MOLECULAR PHYLOGEOGRAPHY OF ABROTHRIX OLIVACEUS (RODENTIA: SIGMODONTINAE) IN CHILE

Transcription

1 Journal of Mammalogy, 87(5): , 2006 MOLECULAR PHYLOGEOGRAPHY OF ABROTHRIX OLIVACEUS (RODENTIA: SIGMODONTINAE) IN CHILE ENRIQUE RODRÍGUEZ-SERRANO, RICARDO A. CANCINO AND R. EDUARDO PALMA* Centro de Estudios Avanzados en Ecología y Biodiversidad and Departamento de Ecología, Facultad de Ciencias Biológicas, Pontificia Universidad Católica de Chile, Casilla 114-D, Santiago , Chile Abrothrix olivaceus is the most abundant and widespread species of sigmodontine rodent in Chile. We evaluated phylogeographic relationships within A. olivaceous based on analyses of 67 specimens collected along most of the distributional range of the species. We used nucleotide sequence data from the hypervariable domain I of the mitochondrial DNA control region. Examination of our data supports a northern origin for this taxon, followed by southward dispersal. Also, we propose the occurrence of a new subspecies of A. olivaceus in the northernmost limit of the species range. We detected a significant influence of recent paleoclimatic events on the current distribution of A. olivaceus, particularly in south-central Chile. The results also support the genetic distinctness of nominal subspecies proposed in the classical literature. Key words: Abrothrix olivaceus, Andean clade, Chile, mitochondrial DNA control region, phylogeography, Quaternary glaciations, Sigmodontinae The subfamily Sigmodontinae (Muroidea) is the 2nd largest mammalian subfamily and includes a vast array of New World mice (Reig 1980; Steppan et al. 2004). The evolutionary history of sigmodontines has been of great interest because of their adaptive radiation, systematic relationships, and diversity of forms (Baskin 1978; D Elía 2003; Engel et al. 1998; Hershkovitz 1966; Patterson and Pascual 1972; Smith and Patton 1993, 1999). One hypothesis proposes a tropical origin for these mice in North America, followed by dispersal to South America over the Panamanian Land Bridge during Plio Pleistocene times (Baskin 1978; Patterson and Pascual 1972; Simpson 1980; Webb 1991). Alternatively, Hershkovitz (1966, 1972), Savage (1974), and Reig (1978, 1986) proposed an early arrival to South America about 20 million years ago, with further diversification in situ. Recent calibrations based on molecular phylogenies suggested an origin before the Panamanian Land Bridge, 5 10 million years ago (Engel et al. 1998; Smith and Patton 1999; Spotorno 1986), with further diversification in South America (Steppan et al. 2004). Despite these uncertainties, it is clear that sigmodontine rodents were widespread in South America by Pleistocene times, occupying most available biomes, particularly in the Andean area where major centers for the diversification of Sigmodontinae have been postulated (Reig 1986; Smith and Patton 1999). * Correspondent: epalma@bio.puc.cl Ó 2006 American Society of Mammalogists The Andean clade (Smith and Patton 1999) is one of the sigmodontine lineages, a group of mice that includes several species formerly assigned to Akodontini and that are mainly restricted to the southwest of the Andes. The Andean clade is thought to be a monophyletic assemblage, but its tribal classification has not been formalized (D Elía 2003; Smith and Patton 1999). Abrothrix olivaceus (Waterhouse, 1837) is one of the most abundant forms within the Andean clade (Reig 1986), commonly known as ratón oliváceo ( olive field mouse ). The species is a small mouse with a short tail (shorter than the head and body), small ears, and grayish brown hair mixed with olive hair (Mann 1978). Traditional literature recognizes the distribution of A. olivaceus along most of Chile and Patagonian Argentina, from 188S to 548S latitude (Mann 1978; Osgood 1943). Thus, the species ranges from the Coastal Desert in the north, southward throughout the Mediterranean region, temperate forests, and Patagonian forests (including Argentina in the latter 2 regions). The species has been recorded from sea level to 1,000 m in central Chile. Mann (1978) proposed that A. olivaceus occurs in Chile from Tarapacá southward to Aisén and recognized 5 subspecies (Fig. 1): A. olivaceus olivaceus from Arica southward to Talca Province; A. olivaceus pencanus from Talca to the Cautín River; A. olivaceus mochae on Mocha Island (388229S, W); A. olivaceus brachiotis in the Valdivian rain forests and southern archipelagoes southward to the Aisén River in the XI Region; and A. olivaceus beatus in the Argentinean provinces of Neuquén, Río Negro, and Chubut. Abrothrix xanthorhinus was recently synonymized with A. olivaceus (Smith et al. 2001), as A. o. xanthorhinus in the Patagonia of 971

2 972 JOURNAL OF MAMMALOGY Vol. 87, No. 5 Chile and Argentina. This was proposed based on morphological differences and genetic similarity across a west east gradient at 418S between the southern coast of Chile and Río Negro in Argentina (Pearson and Smith 1999; Smith et al. 2001). Subspecies have been distinguished on the basis of hair coloration, tail length, and size and shape of tympanic bullae (Mann 1978; Osgood 1943). However, these characters seem insufficient to determine species boundaries within Abrothrix. For example, the morphological features that have traditionally been used to differentiate A. olivaceus from A. xanthorhinus (i.e., external measurements and pelage coloration) do not diagnose these forms as different species (Pearson and Smith 1999). The distribution of A. olivaceus might have been affected by geological events such as the Andean uplift in Plio Pleistocene times that led to the formation of dry, sparsely vegetated areas such as the Andean Altiplano (Puna) and the Atacama and Coastal deserts. In fact, the small mammal fauna of northern Chile is composed mainly of phyllotines and Abrothrix taxa. Both groups are diverse and abundant, ranging from the lowlands of northern Chile through canyons that cross the Atacama Desert, the pre-puna, and up to high elevations in the Andes (Mann 1978; Palma et al. 2005; Redford and Eisenberg 1992). On the other hand, glacial cycles of the Pleistocene promoted genetic differentiation in the biota of southern Argentina and Chile (Clapperton 1994; Villagrán and Hinojosa 1997), mainly because of shifts in the ranges of high-latitude species (Brown and Lomolino 1998; Galbreath and Cook 2004). The major goal of this paper was to evaluate the phylogeography of A. olivaceus along its range between 188S and 528S. Specifically, we evaluated the phylogeographic relationships of populations that inhabit environments as variable as the Coastal Chilean Desert, Mediterranean Chile, temperate and subantarctic forests, and Patagonian steppe. We also evaluated the subspecific status of the forms traditionally recognized within A. olivaceus restricted to continental Chile. We sequenced the hypervariable domain I of the mitochondrial DNA (mtdna) control region in specimens collected along most of the range of the species. This information is discussed in light of historical events such as the formation of the Atacama Desert and Pleistocene glacial cycles that might have affected the genetic structure of Chilean subspecies of A. olivaceus. MATERIALS AND METHODS Samples and study area. Voucher specimens were collected from throughout the latitudinal range of A. olivaceus in Chile (Fig. 1; Appendix I), and were deposited in the Colección de Flora y Fauna Profesor Patricio Sánchez Reyes (SSUC), Departamento de Ecología, Pontificia Universidad Católica de Chile, Santiago, Chile, and the Museum of Southwestern Biology, University of New Mexico, Albuquerque, New Mexico. Specimens were identified following the taxonomic keys of Osgood (1943) and Mann (1978). We followed guidelines of the American Society of Mammalogists (Animal Care and Use Committee 1998) during the collection and handling of animals used in this work. DNA extraction and sequencing analyses. DNA was extracted from frozen liver samples treated with a Wizard Genomic DNA Purification Kit (PROMEGA, Madison, Wisconsin). We amplified the complete hypervariable domain I of the mtdna control region via FIG. 1. Approximate range of distribution of Abrothrix olivaceus in Chile (modified from Mann 1978), with ranges of the continental subspecies. Localities 1 and 5 correspond to Abrothrix andinus (Chungará and Talabre, respectively). Other numbers indicate localities sampled for A. olivaceus: 2) Quebrada de Camarones, 3) Quebrada de Tarapacá, 4) Desembocadura Río Loa, 6) Parque Nacional Llanos de Challe, 7) Parque Nacional Fray Jorge, 8) Quebrada del Tigre, 9) Cerro El Roble, 10) San Carlos de Apoquindo, 11) Tucapel, 12) Temuco, 13) Valdivia, 14) Panguipulli, 15) Chiloé, and 16) Parque Nacional Torres del Paine. polymerase chain reaction (Saiki et al., 1988). We designed a forward primer, LBE08 (59-CAC CCA AAG CTG GAG TTC TCC-39), in the trna thr gene that flanks the control region. The reverse primer, H12S (59-TGG TCT TGT AAA TCA GTA ATG-39), was designed from an alignment of the 12S gene in several sigmodontines. Amplifications were performed with the following parameters: initial step at 948C for 3 min, continued with 28 cycles of 948C (30 s), 598C (14 s), and 728C

3 October 2006 RODRIGUEZ-SERRANO ET AL. PHYLOGEOGRAPHY OF ABROTHRIX OLIVACEUS 973 (45 s). A final extension at 728C for 5 min terminated the reaction. Double-stranded polymerase chain reaction products were purified with a Qiaquik Purification kit (Qiagen Inc., Valencia, California). Cycle sequencing was performed (Murray 1989) using the Big Dye Terminator kit (Perkin-Elmer, Norwalk, Connecticut) and an ABI Prism 310 automated sequencer (Applied Biosystems, Foster City, California). We sequenced a total of 483 base pairs (bp) of the hypervariable domain I of the mtdna control region from 67 individuals and those sequences have been deposited in GenBank (accession numbers: AY AY840081). Sequences were aligned by using the CLUSTAL X program (Thompson et al. 1997) with the default values for all alignment parameters, and by eye, using the flanking trna thr as a reference point. Genealogical analysis. Genealogical analysis was performed using the neighbor joining algorithm in PAUP 4.0 (Swofford 2002). We selected that algorithm because it estimates genealogical relationships by an approximation of differences and clustering using an explicit model of evolution. The best fit substitution model of sequence evolution was identified with Modeltest 3.06 (Posada and Crandall 1998). The Akaike information criterion (AIC Akaike 1974) recognized the general time reversible þ invariable sites þ gamma model (GTRþIþG Rodríguez et al. 1990) as optimal ( lnl ¼ 1, , AIC ¼ 2, ), with base frequencies freqa ¼ , freqc ¼ , freqg ¼ , freqt ¼ ; and substitutions models [A C] ¼ , [A G] ¼ , [A T] ¼ , [C G] ¼ , [C T] ¼ , [G T] ¼ The proportion of invariable sites (I) was I ¼ and the gamma distribution shape parameter (G) was G ¼ The neighbor-joining tree was performed by using a heuristic search with tree-bisectionreconnection branch swapping and using the substitution model suggested by the Modeltest. Confidence values on the nodes were evaluated by performing 1,000 bootstrap replicates with heuristic search (Felsenstein 1985). We rooted the tree with the outgroup criterion using Abrothrix andinus, the sister taxon to A. olivaceus (D Elía 2003; Smith and Patton 1999). We also performed a Bayesian analysis to evaluate genealogical relationships under the best-fit model of evolution obtained with Modeltest. To more thoroughly explore the parameter space we ran metropolis coupled Markov chain Monte Carlo simulations (MCMCMC) using the program MrBayes 3.0 (Huelsenbeck and Ronquist 2001) with 2 incrementally heated chains, using the default values. From a random starting tree we ran generations with the resulting trees sampled at every 100 generations (saving 10,000 trees). We determined when stationarity was reached (to discard burn-in samples) by plotting the log-likelihood scores of sample points against generation time. Three replicate analyses were performed, and stationarity levels were compared for convergence (Huelsenbeck and Bollback 2001). The first 50,000 generations (500 trees) were discarded as burn-in and only the results from the last 950,000 generations (9,500 trees) were used to compute a 50% majority rule consensus tree. The percentage of samples that recovered any particular clade on this tree represented that clades s posterior probability. These are the P-values, and values of P 95% were considered evidence of significant support for a clade (Huelsenbeck and Ronquist 2001). Phylogeographic and genetic structure analyses. Sequence divergence between groups (i.e., subspecies) was evaluated in MEGA 2.0 (Kumar et al. 2001) using the Tamura Nei substitution model (TrN Tamura and Nei 1993). For genetic structure analysis all localities were grouped according to recognized subspecies following Mann (1978), with the exception of A. o. mochae and A. o. beatus, which were not included in this work because of unavailability of samples (A. o. mochae) and because A. o. beatus is considered to be a junior synonym of A. o. brachiotis (Smith et al. 2001). The sequence divergence values were plotted in multidimensional space using multidimensional scaling performed in the STATISTICA 6.0 program (StatSoft, Inc. 2001). The genetic structure of the subspecies was evaluated by calculating genetic divergence (F st ) values using the ARLEQUIN program (Schneider et al. 2000). With DnaSP software (Rozas et al. 2003) we generated a file that included haplotypes and polymorphic sites. To establish the relationships between haplotypes, we constructed a haplotype network using the uncorrected medianjoining values with the NETWORK 4.1 program (Rohl 2000; available at last accessed October 2005). We calculated Tajima s D (Tajima 1989) and Fu s F (Fu 1997) values to check for deviations from neutral evolution of our sequences, as well as to test for demographic expansion. Although these statistics were developed as a test of selective neutrality, they are very powerful for detecting departures from population size equilibrium caused by population expansions or bottlenecks (Aris-Brosou and Excoffier 1996; Fu 1997; Ray et al. 2003; Tajima 1996). ARLEQUIN was used to conduct these tests and calculate the corresponding P-values. These indexes tend to negative significant values when the population is not under mutation drift equilibrium due to sudden expansion. However, Ramos-Onsins and Rojas (2002) demonstrated that Fu s F is a more effective detector of population growth because it detects a greater amount of rare haplotypes in a population sampling. Given the low numbers of individuals of A. o. xanthorhinus, this subspecies was not included in the F st and neutrality test analyses. RESULTS Genealogical analyses. A 483-bp fragment that covered the complete hypervariable domain I of the mtdna control region was sequenced from 67 Abrothrix specimens (including outgroup species). The neighbor joining tree based on a GTRþIþG distance matrix, and the Bayesian analyses produced a congruent topology (Fig. 2). The genealogical analyses recovered several clades that nearly agree with the recognized subspecies of A. olivaceus (Fig. 2). One of these is a basal and moderately supported clade that unites individuals of northernmost Chile (Quebrada de Camarones, Quebrada de Tarapacá, and Desembocadura del Río Loa) with 66 and 0.95 of support for bootstrap and posterior probability values, respectively. We named the group the northern clade. Another grouping clustered individuals from north-central Chile that are within the range of A. o. olivaceus (Llanos de Challe and Fray Jorge) and closely related to individuals from San Carlos de Apoquindo and Quebrada del Tigre. The other major split united several clades representing other nominal subspecies: a south-central Chilean clade was constituted by representatives from Tucapel, Temuco, and Valdivia representing the subspecies A. o. pencanus in close association with the Patagonian forms from Torres del Paine recognized as A. o. xanthorhinus; and another southern clade was composed of populations from Valdivia, Chiloé, and Panguipulli that conform to A. o. brachiotis. In addition, we recovered a clade constituted of representatives from El Roble and San Carlos de Apoquindo that we named as El Roble relict. Population and phylogeographic approaches. The sequence divergence among localities showed a close genetic similarity between southern populations (from Tucapel to Torres del Paine) that varied between 0% and 1% of TrN

4 974 JOURNAL OF MAMMALOGY Vol. 87, No. 5 FIG. 2. Neighbor joining tree based on GTRþIþG substitution model. This topology is congruent with the Bayesian analyses; bootstrap values (1,000 replicates) are above the branches and posterior probability are below the branches. Abbreviations are as follows: QCAMARONES (Quebrada de Camarones), QTARAPACA (Quebrada de Tarapacá), DesRIOLOA (Desembocadura Río Loa), LLANOS (Parque Nacional Llanos de Challe), FRAYJORGE (Parque Nacional Fray Jorge), QTIGRE (Quebrada del Tigre), ELROBLE (Cerro el Roble), SANCARLOS (San Carlos de Apoquindo), and TPAINE (Parque Nacional Torres del Paine). distance, whereas distance values between the rest of Chilean populations varied between 2% and 4% (Table 1). The posterior multidimensional scaling of these values showed 3 groups of populations in the bidimensional space with a stress value of 0.12 (Fig. 3; a stress value lower than 0.1 suggests the occurrence of groups or associations in the multidimensional space Ludwig and Reynolds 1988). These groups were the northern clade (Quebrada de Camarones, Quebrada de Tarapacá, and Desembocadura del Río Loa in fact we made a posteriori delimitation for this northern group given the topology obtained in the neighbor-joining and Bayesian tree; Fig. 2); the central populations (Llanos de Challe, Fray Jorge, Quebrada del Tigre, and San Carlos de Apoquindo); and the southern populations (Tucapel, Temuco, Valdivia, Chiloe, and Torres del Paine) with the addition of the El Roble relict. The position of groups along dimension 1 (x-axis) corresponded

5 October 2006 RODRIGUEZ-SERRANO ET AL. PHYLOGEOGRAPHY OF ABROTHRIX OLIVACEUS 975 TABLE 1. Fixation index (F st, above diagonal) and Tamura Nei (TrN) average distance (below diagonal) values among the nominal subspecies of Abrothrix olivaceus and the northern population group. ** P, Subspecies Northern population group A. o. olivaceus A. o. pencanus A. o. brachiotis Northern population group 0.969** 0.965** 0.980** A. o. olivaceus ** 0.593** A. o. pencanus ** A. o. brachiotis to genetic similarity, whereas dimension 2 (y-axis) grouped localities of similar geographical associations (Fig. 3). Fig. 3 shows a separation between A. o. olivaceus and A. o. brachiotis localities in dimension 1, and as shown in both axes a separation was evident between the northernmost population and remaining localities in Chile. However, individuals from El Roble relict were closely related to A. o. brachiotis. The values of F st evidenced a significant structuring between the northern clade (.0.95, P, 0.001) and localities representing nominal subspecies of A. olivaceus (Table 1). A. olivaceus brachiotis also appeared structured, with significant F st values that varied between and (P, 0.001). Neutrality tests showed nonsignificant values both at the nucleotide (Tajima s D-test) and at the haplotype (Fu s F-test) levels. This was the case for the northern clade, A. o. olivaceus, and A. o. brachiotis, whereas A. o. pencanus showed a nonsignificant value of Tajima s D, but a significant Fu s F value (Table 2). These results suggest that populations represented in the northern clade, A. o. olivaceus, and A. o. brachiotis clades are not in TABLE 2. Values of Tajima s D and Fu s F tests with their P values for subspecies of Abrothrix olivaceus. a Subspecies Tajima s D P (D S, D O ) Fu s F P (F S, F O ) Northern population group A. o. olivaceus A. o. pencanus A. o. brachiotis a Subscript S indicates standard; subscript O indicates observed. demographic expansion, contrary to what could be inferred for populations represented in the A. o. pencanus clade. The haplotype network (Fig. 4) showed 39 haplotypes out of 65 sequences (outgroup excluded), with no shared haplotypes between nominal subspecies. However, some contiguous localities shared haplotypes, including Llanos de Challe Fray Jorge, El Roble San Carlos, and Panguipulli Valdivia. Other sampled localities were strictly monophyletic. The network topology is concordant with that observed in the trees (Fig. 4), where we distinguished several haplogroups that corresponded to subspecies A. o. olivaceus, A. o. pencanus, A. o. brachiotis, and A. o. xanthorhinus. Within A. o. olivaceus we found a subset of individuals from the relictual forests (i.e., Fray Jorge National Park) as differentiated from other individuals of this subspecies. In agreement with the genealogical or phylogenetic tree, the network analysis recovered an isolated haplogroup representing individuals from northernmost Chile or the northern clade. Finally, the individuals from El Roble relict were recovered in close association to A. o. brachiotis haplotypes. FIG. 3. Multidimensional scaling of pairwise distance between populations using the nested substitution model Tamura Nei (TrN) with a stress value of The solid line ellipse and the black diamonds represent the Abrothrix olivaceus olivaceus group; the dashed line ellipse and the black circles constitute the A. o. brachiotis group. Gray squares represent northern populations, the black triangle the El Roble population, and the gray triangle the Tucapel population. Y axis is dimension 2. DISCUSSION In the early to mid-pleistocene, descendants of the Andean clade radiation related to the extant Abrothrix occupied the western Andean slope and the lowlands of the central Andes, where the differentiation between A. andinus and A. olivaceus might have taken place (Palma et al. 2005; Pardiñas et al. 2002). In fact, all individuals from northernmost Chile (e.g., Quebrada de Camarones, Quebrada de Tarapacá, and Desembocadura del Río Loa) are recovered here as a basal clade (the northern clade). The quebradas (corridors) of northern Chile have been postulated as biotic corridors extending from the Puna and pre-puna areas to the lowlands that cross the desert in the western flank of the Andes, and connect the latter with coastal areas (Marquet 1994; Meynard et al. 2002; Palma 1995; Palma et al. 2005). The radiation of abrothricines (consisting of 6 genera sensu Smith and Patton [1999]) has been hypothesized to have occurred in the central and southern Andes of Peru, Bolivia, Chile, and Argentina (Reig 1986; Smith and Patton 1999; Spotorno 1986). The differentiation of A. olivaceus has been proposed to have occurred via peripatric speciation from a widely distributed central Andean abrothricine taxon (e.g., A. andinus) from which peripheral isolates dispersed to the Chilean and Peruvian coastal deserts (Palma et al. 2005). From

6 976 JOURNAL OF MAMMALOGY Vol. 87, No. 5 FIG. 4. Uncorrected median joining network generated in NETWORK 4.0. Labels are haplotype names (see Appendix I). Haplotypes of the same subspecies have the same shading: circles with oblique hatching represents Abrothrix olivaceus olivaceus, circles with crossed hatching represent A. o. tarapacensis ( the northern group ), light gray circles represent A. o. pencanus, white circles represent A. o. brachiotis, and black circles represent A. o. xanthorhinus. Numbers on branches are mutational steps between haplogroups. Note that haplotypes from the locality of El Roble (ROSANC and ROBLE1) are closely related to A. o. brachiotis. Haplotype names: QCAMAR (Quebrada de Camarones); QTARAP (Quebrada de Tarapacá); RIOLOA (Desembocadura Río Loa); LLA1 LLA2, LLAFRA (Parque Nacional Llanos de Challe); FRAY2 (Parque Nacional Fray Jorge); TIGRE1-TIGRE2 (Quebrada del Tigre); ROSANC, ROBLE1 (Cerro El Roble); SANCA1 SANCA4 (San Carlos de Apoquindo); TUCAP1 TUCAP9 (Tucapel); TEMUCO (Temuco); VALD1 VALD7, PAVALD (Valdivia); PANGUI (Panguipulli); CHILOE, VALCHI (Chiloe); TPAIN1 TPAIN3 (Parque Nacional Torres del Paine). the latter areas, further dispersal to the north and south to colonize semiarid and temperate latitudes both in Peru and Chile might have followed. Support for this peripheral isolate hypothesis rests on the fact that A. andinus and A. olivaceus constitute sister taxa and that recent phylogeographic work in A. olivaceus that also included A. andinus recovered the latter as basal to A. olivaceus (Smith et al. 2001). Similar hypotheses of speciation in northern Chile have been postulated for other sigmodontine mice such as the phyllotines (Palma et al. 2005; Steppan 1998). The peripheral isolate scenario might have occurred through the quebradas that connect Puna and pre-puna areas in southern Peru and northern Chile with the lowlands. In northern Chile (and probably southern Peru) A. olivaceus is a form restricted to coastal areas and the canyons or quebradas that traverse the Atacama Desert. As recovered in our phylogeographic and phylogenetic analyses, the northern clade of A. olivaceus constitutes a well-differentiated haplogroup and clade in the neighbor joining and Bayesian phylogenetic tree and in the network topology, respectively. The geographical occurrence of this northern genetically isolated population can be explained because these forms inhabit an extremely fragmented biome, whose boundaries are defined by the Coastal Desert of Chile Peru and the Atacama Desert. Further expeditions in northern Chile s region of Antofagasta, between the mouth of the Rio Loa (Desembocadura del Río Loa, S, W) and Llanos de Challe (288119S, W), have been unsuccessful in capturing Abrothrix (Jaksic et al. 1999), suggesting that the Atacama Desert constitutes a natural barrier in the distributional range of A. olivaceus. Given the geographical isolation and elevated F st values of the northern clade populations, it is reasonable to propose that those forms constitute a different subspecies of A. olivaceus, for which we propose a new name (see below). We thus restrict the range of A. o. olivaceus between 288S and 358S (Atacama and Maule regions) of Chile. The phylogenetic position of the clades representing A. olivaceus subspecies agrees with the proposed sequential geographical colonization of the species, given that A. o. olivaceus from north-central Chile appears more basal than southern A. o. brachiotis. A similar reconstruction was obtained with cytochrome-b sequences by Smith et al. (2001). These results suggest that colonization of A. olivaceus through Chile occurred along a north-to-south route. Smith et al. (2001) also showed that the species moved from coastal refuge areas in Chile and reached the Argentinean side through lower elevations of the southern Andes (about latitude 408S). This west east passage to Argentinean Patagonia involved the transition from temperate forests to a drier steppe environment,

7 October 2006 RODRIGUEZ-SERRANO ET AL. PHYLOGEOGRAPHY OF ABROTHRIX OLIVACEUS 977 leading to the morphological divergence of A. o. xanthorhinus in the Argentinean and Chilean Patagonia. Compared to other localities of the Mediterranean zone (e.g., San Carlos de Apoquindo), the genealogical analysis showed that Abrothrix populations from Fray Jorge are slightly differentiated. This locality represents a relictual forest that occurs discontinuously along the coast of central Chile that shares floristic affinities with the southern Valdivian rain forests (Villagrán and Armesto 1980). These forests are remnants of an ancient and wider distribution that occupied the south-central part of the country (Villagrán and Armesto 1980). Interestingly, the olive mouse populations of central Chile that inhabit the relictual forests shared a unique haplotype (Fig. 4), suggesting that they diverged in isolation as a result of recent forest fragmentation. Consistent with previous studies, the diversification between A. o. pencanus and A. o. brachiotis seems to be of the same nature as that between A. o. xanthorhinus and A. o. brachiotis (Pearson and Smith 1999; Smith et al. 2001). The effect of glaciations in the southern Andes triggered several cycles of migration of biota toward the north followed by further southern retraction in belts, taking a V-shape (Moreno et al. 1994; Villagrán and Hinojosa 1997). The advance of ice over the western slope of the Andes in the south central region of Chile (Hollin and Schilling 1981) established populations that have been isolated near the coast of Valdivia (e.g., A. o. brachiotis). The retractions of ice allowed the rapid expansion of the isolated populations toward Patagonia (Smith et al. 2001) and the eastern edge of the central valley of Chile. This expansion followed an environmental gradient from the Valdivian rain forest to the southern limit of the Mediterranean zone, establishing the form A. o. pencanus. The subspecies of A. olivaceus were recovered in almost all analyses, suggesting local adaptation of populations along their range from semiarid, to Mediterranean, to forest environments in the southern part of the country. A. o. olivaceus possesses morphologically distinctive characteristics, such as small size and lighter hair coloration compared with the subspecies of southern Chile (Mann 1978). If these morphologic characteristics are correlated with the level of genetic differentiation found for A. o. olivaceus, it is reasonable to postulate an early directional selective process toward small size (Daan et al 1990). The multidimensional scaling grouped the genetic distances concordant with the 1st split of the tree. This implies that A. o. olivaceus and A. o. brachiotis were derived from the original stock of A. olivaceus that occupied the Chilean territory, reaching the Mediterranean region of central Chile in early to mid-pleistocene (Palma et al. 2005). Further dispersal to southern latitudes might have followed, producing the structuring of subspecies by the reduction of gene flow (Table 1). Interestingly, individuals from Cerro El Roble, in central Chile, as well as some representatives from San Carlos de Apoquindo, were very closely related to A. o. brachiotis. The former locality constitutes the northernmost limit of Nothofagus macrocarpa (Nothofagaceae) distribution. This area represents an isolated oak forest, which has retained its structure and composition since the end of last glaciation about 10,000 years ago (Armesto et al., in press). The fauna associated with these forests also is evolutionarily derived from southern latitudes in Chile. The strong relationship of genetic distances shown in the multidimensional scaling analysis, the direct relationship in the network tree, and the phylogenetic position of individuals of Cerro El Roble and San Carlos de Apoquindo with the localities where A. o. brachiotis is distributed agree with the hypothesis that populations had a southern origin and that their occurrence in central forests is a product of migrations of fauna associated with the forests from southern Chile. Such results suggest that the establishment of A. o. olivaceus and A. o. brachiotis occurred before the most recent glaciation events. Today the haplotypes of Cerro El Roble are shared with nearby localities such as San Carlos de Apoquindo in central Chile. We thus conclude that the subspecies of A. olivaceus, as understood by Mann (1978), are supported by genetic and geographical evidence. They may have diverged as part of recent climatic events, constituting a unique lineage distributed from the coastal canyons of northermost Chile, to the subantarctic islands of Chile and Argentinean Patagonia. Abrothrix olivaceus tarapacensis, new subspecies Holotype. Adult female, skin (adult pelage), skull and tissues (liver and kidney) under collection number SSUC- MA275 (Colección de Flora y Fauna Profesor Patricio Sánchez Reyes, Departamento de Ecología, Pontificia Universidad Católica de Chile), obtained on 17 December 1999 by R. E. Palma under field catalogue number EP 413 from Quebrada de Tarapacá, región I Tarapacá, Chile. Distribution. Currently known from transverse canyons in northernmost Chile: Arica Province, Quebrada de Camarones (198119S, W); Iquique Province, Quebrada de Tarapacá (198569S, W), and Antofagasta Province, Desembocadura del Río Loa (218259S, W). Diagnosis. Size relatively smaller than A. o. olivaceus in most external measurements; pelage of dorsum varying from pale brown to light gold mixed with ochraceous gray; venter laterally gray and centrally as the dorsum, with hairs blackish at bases. Description. The subspecies occurs in the transverse canyons of northernmost Chile, in communities of xeric vegetational assemblages composed mainly of shrubs (Atriplex atacamensis and Baccharis petiolata) and grasses (Distichlis spicata). These canyons are irrigated by waters of the Andes. There are no other abrothricine species in the area. A. o. tarapacensis is the sister taxon of A. o. olivaceus, distinguished by size and coloration. Measurements. External measurements for the holotype are total length, 164 mm; length of tail, 75 mm; length of hind foot, 20 mm; and ear from notch, 16 mm; body weight 23 g. Means for external measurements of the holotype and 14 topotypes are total length, 156 mm; length of tail, 71 mm; length of hind foot, 20 mm; and ear from notch, 15 mm. Means for external measurements of 14 paratypes from Quebrada de Camarones and Desembocadura del Río Loa are total length,

8 978 JOURNAL OF MAMMALOGY Vol. 87, No mm; length of tail, 76 mm; length of hind foot, 21 mm; and ear from notch, 14 mm. Etymology. The subspecies name is derived from the Quechuan word Tarapacá, referring to the homonymous region of Chile. RESUMEN Abrothrix olivaceus es la especie más abundante y de mayor distribución geográfica del Clado Andino que habita el territorio chileno. Evaluamos las relaciones filogeográficas a nivel intra-específico utilizando datos de secuencias nucleotídicas del dominio hipervariable I de la región control del DNA mitocondrial, en 67 especimenes a lo largo de la mayor parte del rango de distribución de la especie. Reconocimos un origen nortino para este taxon seguido de una dispersión hacia el sur. Asimismo, proponemos la existencia de una nueva subespecie de A. olivaceus en el límite más septentrional de su distribución. Detectamos una profunda influencia de los acontecimientos paleo-climatológicos recientes, en la conformación de distribución actual de poblaciones de A. olivaceus en Chile Mediterráneo. ACKNOWLEDGMENTS We acknowledge the valuable comments of F. Torres-Pérez and D. A. Kelt, who helped us to improve early versions of this manuscript. Special thanks to the Hanta and Peridomestic crews of the Chilean Hantavirus Project for field collection of specimens: S. Alvarado, A. Charrier, M. Eherenfeld, J. González, G. Lobos, R. Medina, J. Navarrete-Droguett, and H. Vallejos. We also are grateful for the laboratory assistance of D. Boric-Bargetto. I. Barría provided us with the map used in Fig. 1. We also thank Corporación Nacional Forestal (CONAF) and Servicio Agrícola y Ganadero (SAG) for allowing us to capture mice in protected and unprotected areas of Chile. We greatly appreciate financial support from grants National Fund of Science and Technology (FONDECYT ; Chile); National Fund of Prioritary Areas Center for Advanced Studies in Ecology and Biodiversity (FONDAP-CASEB ), Program 2; and Hantavirus Ecology and Disease in Chile (National Institute of Health- ICIDR 1 U19 AI to The University of New Mexico and to the Pontificia Universidad Católica de Chile, Chile). ER-S acknowledges a DIPUC (Research Office, Pontificia Universidad Católica de Chile) doctoral fellowship. LITERATURE CITED AKAIKE, H A new look at the statistical model identification. IEEE Transactions on Automatic Control 19: ANIMAL CARE AND USE COMMITTEE Guidelines for the capture, handling, and care of mammals as approved by the American Society of Mammalogists. Journal of Mammalogy 79: ARIS-BROSOU, S., AND L. EXCOFFIER The impact of population expansion and mutation rate heterogeneity on DNA sequence polymorphism. Molecular Biology and Evolution 13: ARMESTO, J. J., M. T. K. ARROYO, AND L. F. HINOJOSA. In press. The Mediterranean environment. In The physical geography of South America (T. Veblen, K. Young, and A. Orne, eds.). Oxford University Press, New York. BASKIN, J. A Bensonomys, Calomys, and the origin of the phyllotine group of neotropical cricetines (Rodentia, Cricetidae). Journal of Mammalogy 59: BROWN, J. H., AND M. V. LOMOLINO Biogeography. 2nd ed. Sinauer Associates, Inc., Publishers, Sunderland, Massachusetts. CLAPPERTON, C. M The Quaternary glaciation of Chile: a review. Revista Chilena de Historia Natural 67: DAAN, S., D. MASMAN, AND A. GROENEWOLD Avian basal metabolic rates: their association with body composition and energy expenditure in nature. American Journal of Physiology 259:R333 R340. D ELÍA, G Phylogenetics of Sigmodontinae (Rodentia, Muroidea, Cricetidae), with special reference to the akodont group, and with additional comments on historical biogeography. Cladistics 19: ENGEL, S. R., K. M. HOGAN, J.F.TAYLOR, AND S. C. DAVIS Molecular systematics and paleobiogeography of South American sigmodontine rodents. Molecular Biology and Evolution 15 (1): FELSENSTEIN, J Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: FU, Y. X Statistical tests of neutrality of mutations against population growth, hitchhiking and background selection. Genetics 147: GALBREATH, K., AND J. COOK Genetic consequences of Pleistocene glaciations for the tundra vole (Microtus oeconomus) in Beringia. Molecular Ecology 13: HERSHKOVITZ, P Mice, land bridges and Latin American faunal interchange. Pp in Ectoparasites of Panama (R. L. Wenzel and V. J. Tipton, eds.). Field Museum of Natural History, Chicago, Illinois. HERSHKOVITZ, P The recent mammals of the neotropical region: a zoogeographic and ecological review. Pp in Evolution, mammals and southern continents (A. Keast, F. C. Erk, and B. Glass, eds.). State University of New York Press, Albany. HOLLIN, J. T., AND D. H. SCHILLING Late Wisconsin Weichselian mountain glaciers and small ice caps. Pp in The late great ice sheets (G. H. Denton and T. J. Hughes, eds.). John Wiley & Sons, Inc., New York. HUELSENBECK, J. P., AND J. P. BOLLBACK Empirical and hierarchical Bayesian estimation of ancestral states. Systematic Biology 50: HUELSENBECK, J. P., AND F. RONQUIST MrBayes: Bayesian inference of phylogenetic trees. Bioinformatics 17: JAKSIC, F., J. C. TORRES-MURA, C.CORNELIUS, AND P. A. MARQUET Small mammals of the Atacama Desert (Chile). Journal of Arid Environments 42: KUMAR, S., K. TAMURA, AND M. NEI MEGA 2.0: molecular evolutionary genetics analysis, version 2.0. Pennsylvania State University, University Park. LUDWIG, J. A., AND J. F. REYNOLDS Statistical ecology: a primer on methods and computing. John Wiley & Sons, Inc., New York. MANN, G Los pequeños mamíferos de Chile. Gayana, Zoología 40: MARQUET, P. A Diversity of small mammals in the Pacific coastal desert of Peru and Chile and in the adjacent Andean area: biogeography and community structure. Australian Journal of Zoology 42: MEYNARD, A. P., R. E. PALMA, AND E. RIVERA-MILLA Filogeografía de las llacas chilenas del género Thylamys (Masupialia, Didelphidae) en base a secuencias del gen mitocondrial citocromo b. Revista Chilena de Historia Natural 75: MORENO, P. I., C. VILLAGRÁN, P.A.MARQUET, AND L. G. MARSHALL The Quaternary glaciation of Chile: a review. Revista Chilena de Historia Natural 67:

9 October 2006 RODRIGUEZ-SERRANO ET AL. PHYLOGEOGRAPHY OF ABROTHRIX OLIVACEUS 979 MURRAY, V Improved double-stranded DNA sequencing using the linear polymerase chain reaction. Nucleic Acids Research 17:8889. OSGOOD, W. H The mammals of Chile. Field Museum of Natural History, Zoology Series 30: PALMA, R. E Range expansion of two South American mouse opossums (Thylamys, Didelphidae) and their biogeographic implications. Revista Chilena de Historia Natural 68: PALMA, R. E., P. A. MARQUET, AND D. BORIC-BARGETTO Interand intraspecific phylogeography of small mammals in the Atacama Desert and adjacent areas of northern Chile. Journal of Biogeography 32: PARDIÑAS, U. F. J., G. D ELÍA, AND P. E. ORTIZ Sigmodontinos fósiles (Rodentia, Muroidea, Sigmodontinae) de América del Sur: estado actual de su conocimiento y prospectiva. Journal of Neotropical Mammalogy 9: PATTERSON, B., AND R. PASCUAL The fossil mammal fauna of South America. Pp in Evolution, mammals and southern continents (A. Keast, F. C. Erk, and B. Glass, eds.). State University of New York Press, Albany. PEARSON, O. P., AND M. F. SMITH Genetic similarity between Akodon olivaceus and Akodon xanthorhinus (Rodentia: Muridae) in Argentina. Journal of Zoology (London) 247: POSADA, D., AND K. A. CRANDALL Modeltest: testing the model of DNA substitution. Bioinformatics 14: RAMOS-ONSINS, S. E., AND J. ROJAS Statistical properties of new neutrality tests against population growth. Molecular Biology and Evolution 19: RAY, N., M. CURRAT, AND L. EXCOFFIER Intra-deme molecular diversity in spatially expanding populations. Molecular Biology and Evolution 20: REDFORD, K. H., AND J. F. EISENBERG Mammals of the Neotropics: the southern cone, Chile, Argentina, Uruguay, Paraguay. University of Chicago Press, Chicago, Illinois. REIG, O. A Roedores cricétidos del Plioceno superior de la Provincia de Buenos Aires (Argentina). Publicaciones del Museo Municipal de Ciencias Naturales de Mar del Plata Lorenzo Scaglia 2: REIG, O. A A new fossil of South American cricetids rodents allied to Wiedomys, with an assessment of the Sigmodontinae. Journal of Zoology (London) 192: REIG, O. A Diversity patterns and differentiation of high Andean rodents. Pp in High altitude tropical biogeography (F. Vuilleumier and M. Monasterio, eds.). Oxford University Press, New York. RODRÍGUEZ, F., J. F. OLIVER, A.MARÍN, AND J. R. MEDINA The general stochastic model of nucleotide substitution. Journal of Theoretical Biology 142: ROHL, A Network: a program package for calculating networks, version Mathematisches Seminar, University of Hamburg, Hamburg, Germany. ROZAS, J., J. C. SÁNCHEZ-DELBARRIO, X.MESSEGUER, AND R. ROZAS DnaSP, DNA polymorphism analyses by the coalescent and others methods. Bioinformatics 19: SAIKI, R. K., ET AL Primer directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science 239: SAVAGE, J. M The isthmian link and the evolution of neotropical mammals. Contributions in Science, Natural History Museum of Los Angeles County 260:1 51. SCHNEIDER, S., D. ROESSLI, AND L. EXCOFFIER Arlequin ver. 2000: a software for population genetics data analysis. Genetics and Biometry Laboratory, University of Geneva, Geneva, Switzerland. SIMPSON, G. G Splendid isolation: the curious history of South American mammals. Yale University Press, New Haven, Connecticut. SMITH, M. F., D. A. KELT, AND J. L. PATTON Testing models of diversification in mice in the Abrothrix olivaceus/xanthorhinus complex in Chile and Argentina. Molecular Ecology 10: SMITH, M. F., AND J. L. PATTON The diversification of South American murid rodents: evidence from mitochondrial DNA sequence data for the akodontine tribe. Biological Journal of the Linnean Society 50: SMITH, M. F., AND J. L. PATTON Phylogenetic relationships and radiation of sigmodontine rodents in South America: evidence from cytochrome b. Journal of Mammalian Evolution 6: SPOTORNO, A. E Systematics and evolutionary relationships of Andean phyllotine and akodontine rodents. Ph.D. dissertation, University of California, Berkeley. STATSOFT, INC STATISTICA for Windows [computer program manual]. StatSoft, Inc., Tulsa, Oklahoma. STEPPAN, S. J Phylogenetic relationships and species limits within Phyllotis (Rodentia: Sigmodontinae): concordance between mtdna sequence and morphology. Journal of Mammalogy 79: STEPPAN, S. J., R. M. ADKINS, AND J. ANDERSON Phylogeny and divergence date estimate of rapid radiation in muroid rodents based on multiple nuclear genes. Systematic Biology 53: SWOFFORD, D. L PAUP*: phylogenetic analyses using parsimony (* and other methods). Version 4.0. Sinauer Associates, Inc., Publishers, Sunderland, Massachusetts. TAJIMA, F The effect of change in population size on DNA polymorphism. Genetics 123: TAJIMA, F The amount of DNA polymorphism maintained in a finite population when the neutral mutation rate varies among sites. Genetics 143: TAMURA, K., AND M. NEI Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Molecular Biology and Evolution 10: THOMPSON, J. D., T. J. GIBSON, F.PLEWNIAK, F.JEANMOUGIN, AND D. G. HIGGINS The CLUSTAL_X Windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Research 25: VILLAGRÁN, C., AND J. J. ARMESTO Relaciones florísticas entre las comunidades relictuales del norte chico y la zona central con el bosque del sur de Chile. Boletín del Museo Nacional de Historia Natural 37: VILLAGRÁN, C., AND L. F. HINOJOSA Historia de los bosques de Sudamérica, II: Análisis fitogeográfico. Revista Chilena de Historia Natural 70: WEBB, S. D Ecogeography of the Great American Interchange. Paleobiology 17: Submitted 25 Novemeber Accepted 9 March Associate Editor was Carey Krajewski. APPENDIX I List of localities sampled for Abrothrix taxa in Chile, for which we sequenced the hypervariable domain I of the mitochondrial DNA control region. See Materials and Methods for GenBank accession numbers. Abbreviations for museum and collections are as follows: NK is a numbering system used by the Museum of Southwestern Biology (MSB), University of New Mexico, Albuquerque, New

10 980 JOURNAL OF MAMMALOGY Vol. 87, No. 5 Mexico; SSUC refers to Colección de Flora y Fauna Profesor Patricio Sánchez Reyes, Departamento de Ecología, Pontificia Universidad Católica de Chile, Santiago, Chile; EP refers to the field catalog of one of us (REP). Abrothrix andinus (2). CHILE: Tarapacá, Parinacota, Lago Chungará, S, W (EPA04); Talabre, S, W. Abrothrix olivaceus (65). CHILE: Tarapacá, Arica, Quebrada de Camarones, S, W (EP 432, EP 435, EP 438); Iquique, Quebrada de Tarapacá, S, W (EP 410, EP 413, EP 417); Antofagasta, Tocopilla, Desembocadura Río Loa, S, W (specimen SSUC 1884); Atacama, Huasco, Parque Nacional Llanos de Challe, S, W (NK , NK , NK , NK , NK , NK ); Coquimbo, Limarí, Parque Nacional Fray Jorge, S, W (NK , NK ); Valparaíso, Petorca, Quebrada del Tigre, S, W (NK 96760, NK 96766, NK 96767, NK 96778); Región Metropolitana, Chacabuco, Cerro El Roble, S, W (NK , NK , NK , NK ); Santiago, San Carlos de Apoquindo, S, W (NK 95341, NK 95549, NK 95671, NK 95672, NK 95720, NK 95812, NK 96309, NK 96348, NK , NK ); Bío-Bío, Los Angeles, Tucapel, S, W (NK , NK , NK , NK , NK , NK , NK , NK , NK , NK , NK , NK , NK ); Araucanía, Cautín, Temuco, S, W (NK 95575); Los Lagos, Valdivia, Valdivia, S, W (NK , NK , NK , NK , NK , NK , NK , NK , NK , NK , NK ); Panguipulli, S, W (NK , NK ); Chiloé, Senda Darwin, S, W (NK 95647, NK 95650); Magallanes, Palena, Parque Nacional Torres del Paine, S, W (NK , NK , NK ).