OUP user menu

Phylogeography of Loxodontomys micropus with comments on the alpha taxonomy of Loxodontomys (Cricetidae: Sigmodontinae)

Carola Cañón, Guillermo D'Elía, Ulyses F. J. Pardiñas, Enrique P. Lessa
DOI: http://dx.doi.org/10.1644/10-MAMM-A-027.1 1449-1458 First published online: 16 December 2010

Abstract

Loxodontomys micropus is a rodent that is widely distributed in Andean and Patagonian Argentina and Chile. This range covers a heterogeneous area that has been influenced by geologic and palaeoclimatic events, such as the glaciations during the Neogene. To investigate the genetic structure, phylogeographic pattern, and biogeographic history of this sigmodontine rodent we analyzed a 801-base-pair fragment of the mitochondrial genome (cytochrome-b gene) of 87 specimens from 24 localities from Argentina and Chile. Results indicate that L. micropus has a shallow genealogy that is geographically structured and is a taxon characterized by an historical population expansion. We discuss the distribution of the genetic variation of L. micropus in relation to population history and the concordance with other codistributed sigmodontine rodents. On the basis of molecular evidence, we suggest that the L. pikumche, corresponding to the second extant species of the genus, could be a junior synonym of L. micropus.

Key words
  • Andes
  • cytochrome b
  • genetic structure
  • Loxodontomys pikumche
  • Patagonia
  • phylogeography

Current species distributions are, in part, a result of changes and fluctuations in historical processes. The nature of these processes can be inferred from the study of current patterns of genetic variation within and among populations (Avise 2000). In this sense, southern South America is a geographic area of great interest because during the Neogene, it experienced a complex palaeoclimatic and geomorphological history. These include the uplift of the southern Andes, starting at the beginning of the Miocene (∼23 million years ago [mya]) and continuing to the present (Ramos 1989), and several glaciations during the Pliocene-Pleistocene (∼5 mya to 18,000 years ago) that produced multiple landscape and sea-level shifts. Geological evidence suggests that during the last glacial maximum (LGM) glaciers covered most of southern Chile and adjacent Argentina (Rabassa 2008), and extended to the north as a single continuous unit along the higher elevations of the Andes to approximately 38°S (Hulton et al. 2002). These events would have promoted the genetic differentiation of the biota of southern Argentina and Chile (Allnutt et al. 2003; Hinojosa and Villagrán 1997; Lessa et al. 2010; Morando et al. 2007; Palma et al. 2005; Rodríguez-Serrano et al. 2006; Ruzzante et al. 2008).

Loxodontomys micropus (Waterhouse 1837), the southern pericote, is a medium-size sigmodontine rodent that occurs in forest, shrubs, and forest-steppe shrub ecotonal areas along the Andes of Argentina and Chile from about 35°S to the Strait of Magellan, including Chiloé Island (Pardinas et al. 2008; Teta et al. 2009; Fig. 1). Most of this distribution is in the Patagonian region. In the last decade owl pellet analyses that have uncovered its presence in isolated populations to the east of the main range have been found in the Patagonian steppes of the Argentinean provinces of Chubut, Río Negro, and Santa Cruz (Teta et al. 2002; Udrizar Sauthier 2009; Udrizar Sauthier et al. 2008). Additionally, fossils have been found in archeological deposits of the Pleistocene and Holocene in Argentinean and Chilean Patagonia (Pardiñas 1998; Pearson and Pearson 1993; Teta et al. 2005; Udrizar Sauthier 2009). Some of these fossil localities lie outside of the current range of the species, showing that the distribution of L. micropus has fluctuated greatly in the recent past, including regional extinctions over more than 300 km (Pardiñas 1999; Udrizar Sauthier 2009).

Fig. 1

Geographic distribution (gray area) of Loxodontomys micropus (modified from Teta et al. 2009) showing the collecting localities of the specimens analyzed in this study. Locality numbers follow those of Appendix I. Polygons indicate the 4 main clades found in the genealogical analysis (Fig. 2).

Previous studies of codistributed sigmodontines, such as Abrothrix olivaceus (Rodríguez-Serrano et al. 2006; Smith et al. 2001), A. longipilis (Lessa et al. 2010), Oligoryzomys longicaudatus (Belmar-Lucero et al. 2009; Palma et al. 2005), and Phyllotis xanthopygus (Kim et al. 1998), showed that only A. longipilis and P. xanthopygus were structured phylogeographically. In contrast, low genetic divergence and geographical homogeneity was found within A. olivaceus and O. longicaudatus across Patagonia.

A recent multispecies survey of the phylogeography of Patagonian sigmodontines (Lessa et al. 2010) included a few specimens of L. micropus and failed to detect significant geographical structure. We documented the pattern of geographical structure and demographic history of the species on the basis of a more complete coverage of its geographical distribution. Additionally, we assessed the recent evolutionary history of L. micropus with respect to variation in other codistributed sigmodontines. We also suggest the need of taxonomic revision for the genus Loxodontomys Osgood, 1947.

Materials and Methods

Sampling.—Geographic coverage spans most of the species range with samples from 87 specimens of L. micropus collected at 24 localities in Argentina and Chile (Fig. 1; Appendix I). Animal care and use procedures followed guidelines approved by the American Society of Mammalogists (Gannon et al. 2007). Sequences of two specimens were retrieved from GenBank (AF108690 and AY275122). Ten specimens from the northern part of the Argentinean and Chilean range currently are assigned to L. pikumche Spotorno et al., 1998, the other extant species in the genus.

Laboratory procedures.—We extracted genomic DNA using the protocol of Wizard Genomic DNA Purification System (Promega, Madison, Wisconsin). We then amplified a 801-base-pair fragment of the mitochondrial genome corresponding to the first portion of the protein-coding cytochrome-b gene using primers MVZ05 and MVZ16 (da Silva and Patton 1993) with the following conditions: 94°C for 3 min; 35 cycles of 94°C for 20 s, 45°C for 15 s, 72°C for 60 s, and 72°C for 7 min. All reactions included a negative control to check for contamination. Amplicons were purified and sequenced by Macrogen, Inc. (Seoul, Korea). All sequences were deposited in GenBank (GU553838-GU553922).

Alignment, descriptive, and genealogical analysis.—Sequence alignment was performed with Clustal X (Thompson et al. 1997) using the default parameter values. Haplotype and population comparisons were performed with MEGA 4.0 (Tamura et al. 2007) in the form of observed genetic distances. Haplotype and nucleotide diversity were assessed with Arlequin (Schneider et al. 2000).

Only nonredundant haplotypes previously selected with the program DNAsp (Librado and Rozas 2009) were used for genealogical reconstructions. Haplotype relationships were inferred using Bayesian inference. Genealogies were rooted using the outgroup criterion with sequences of the closely related phyllotines (D'Elía 2003) Auliscomys pictus, Graomys griseoflavus, and Phyllotis anitae that are on GenBank (APU03545, AY275117, and AY627299, respectively). Bayesian analysis was conducted in MrBayes 3.1 (Ronquist and Huelsenbeck 2003), with 2 independent runs, each consisting of 3 hot and 1 cold Markov chains. The model used included 6 categories of base substitution, a gamma-distributed rate parameter, and a proportion of invariant sites estimated in MrBayes. Uniform-interval priors were assumed for all parameters except base composition and generalized time-reversible parameters, which assumed a Dirichlet prior process. Runs were allowed to proceed for 6,000,000 generations with trees sampled every 500 generations per chain. To check for convergence on a stable log-likelihood value we plotted the log-likelihood values against generation time. The first 25% of the trees were discarded as burn-in, and the remaining trees were used to compute a 50% majority rule consensus tree and obtain posterior probability (PP) estimates for each clade.

Population structure.—FST values, previously obtained in Arlequin (Schneider et al. 2000), were plotted in multidimensional space using nonmetric multidimensional scaling (MDS—Lessa 1990) performed in the XLSTAT software (Addinsoft 2009). Given the nature of the data, the analysis was implemented with the ordinal 2 model and from 3 to 2 dimensions to evaluate the distortion associated with the decrease of dimensions. A Kruskal stress value <0.1 suggests the occurrence of significant associations in the multidimensional space (Ludwig and Reynolds 1988). This analysis was used to visualize and test the existence of discrete groups as a function of genetic similarities and geographic proximity and to evaluate the concordance with main clades recovered in the genealogy.

XLSTAT software also was used to assess the relationship between FST values and geographic distances among population pairs via a Mantel test. Geographical distances were determined in kilometers from the latitudinal and longitudinal coordinates using a distance calculator available from the National Weather Service (http://www.nhc.noaa.gov/gccalc.shtml). The significance of the relationship between genetic and geographic distances was tested using a nonparametric permutation approach (Addinsoft 2009). A pattern of isolation by distance is anticipated when populations have been stable over a long period and when gene flow is higher among geographically proximate populations.

As an additional method to determine how the genetic variation was geographically structured, we performed hierarchical analyses of molecular variance (AMOVA— Excoffier et al. 1992) implemented in Arlequin (Schneider et al. 2000). Haplotypes were grouped in 4 different hierarchical schemes according to the genealogical results and geography. These analyses partition the total variance observed into covariance components that result from the variance among groups, among populations within groups, and within populations. The covariance components then are used to calculate fixation indices that are equivalent to Wright's F statistics (Wright 1978). The significance of the fixation indices is tested using a nonparametric permutation approach (Excoffier et al. 1992).

Molecular diversity and historical population dynamics.— To infer population demographic history we conducted several analyses in Arlequin. Frequency distributions of pairwise differences between haplotype pairs (Rogers and Harpending 1992) and the associated sum-of-square deviations (SSD) and raggedness indices were calculated. A unimodal distribution is expected when the sample is drawn from populations that have experienced recent demographic expansion, whereas samples from populations at demographic equilibrium typically show multimodal distributions. We calculated Fu's Fs and Tajima's D statistical tests of neutrality (Fu 1997; Tajima 1989), which are based on an infinite-sites model without recombination. These tests evaluate the probability of observing a random neutral sample with a number of alleles that is equal to or smaller than the observed number of alleles. Negative and significant Fs and D values are obtained from samples corresponding to populations that have undergone recent demographic expansion or show certain departures from neutrality.

Results

Mitochondrial diversity.—We analyzed sequences from 87 specimens of L. micropus from 24 localities (Fig. 1) and identified 72 polymorphic sites that define 68 haplotypes. Overall haplotype diversity is 0.99 (±0.009 SD) and nucleotide diversity of 0.011 ± 0.0056. Two haplotypes (H1 and H39) are shared among individuals from pairs of localities in Argentina and Chile. Haplotype H1 is shared among individuals collected over a large geographic area: Las Leñas, Mendoza Province, and La Angostura and Laguna Epulafquen, Neuquén Province, in Argentina; and Lircay, Región del Maule and Quilleco, Región del Bio Bio in Chile. This shared haplotype was recovered from individuals previously assigned to L. pikumche and specimens from the distributional range of L. micropus. Haplotype H39 is present in individuals collected in southern Patagonia at Estancia Ensenada, Santa Cruz Province, Argentina and El Manzano, Aysén, Chile.

Average pairwise difference among haplotypes is 1.1%. The observed divergence value among Argentinean samples is 1.12%, and Chilean samples differ, on average, by 0.97%. Intrapopulational divergence ranges between 0.08% and 1.13% for Quilleco and Lucaschewsky localities, respectively, both in Chile (Fig. 1).

Genealogical reconstruction.—The recovered genealogy is shallow, but geographically structured (Fig. 2). Four main, but weakly supported, clades replace each other latitudinally. The first clade (North 1: N1; 0.58 PP) contains 17 haplotypes from the northern Argentinean (Chubut, Mendoza, and Neuquén provinces) and Chilean (Araucanía, Bio Bio, Maule, and Los Lagos regions) distributional range. Haplotypes from specimens collected at populations currently assigned to L. pikumche are in this clade. The second clade (North 2: N2; 0.94 PP) consists of 10 haplotypes recovered from specimens collected in northern Argentinean Patagonia (Chubut and Río Negro Provinces). The third clade (South 1: S1; 0.8 PP) includes 11 haplotypes from specimens collected in central Argentinean (southern Chubut) and Chilean (Aysén) Patagonia. The fourth clade (South 2: S2; 0.8 PP) includes 30 haplotypes recovered from specimens from southern Argentinean (Santa Cruz) and Chilean (Aysén and Magallanes) Patagonia. Clades N1 and N2 overlap at El Maitén in northern Chubut Province. Similarly, clades S1 and S2 overlap at El Manzano, Aysén. Intraclade average divergence is low and similar in the four clades, ranging from 0.31% in clade N1 to 0.73% for clade S2 (Table 1).

Fig. 2

Genealogical relationship of 68 haplotypes of Loxodontomys micropus. Bayesian posterior probabilities (>0.50) indicating nodal support of the main clades are shown above branch lines. * indicates haplotypes recovered from specimens previously assigned to L. pikumche (see text). indicates haplotypes recovered from topotypic specimens of L m. alsus. Numbers within parentheses indicate the number of specimens sharing each haplotype. Numbers within brackets indicate localities from where haplotypes were recovered (see Appendix I and Fig. 1).

View this table:
Table 1

Genetic variation of cytochrome b for the 4 main clades of Loxodontomys micropus found in the genealogical analysis (Fig. 2). Locality numbers are those of Fig. 1 and Appendix 1. Sample size (n), number of haplotypes found (h), nucleotide diversity (π), haplotype diversity (Hd), and within- and between-clade observed genetic distances (p) for each clade.

Genetic distance
Between groups
CladeLocalitiesnh%SD)Hd (±SD)Within groupsNlN2SI
N11, 2, 3, 4, 6, 7, 8, 13, 14, 15, 16, 1725171.000 ± 0.01130.0031 ± 0.001930.31
N25, 610101.000 ± 0.04470.0038 ± 0.002490.391.13
S19, 10, 1812111.000 ± 0.03400.0038 ± 0.002400.381.171.34
S211, 12, 18, 19, 20, 21, 22, 23, 2440301.000 ± 0.00560.0073 ± 0.003950.731.371.501.30

Relationships among the four clades are poorly resolved. At the base of the Lodoxontomys clade is a trichotomy formed by clades N1, N2, and a clade (0.52 PP) formed by clades S1 and S2. Divergence among clades is also low and ranges from 1.13% for the pair N1−N2 to 1.3% for the comparison between S1 and S2 (Table 2). Maximum parsimony analysis recovers the same four main clades, but relationships among them are slightly different: ([{S1, S2} N1] N2) and poorly supported. Values of haplotype (Hd) and nucleotide diversity (π) for each of the main clades identified in the phylogenetic analysis are presented in Table 1.

View this table:
Table 2

Results of 4 analyses of molecular variance (AMOVA) with different arrangements of populations into groups.

Nl versus N2 versus SI versus S2(NI + N2) versus (SI versus S2)Nl versus N2 versus (SI + S2)Argentina versus Chile
AMOVAVariance componentsPercentage explainedVariance componentsPercentage explainedVariance componentsPercentage explainedVariance componentsPercentage explained
Among groups3.1858.812.0737.432.6647.340.028450.63
Among populations within groups0.5379.941.7732.031.2722.652.6393958.06
Within populations1.6931.251.6930.531.6930.011.8782341.32
0.687*0.694*0.699*0.587*
  • * indicates P < 0.0001

Geographic structure and demography.—Multidimensional scaling recovers 2 nonoverlapping populations from dimensions 1 and 3 (Fig. 3). One group includes populations from the northern half of the range that in the genealogical analysis form clades N1 and N2. The second group includes those southern populations that form clades S1 and S2. Populations in which clades N1 and N2 (El Maitén, Chubut Province) and clades S1 and S2 (El Manzano, XI region) overlap fall in an intermediate position within their respective groups. Populations from each of the four main clades tend to group together, agreeing with the genealogy.

Fig. 3

Nonmetric multidimensional scaling analysis on the basis of FST values among 24 localities of Loxodontomys micropus (Kruskal stress value = 0.184). Solid lines delimit the geographic groups discussed in the text. Populations marked by arrows are those in the genealogical analysis (Fig. 2) that fall in 2 different clades of the 4 main clades that were recovered. Locality numbers are those of Appendix I.

The Mantel test indicates no correlation between genetic and geographical distances among populations of L. micropus (r = 0.059, n = 24, P = 0.256). Therefore, these populations do not reflect a pattern of isolation by distance.

Results of 4 different geographical arrangements of AMOVA (Table 2) show that the variation among groups is maximized (58.8%) when localities are grouped according to the 4 main clades recovered in the genealogical analysis (i.e., N1 versus N2 versus S1 versus S2). This component has a minimum value (0.63%) when localities are grouped into an Argentinean and a Chilean group (i.e., eastern and western groups). When localities are grouped into 2 (i.e., [N1 + N2] versus [S1 +) or 3 (i.e, N1 versus N2 versus [S1 +) groups, variation explained by differences among groups has intermediate values (37.43% and 47.34%, respectively). Furthermore, percentage of variation explained by differences among populations within groups is minimized (9.94%) when the localities are grouped into the 4 main genealogical groups and maximized (58.06%) when populations are pooled into an Argentinean and a Chilean group.

Mismatch distributions (Fig. 4) for groups N1, N2, and S2 are unimodal, suggesting recent historical population growth.

Fig. 4

Observed and expected patterns of the mismatch distribution of the 4 main clades of Loxodontomys micropus (Fig. 2). The y axis shows the frequency values of the number of nucleotide differences within haplotype pairs. For additional details see Table 3.

Although the S1 group appears slightly bimodal, the tests of the observed SSD and raggedness index failed to reject the hypothesis of recent population expansion. Similarly, Tajima's and Fu's neutrality test values are negative and, in the latter case, significant (Table 3), suggesting a population expansion for L. micropus along the entire distributional range.

View this table:
Table 3

Demographic parameters for the 4 main clades of haplotypes of Loxodontomys micropus recovered in the genealogical analysis (Fig. 2), including estimates of Tajima's D (Tajima 1989), Fu's Fs (Fu 1997) neutrality tests, and parameters of the mismatch distributions: sum-of-square deviations (SSD), Harpending's raggedness index, and P (value of the simulated raggedness ≥ observed raggedness) of goodness-of-fit test (Rogers and Harpending 1992; Schneider and Excoffier 1999).

CladeTajima DFu's FsSSDHarpending's raggedness indexP
N1−1.14927−26.6472*0.008630.024930.95
N2−1.20399−8.12019*0.010460.049380.74
S1−1.23917−11.21354*0.142650.026171.00
S2−1.14939−25.31466*0.005340.013540.40

Discussion

Phylogeographic structure.—Loxodontomys micropus is a phyllotine rodent widely distributed along the central and southern Andes of Argentina and Chile. As also reported by Lessa et al. (2010), limited divergence exists among haplotypes of this species. However, we document haplotypes from 4 main clades that are latitudinally differentiated (Figs. 1 and 2). The phylogeny shows a monophyletic southern group, which is composed of 2 main clades (S1, S2) and 2 clades (N1, N2) restricted to the northern part of the distribution. Relationships among S1−S2, N1, and N2 are unresolved. Of these 4 clades only clade N2 is restricted to 1 side of the Andes in the eastern Argentinean provinces of Ríio Negro and Chubut. The 3 others clades (N1, S1, S2) are distributed across the Andes in both Argentinean and Chilean localities. Results of an AMOVA, in which haplotypes and localities were grouped into eastern and western groups (Table 2), indicate that differences between both groups is minimal (0.63%) and as such, most of the variation lies among populations within the groups (58.06%), followed by differences within localities (41.32%). Thus, the Andean mountain range does not play a relevant role in structuring the genetic variation of L. micropus. Similar results have been found in the codistributed sigmodontines Abrothrix longipilis (Lessa et al. 2010), A. olivaceus (Rodríguez-Serrano et al. 2006; Smith et al. 2001), and O. longicaudatus (Palma et al. 2005).

Studies on A. olivaceus (Rodríguez-Serrano et al. 2006; Smith et al. 2001) and O. longicaudatus (Palma et al. 2005) have shown that these species have reduced genetic divergence lacking phylogeographic structure, suggesting that they originated from a single refugium. Their current distributional area was covered by ice during the LGM (Hulton et al. 2002). In the case of A. olivaceus the refuge was suggested to be located in central Chile, and the species would have dispersed south and east of the Andes after the ice shield melted (Smith et al. 2001). For O. longicaudatus Palma et al. (2005) suggested a scenario where the species would have crossed the Andes at southern latitudes from Argentina to Chile with a latter expansion to the north; however, the evidence for this is not conclusive. In contrast, genetic variation of L. micropus is geographically structured. Therefore, it is possible to suggest that L. micropus colonized its current geographical range from more than 1 refuge. The number of refuges to be invoked would vary depending on whether contact zones between clades N1 and N2 and between clades S1 and S2 are considered areas of primary or secondary contact. Currently, it is not possible to distinguish between these 2 scenarios.

The geographic placement of refugia of L. micropus remains unclear; however, several lines of evidence suggest that refugia might have been outside of our sampling area. First, much of the range that we sampled was mostly covered by ice during the LGM (Hulton et al. 2002; Rabassa 2008), although recent studies indicate that the ice sheet was not continuous, allowing the existence of refugia surrounded by ice (Xu et al. 2009). Second, the 4 main clades show signs of recent population expansion (Table 3, Fig. 4), suggesting that the sampled populations are of recent origin. Third, the fossil record (Pardiñas and Teta 2008) shows that L. micropus has experienced shifts in its range during the Pleistocene and Holocene, previously occupying areas that now are locally extinct. Similarly, isolated populations in the Argentinean Patagonian steppe (Teta et al. 2009) may be considered relicts of a past, more extensive eastern distribution. In the area of Chacays, a vast basaltic plateau in central Patagonia, a single recording locality of L. micropus is about 250 km east of the present continuous range (Udrizar Sauthier et al. 2008).

On the basis of the information just described, we suggest the following scenario for the recent history of L. micropus. During the LGM the range of L. micropus was fragmented into at least 2, and possibly 4, areas. The location of these isolates is uncertain, but they likely were outside the area surveyed in this study, probably in the Argentinean Patagonian steppe to the east of the current range. However, the possibility of a refugium in central Chile, north of the sampled area, cannot be disregarded. Later, L. micropus colonized its current range after the ice retreated. Under this scenario, current steppe isolates (Teta et al. 2009; Fig. 1), which were not sampled in this study because they are based only on remains recovered from owl pellets, are interpreted as relicts of a past distribution that as a whole was more to the east than the current distribution. If this is the case, these isolates are expected to be more structured and show less or no evidence of demographic expansion, relative to populations in recently colonized areas. Therefore, the inclusion of these isolates would constitute a partial test of this biogeographic scenario. This, together with the analysis of unlinked loci, would help to further our understanding of the effect of the Neogene climatic fluctuations on L. micropus and, by extension, of the Patagonian mammal assemblage.

Taxonomic implications.—Results of the present study have several implications relative to the alpha diversity of Loxodontomys. The first relates to the distinctiveness and validity of L. pikumche, and the second concerns the 3 currently recognized subspecies of the southern pericote.

Loxodontomys pikumche was described by Spotorno et al. (1998) mostly on the basis of presumably fixed karyotypic differences with L. micropus. Although L. micropus has a diploid complement of 2n = 34 and a fundamental number of FN = 36, L. pikumche has 2n = 32 and FN = 34. Morphological differences between both forms are subtle; Spotorno et al. (1998) identified a few dental traits (e.g., incisor orientation varying from hyperopisthodont to opisthodont, upper incisor dentine fissure long or comma-shaped) as differing in both taxa. However, cranial and dental variability within populations of Loxodontomys is high, as noted by Hershkovitz (1962), Pearson (1995), and Spotorno et al. (1998). Several populations assigned to L. micropus have individuals with character states used to diagnose L. pikumche. For example, populations of L. micropus studied by Pearson (1995) include specimens with molar root numbers considered to be diagnostic of L. pikumche. Both forms are distributed parapatrically in central and southern Chile; however, their geographic limits have not been clearly established. They supposedly replace each other at about the latitude of Chilian (∼35.5°S) in central Chile (Spotorno et al. 1998). Recently, Novillo et al. (2009) reported the first putative Argentinean record of L. pikumche. Specimens were collected in Las Lenas, Mendoza province. They have a third karyomorph, 2n = 32 and FN = 32, but were referred to L. pikumche on the basis of morphological characters. However, none of the dental features used by Spotorno et al. (1998) to diagnose L. pikumche was discussed. In addition, the character state added by Novillo et al. (2009) to discriminate between L. pikumche and L. micropus was based on the examination of 1 specimen of each taxon and did not take into account intrapopulation variation (e.g., zygomatic plate morphology—Hershkovitz 1962).

Our sampling includes 10 specimens assignable to L. pikumche from 1 Argentinean (n = 1: Las Leñas, Mendoza) and 2 Chilean (n = 1: Las Trancas, VIII Region; n = 8: Lircay, VII Region) localities. The Argentinean specimen is one used by Novillo et al. (2009) to cite the species for Argentina. Among these 10 specimens 5 haplotypes were recovered. One of these haplotypes was recovered not only from specimens from the three localities of L. pikumche but also from individuals morphologically recognized as L. micropus from the Argentinean localities of Laguna Epulafquen, and La Angostura localities, in the Neuquén Province. This widely distributed haplotype falls in clade N1 and is remarkably similar (0.05%) to the other haplotypes assignable to L. pikumche and to those of L. micropus forming the same clade. We cannot rule out that haplotype sharing among taxa may be the result of a selective process or introgression after hybridization. However, given the marked morphological similarities between L. pikumche and L. micropus, we suggest that both taxa might represent the same biological species and as such the former may constitute a junior synonym of the latter. Under this scenario, L. micropus would be an example of another known karyotypically variable sigmodontine species in addition to, among others, Akodon cursor (Geise et al. 1998; Nogueira and Fagundes 2008), Holochilus brasiliensis (Nachman 1992), and O. flavescens (Weksler and Bonvicino 2005).

Three subspecies are currently recognized within L. micropus (Teta et al. 2009). The type locality of L. m. micropus (Waterhouse 1837) was recorded as “Interior plains of Patagonia in lat. 50°, near the banks of the Santa Cruz” (Waterhouse 1839), Santa Cruz Province, Argentina. However, uncertainty remains about the exact location where Charles Darwin collected the holotype (Teta et al. 2009). The second subspecies is L. m. alsus (Thomas 1919), and the holotype comes from “Maitén, W. Chubut. 700 meters” (Thomas 1919), Chubut Province, Argentina (=El Maitén; 42°03′S, 71°10, W). Finally, L. m. fumipes (Osgood 1943) was described from “Quellón, Chiloé Island, Chile” (Osgood 1943—43°07′S, 73°35′W, Region de Los Lagos). Differences among these 3 forms related mostly to pelage color, and their degree of differentiation has not been assessed with contemporary approaches. Similarly, the geographic distribution of these forms is unclear.

Our sampling includes haplotypes from topotypic specimens of L. micropus alsus and from near the type locality of L. m. micropus (i.e., Estancia Ensenada and Lago Viedma localities, Santa Cruz). Unfortunately, we were unable to include specimens from Chiloé Island. Haplotypes assignable to each nominal form fall in different clades. Those of L. m. alsus fall in clades N1 and N2 of the northern group, and those of L. m. micropus belong to clade S2 of the southern group. Therefore, our molecular data indicate that L. m. micropus and L. m. alsus could represent distinct evolutionary units within the main lineage of L. micropus that might be considered subspecies. Under this taxonomic arrangement L. pikumche is a synonym of L. m. alsus.

Final considerations.—Additional studies combining broader geographic coverage (especially Chiloé Island, central Chile and isolated eastern populations) and the analysis of nuclear loci, karyotypes, and morphological characters are needed. Such investigations will be important in determining the validity of the biogeographic and taxonomic hypotheses presented in this study.

Acknowledgments

We thank the following people and institutions for providing tissue samples: Agustina Novillo and Agustina Ojeda (CRICYT), Joe Cook (MSB), Jonathan Guzmán (UdeC), Milton Gallardo (IEEUACH), Richard Sage (MVZ), and Robert Baker (TTU). Assistance during field work was provided by Daniel Udrizar Sauthier, Jonathan Guzman, Dario Podestá, Marta Lizarralde and her team, and Sebastian Cirignoli. The suggestions of 2 anonymous reviewers contributed to the improvement of this study. Financial support was given by NGS 7813-05, CONICET PIP 6179, Agencia PICT 2008-0547, and FONDECYT 11070157.

Appendix I

List of specimens of Loxodontomys micropus used in the present study, including: collection locality, identification number, GenBank accession number (those 2 retrieved from GenBank are indicated by **), haplotype number (see Fig. 2), and geographic groups (N1: North 1; N2: North 2; S1: South 1; S2: South 2). Locality numbers correspond to those of Fig. 1. Museum collection acronyms and personal field numbers are as follows. Argentina: Colección de Mamíferos del Centro Nacional Patagónico, Puerto Madryn (CNP); Agustina Novillo (AN, vouchers will be deposited at Colección Mastozoológica del Instituto Argentino de Zonas Áridas, IADIZA). Chile: Instituto de Ecología y Evolución, Universidad Austral de Chile, Valdivia (IEEUACH); Guillermo D'Elía (GD, vouchers will be deposited at Museo de Zoología, Universidad de Concepción); Jonathan Guzmán (JG, voucher will be deposited at Museo de Zoología, Universidad de Concepción). United States: Museum of Southwestern Biology, Albuquerque (NK: tissue accession prefix);

The Museum, Texas Tech University, Lubbock (TK: tissue accession prefix); Richard D. Sage (RDS, vouchers will be deposited at the Museum of Vertebrate Zoology, Berkeley).

ARGENTINA: 1) Mendoza Province, Las Leñas, 35°06′S, 70°05′W (AN 60: GU553838: 1: N1). 2) Neuquén Province, Laguna Epulafquen, 36°49′S, 71°04′W (CNP 1478: AY275122**: 1: N1). 3) Neuquén Province, Alumine, 3.8 km W, 0.7 km N Quillén, 39°22′S, 70.95°W (RDS 18031: GU553839: 7: N1). 4) Neuquén Province, Los Lagos, 12.2 km N, 4.8 km W, Villa La Angostura, 40°39′S, 71°42′W (RDS 18113: GU553840: 1: N1). 5) Río Negro Province, Bariloche, 41°14′S, 71°32′W (MVZ 182661: AF108690**: 22: N2; RDS 17829: GU553841: 26: N2; RDS 17830: GU553842: 27: N2; RDS 17834: GU553843: 20: N2; RDS 17839: GU553844: 21: N2; RDS 18129: GU553845: 23: N2; RDS 18140: GU553846: 24: N2). 6) Chubut Province, Estancia El Maitén, 42°03′S, 71°09′W (CNP 1980: GU553847: 18: N2; CNP 1981: GU553848: 25: N2; CNP 1982: GU553849: 10: N1; CNP 1983: GU553850: 19: N2). 7) Chubut Province, Laguna Larga, 42°53”S, 71°34”W (CNP 1046: GU553851: 15: N1). 8) Chubut Province, Futaleufú, 5.5 km S and 2.8 km W Villa Futalaufquén, 42°56′S, 71°34′W (RDS 17861: GU553852: 12: N1; RDS 17866: GU553853: 16: N1; RDS 17894: GU553854: 11: N1; RDS 18069: GU553855: 17: N1; RDS 18078: GU553856: 13: N1). 9) Chubut Province, Lago Blanco, W-SW end, 45°55′S, 71°19′W (CNP 299: GU553857: 32: S1; CNP 312: GU553858: 35: S1; CNP 317: GU553859: 32: S1; CNP 398: GU553860: 30: S1; CNP 419: GU553861: 34: S1; CNP 1221: GU553862: 37: S1; CNP 1984: GU553863: 33: S1). 10) Chubut Province, Lago Blanco, Estancia Valle Huemules, 45°56′S, 71°31′W (CNP 399: GU553864: 29: S1; CNP 482: GU553865: 28: S1; CNP 1165: GU553866: 38: S1; CNP 1585: GU553867: 31: S1). 11) Santa Cruz Province, Estancia La Ensenada, 48°21′S, 72°05′W (CNP 536: GU553868: 64: S2; CNP 1985: GU553869: 47: S2; CNP 1986: GU553870: 60: S2; CNP 1987: GU553871: 61: S2; CNP 1988: GU553872: 48: S2; CNP 1989: GU553873: 64: S2; CNP 1990: GU553874: 48: S2; CNP 1991: GU553875: 39: S2; CNP 1992: GU553876: 64: S2; CNP 1993: GU553877: 64: S2; CNP 1994: GU553878: 39: S2; CNP 1995: GU553879: 49: S2; CNP 1996: GU553880: 64: S2). 12) Santa Cruz Province, Lago Viedma, Bahía Túnel, 49°12′S, 72°58′W (CNP 1997: GU553881: 63: S2).

CHILE: 13) VII región del Maule, Altos del Lircay, 35°35′S, 71°02′W (GD 1301: GU553882: 2: N1; GD 1302: GU553883: 3: N1, GD 1303: GU553884: 3: N1; GD 1304: GU553885: 3: N1; GD 1305: GU553886: 4: N1; GD 1328: GU553887: 3: N1; GD 1329: GU553888: 5: N1; GD 1330: GU553889: 1: N1). 14) VIII region del Bio Bío, Chilián, Las Trancas, 36°55′S, 71°30′W (GD 1300: GU553890: 1: Nl). 15) VIII región del Bío Bío, Quilleco, Hacienda San Lorenzo, 37°32′S, 71°27′W (NK 120134: GU553891: 6: N1; NK 120138: GU553892: 6: N1; NK 120139: GU553893: 8: N1). 16) IX región de la Araucanía, Curacautín, Piedra Santa, Malalcahuello, 38°28′S, 71°35′W (IEEUACH 4897: GU553894: 9: N1). 17) X región de los Lagos, Palena, El Encuentro, 43°37′S, 71°45′W (NK 129189: GU553895: 14: Nl). 18) XI región de Aysén del General Carlos Ibáñez del Campo, Aysén, El Manzano, 47°09′S, 72°39′W (GD 928: GU553896: 36: S1; GD 929: GU553897: 62: S2; GD 930: GU553898: 42: S2; GD 944: GU553899: 54: S2; GD 997: GU553900: 44: S2; GD 998: GU553901: 56: S2; GD 1006: GU553902: 45: S2; JG 085: GU553903: 46: S2). 19) XI región de Aysén del General Carlos Ibáñez del Campo, Aysén, Predio Lucaschewsky, 47°11′S, 72°44′W (JG 023: GU553904: 40: S2; JG 086: GU553905: 55: S2). 20) XI región de Aysén del General Carlos Ibáñez del Campo, Aysén, Sur de Cochrane, 47°19′S, 72°36′W (GD 963: GU553906:43: S2; GD 1000: GU553907: 39: S2). 21) XI región de Aysén del General Carlos Ibáñez del Campo, Aysén, Confluencias ríos Colonia-Baker, 47°18′S, 72°50′W (JG 084: GU553908: 41: S2). 22) XI región de Aysén del General Carlos Ibáñez del Campo, Aysén, Sector Barrancoso, 47°29′S, 72°48′W (GD 967: GU553909: 55: S2). 23) XII región de Magallanes, Torres del Paine, 50°56′S, 72°58′W (TK 110006: GU553910: 52: S2; TK 110007: GU553911: 59: S2; TK 110009: GU553912: 58: S2; TK 110027: GU553913: 50: S2; TK 110028: GU553914: 53: S2; TK 110033: GU553915: 65: S2; TK 110037: GU553916: 57: S2; TK 110042: GU553917: 51: S2). 24) XII región de Magallanes, Punta Arenas, Puerto del Hambre, 53°36′S, 70°56′W (NK 129278: GU553918: 66: S2; NK 129295: GU553919: 67: S2; NK 129297: GU553920: 67: S2; NK 129298: GU553921: 68: S2; NK 129308: GU553922: 67: S2).

Footnotes

  • Associate Editor was Burton K. Lim.

Literature Cited

View Abstract