areas where the presum-
been affected by Pliocene mountain building and ably lesser glacier development may have fragmented

© 2018 The Linnean Society of London, Biological Journal of the Linnean Society, 2018, 123, 712–727
724  F. M. ALFARO ET AL.

rather than displacing its habitat, facilitating popula- was funded by FONDECYT project 1140540 (M.A.M.)
tion isolation and divergence. In contrast, more con- and CONICYT fellowships no. 2013-22130123 (F.M.A.)
tinuous and widespread ice coverage in southern areas and no. 21110367 (A.Z.-R.). We thank three anonym-
may have prevented isolation in multiple isolated ous reviewers for their helpful comments that helped
refugia; hence, it prevented the diversification of the to improve the manuscript and Lafayette Eaton
southern lineage. Interestingly, our dated genealogy for proofreading the English version. We are espe-
indicates that divergence times for all major clades cially indebted to Luis Pastenes, Paola Sáez, Pamela
of the northern lineage (clades A–D; Fig. 2) date back Morales, Franco Cruz, Dusan Boric, Einer Sepúlveda,
to 1.1–0.84 Mya (Table 2), matching the period of the Daniela Mardones, Jaime Pizarro-Araya, Andrés
most extensive Andean glaciation (Largest Patagonian Ojanguren-Affilastro, Ricardo Botero-Trujillo, Hernán

Downloaded from https://academic.oup.com/biolinnean/article-abstract/123/4/712/4925800 by guest on 23 July 2019

Glaciation; Singer, Ackert & Guillou, 2004). These Iuri, Jaime Zuñiga and Fermín Alfaro Contreras for
results support a role of glaciations in the diversifica- their invaluable help in the laboratory and during the
tion of the C. armatus complex in the northern range collecting trips.
of its distribution and add evidence to a growing body
of knowledge about the synergistic role of glaciations
and mountain ranges on the evolution of biodiversity
during the Pleistocene (Wallis et al., 2016). REFERENCES
It is also possible that other factors could have con-
Alfaro FM, Zuñiga-Reinoso A, Muñoz-Ramírez C,
tributed to the diversification of the northern lineage.
Elgueta M. 2015. New records reveal the actual dis-
The broad match between clade limits and several tribution of Cratomelus meritus Gorochov (Orthoptera:
Andean river basins (see Fig. 3) may suggest a role Anostostomatidae), a giant red cricket from Chile. Neotropical
of river basins in the phylogeographical structure of Entomology 44: 194–196.
C. armatus in Central Chile, similar to that observed Allegrucci G, Massa B, Trasatti A, Sbordoni V. 2013. A
in some terrestrial vertebrates (Lamborot & Eaton, taxonomic revision of western Eupholidoptera bush crickets
1997; Lamborot, Eaton & Carrasco, 2003; Torres-Pérez (Orthoptera: Tettigoniidae): testing the discrimination power
et al., 2007; Vásquez, Torres-Pérez & Lamborot, 2007; of DNA barcode. Systematic Entomology 39: 7–23.
Sallaberry-Pincheira et al., 2011) and aquatic taxa Allegrucci G, Trucchi E, Sbordoni V. 2011. Tempo and
(Ruzzante et al., 2006; Unmack et al., 2009; Muñoz- mode of species diversification in Dolichopoda cave crickets
Ramírez et al., 2014; Muñoz-Ramírez, Victoriano & (Orthoptera, Rhaphidophoridae). Molecular Phylogenetics
Habit, 2015). However, it is difficult to postulate riv- and Evolution 60: 108–121.
ers as the main factor responsible for diversification in Angulo AO. 2001. Anatomy, development and behaviour of
Cratomelus for several reasons. First, most of the larg- the Chilean red cricket, Cratomelus armatus. In: Field LH,
est river basins of Chile are located in the south, where ed. The biology of the wetas, king crickets and their allies.
the southern lineage is present. However, the south- Wallingford: CABI Publishing, 205–222.
ern lineage does not show internal phylogeographical Arif S, Adams DC, Wicknick JA. 2007. Bioclimatic modelling,
breaks as would be expected if rivers played a role in morphology, and behaviour reveal alternative mechanisms
diversification. Second, central and southern Chile regulating the distributions of two parapatric salamander
species. Evolutionary Ecology Research 9: 843–854.
are longitudinally dissected by many rivers, so it is
Avise JC. 2000. Phylogeography: the history and formation of
expected that any latitudinal phylogeographical break
species. Cambridge, MA: Harvard University Press.
will match a river to some extent, even if rivers were
Azpilicueta MM, Marchelli P, Gallo LA. 2009. The effects of
not responsible for the breaks. Therefore, evidence for
Quaternary glaciations in Patagonia as evidenced by chloro-
the impact of rivers on the diversification of the north plast DNA phylogeography of Southern beech Nothofagus
lineage remains elusive. We hypothesize that both riv- obliqua. Tree Genetics & Genomes 5: 561–571.
ers and multiple glacial refugia could have shaped pat- Bandelt HJ, Forster P, Röhl A. 1999. Median-joining net-
terns of diversity seen today in the north lineage of works for inferring intraspecific phylogenies. Molecular
C. armatus. However, glacial refugia could have played Biology and Evolution 16: 37–48.
a greater role in the origin of the diversity of the north Beheregaray LB. 2008. Twenty years of phylogeography:
lineage, whereas rivers may have played a role in their the state of the field and the challenges for the Southern
maintenance. Hemisphere. Molecular Ecology 17: 3754–3774.
Bouckaert R, Heled J, Kühnert D, Vaughan T, Wu CH,
Xie D, Suchard MA, Rambaut A, Drummond AJ. 2014.
BEAST 2: a software platform for Bayesian evolutionary
ACKNOWLEDGEMENTS
analysis. PLoS Computational Biology 10: e1003537.
We are grateful to CONAF (Corporación Nacional Brettschneider H, Chimimba CT, Scholtz CH, Bastos
Forestal, Chile) for allowing us to work in the National ADS, Bateman PWN. 2009. The tusked king cricket,
System of Wild Protected Areas (SNASPE). This study Libanasidus vittatus (Kirby, 1899) (Anostostomatidae), from

© 2018 The Linnean Society of London, Biological Journal of the Linnean Society, 2018, 123, 712–727
 PHYLOGEOGRAPHY OF CRATOMELUS ARMATUS  725

South Africa: morphological and molecular evidence sug- Jowett T. 1986. Preparation of nucleic acids. In: Roberts DB,
gest two cryptic species. Insect Systematics & Evolution 40: ed. Drosophila: a practical approach. Oxford: IRL Press,
85–103. 275–286.
Bulgarella M, Trewick SA, Minards NA, Jacobson MJ, Kaya S, Çiplak B. 2016. Budding speciation via peripheral iso-
Morgan-Richards M. 2014. Shifting ranges of two tree lation: the Psorodonotus venosus (Orthoptera, Tettigoniidae)
weta species (Hemideina spp.): competitive exclusion and species group example. Zoologica Scripta 45: 521–537.
changing climate. Journal of Biogeography 41: 524–535. Lamborot M, Eaton L. 1997. The Maipo River as a biogeo-
Chappell EM, Trewick SA, Morgan-Richards M. 2012. graphical barrier to Liolaemus monticola (Tropiduridae) in
Shape and sound reveal genetic cohesion not speciation in the mountain ranges of central Chile. Journal of Zoological
the New Zealand orthopteran, Hemiandrus pallitarsis, des- Systematics and Evolutionary Research 35: 105–111.
pite high mitochondrial DNA divergence. Biological Journal Lamborot M, Eaton L, Carrasco BA. 2003. The Aconcagua

Downloaded from https://academic.oup.com/biolinnean/article-abstract/123/4/712/4925800 by guest on 23 July 2019

of the Linnean Society 105: 169–186. River as another barrier to Liolaemus monticola (Sauria:
Clapperton CM. 1993. Quaternary geology and geomorph- Iguanidae) chromosomal races of central Chile. Revista
ology of South America. Amsterdam: Elsevier. Chilena de Historia Natural 76: 23–34.
Drummond AJ, Suchard MA, Xie D, Rambaut A. 2012. Lefébure T, Douady CJ, Gouy M, Gibert J. 2006.
Bayesian phylogenetics with BEAUti and the BEAST 1.7. Relationship between morphological taxonomy and molecu-
Molecular Biology and Evolution 29: 1969–1973. lar divergence within Crustacea: proposal of a molecu-
Elgueta M, Cammouseight A, Carbonnel CS. 1999. lar threshold to help species delimitation. Molecular
Catálogo de Orthoptera (Insecta) de Chile. Publicación Phylogenetics and Evolution 40: 435–447.
Ocasional, Museo Nacional de Historia Natural (Chile) 54: Leigh JW, Bryant D. 2015. POPART: full-feature software
1–60. for haplotype network construction. Methods in Ecology and
Excoffier L, Lischer HE. 2010. Arlequin suite ver 3.5: a new Evolution 6: 1110–1116.
series of programs to perform population genetics analyses Lessa EP, D’Elía G, Pardiñas UF. 2010. Genetic foot-
under Linux and Windows. Molecular Ecology Resources 10: prints of late Quaternary climate change in the diversity
564–567. of Patagonian-Fueguian rodents. Molecular Ecology 19:
Fu YX. 1997. Statistical tests of neutrality of mutations 3031–3037.
against population growth, hitchhiking and background Librado P, Rozas J. 2009. DnaSP v5: a software for compre-
selection. Genetics 147: 915–925. hensive analysis of DNA polymorphism data. Bioinformatics
Gorochov AV. 1999. A new species of Cratomelus from 25: 1451–1452.
Chile (Orthoptera: Anostostomatidae: Cratomelinae). Mantel N. 1967. The detection of disease clustering and a gen-
Zoosystematica Rossica 8: 242. eralized regression approach. Cancer Research 27: 209–220.
Hajibabaei M, Janzen DH, Burns JM, Hallwachs W, Miller MP. 2005. Alleles In Space (AIS): computer software
Hebert PDN. 2006. DNA barcodes distinguish species of for the joint analysis of interindividual spatial and genetic
tropical Lepidoptera. Proceedings of the National Academy of information. Journal of Heredity 96: 722–724.
Sciences of the United States of America 103: 968–971. Morgan-Richards M, Trewick SA, Wallis GP. 2001.
Hall TA. 1999. BioEdit: a user-friendly biological sequence Chromosome races with Pliocene origins: evidence from
alignment editor and analysis program for Windows 95/98/ mtDNA. Heredity 86: 303–312.
NT. Nucleic Acids Symposium Series 41: 95–98. Morrone JJ. 2006. Biogeographic areas and transition zones
Hewitt GM. 2000. The genetic legacy of the Quaternary ice of Latin America and the Caribbean islands based on pan-
ages. Nature 405: 907–913. biogeographic and cladistic analyses of the entomofauna.
Hewitt GM. 2004. The structure of biodiversity – insights Annual Review of Entomology 51: 467–494.
from molecular phylogeography. Frontiers in Zoology 1: 4. Mortyn P, Charles C, Ninnemann U, Ludwig K, Hodell D.
Himes CMT, Gallardo MH, Kenagy GJ. 2008. Historical bio- 2003. Deep sea sedimentary analogs for the Vostok ice core.
geography and post-glacial recolonization of South American Geochemistry, Geophysics, Geosystems 4: 1525–2027.
temperate rain forest by the relictual marsupial Dromiciops Muñoz-Ramírez CP, Unmack PJ, Habit E, Johnson JB,
gliroides. Journal of Biogeography 35: 1415–1424. Cussac VE, Victoriano P. 2014. Phylogeography of the
Jaksić FM, Jiménez JE, Medel RG, Marquet PA. 1990. ancient catfish family Diplomystidae: biogeographic, system-
Habitat and diet of Darwin’s fox (Pseudalopex fulvipes) atic, and conservation implications. Molecular Phylogenetics
on the Chilean mainland. Journal of Mammalogy 71: and Evolution 73: 146–160.
246–248. Muñoz-Ramírez C, Victoriano P, Habit E. 2015. Inter-basin
Jankowski JE, Robinson SK, Levey DJ. 2010. Squeezed at dispersal through irrigation canals explains low genetic
the top: interspecific aggression may constrain elevational structure in Diplomystes cf. chilensis, an endangered fresh-
ranges in tropical birds. Ecology 91: 1877–1884. water catfish from Central Chile. Limnologica 53: 10–16.
Jimenez JE, Marquet PA, Medel RG, Jaksić FM. 1990. Nei M, Kumar S. 2000. Molecular evolution and phylogenet-
Comparative ecology of Darwin’s fox (Pseudalopex fulvipes) ics. Oxford: Oxford University Press.
in mainland and island settings of southern Chile. Revista Nespolo RF, Artacho P, Castañeda LE. 2007. Cyclic gas-
Chilena de Historia Natural 63: 177–186. exchange in the Chilean red cricket: inter-individual variation

© 2018 The Linnean Society of London, Biological Journal of the Linnean Society, 2018, 123, 712–727
726  F. M. ALFARO ET AL.

and thermal dependence. The Journal of Experimental Song H, Moulton MJ, Whiting MF. 2014. Rampant
Biology 210: 668–675. nuclear insertion of mtDNA across diverse lineages within
Nevo E, Naftali G, Guttman R. 1975. Aggression patterns Orthoptera (Insecta). PLoS One 9: e110508.
and speciation. Proceedings of the National Academy of Sunnucks P, Hales DF. 1996. Numerous transposed sequences
Sciences of the United States of America 72: 3250–3254. of mitochondrial cytochrome oxidase I–II in aphids of the
Palma RE, Boric-Bargetto D, Torres-Pérez F, Hernández genus Sitobion (Hemiptera: Aphididae). Molecular Biology
CE, Yates TL. 2012. Glaciation effects on the phylogeo- and Evolution 13: 510–524.
graphic structure of Oligoryzomys longicaudatus (Rodentia: Torres-Pérez F, Lamborot M, Boric-Bargetto D,
Sigmodontinae) in the southern Andes. PLoS One 7: e32206. Hernández C, Ortiz JC, Palma R. 2007. Phylogeography
Palma RE, Rivera Milla E, Salazar Bravo J, Torres Pérez of a mountain lizard species: an ancient fragmentation pro-
F, Pardiñas UFJ, Marquet PA, Spotorno Oyarzún cess mediated by riverine barriers in the Liolaemus monti-

Downloaded from https://academic.oup.com/biolinnean/article-abstract/123/4/712/4925800 by guest on 23 July 2019

ÁE, Meynard AP, Yates TL. 2005. Phylogeography of cola complex (Sauria: Liolaemidae). Journal of Zoological
Oligoryzomys longicaudatus (Rodentia: Sigmodontinae) in Systematics and Evolutionary Research 45: 72–81.
temperate South America. Journal of Mammalogy 86: 191–200. Trewick SA. 2008. DNA barcoding is not enough: mismatch
Papadopoulou A, Anastasiou I, Vogler AP. 2010. Revisiting of taxonomy and genealogy in New Zealand grasshoppers
the insect mitochondrial molecular clock: the mid-Aegean (Orthoptera: Acrididae). Cladistics 24: 240–254.
trench calibration. Molecular Biology and Evolution 27: Trewick SA, Morgan-Richards M. 2004. Phylogenetics
1659–1672. of New Zealand’s tree, giant and tusked weta (Orthoptera:
Posada D. 2008. jModelTest: phylogenetic model averaging. Anostostomatidae): evidence from mitochondrial DNA.
Molecular Biology and Evolution 25: 1253–1256. Journal of Orthoptera Research 13: 185–196.
Pratt RC, Morgan-Richards M, Trewick SA. 2008. Trewick SA, Wallis GP, Morgan-Richards M. 2000.
Diversification of New Zealand weta (Orthoptera: Ensifera: Phylogeographical pattern correlates with Pliocene moun-
Anostostomatidae) and their relationships in Australasia. tain building in the alpine scree weta (Orthoptera, anostos-
Philosophical Transactions of the Royal Society B: Biological tomatidae). Molecular Ecology 9: 657–666.
Sciences 363: 3427–3437. Turchetto-Zolet AC, Pinheiro F, Salgueiro F, Palma-Silva
Puillandre N, Lambert A, Brouillet S, Achaz G. 2012. C. 2013. Phylogeographical patterns shed light on evolu-
ABGD, Automatic Barcode Gap Discovery for primary spe- tionary process in South America. Molecular Ecology 22:
cies delimitation. Molecular Ecology 21: 1864–1877. 1193–1213.
Rambaut A. 2008. FigTree v112. Available at: http:// Unmack PJ, Bennin AP, Habit EM, Victoriano PF,
treebioedacuk/software/figtree Johnson JB. 2009. Impact of ocean barriers, topog-
Rambaut A, Drummond AJ. 2009. Tracer v1.5.0. Available raphy, and glaciation on the phylogeography of the catfish
at: http://beast.bio.ed.ac.uk/Tracer Trichomycterus areolatus (Teleostei: Trichomycteridae) in
Rodríguez-Serrano E, Cancino R, Palma RE. 2006. Chile. Biological Journal of the Linnean Society 97: 876–892.
Molecular phylogeography of Abrothrix olivaceus (Rodentia: Vásquez M, Torres-Pérez F, Lamborot M. 2007. Genetic
Sigmodontinae) in Chile. Journal of Mammalogy 87: 971–980. variation within and between four chromosomal races of
Ruzzante DE, Walde SJ, Cussac VE, Dalebout ML, Liolaemus monticola (Tropiduridae) in Chile. Herpetological
Seibert J, Ortubay S, Habit E. 2006. Phylogeography of Journal 17: 149–160.
the Percichthyidae (Pisces) in Patagonia: roles of orogeny, Vera-Escalona I, D’Elia G, Gouin N, Fontanella F, Muñoz
glaciation, and volcanism. Molecular Ecology 15: 2949–2968. C, Sites J, Victoriano P. 2012. Lizards on ice: evidence for
Sallaberry-Pincheira N, Garin CF, González-Acuña D, multiple refugia in Liolaemus pictus (Liolaemidae) during
Sallaberry MA, Vianna JA. 2011. Genetic divergence of the last glacial maximum in the Southern Andean beech for-
Chilean long-tailed snake (Philodryas chamissonis) across ests. PLoS One 7: e48358.
latitudes: conservation threats for different lineages. Vianna JA, Medina-Vogel G, Chehébar C, Sielfeld W,
Diversity and Distributions 17: 152–162. Olavarría C, Faugeron S. 2011. Phylogeography of the
Sérsic AN, Cosacov A, Cocucci AA, Johnson L, Pozner R, Patagonian otter Lontra provocax: adaptive divergence to
Avila LJ, Sites JW, Morando M. 2011. Emerging phylo- marine habitat or signature of southern glacial refugia?
geographical patterns of plants and terrestrial vertebrates BMC Evolutionary Biology 11: 53.
from Patagonia. Biological Journal of the Linnean Society Victoriano PF, Ortiz JC, Benavides E, Adams BJ, Sites
103: 475–494. JW Jr. 2008. Comparative phylogeography of codistributed
Simon C, Frati F, Beckenbach A, Crespi B, Liu H, Flook species of Chilean Liolaemus (Squamata: Tropiduridae) from
P. 1994. Evolution, weighting and phylogenetic utility of the central-southern Andean range. Molecular Ecology 17:
mitochondrial gene sequences and a compilation of con- 2397–2416.
served polymerase chain reaction primers. Annals of the Victoriano PF, Vera I, Olmos V, Dib M, Inzunza B, Muñoz-
Entomological Society of America 87: 651–701. Ramírez C, Montoya R, Jara A, Habit E. 2012. Patrones
Singer BS, Ackert RP, Guillou H. 2004. 40Ar/39Ar and K–Ar idiosincráticos de diversidad genética de peces nativos del
chronology of Pleistocene glaciations in Patagonia. Geological Río San Pedro (Cuenca del Río Valdivia), un sistema de la
Society of America Bulletin 116: 434–450. región glaciada del sur de Chile. Gayana 76: 71–85.

© 2018 The Linnean Society of London, Biological Journal of the Linnean Society, 2018, 123, 712–727
 PHYLOGEOGRAPHY OF CRATOMELUS ARMATUS  727

Wallis GP, Waters JM, Upton P, Craw D. 2016. Transverse freshwater crab Aegla alacalufi in Chilean Patagonia.
Alpine speciation driven by glaciation. Trends in Ecology & Molecular Ecology 18: 904–918.
Evolution 31: 916–926. Zemlak TS, Habit EM, Walde SJ, Carrea C, Ruzzante
Xia X, Xie Z. 2001. DAMBE: software package for data ana- DE. 2010. Surviving historical Patagonian landscapes and
lysis in molecular biology and evolution. Journal of Heredity climate: molecular insights from Galaxias maculatus. BMC
92: 371–373. Evolutionary Biology 10: 67.
Xia X, Xie Z, Salemi M, Chen L, Wang Y. 2003. An index Zúñiga-Reinoso A, Jerez V, Avaria-Llautureo J,
of substitution saturation and its application. Molecular Hernández CE. 2016. Consequences of the Last Glacial
Phylogenetics and Evolution 26: 1–7. Maximum on Nyctelia confusa (Coleoptera: Tenebrionidae)
Xu J, Pérez-Losada M, Jara CG, Crandall KA. 2009. in Patagonia. Biological Journal of the Linnean Society 117:
Pleistocene glaciation leaves deep signature on the 705–715.

Downloaded from https://academic.oup.com/biolinnean/article-abstract/123/4/712/4925800 by guest on 23 July 2019

SUPPORTING INFORMATION
Additional Supporting Information may be found in the online version of this article at the publisher’s web-site:

Figure S1. Histograms resulting from the species delineation analyses depicting pairwise genetic distance
between individuals. Note that the histograms are bimodal for all three types of genetic distances, and the gap
between modes would indicate the gap between intraspecific (left mode) and interspecific (right mode) pairwise
comparisons.
Figure S2. Relationship between genetic diversity (π and Hd) and latitude, calculated for each locality of
Cratomelus armatus. A, B, results using all localities. C, D, results excluding localities with admixed lineages. All
these results included only localities with at least five individuals.
Figure S3. Outlier analyses for genetic diversity (π) for populations of Cratomelus armatus.
Figure S4. Relationship between genetic distance and latitude, calculated for southern areas of Cratomelus
armatus.
Figure S5. Bayesian skyline plot of estimated effective population size (logarithmic scale) across time (years)
for the north (A) and south (B) lineages of Cratomelus armatus. The black continuous lines represent the median
effective population size. The grey shaded areas are the 95% high posterior density (HPD) intervals. The dashed
vertical lines represent the approximated time span of the OIS 6 event, one of the most extensive glaciations in
the past 400 kya (Mortyn et al., 2003).

© 2018 The Linnean Society of London, Biological Journal of the Linnean Society, 2018, 123, 712–727