SEARCH

SEARCH BY CITATION

Keywords:

  • Akodontini;
  • Cricetidae;
  • cyt b gene;
  • Deltamys kempi;
  • divergence time;
  • phylogeography;
  • RAG2 gene;
  • Sigmodontinae;
  • South America

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendix

The rodent Deltamys kempiThomas, 1917 is found on the Coastal Plain – a recently formed geographic region located on Brazil’s south-east coast. Considering that Deltamys is the only South American sigmodontine with a sex chromosome system of type X1X1X2X2/X1X2Y, this investigation was focused on the phylogeographic history of this taxon by using gene sequence analysis, trying to clarify when Deltamys differentiated, what was its centre of diversification, and what were the probable routes it used to reach its present distribution. We analysed sequences of the mitochondrial cytochrome b gene and nuclear recombination activating gene 2, performed cranial measurements and searched for centric fusions in individuals collected in distinct localities. The results, clearly demonstrate that D. kempi, on the Coastal Plain, divided into two groups, one occupying a small portion to the north of this region and the other spreading widely to the south. In this process, the phenomena of marine transgression and regressions which moulded its habitat, together with the occurrence of successive chromosomal rearrangements, were certainly the fundamental factors in shaping D. kempi diversification.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendix

The Coastal Plain consists of an area of lowlands (mostly occupied by lakes and coastal lagoons), which encompasses the Atlantic coastline of the State of Rio Grande do Sul in Brazil. Palaeotemperature studies show that marine transgression/regression events originated a Lagoon System, formed by Lagoa dos Patos, the largest lagoon in South America, and by a group of smaller interconnected lagoons on the northern coastline of this Brazilian state (Fig. 1; Villwock et al., 1986).

image

Figure 1.  (a) Shows the distribution of the Deltamys kempi on a map of the South American continent. (b) Map of the Coastal Plain of southern Brazil, Uruguay and Northern Argentina showing the Holocene transgressions. The numbers correspond to the localities of the samples of D. kempiThomas, 1917 used in the analyses: 1 = Torres, 2 = Osório, 3 = Tramandaí, 4 = Charqueadas, 5 = Tapes, 6 = Taim, 7 = San José and 8 = Buenos Aires. Asterisks correspond mainly to the sites where Deltamys specimens were recorded (González & Massoia, 1995;Udrizar Sauthier et al., 2005). (c) Geological map of Rio Grande do Sul State Coastal Plain (Brazil) showing the distribution of the Multiple Barrier system (adapted from Tomazelli & Villwock, 1996).

Download figure to PowerPoint

These plains are inhabited by several endemics, including the rodent Deltamys kempi. The Cricetidae rodents of genus DeltamysThomas, 1917 belong to the tribe Akodontini, whose representatives occupy most of South America (Musser & Carleton, 2005). Its unique species, D. kempi (=Ratón aterciopelado of Massoia, 1964 or Kemp’s Akodont of Musser & Carleton, 2005) was described in 1917, based on four individuals collected near Buenos Aires, but whose type locality had not been accurately determined (Pardiñas et al., 2007). Later, in 1984, the species was found more than 1000 km north-west of that area (on the Brazilian coast) by Sbalqueiro et al. In the meantime, the species had been captured only in a small area of Buenos Aires, Argentina and in Montevideo, Uruguay (Massoia, 1964; Fronza et al., 1981). More recently, González & Massoia (1995) found D. kempi in some other Uruguayan localities. Furthermore, this rodent is but rarely captured – since the first capture made by R. Kemp in the beginning of the 20th century until the last article published about it in 2005 (Udrizar Sauthier et al., 2005), only 117 D. kempi specimens have been reported (Thomas, 1917; Massoia, 1964; Fronza et al., 1981; Sbalqueiro et al., 1984; Castro et al., 1991; Bianchini & Delupi, 1994; González & Massoia, 1995; D’Elia et al., 2003), 42 of them trapped in Brazil. The geographical distribution of D. kempi in Brazil is limited to a region on the south-eastern coast known as the Coastal Plain, where this species preferentially inhabits lowland areas frequently swamped by the sea, although it does not show any specialized external adaptations for aquatic life (Sbalqueiro et al., 1984; Castro et al., 1991). Considering the fact that it can be trapped using many different kinds of bait, this rodent probably is a generalist (Massoia, 1964). No fossil records of Deltamys are known (Pardiñas et al., 2002).

The first cytogenetic studies in D. kempi were carried out by Fronza et al. (1981) who, studying the karyotype of a male and a female from Buenos Aires, Argentina, found that both individuals bore translocations. Subsequently, Sbalqueiro et al. (1984) showed that all the males had 2n = 37 and the females 2n = 38. These differences in the karyotypes were interpreted as being due to a centric fusion between the Y chromosome and an autosome (Y/A), originating a unique multiple sex chromosome system of type X1X1X2X2/X1X2Y. Afterwards, Castro et al. (1991), revising the other autosomal fusions reported for the species, found that D. kempi bears not only fusion (Y/A), plesiomorphic to the males of the taxon, but also an extensive system of autosomal apomorphic fusions characterized by being local/specific which warrant future investigations (Musser & Carleton, 2005).

By using gene sequence analysis, D’Elia (2003) and D’Elia et al. (2003) reaffirmed the monophyly of the genus and Smith and Patton (2007) found that it grouped within the genus Akodon.

Taking into account the peculiarities shown by the biological model –Deltamys is the only South American sigmodontine with a multiple sex chromosome system of type X1X1X2X2/X1X2Y – this study was designed to reconstitute the phylogeographic history of this taxon by using gene sequence analysis, seeking to clarify when Deltamys differentiated, what was its centre of diversification and which routes it probably used to reach its present distribution. Therefore, we analysed sequences of the mitochondrial cytochrome b gene (cyt b) and nuclear recombination activating gene 2 (RAG2), performed cranial measurements and looked for the presence of different centric fusions in individuals collected in various localities in its distribution area. Finally, we investigated how the geological events that resulted in the formation of the Coastal Plain might be related to the present genetic structure of D. kempi.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendix

Specimens analysed

We analysed the complete sequence of the cyt b gene (1140 bp) obtained from 28 D. kempi specimens collected in the Coastal Plain of Rio Grande do Sul State (Table 1; Fig. 1b). We included also the only two complete sequences of this species deposited in GenBank – one from Argentina (AY195860) and the other from Uruguay (AY195862). As outgroups, we chose the same akodontini species of cyt b genes which were analysed with the RAG2 gene also from Akodon cursor (Winge, 1887) distributed throughout the Atlantic forest formations from Paraíba State to Paraná State, Brazil; A. montensis Thomas, 1913 inhabiting east Paraguay, north-east Argentina and south-east and south Brazil; and Thaptomys nigrita (Lichtenstein, 1829) which occupies south-east Brazil (from Bahia to Rio Grande do Sul), east Paraguay and north-east Argentina.

Table 1.   Taxa studied localities, coordinates, number of specimens, number of exemplar measured, cyt b (Arabic) and RAG2 (Roman) haplotypes (H) and GenBank Accession numbers of the samples investigated.
Locality*Coordinates Specimens (N) Cranial measures (N)Cyt b haplotypesGenBank accession numbersRAG2 haplotypesGenBank accession numbers
  1. *Numbers correspond to those of Fig. 1.

Deltamys kempiThomas, 1917
 Torres (1)29°19′S; 49°46′W1 H5EF206809HIEF206816
 Osório (2)29°53′S; 50°16′W12H7EF206811HIEF206816
 Tramandaí (3)29°54′S; 50°16′W1841H5, H6EF206809, EF206810HIEF206816
 Charqueadas (4)29°57′S; 51°37′W2 H4EF206808HIIEF206817
 Tapes (5)30°40′S; 51°23′W1 H8EF206812HIEF206816
 Taim (6) 32°29′S; 52°34′W518H1, H2, H3EF206805, EF206806, EF206807HIIEF206817
 San José, UR (7)34°21′S; 57°06′W1 H9AY195862  
 Buenos Aires, AR (8)34°54′S; 57°45′W1 H10AY195860  
Outgroups
Akodon montensis Thomas, 1913
 Tainhas29°16′S; 50°18′W1 H11EF206813HIIIEF206819
Akodon cursor (Winge, 1887)
 João Pessoa7°06′S; 34°51′W1 H12EF206814HIVEF206820
Thaptomys nigrita (Lichtenstein, 1829)
 Monte Verde19°53′S; 41°57′W1 H13EF206815HVEF206818

The samples were collected in the Coastal Plain, an area located approximately between latitudes 29°S and 33°S (Fig. 1a,b) and covering the coastline of south Brazil, extending to the coasts of Uruguay and Argentina (Villwock et al., 1986). Palaeotemperature studies demonstrated that the temperatures dropped slowly and gradually in the glacial periods but rose rapidly in the interglacial periods (Imbrie et al., 1984). Such alternations caused considerable variation in sea level, opening and closing areas of communication with the Atlantic Ocean, building a system named Multiple Barrier (Urien et al., 1981; Villwock, 1984), formed mainly by four major depositional events (Barrier I–IV) that extended from 400 to 5 kyr, and led to the formation of large lagoons.

The earlier of these depositional events (Barrier I, Fig. 1c) developed as a result of the first transgression–regression of the sea in the Pleistocene (about 400 ka) and shaped what is called the Lombas Barrier. Lombas Barrier is a strip that runs from Osório to Tapes and that isolated the ‘Guaiba–Gravataí’ lagoon system from the sea. The next event, which occurred about 325 ka, formed Barrier 2 (Fig. 1c) and connected Porto Alegre to the Continent by sand banks, and flooded areas. The system called Barrier 3 (Fig. 1c) is formed by deposits that run from Torres to Chuí– its central portion created a barrier that isolated the Lagoa dos Patos. According to Martin et al. (1982), these deposits are around 120 kyr old. The depositional system called Barrier 4 (Fig. 1b,c) developed during the Holocene, more or less 5 ka. In this event, the sea reached roughly 5 m above the present level and formed a barrier along the coastline from Torres to Argentina. The space between Barrier 4 and Barrier 3 was occupied during the transgressive peak of the Holocene by great bodies of brackish water and a rosary of smaller interconnected lagoons on the northern coastline of the Coastal Plain of Rio Grande do Sul and the Uruguayan coast. It may also have formed the Paraná delta in Argentina (Urien et al., 1980; Tomazelli et al., 2000).

The specimens sequenced here were collected at six Brazilian sites located in the depositional system described above: Barrier I (samples from Tapes, Fig. 1, locality 5), II (Taim, locality 6), III (Osório, locality 2), IV (Torres and Tramandaí, localities 1 and 3 respectively) and Charqueadas (locality 4), this latter, aged about 65–10 Myr, situated in the sedimentary sequence of the Paraná basin. The GenBank sequences included in the analyses came from Route 1 near Cufre river in San José Department (Uruguay, locality 7) and La Balandra, Buenos Aires (Argentina, locality 8), sites that correspond to Barrier IV (Table 1; Fig. 1).

Our study also included the sequence analysis of the RAG2, for which we selected individuals to represent all haplotypes found for cyt bfrom the localities investigated in Brazil (one from Charqueadas, three from Taim, two from Tramandaí, one from Torres, one from Tapes and one from Osório). The outgroups were also sequenced.

All the animals sequenced for both genes were karyotyped, the karyotypes being organized as proposed in Sbalqueiro et al. (1984) and Castro et al. (1991). The craniums and skins of the specimens used in this study are deposited in the mammal collection of the National Museum in Rio de Janeiro, Brazil (Appendix).

Morphological measurements

We analysed 61 specimens from Taim (n = 18), Tramandaí (n = 41) and Osório (n = 2). Twenty cranial measurements (Fig. 2) were made with a digital caliper accurate to 0.01 mm according to DeBlase & Martin (1981). All measurements were log-transformed before statistical analysis. Sexual dimorphism in cranial morphometric variables was evaluated (t-test) and no significant differences (P > 0.10) were detected between males and females. One missing value (mandible length) was estimated by using the mean. Discriminant function and principal component analyses of the craniometrical data, using the spss® software (SPSS Inc. Software Products, Chicago, USA), are performed to compare the groups of individuals produced by using the molecular analysis (M.A. Montes, L.F.B. Oliveira, S.M. Coliegari-Jacques & M.S. Mattavi, unpublished data). Preliminary results are shown in this paper.

image

Figure 2.  Cranial characters used in morphometric analyses of Deltamys: LIF, length of incisive foramina; LMR, crown length of maxillary tooth row; NL, nasal length; BR, breadth of rostrum, TBL, tympanic bullae length; BB, breadth of braincase; BZB, breadth of zygomatic bone; LIB, least interorbital breadth; LR, length of rostrum; DFM, distance between first molars; BFM, breadth of first upper molar; CIL, condylo-incisive length; PL, palatal length; LD, length of diastema; OL, orbital length. Not shown: mandible length, height of braincase, breadth of zygomatic plate and depth of upper incisive.

Download figure to PowerPoint

Analysis of nucleotide sequences

DNA was extracted from kidneys, heart, liver and muscles (preserved at −20 °C or in 70% ethanol), using the standard protocol described by Medrano et al. (1990). The sequences of the cyt b were isolated by using polymerase chain reaction (PCR), using two sets of primers. For the first part of the cyt b amplification, primers MVZ 05 (light-strand) and MVZ 16 (heavy-strand) were used, whereas for the second part we used primers MVZ 26 (light-strand) and MVZ 14 (heavy-strand), as suggested in Smith & Patton (1993). For the amplification of RAG2, the primers were RAG2F1 (light-strand) and RAG2R1 (heavy-strand), both as suggested in Baker et al. (2000).

The PCR products were purified with exonuclease I and shrimp alkaline phosphatase (Amersham Biosciences, Ltd, Little Chalfont, Buckinghamshire, UK). All the specimens were sequenced directly from the purified PCR products using the primers mentioned above and the DYEnamic ET Dye Terminator Cycle Sequencing Kit (Amersham Biosciences) in a MegaBase 1000 automated sequencer (Amersham Biosciences) following the manufacturer’s protocols. The sequences of all the haplotypes are available in GenBank, as shown in Table 1.

Data analysis

The sequences obtained were read with the Chromas 1.45 software and aligned with the Clustal X 1.81 program (Thompson et al., 1997), using the default setting costs and manually corrected with the bioedit program (Hall, 1999). The saturation test of substitutions was carried out using the Data Analysis in Molecular Biology and Evolution software (dambe; Xia & Xie, 2001). The Kimura 2-parameter distance (K2p) and the base composition were obtained with the Molecular Evolution Genetics Analysis software (MEGA 3; Kumar et al., 2004).

The phylogenetic analysis was carried out using the neighbour-joining (NJ), maximum likelihood (ML) and maximum parsimony (MP) algorithms, with the paup* v.4.0b10 program (Swofford, 2001). The networks using the median-joining (MJ) method were obtained with the network v.4.1.0.0 software (Bandelt et al., 1999).

Before performing the analyses, the nucleotide substitution model appropriate for the ML and NJ analyses was determined using the modeltest 3.06 program (Posada & Crandall, 1998). To estimate the ML tree (with 1000 bootstrap replications), we chose a heuristic search with as-is, tree bisection–reconnection (TBR), branch swapping and the MULPARS option. The support values for the tree branches were estimated by using bootstrap analysis as described in Xiang et al. (2002).

The MP analysis was performed by using heuristic search with TBR, branch swapping, the MULPARS option and 100 random-addition replicates. The statistical support for bootstrap (Felsenstein, 1985) was executed with 1000 replications of the heuristic search and the addition of a simple taxon, saving all the trees in a file.

The Bayesian inference (BI) was conducted with the MrBayes 3.0b4 program (Huelsenbeck & Ronquist, 2001) to generate the distribution of the posterior probabilities, according to the Markov Chain Monte Carlo method (MCMC). No a priori assumptions about the topology of the tree were made and all searches were provided with a uniform prior. The DNA substitution model used was as estimated by using the modeltest. The MCMC process was adjusted so that four chains ran simultaneously for a million generations, displaying trees every 100 generations, for a total of 10 000 trees. The first 100 000 generations were considered a burn-in and thus discarded. To calculate the posterior probability of each bipartition, a 50% majority-rule consensus tree was constructed from the remaining trees using paup. A partition-homogeneity test (incongruence length difference, ILD) was computed to detect the presence of conflictive characters between the cyt b and RAG2 genes, as described by Farris et al. (1994, 1995) and was implemented in the winclada version 0.9.9 + (BETA) (Nixon, 1999).

All topologies were compared using the Templeton (Templeton, 1983) and Shimodaira–Hasegawa (Shimodaira & Hasegawa, 1999) tests performed under the parsimony and the likelihood criteria respectively.

The population genetics software package arlequin version 2000 (Scheneider et al., 2000) was used to examine patterns of mismatch distribution using the observed number of differences between pairs of haplotypes, and the hierarchy analyses of genetic diversity of populations were carried out using amova (Excoffier et al., 1992). We tested the selective neutrality of the sequences by estimating the sequences’F-values (Fu & Li, 1993). In addition, we assessed neutral evolution bias by the presence of significant values of parameter D described by Tajima (1989) and the Ramos-Onsins and Rozas test (R2) design for small sample sizes, using the DnaSP program, version 4.1 (Rozas et al., 2003).

The molecular clock hypothesis was tested for complete cyt b gene using the likelihood ratio test (LRT) (Huelsenbeck & Rannala, 1997) with the paup* v.4.0b10 program (Swofford, 2001). For the LRT, the likelihoods of the trees with and without a molecular clock enforced were compared. As the clock hypothesis is a simpler model, the likelihood statistic is estimated as 2(ln Lclock − ln Lno clock). This statistic follows a chi-squared distribution with n − 2 degrees of freedom, where n is the number of sequences. Divergence time was calculated using mega 3.1 (Kumar et al., 2004) taking into account all the mutations (transitions and transversions) in the sequences, including all codon positions (Palma et al., 2005). The estimate of the separation of genera Akodon/Thaptomys-Necromys, considered as 3.5 (Reig, 1987; Pardiñas et al., 2002), was used as calibration for the analysis of the divergence time of our sequences.

To calculate the divergence time, three approaches were employed: one used Kimura 2-parameter model (K2p, Kimura, 1980), according to Smith & Patton (1993); the other employed Tamura–Nei model (TN, Tamura & Nei, 1993), as Palma et al. (2005); and the third utilized transversions in the third position of the codon to calibrate the differentiation between Mus and Rattus.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendix

Characteristics of the sequences

The analysis of the mitochondrial cyt b gene included sequences of 30 D. kempi individuals (28 sequenced by us and another two specimens from GenBank) from eight locations in South America and the three outgroups (Table 1).

The average cyt b sequence composition presented low guanine content (0.1655) and similar A, T and C contents (0.296, 0.297 and 0.282 respectively). Such nucleotide compositions have been reported in Deltamys (D’Elia et al., 2003), in other sigmodontine rodents (Smith & Patton, 1991, 1993, 1999), and this strong bias in base composition is a characteristic of the cyt b gene in mammals (Irwin et al., 1991). The guanine deficiency observed was more marked in the second and third codon positions (0.122 and 0.034 respectively), whereas the frequencies of the different bases in the first codon positions were similar. The graph representing the transition/transversion vs. the pairwise distance indicated an absence of substitution saturation (data not shown).

A total of 10 haplotypes exclusive to Deltamys and one haplotype for each outgroup were obtained for the cyt b gene (Table 1). Haplotypic diversity was 0.0098 and nucleotide diversity was 0.0071. Only one haplotype was shared between two geographically close locations (Tramandaí and Torres). The analysis of these 10 haplotypes revealed 32 polymorphic sites, of which 26 were transitions and six transversions. In the transitions, 23 were silent, two located in the first and 21 in the third codon positions. The remaining three were nonsilent – two in the first and one in the second codon positions. As for transversions, three were silent in the third codon positions and three were nonsilent – two of these being in the first and one in the third codon position.

As regards the RAG2 gene, an 801-bp fragment was generated, and the sequences resulted in two haplotypes (Table 1) in D. kempi, two polymorphic sites being two silent transitions, one in the first and the other in third codon position. Three other haplotypes were found in the outgroups.

Phylogeographic analysis

As the ILD test did not reveal significant incongruence between the mitochondrial and nuclear genes (S = 0, W = 5, P = 1), both were brought together in a combined analysis, resulting in a 1941-bp fragment. However, because the representatives from GenBank of Argentina and Uruguay were studied through cyt b only, aiming at obtaining information about all the distribution area of the taxon, we performed an analysis including this gene separately.

The proportion of invariable sites (I) was 0.6588. The analysis of the cyt b 1140-bp fragment using the MP, NJ, ML and BI, generated trees with highly similar topologies, as neither the Templeton or the Shimodaira–Hasegawa tests detected statistically significant differences between any of them (P = 0.3422 and P = 0.2371 respectively). The ML tree is shown in Fig. 3a. It could be observed that there are two groups within D. kempi, one which we have called the Tramandaí Group (with bootstrap values of 61 for NJ, 96 for MP, 86 for ML and 1.00 for BI posterior probability), formed by the Torres, Tramandaí, Osório and Tapes Brazilian localities; and a second group, denominated the Taim Group (with bootstrap values 70 for NJ, 93 for MP, 57 for ML and 0.99 for BI posterior probability), which includes Charqueadas and Taim (Brazil), San José (Uruguay) and Buenos Aires (Argentina). The average genetic distance, using K2p, measured between the groups was 0.010. Within the Tramandaí Group, the average genetic distance was 0.005, varying from 0 between the Tramandaí and Torres haplotypes to 0.008 between Osório and Tapes haplotypes. For the Taim Group, the average genetic distance was 0.006, varying from 0 for the Charqueadas and some Taim haplotypes to 0.009 between haplotypes from Charqueadas and Uruguay (Table 2).

image

Figure 3.  (a) Maximum likelihood tree based on the TVM + I model (−ln L = 3087, 7561, AIC = 6191, 5122) derived from cyt b sequences for 10 haplotypes found among eight populations of Deltamys kempi. Akodon cursor (Winge, 1887), A. montensis Thomas, 1913 and Thaptomys nigrita (Lichtenstein, 1829) were used as outgroup. Bootstrap values for maximum likelihood are given below each branch. The rectangles indicate the two divergent groups inferred: the Tramandaí and Taim groups. st/st = standard homozygote karyotype; st/t(2;3) = 2;3 fusion heterozygote; st/t(9;15) = 9;15 fusion heterozygote; t(9;15)/t(9;15) homozygote; st/t(1;13) = 1;13 fusion heterozygote. (b) Median-joining network showing the phylogenetic relationships between D. kempi cyt b haplotypes. Circular areas are proportional to the number of individuals bearing a particular haplotype. Branch lengths are proportional to the number of mutations involved between haplotypes. Mutations are indicated with tick marks and the longest branches are represented by numbers.

Download figure to PowerPoint

Table 2.   Genetic distances (Kimura 2-parameter model) between the haplotypes of Deltamys kempi of the different localities and outgroups.
 Osório (Tramandaí)Tramandaí/Torres (Tramandaí)Tramandaí (Tramandaí)Tapes (Tramandaí)Uruguay (Taim)Argentina (Taim)Taim 1 (Taim)Taim 2 (Taim)Taim 3 (Taim)Charqueadas (Taim)Akodon cursorAkodon montensisThaptomys nigrita
  1. Superior diagonal cyt b and RAG2 concatenates gene distances and inferior diagonal cyt b gene distances. The model selected for cyt b in the modeltest (Posada & Crandall, 1998) was TVM + I, and the values were −ln L = 3087.7561, AIC = 6191.5122. Base frequencies were A = 0.2902; C = 0.2970; G = 0.1198 and T = 0.2929. The rate matrix was R(a) [A–C] = 2.0616; R(b) [A–G] = 25.7892; R(c) [A–T]=3.7890; R(d) [C–G] = 1.0642; R(e) [C–T] = 25.7892; R(f) [G–T] = 1.0000. The proportion of invariable sites (I) was 0.6588. The model selected for cyt b and RAG2 concatenated analysis, using the modeltest (Posada & Crandall, 1998), was GTR + G and the values were −ln L = 4454.4136 and AIC = 8926.8271. Base frequencies were A = 0.2923; C = 0.2599; G = 0.1637 and T = 0.2841. The rate matrix was R(a) [A–C] = 1.6701; R(b) [A–G] = 8.7551; R(c) [A–T] = 2.1787; R(d) [C–G] = 0.4263; R(e) [C–T] = 18.2225; R(f) [G–T] = 1.0000. The gamma distribution parameter (G) was 0.0653. This model was employed for the NJ, ML and BI analyses.

Osório (Tramandaí) 0.0020.0030.0050.0130.0150.0110.0090.3670.4310.435
Tramandaí/Torres (Tramandaí)0.003 0.0010.0030.0110.0130.0090.0090.3500.4270.422
Tramandaí (Tramandaí)0.0040.002 0.0030.0110.0130.0090.0070.3500.4100.414
Tapes (Tramandaí)0.0080.0050.005 0.0110.0130.0090.0090.3520.4300.394
Uruguay (Taim)0.0180.0150.0150.015 
Argentina (Taim)0.0170.0140.0140.0140.001 
Taim 1 (Taim)0.0180.0150.0150.0150.0040.003 0.0030.0030.0040.3460.4300.442
Taim 2 (Taim)0.0200.0170.0170.0170.0070.0060.005 0.0040.0060.3520.4390.434
Taim 3 (Taim)0.0150.0120.0120.0120.0060.0050.0040.006 0.0030.3260.4060.434
Charqueadas (Taim)0.0120.0100.0120.0120.0090.0080.0070.0090.004 0.3360.3950.414
Akodon cursor0.1590.1550.1550.1550.1590.1580.1540.1550.1510.153 0.1350.385
Akodon montensis0.1650.1610.1640.1640.1640.1650.1640.1650.1600.1590.096 0.405
Thaptomys nigrita0.1730.1690.1700.1660.1750.1740.1730.1720.1720.1690.1550.156 

The MP analysis of ingroup plus outgroups included 152 informative sites and generated a tree (not shown) of 349 evolutionary steps, with a consistency index (CI) of 0.8796 and a retention index (RI) of 0.8384. The MJ analysis demonstrated an arrangement similar to that of the NJ, MP, ML and BI trees, with a network tree (Fig. 3b) differentiating Tramandaí and Taim groups by a separation of 10 mutational events.

As regards the cyt b and RAG2 concatenated analysis, MP produced two equally parsimonious trees (including the outgroups), with a length of 366 steps, and with CI = 0.8880 and RI = 0.8400. The tree topologies generated by using these methods were very similar, as neither the Templeton test (P = 0.3173) nor the Shimodaira–Hasegawa test (P = 0.2100) detected significant differences between them, and discriminated two groups, Tramandaí and Taim. The Bayesian tree is presented in Fig. 4a. Similar to the cyt b individual analysis, the Tramandaí Group is formed by the Torres, Tramandaí, Osório and Tapes locations, with bootstrap values 93 for NJ, 95 for MP and ML, and 1.00 for BI posterior probability. The Taim Group includes Taim and Charqueadas locations, with bootstrap values of 88 for NJ and MP, 84 for ML and 1.00 for BI posterior probability. The average genetic distance between both groups was 0.014. Within the Tramandaí Group, the average genetic distance was 0.005, varying from 0 between the Tramandaí and Torres haplotypes up to 0.008 between Osório and Tapes haplotypes. In the Taim Group, the average genetic distance was 0.006, varying from 0 in the Charqueadas and in some of the specimens from Taim, up to 0.009 between Taim and Charqueadas haplotypes (Table 2).

image

Figure 4.  (a) Phylogenetic tree based on the combined analysis of the nuclear RAG2 and cyt b genes of Deltamys kempi using Bayesian analysis (GTR + G model of evolution). Akodon cursor, A. montensis and Thaptomys nigrita were used as outgroups. Posterior probabilities are shown below the branches. The rectangles indicate the two divergent groups inferred: the Tramandaí and Taim groups. (b) Median-joining network showing the phylogenetic relationships of the D. kempi based on the combined analysis of the nuclear RAG2 and cyt b genes. The circular areas are proportional to the number of individuals bearing a particular haplotype. Branch lengths are proportional to the number of mutations involved between the haplotypes. Mutations are indicated by tick marks and the longest branches are represented by numbers.

Download figure to PowerPoint

In the MJ analysis with the two concatenated genes, a pattern similar to that presented by using the NJ, MP, ML and BI trees can be observed (Fig. 4b), with the Tramandaí and Taim groups separated by 12 mutational events.

The amova including all the individuals of the eight localities investigated for a complete cyt b gene (1140 bp), revealed a considerable differentiation between the localities, resulting in a significant FST (0.8696; P < 0.0001). As expected in view of the fact that all the trees generated indicated the formation of two groups, we found a statistically significant differentiation between both (FCT = 0.6634; P = 0.0282) using a hierarchical amova.

As regards the karyotypes found in the different locations (Fig. 3a), four centric fusions were observed between various autosomes each one restricted to a specific site: t(2;3) from Tapes, t(9;15) in Tramandaí (of which both the homozygotes and the heterozygotes were found), t(1;13) in Taim and another distinct fusion of this kind which occurred in Buenos Aires (Fronza et al., 1981).

The t-test of the discriminant function of 20 cranial measurements obtained from individuals in both groups showed significant differences between the two groups (P > 0.05), but in the case of principal components analysis significant differences were observed in the second component only (PC1: P = 0.383; PC2: P = 0.000).

Divergence time

The molecular clock hypothesis for the cyt b gene was not rejected by using the LRT test (D = 9.2154; d.f. = 12; P > 0.20) and the time of differentiation of this rodent on the Coastal Plain was calculated. The neutral evolution of the sequences were investigated with the F′-value (Fu & Li, 1993), with the Tajima (Tajima, 1989) and R2 (Ramos-Onsins & Rozas, 2002) tests. The values obtained are depicted in Table 3, and they show that the neutrality for the sequences studied was not rejected.

Table 3. Deltamys kempi summary statistics.
 Fu and Li testTajima testRamos-Onsins and Rozas’s test
Deltamys kempi2.475 (P = 0.846)0.015 (P = 0.579)0.159 (P = 0.000)

The evolutionary rate obtained between Akodon and Necromys/Thaptomys (employing transitions and transversions in all positions of codon) by using the K2p model was 0.02366 per Myr, a value very similar to that found with the TN model, which was 0.0232 per Myr. Smith & Patton (1993), calculating the transversions in the third codon positions for Mus and Rattus, obtained a differentiation value of 0.0170 per Myr.

Using these values in the calculation of the Deltamys divergence time from the other Akodontini, we estimated that this genus differentiated approximately 3.2 and 3.5 Ma (Table 4). The same procedure was utilized to estimate the period spent by Deltamys to spread in the Costal Plain (from Torres to Buenos Aires, assuming a molecular clock already tested): c. 200 000 years ago (Table 4).

Table 4.   Divergence time estimations (in Myr).
Divergence periodK2p model (all positions)TN model (all positions with)Split Mus/Rattus (third position transversions)
Between Deltamys and others Akodontini3.23.33.5
Spread of Deltamys in the Costal Plain0.21130.21500.2941

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendix

All the analyses, whether using the cyt b on its own or in combination with RAG2, have shown the Deltamys genus to be a monophyletic taxon, with a high level of bootstrap and posterior probability support. This monophyly was expected when the Deltamys chromosomal analysis demonstrated that the genus was a bearer of unique one-character apomorphic to the Sigmodontinae – the multiple sex chromosome system in the X1X1X2X2/X1X2Y form. The meiotic configuration of this system prevents its bearers from mating with other taxa that are bearers of other sex chromosome systems, including the XX/XY.

Both the phylogenetic trees and the haplotype networks generated have demonstrated, with high congruence, the formation of two groups of great statistical confidence (bootstrap and posterior probability). As can be seen in Fig. 5, the Tramandaí Group includes the Brazilian localities of Torres, Tramandaí, Osório and Tapes, and the Taim Group formed by the localities of Charqueadas, Taim (Brazil), San José (Uruguay) and Buenos Aires (Argentina). As regards the Tramandaí Group, two sister groups were formed, one including the haplotype of Tapes and the other putting together the haplotypes of Osório and Tramandaí (Figs 3a and 4a). The position of Tapes on the Coastal Plain corresponds to the ‘Barrier I’ geological formation formed by the first marine transgression (Fig. 1b) around 0.4 Ma (Tomazelli & Villwock, 1996). As a result, the subsequent regression linked Tapes to Tramandaí and Osório. The two latter locations, together with Torres, originated in the third and fourth marine transgressions which took place about 0.12 and 0.05 Ma respectively (Villwock & Tomazelli, 1998).

image

Figure 5.  Map of the Coastal Plain of southern Brazil, Uruguay and Northern Argentina and superimposed maximum likelihood tree based on the TVM + I model showing the geographical association of the different haplotypes.

Download figure to PowerPoint

The region of the Taim Group corresponds to the second, third and fourth marine transgressions. Of these localities, Taim is the result of the second and fourth transgressions (Tomazelli et al., 2000), whereas San José and Buenos Aires were formed only in the last Holocene transgression (Fig. 1b) and are therefore geologically more recent (Urien et al., 1980).

Charqueadas, in geological terms, is located in the Paraná basin sedimentary sequences, a site which originated before the marine transgressions of the Pleistocene. When only the haplotypes of Taim Group (Table 2, in grey) were considered, the genetic distances show that the haplotype of Charqueadas locality is, most of the time (two out of three), the closest to the outgroups. This fact, together with the relatively old geological age of this locality, suggests that the region that encloses the Tapes locality could also be the centre from which this species started its dispersion across the Coastal Plain. However, this scenario do not eliminate the possibility that Charqueadas would be occupied recently by individuals migrated from south.

This geological background, with localities of different ages originated by the sequential marine transgressions in the Pleistocene/Holocene that shaped the southern coasts of Brazil and Uruguay and the basin of the River Plate in Argentina, may explain the topology of the molecular trees found for Deltamys, in which the most derived sequences are located in the most recent land masses.

This organization into Tramandaí and Taim groups becomes very clear in the network analyses (Figs 3b and 4b), where it can be seen that 10 and 12 mutations, for the cyt b gene and for the combined analysis of cyt b and RAG2, respectively, separate the two groups. The amova analysis also indicates a significant differentiation between the Tramandaí and Taim groups. It is observed from the mismatch distribution (not shown) that the whole species (Tramandaí group + Taim group) does not present a signal of a population size expansion as it shows a ragged distribution. However, when the two groups were analysed separately, we found only a sign of expansion in the Taim group, as it presents a single wave pattern. This result seems to be a reflex of the dispersion pattern of Deltamys in a southerly direction, from Charqueadas (Brazil) to the River Plate, near Buenos Aires. This dispersion may have progressed along with the formation of brackish coastal lagoons in this part of the South American coast – a habitat characteristic of Deltamys. Another factor, which stands out, is that this territorial occupation was apparently accompanied by a number of rearrangements of the autosomal chromosomes specific to each location.

The rearrangement of the sex chromosome pair may be very old and was perhaps the mechanism that triggered the differentiation of the taxon, as this rearrangement was found in all the individuals from all the localities we studied. The differences in the autosomal rearrangements between the localities are accompanied by high and significant FST values, indicating that these locations are structured.

In the case of this rodent, it is worth noting, however, that living in small populations seems to be a feature of its natural history. This characteristic may one of the factors responsible, at least in part, by the fact that we obtained only a few individuals in certain locations, notwithstanding the fact that we have searched the Coastal Plain intensively for specimens during the past 10 years – for instance, the absence of Deltamys in the several collections performed between the site numbers 5 and 6 of Fig. 1. On the other hand, both in our own research and in the revision made by González & Pardiñas (2002), it was observed that in certain locations (Tramandaí, for instance) Deltamys seems to be relatively common.

The analysis of the cranial measurements of animals from the two groups for which we have representative material (Osório and Tramandaí in the Tramandaí Group and Taim from the Taim Group) revealed a significant separation between the two. This means that the molecular differentiation observed in this study was also accompanied by modifications in the morphology of the animals’ craniums, and these differences suggest that Deltamys is divided into two taxa – one to the North of the Lagoa dos Patos and the other to the south (extending as far as Buenos Aires). We observed that the craniums of the specimens of Taim group (south) were bigger (according to the Bergmann’s rule?), but a larger sample analysis is necessary. González & Massoia (1995) suggested that Deltamys could support two subspecies, one of which, D. k. langguthi, would correspond to the individuals inhabiting Brazil and Uruguay, whereas the remainder would be D. k. kempi. The fact that we found both molecular and morphological differences significant only between the individuals belonging to the groups we denominated Tramandaí and Taim indicates that, if Deltamys is in fact differentiated into two taxa, this differentiation is true only between the representatives of these two groups.

The molecular analyses conducted in this study reveal that Deltamys is a taxon whose differentiation as an akodontine genus must have occurred in the Tertiary, a little more than 3 Ma, a period of time that corresponds to the differentiation of the other akodontines (Reig, 1987; Smith & Patton, 1999). The molecular analyses also show that this rodent spread out over the Coastal Plain after the first marine transgression in the Pleistocene (which was about 400 ka) a period which agrees perfectly with the geological age estimated for this region of the South American Atlantic coastline.

The trees generated in the NJ, MP, ML, BI and MJ analyses, the hierarchical amova test and the cranial measurements demonstrate clearly that D. kempi, on the Coastal plain, initially divided into two groups probably in the region of Charqueadas: one occupying a small portion to the north of this region and the other spreading widely to the south – from Charqueadas in Brazil, along the coast of Uruguay, to the River Plate in Argentina. In this process, the phenomena of marine transgression and regressions which moulded its habitat, together with the occurrence of successive chromosomal rearrangements, were certainly the fundamental factors in D. kempi diversification.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendix

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), Financiadora de Estudos e Projetos (FINEP), Fundação de Amparo à Pesquisa do Estado do Rio Grande do Sul (FAPERGS) and the Organization of American States (OAS) have supported this study. The authors are grateful to Luciano S. Silva for technical help, to Dr Mara H. Hutz for several facilities provided, to Dr Jorge A. Villwock for the discussion of the geology of the Coastal Plain, to Dr Ana L. Garcia for help in preparing the figures and to Valeria C. Muschner for help in the molecular analyses.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendix
  • Baker, R.J., Porter, C.A., Patton, J.C. & Van Den Bussche, R.A. 2000. Systematic of bats of the Family Phyllostomidae based on Rag2 DNA sequences. Occas. Pap. Tex. Tech. Univ. Mus. 203: 116.
  • Bandelt, H.J., Forster, P. & Röhl, A. 1999. Median-joining networks for inferring intraspecific phylogenies. Mol. Biol. Evol. 16: 3748.
  • Bianchini, J. & Delupi, L. 1994. Consideraciones sobre el estado sistemático de Deltamys kempi Thomas, 1917 (Cricetidae: Sigmodontinae). Physis 49: 2735.
  • Castro, E., Mattevi, M.S., Maluf, S. & Oliveira, L.F.B. 1991. Distinct centric fusions in different populations of Deltamys kempi (Rodentia: Cricetidae) from South America. Cytobios 68: 153159.
  • D’Elia, G. 2003. Phylogenetics of Sigmodontinae (Rodentia, Muroidea, Cricetidae), with special reference to the akodont group, and with additional comments on historical biogeography. Cladistics 19: 307323.
  • D’Elia, G., González, E.M. & Pardiñas, U.F.J. 2003. Phylogenetic analysis of sigmodontine rodents (Muroidea), with special reference to the akodont genus Deltamys. Mamm. Biol. 68: 351364.
  • DeBlase, A.F. & Martin, R.E. 1981. A Manual of Mammalogy with Keys to Families of the World, 2nd edn. Brown Company Publishers, Dubuque, IA.
  • Excoffier, L., Smouse, P.E. & Quatro, J.M. 1992. Analysis of molecular variance inferred from metric distances among DNA haplotypes application of human mitochondrial DNA restriction data. Genetics 131: 479491.
  • Farris, J.S., Källersjö, M., Kluge, A.G. & Bult, C. 1994. Testing significance of incongruence. Cladistics 10: 315319.
  • Farris, J.S., Källersjö, M., Kluge, A.G. & Bult, C. 1995. Constructing a significance test for incongruence. Syst. Biol. 44: 570572.
  • Felsenstein, J. 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783791.
  • Fronza, M., Wainberg, R. & Cataleo, G. 1981. Multiple sex chromosomes in Deltamys kempi (Rodentia: Cricetidae): Preliminary steps towards the establishment of the XY1Y2/X1X1X2X2 system. Caryologia 34: 457466.
  • Fu, Y.X. & Li, W.H. 1993. Statistical test of neutrality of mutations. Genetics 133: 693709.
  • González, E.M. & Massoia, E. 1995. Revalidación del género Deltamys Thomas, 1917, con la descripción de una nueva subespecie de Uruguay y sur de Brasil (Mammalia: Rodentia: Cricetidae). Comunic. Zool. Mus. Hist. Nat. Montevideo 12: 18.
  • González, E.M. & Pardiñas, U.F.J. 2002. Deltamys kempi. Mamm. Spec. 711: 14.
  • Hall, T.A. 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucleic Acids Symp. Ser. 41: 9598.
  • Huelsenbeck, J.P. & Rannala, B. 1997. Phylogenetic methods come of age: testing hypothesis in an evolutionary context. Science 276: 227232.
  • Huelsenbeck, J.P. & Ronquist, F. 2001. MrBayes: Bayesian inference of phylogenetic trees. Bioinformatics 17: 754755.
  • Imbrie, J., Hays, J., Martinson, D., McIntyre, A., Mix, A., Morley, J., Pisias, N., Prell, W. & Shackleton, N. 1984. The orbital theory of Pleistocene climate: support from a revised chronology of the marine δ18O record. In: Milankovitch and Climate (A.Berger, J.Imbrie, J.Hays, G.Kukla & B.Saltzman, eds), pp. 269305. Reidel Publishing Co., Dordrecht.
  • Irwin, D., Kocher, T. & Wilson, A. 1991. Evolution of the cytochrome b gene of mammals. Mol. Ecol. 32: 128144.
  • Kimura, M. 1980. A simple method for estimating evolutionary rates of base substitutions through comparative studies of nucleotide-sequences. J. Mol. Evol. 16: 111120.
  • Kumar, S., Tamura, K. & Nei, M. 2004. Mega3: integrated software for molecular evolutionary genetics analysis and sequence alignment. Brief Bioinform. 5: 150163.
  • Martin, L., Bittencourt, A.C.S.P. & Vilas-Boas, G.S. 1982. Primeira ocorrência de corais pleistocênicos da costa brasileira: datação do máximo da penúltima transgressão. Ciência da Terra 1: 1617.
  • Massoia, E. 1964. Sistemática, distribución geográfica y rasgos etoecológicos de Akodon (Deltamys) kempi (Rodentia: Cricetidae). Physis 24: 299305.
  • Medrano, J.F., Aasen, E. & Sharrow, L. 1990. DNA extraction nucleated red blood cells. Biotechniques 8: 43.
  • Musser, G.G. & Carleton, M.D. 2005. Superfamily Muroidea. In: Mammal Species of the World: A Taxonomic and Geographic Reference (D.E.Wilson & D.M.Reeder, eds), pp. 8941531. Johns Hopkins University Press, Baltimore, MD.
  • Nixon, K.C. 1999. winclada (beta) Version 0.9.9. Published by the author, Ithaca, NY.
  • Palma, R.E., Marquet, P.A. & Boric-Bargetto, D. 2005. Inter- and intraspecific phylogeography of small mammals in the Atacama Desert and adjacent areas of northern Chile. J. Biogeogr. 32: 19311941.
  • Pardiñas, U.F.J., D’Elia, G. & Ortiz, P.E. 2002. Sigmodontinos fósiles (Rodentia, Muroidea, Sigmodontinae) de América del Sur: estado actual de su conocimiento y prospectiva. J. Neotrop. Mammal. 9: 209252.
  • Pardiñas, U.F.J., Teta, P., D’Elia, G., Cirignoli, S. & Ortiz, P.E. 2007. Resolution of some problematic type localities for sigmodontine rodents (Cricetidae, Sigmodontinae). In: The Quintessential Naturalist: Honoring The life and Legacy of Oliver P. Pearson (D.A.Kelt, E.P.Lessa, J.Salazar-Bravo & J.L.Patton, eds), pp. 391416. University of California Press, Berkeley, Los Angeles, London.
  • Posada, D. & Crandall, K.A. 1998. modeltest: testing the model of DNA substitution. Bioinformatics 14: 817818.
  • Ramos-Onsis, S.E. & Rozas, J. 2002. Statistical properties of new neutrality tests against population growth. Mol. Biol. Evol. 19: 20922100.
  • Reig, O.A. 1987. An assessment of the systematics evolution of Akodontini, with the description of the new fossil species of Akodon (Rodentia: Cricetidae). Field. Zool. 39: 347399.
  • Rozas, J., Sánchez-DelBarrio, J.C., Messeguer, X. & Rozas, R. 2003. DnaSP, DNA polymosphism analysis by the coalescence and other methods. Bioinformatics 19: 24962497.
  • Sbalqueiro, I.J., Mattevi, M.S. & Oliveira, L.F.B. 1984. An X1X1X2X2/X1X2Y mechanism of sex determination in a South American rodent, Deltamys kempi (Rodentia: Cricetidae). Cytogenet. Cell Genet. 38: 5055.
  • Scheneider, S., Roessli, D. & Excoffier, L. 2000. Arlequin Version 2.0: A Software for Population Genetic Data. Genetics and Biometry Laboratory, University of Geneva, Geneva, Switzerland.
  • Shimodaira, H. & Hasegawa, M. 1999. Multiple comparisons of log-likelihoods with applications to phylogenetic inference. Mol. Biol. Evol. 16: 11141116.
  • Smith, M. & Patton, J. 1991. Variation in mitochondrial cytochrome b sequence in natural populations of South American akodontine rodents (Muridae, Sigmodontinae). Mol. Biol. Evol. 8: 149177.
  • Smith, M. & Patton, J. 1993. The diversification of South American murid rodents: evidence from mitochondrial DNA sequence data for the akodontine tribe. Biol. J. Linn. Soc. 50: 149177.
  • Smith, M. & Patton, J. 1999. Phylogenetic relationships and the radiation of sigmodontine rodents in south America: evidence from cytochrome b. J. Mamm. Evol. 6: 89128.
  • Smith, M. & Patton, J. 2007. Molecular phylogenetics and diversification of South American grass mice, genus Akodon. In: The Quintessential Naturalist: Honoring the Life and Legacy of Oliver P. Pearson (D.A.Kelt, E.P.Lessa, J.Salazar-Bravo & J.Patton, eds), pp. 827858. University of California Publications in Zoology, University of California Press.
  • Swofford, D.L. 2001. paup*: Phylogenetic analysis using parsimony (* and other methods), Version 4.0b10, Sinauer Associates, Inc., Publishers, Sunderland, MA.
  • Tajima, F. 1989. The effect of change in population size on DNA polymorphism. Genetics 123: 597601.
  • Tamura, N. & Nei, M. 1993. Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Mol. Biol. Evol. 10: 512526.
  • Templeton, A.R. 1983. Convergent evolution and non-parametric inferences from restriction fragment and DNA sequence data. In: Statistical Analysis of DNA Sequence Data (B.Weir, ed.), pp. 151179. Marcel Dekker, New York.
  • Thomas, O. 1917. On small mammals from the Delta del Paraná. J. Nat. Hist. (=Ann. Mag. Nat. Hist.) 20: 95100.
  • Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F. & Higgins, D.G. 1997. The Clustal X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res. 24: 48764882.
  • Tomazelli, L.J. & Villwock, J.A. 1996. Quaternary geological evolution of Rio Grande do Sul coastal Plain, southern Brazil. An. Acad. Bras. Cienc. 68: 373382.
  • Tomazelli, L.J., Dillenburg, S.R. & Villwock, J.A. 2000. Late quaternary geological history of Rio Grande do Sul coastal Plain, southern Brazil. Rev. Bras. Geoc. 30: 470472.
  • Udrizar Sauthier, D.E., Abba, A.M., Pagano, L.G. & Pardiñas, U.F.J. 2005. Ingreso de micromamíferos brasílicos en la provincia de Buenos Aires, Argentina. Mastozool. Neotrop. 12: 9195.
  • Urien, C.M., Martins, L.R. & Martins, I.R. 1980. Modelos deposicionais na plataforma continental do Rio Grande do Sul (Brasil), Uruguai e Buenos Aires (Argentina). Rev. Notas Técnicas 2: 1326.
  • Urien, C.M., Martins, L.R. & Martins, I.R. 1981. Evolução geológica do quaternário do litoral atlântico uruguaio e regiões vizinhas. Rev. Notas Técnicas 3: 743.
  • Villwock, J.A. 1984. Geology of the Coastal province of Rio Grande do Sul, southern Brazil. A synthesis. Pesquisas 16: 549.
  • Villwock, J.A. & Tomazelli, L.J. 1998. Holocene Coastal Evolution in Rio Grande do Sul, Brazil. In: Quaternary of South America and Antartic Peninsula (J.Rabassa, ed.), pp. 283296. Balkema, Rotterdam.
  • Villwock, J.A., Tomazelli, L.J., Loss, E.L., Dehnhardt, E.A., Fo, N.O.H. & Bachi, F.A. 1986. Geology of the Rio Grande do Sul Coastal Province. In: Quaternary of South America and Antartic Peninsula (J.Rabassa, ed.), pp. 7997. Balkema, Rotterdam.
  • Xia, X. & Xie, Z. 2001. DAMBE: data analysis in molecular biology and evolution. J. Hered. 92: 371373.
  • Xiang, Q.Y., Moody, M.L., Soltis, D.E., Fan, C. & Soltis, P.S. 2002. Relationships within Cornales and circumscription of Cornaceae-matK and rbcL sequence data and effects of outgroups and long branches. Mol. Phylogenet. Evol. 24: 3557.

Appendix

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendix

Specimens examined of Deltamys kempi. Brazil, Rio Grande do Sul State, Torres City, ZE13*; Osório City, MN42081; Tramandaí City, MN42085, MN42073, MN42079, MN42072, MN42065, MN42064, MN42056, MN42057, LF587* (MN42046 = LF994), MN42041, LF589* (MN42048 = LF1060), LF591* (MN42047 = LF1021), MN42054, MN42070, MN42039, MN42059, MN42045; Charqueadas City, LF2371*, LF2392*; Tapes City, LF421*; Taim City, MN42078, MN42058, MN42071, MN42080, MN42084.

*Samples without vouchers were from rodents released following capture.