Evolutionary and systematic relationships among tuco-tucos of the Ctenomys pundti complex (Rodentia: Octodontidae): a cytogenetic and morphological approach

We present new cytogenetic, morphometric, and sperm morphology data of eight populations belonging to the C. pundti complex from Southern Córdoba and Eastern La Pampa Provinces in Argentina. The diploid numbers ranged from 2n = 44 to 2n = 50, and C- bands revealed a pattern of centromeric and pericentromeric heterochromatin. Comparisons of G-banded karyotypes revealed that the 2n = 44 (Holmberg, Santa Catalina, Sampacho), 2n = 46 (Realicó), 2n = 48 (El Guanaco, Guatraché), 2n = 46-48 (Vicuña Mackenna), and 2n = 50 (Puente Olmos) karyotypes, are closely related. In addition, these karyotypes show a high degree of homology (95%) with C. talarum talarum, despite the fact that five chromosomal rearrangements differentiate both taxa. Discriminant Function Analysis of morphometric data allows to distinguish three clusters: i) the C. mendocinus species group, ii) C. t. talarum, and iii) populations of the C. pundti complex proposed herein. The close phylogenetic relationship between C. talarum and the C. pundti complex, which undoubtedly belong to the same evolutionary lineage, is well supported by two different kinds of evidence: the extensive chromosomal homology and the same symmetric type of sperm. The morphological and chromosomal differences show that these two forms have diverged recently.


INTRODUCTION
The fossorial rodents of the genus Ctenomys, commonly known as tuco-tucos, are characterized by an extensive karyotypic heterogeneity, ranging from 2n = 10 in C. steinbachi to 2n = 70 in C. pearsoni and C. dorbignyi (Reig & Kibliski 1969;Anderson et al. 1987;Cook et al. 1990;Ortells et al. 1990;Reig et al., 1990;Gallardo 1991;Ortells 1995). However, some species from western and central Argentina, referred to as the C. mendocinus group (Massarini et al. 1991;Freitas 1994;D'Elía et al. 1999), show a similar karyotype of 2n = 47 to 48, and recently has been described de 2n = 46 karyomorph for two population of C. azarae (Massarini et al. 1998). Moreover, some species exhibit a wide array of chromosomal polymorphisms, and interpopulation variation (Massarini et al. 1991;Freitas 1995;Massarini et al. 1998). The significance of this extensive intra and interspecific chromosomal variation has been the subject of considerable debate (Baker et al. 1983;Reig 1989;Nevo 1991;Ortells & Barrantes 1994;Freitas 1994;Freitas 1997). In one of the best studied species, Ctenomys talarum, two subspecies are currently recognized, C. t. recessus and C. t. talarum which include several populations distributed along the Atlantic coast of Buenos Aires Province, varying in diploid number from 2n = 46 to 2n = 50, as well as other populations in central and western areas of the species' distribution range, whose karyotypes have not been yet studied (Contreras & Reig 1965;Vidal Rioja 1985). Moreover, C. talarum recessus is, in certain areas, sympatric with C. australis, which belongs to the C. mendocinus group. Sperm morphology has also been used to discriminate between species groups in the genus.

MOTS-CLÉS
Ctenomys, cytogénétique, morphométrie, systématique, relations évolutives. Three types of sperm have been described: symmetric, simple-asymmetric and complexasymmetric (Vitullo et al. 1988;Vitullo & Cook 1991;Freitas 1995). C. talarum has symmetric sperm whereas species of the C. mendocinus group have simple asymmetric sperm. In Northern Córdoba province, new species with assymetric sperm have recently been described (Giménez et al. 1999), but in Southern Córdoba and Eastern La Pampa provinces, there is a suite of Ctenomys populations with symmetric sperm that has not been cytogenetically characterized yet. Although, C. azarae was described from material collected in La Pampa Province (Thomas 1903), in southern Córdoba the only described taxa is Ctenomys pundti (Nehring, 1900, type locality: Alejo Ledesma, Marcos Juárez Department). These two species can be easily distinguished in the field by external morphology. Ctenomys azarae is larger and more robust, having a bigger hindfoot, than Ctenomys pundti, which is more similar to C. talarum recessus. The skull of C. azarae is larger with wide auditory bullae, while C. pundti and C. talarum have narrower bullae. Here we report new cytogenetic, morphometric, and sperm type data of Ctenomys populations from Southern Córdoba and Eastern La Pampa, which may help to interpret the phylogenetic relationships among these populations and other taxa of the genus of central Argentina.

MATERIAL AND METHODS
A morphometric comparative study based on 24 craniometric variables of 241 adult specimens from a total of 14 localities, 10 in male samples and 12 in female samples, was performed on data from Ortells (1990) and Massarini (1992) ( Table 1). The studied craniometric variables are shown in Fig. 1. The data were analyzed by means of Discriminant Function Analysis (DFA), using the program STATISTICA. This kind of analysis allows to find a linear combination of variables that maximize the differences among previously defined groups. Groups were defined according to the geogrphapic sampling locality, which corresponds to local populations. Chromosomal preparations of C. pundti and C. t. talarum specimens were obtained from bone marrow of animals injected with yeast one day before sacrifice (Lee & Elder 1980), and G-and C-banded karyotypes by the methods of Seabright (1971) and Hsu (1974), respectively. Chromosomal nomenclature follows Levan et al. (1964), and nomenclature relating to the chromosomal size and arrangement of chromosome  Number of males and females used in morphometric analysis, localities of collection with its coordinates, and species to which they belong.  Table 2. Ctenomys t. talarum specimens included in this study were collected at the type locality (Magdalena, Buenos Aires Province). Studied specimens were deposited in the Museo de Ciencias Naturales "Lorenzo Scaglia" from Mar del Plata, Buenos Aires Province, Argentina.
All specimens analysed in this study, showed the symmetric type of sperm morphology. The discriminant function analysis of morphometric data (Table 3A & 3B) showed that for both sexes it is possible to distinguish clusters of populations that correspond to different species. The first discriminant function, which accounts for 51 and 59% of total variance in males and females respectively, shows the presence of two clusters. One of the clusters includes populations of C. t. talarum and C. pundti, with negative values; and the second cluster, species of the C. mendocinus group, with values greater or equal to 0. In males, the second discriminant function (which explains 22% of total variance) separates, within the first group, populations of C. t. talarum, with negative values, from those of the C. pundti complex, with positive values. In females, C. t. talarum and C. pundti presented positive values and negative values, respectively, for the second discriminant function function (which accounts for 21% of total variance). In the mendocinus group the second function allows to distinguish in males, populations of C. azarae, with high positive values, from those of C. porteousi and Tiranti S. I. et al. 76  and 2n = 48 karyotypes, involve a complex set of mutation events that suggests a significant distinction between these two forms. However, the existence of the 2n = 47 heterokaryotype in Vicuña Mackenna points out that this incipient chromosomal differentiation does not involve reproductive isolation. The 2n = 50 karyotype from Puente Olmos is the more divergent. With the data presented so far, it was not possible to establish the nature of the chromosomal rearrangements involved in karyotypic divergence, but it is expected that high resolution G-banding would provide an answer to this question. The close phylogenetic relationships between C. talarum and the C. pundti complex, is well supported by two different kind of evidence: the great chromosomal homology and the sharing of the same type of sperm. Although, they undoubtedly belong to the same evolutionary lineage, morphological and cytogenetic differentiation suggest that these are recently diverged species. Cytogenetic and sperm morphology evidence along with the morphometric analysis support the homogeneity within lineages of Ctenomys. On one side the C. mendocinus group shares the asymmetric sperm type, and a high degree of chromosomal uniformity, but these taxa are morphologically differentiated from each other. On the other side, the populations of the C. pundti complex and C. talarum, also share the same symmetric sperm morphology and a significant degree of chromosomal homology. Morphometric analysis allows to distinguish C. t. talarum from the C. pundti complex, but populations of the latter, though variable in chromosome number, cannot be discriminated by means of morphometric criteria. In this sense, it is noteworthy that the level of morphological differentiation that exists between populations of the C. pundti complex, is comparable to differentiation among C. azarae populations. Based solely on morphology, Justo (1992) described a new subspecies of C. talarum (C. t. occidentalis) from La Pampa and included samples from El Guanaco. However, our present study shows that the latter is better referred to as belonging to the C. pundti complex rather than Only centroids of each a priori defined group are plotted. Circles assemble populations belonging to the same species. In females (A) population 1 belongs to C. australis, 2 to C. mendocinus, 3 to C. porteusi, 4-5 to C. t. talarum, 6-10 to C. pundti and 11-12 to C. azarae. In males (B) populations 1-3 belong to C. australis, 4 to C. mendocinus, 5 to C. porteusi, 6 to C. talarum, 8-9 to C. pundti and 10 to C. azarae.

A B
069_080_TIRANTI 3/03/05 13:25 Page 78 P R O O F to C. t. talarum, on grounds of cytogenetic and morphometric criteria . Similarly, we propose that rest of the populations studied by Justo should be assigned to the pundti complex, which conform a closely related but distinguishable unit from C. talarum.
Despite the growing amount of information available for these forms, there exists a nomenclatorial problem regarding C. pundti. Nowadays, in the type locality of this species (Alejo Ledesma), there are no tuco-tucos, due to the intense ecosystem modification brought upon by agriculture.
The nearest population of tuco-tucos is in Puente Olmos, which is 50 km W of Alejo Ledesma (Reig et al. 1992). Therefore, we propose to use the name C. pundti for the 2n = 50 population from Puente Olmos. Additionally, this one and the rest of the populations (2n = 44, 46, 47, 48) should be included in the C. pundti complex until more information is available to further solve the taxonomy and systemtics of these comlex forms from Central Argentina.