Journal of Heredity 2003:94(4)
© 2003 The American Genetic Association 94:315-328
Insight Into the Origin of Endemic Mediterranean Ichthyofauna: Phylogeography of Chondrostoma Genus (Teleostei, Cyprinidae)
From CNRS UMR 5023, Ecologie des Hydrosystèmes Fluviaux, Université Claude Bernard Lyon 1, 43 Bd du 11 Novembre 1918, 69622,Villeurbanne Cedex, France (Durand and Laroche); Department of Zoology, Universita di Napoli "Frederico II," 8 Via Mezzocannone I-80134, Napoli, Italy (Bianco); and Laboratoire d'Hydrobiologie, EA 2202, case 36, Université de Provence, Pl. V. Hugo, 13331 Marseille, Cedex 3, France (Gilles). Durand is currently at Centre IRD de Bel Air, BP 1386, Dakar, Senegal. Laroche is currently at IUEM, Ressources Halieutiques–Poissons Marins, Place Nicolas Copernic Technopôle Brest-Iroise, 29 280 Plouzané, France.
Address correspondence to J.-D. Durand at the address above, or e-mail: durandjd{at}ird.sn.
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
The Chondrostoma genus is widespread in Europe, with numerous endemic species in northern Mediterranean rivers. We reconstructed the phylogenetic relationships of this genus, using the whole cytochrome b sequence and compared the two freshwater fish dispersion hypotheses: (1) dispersion around the Mediterranean Sea during the Lago Mare phase of the Messinian salinity crisis (Bianco's hypothesis) and (2) an older and more gradual colonization of the Mediterranean rivers (Banarescu's hypothesis). All phylogenetic analyses identified two levels of divergences, implying two radiation events in the Chondrostoma genus. The first radiation mainly concerned Mediterranean species, whereas the second one includes Danubian and Mesopotamian species. This phylogeographic pattern was already observed for the genus Squalius, which exhibits a similar geographic range distribution in Europe and probably is shared with several other Mediterranean genera, such as Scardinius, Rutilus, and Telestes. Furthermore, assuming a molecular clock of 1% per million years, the first radiation appears consistent with a Messinian dispersion during the Lago Mare, 5.3 million years ago, whereas the second one may correspond to a Mesopotamian dispersion through the Black Sea to the Danube system. According to our results, the Lago Mare theory is strengthened, and a more recent and pre-Pleistocene colonization of the Danube from Mesopotamian freshwater fishes is suggested.
The subfamily Leuciscinae is represented within the Cyprinidae family by numerous species distributed throughout mainland Eurasia (excluding Arabia, India, and Southeast Asia), Japan, and North America (Howes 1991). In Europe, basing their conclusions on molecular evidence, Zardoya and Doadrio (1999), and Durand et al. (2002) identified 10 Leuciscinae lineages: Pachychilon; Scardiniusi; Tropidophoxinellusi; Abramis + Vimba + Blicca + Acanthalburnus + Acanthobrama; Anaecypris + Leucaspius + Chalcalburnus + Alburnus; Alburnoides; Leuciscus (Squalius) + Ladigesocypris; Rutilus; Leuciscus (Leuciscus) + Aspius; and Telestes + Chondrostoma. Several Leuciscinae lineages display a range distribution that encompasses the Danubian district, as well as some Mediterranean and Mesopotamian districts (Bianco 1990). Although many endemic species are found in Mediterranean and Mesopotamian areas, central Europe shows a relatively uniform ichthyofauna. This quite unusual species distribution has quickly led to several biogeographic scenarios that can be summarized in two alternative hypotheses, mainly differing in the geological events allowing the fish dispersion.
According to the most standard hypothesis (Banarescu 1960, 1992), the colonization of European drainages occurred by repeated waves of dispersion from Siberia to Northern Europe. This dispersal process might have begun in the Oligocene (about 35 million years ago) and lasted until the late Pliocene (about 1.7 million years ago). Mediterranean rivers were colonized through river captures from central Europe by old Siberian lineages, whereas present Euro-Siberian species in central and northern Europe descended from more recent dispersal events. This theory assumes a succession of gradual dispersion over time and space, functions of the hydrographic evolution and the geotectonic history of the European landmass.
In contrast, according to Bianco (1990), the first penetration in the western Palearctic region is mainly the result of dispersal events, namely cyprinids and other primary freshwater fishes, which colonized the western Palearctic area from eastern Asia during the Miocene period, following the progressive salinity fall of the Paratethys. The Mediterranean Sea, an evaporative basin, was at that time hypersaline, preventing any dispersion of primary freshwater fish. After the Messinian period (5.5 million years ago), the water salinity decreased during the 100,000 years of desertification as Paratethys waters drained to the Mediterranean basin (Hsü et al. 1973, 1977). Colonization of the Mediterranean basin by primary fish species of Paratethys origin was hypothesized to have taken place only during the freshwater Lago Mare period, termed the "Lago Mare" theory by Bianco (1990). As supported by much geological, paleontological, and palaeo-ecological data, this theory suggests that the Paratethys was an important environment in the dispersal movement of most freshwater species. The opening of the Straits of Gibraltar, following the "Lago Mare" phase, led to the invasion of Atlantic marine waters that represents a strong vicariant event responsible for the speciation and the origin of the present endemic taxa in southern peninsular Europe. This alternative scenario may be the "vicariant point of view" of the freshwater biogeography in Europe.
Although the Lago Mare theory has been extensively used in recent studies to justify several ancient divergences in phylogenies of some close Mediterranean freshwater fishes (Apostolidis et al. 1997; Brito et al. 1997; Durand et al. 2000, 2002; Gilles et al. 1998a, b; Machordom and Doadrio 2001), few have really attempted to estimate the validity of such a scenario. Only Penzo et al. (1998) and Zardoya and Doadrio (1999) have tried to appreciate the importance of the Messinian salinity crisis in the origin of freshwater gobies and Mediterranean Cyprinids. Penzo et al. (1998) do not find any clear evidence concerning the impact of the Messinian salinity crisis in the goby evolutionary process. However, Penzo et al. (1998) did not really test the importance of the Lago Mare in the freshwater fish dispersion in the Mediterranean area, but instead tested the importance of the salinity crisis in the adaptation of a marine species to freshwater conditions, which is quite far from the original theory of Bianco (1990). Zardoya and Doadrio (1999) demonstrate the existence of a Mediterranean ichtyological district independent of the central European route, as already suggested by Bianco (1990), and thus highlighted the discrepancy in the classical hypothesis of Banarescu (1992). However, they proposed only a few explanations concerning the European Leuciscinae contemporary distribution and failed to conclude about the validity of the Lago Mare scenario. Furthermore, they have not analyzed any Leuciscinae fishes from the Middle East, even though this geographic area is known as an important route of cyprinids migrations (Howes 1991). Mesopotamia is a migration crossroads for many taxa in addition to freshwater fishes. Many authors (Banarescu 1992; Coad 1996; Kosswig 1955; Por and Dimetman 1985) have shown the importance of this geographic area, which is a border between Europe, Africa, and Asia. A more exhaustive study of the phylogenetic relationships of Turkish and European Leuciscinae may result in a better comprehension of the biogeography of this subfamily in Europe.
We have chosen the Chondrostoma genus (Chondros means cartilaginous; Stoma means mouth), which is characterized by a mouth clearly subterminal (which is contrary to the etymology but does not consist of keratinized cells), with transverse or arched slit, without barbell, and with the upper jaw forming a muzzle well arched, with very hard oral lips and a sharp boarder. This genus is present in almost all of Europe (except the northern countries), from the Atlantic to the Caspian and Mediterranean Seas, to the North and the Baltic Seas, as well as in minor Asia, the Caucasus, and Mesopotamia (northwestern Asia). In accordance with the bibliographical data (Collares-Pereira 1978, 1980; Elvira 1987, 1990, 1991; Mathias 1921; Mesquita et al. 2001; Nelva et al. 1988), we distinguish more than 23 forms that correspond to 23 species constituting a clade (based on morphological differences). We take into account Telestes species complex, which is the sister group of Chondrostoma.
This study proposes to establish the phylogenetic relationships among Chondrostoma species endemic to different Mediterranean Rivers, from Spain to Turkey, by means of nucleotide sequence analyses of the cytochrome b mitochondrial gene. Our primary objective is to clarify the timing and mode of speciation or lineage divergence within this genus, with the goal of highlighting the role played by the Lago Mare in the evolutionary process of Mediterranean Leuciscinae.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
The specimens studied come from the perimediterranean area (Figure 1; Table 1). We include virtually all the species of the genus Chondrostoma. Morphological identification was done in the laboratory (Bianco PG, personal collection).
|
|
mtDNA Amplification and Sequencing
We extracted total genomic DNA from scales or muscle, following the protocol described by Kocher et al. (1989), with some simplifications (Briolay et al. 1998). Amplification conditions for the polymerase chain reaction (PCR; Saiki et al. 1988) were 30 cycles of denaturation at 94°C for 1 min., annealing at 55°C for 30 s, and extension at 72°C for 1 min. The PCR was performed in a 50 µl reaction volume containing 12.5 nM MgCl2, 10 nM of each of the four deoxynucleotides, 1 µl of each 10 µM primer, 10x PCR buffer II (Perkin Elmer–Cetus) and 2 U Amplitaq (Promega) in a programmable thermal cycler (Perkin Elmer–Cetus, Model 9600). A fragment of 1,300 bp of the 5' cytochrome b extremity was amplified. PCR primers (Briolay et al. 1998) used were L15267 and H16526. Primer names indicate the light (L) versus heavy (H) DNA strand, and numbers represent the position of the 3' base of the oligonucleotide in the complete mitochondrial sequence of Cyprinus carpio (accession number X61010). Double-stranded PCR products were purified with Qiaquick (QIAGEN) columns. Direct sequencing was carried out with T7 DNA polymerase kits (Pharmacia). The primers used for sequencing were L15639, L15840 and L15267, H15512, H15149, and H16461 (Briolay et al. 1998). We used the C. carpio sequence (Chang et al. 1994) as outgroup, because this species belonged to the Cyprininae (which constitute a monophyletic grou) and was sufficiently distant from the subfamily of the Leuscicinae, thus avoiding any background noise resulting from introgression. Sequences were aligned with MUST (Philippe 1993). For the studies on maximum parsimony (MP), we used the unweighted parsimony and a weighting parsimony. Transversion-transition ratio was equal to 2:1 at the third position. The parsimony algorithm necessitated long calculation times using a heuristic search of the program PAUP* (4.0d64 provided by D. Swofford), which limited us to 100 replicas, with a TBR swapping (Tree Bisection and Reconnection) of 100. For this reason, to obtain a more significant bootstrap test (Felsenstein 1985; Hedges 1992; Hillis and Bull 1993; Lecointre et al. 1993; Zharkikh and Li 1992), the fast search option of the program PAUP* with 1,000 replicas was used. For the unweigted parsimony we used the heuristic search of the program PAUP* with a TBR swapping of 100.
For the neighbor-joining studies (NJ; Saitou and Nei 1987) the PAUP* program was used. The distance was estimated, on a Juke and Cantor distance, on a Tamura-Nei distance with alpha parameter for the gamma shape equal to 0.21, and on a maximum likelihood (ML) distance (as described below). The topological difference between the trees was tested with Templeton's test (Wilcoxon sign-rank tests; Templeton 1983), and the likelihood ratio test (Kishino and Hasegawa 1989). When the different models produced the same topology (not significantly different with the likelihood ratio test and Templeton test), the simplest model was performed. Furthermore, we tested the impact of the species and the different cytochrome b domains (17, following Kocher and Stepien 1997: 5 domains on the inner surface of the mitochondrial membrane, 4 domains on the outer surface, and 8 transmembrane domains) on branch support by jackknifing both separately.
The ML estimation was performed using the following model of evolution retained by Modeltest (Posada and Crandall 1998), with nucleotide frequencies of p(A) = 0.2840, p(C) = 0.3001, p(G) = 0.1354, and p(T) = 0.2804. The rate variation among sites was modeled using a gamma distance and the gene-specific shape parameters 
= 0.96, with a proportion of invariant sites Pinvar = -/523 and a rate matrix A–C = 1.068, A–G = 30.852, A–T = 0.260, C–G = 1.577, C–T = 7.420, and G–T = 1.000. The procedure of Rogers and Swofford (1998) was used to reduce preliminary computing time.
We tested the effects of species and sites on the tree reconstruction for the unresolved branches. We used the ML reconstruction implemented in the Tree-Puzzle program (Strimmer and von Haeseler, 1996) to evaluate all the quartet puzzling (this is the number of groups constituted by four species out of a total number of 42 species). Indeed, to calculate the impact of the species in using sub-sampling of species, we visualized the species responsible for the unresolved quartets. An unresolved quartet is a quartet for which the maximum likelihood values for each of the three possible quartet topologies are not significantly different. Of the 11 1930 quartets analyzed (using the ML methods), 527 quartets were unresolved.
We analyzed the pattern of substitution for the four branches (as described in Figure 2) in observing the rate of substitution for the different positions. With this goal, we computed the rate of substitution for each position, using the PAUP* program and Modeltest 3.0 (Posada and Crandall 1998) on the complete cytochrome b (Figure 3). We used the PAUP* program to obtain the list of apormorphic sites involved in the four branches. For each branch, we analyzed the distribution of the substitution pattern, the overall consistency index (CI) value, and their impact.
|
|
The Relative Rate Test was performed with Phyltest (Kumar 1996) in order to determine if the substitution rate of the different sequences was homogeneous. Sequences whose substitution pattern was significantly different from the mean pattern were not included in the estimation. The molecular clock was calibrated using the scale established by Smith et al. (2002) and Dowling et al. (2002). These authors used fossil data (including several species of cyprinids) and mitochondrial DNA (cytochrome b). They found for this group a rate of substitution equal to 1.05% sequence divergence per pairwise comparison per million years (0.53% per lineage per million years, with an unconstrained regression), which we approximated to 1% per million years. The authors took into account the problem of past hybridization to calibrate the rate of substitution.
Separate analyses were performed on the complete data set for the Chondrostoma genus (528 pairwise comparisons) and on a partial data set comprising only one specimen representing a Chondrostoma species (153 pairwise comparisons). The distribution of mean sequence divergence estimates within groups was tested. The normality was tested using the 
2 test.
The TreeMap program (Page 1994b) was used to reconstruct the gene-district assemblage. (considering only vicariance phenomenon, the species coming from the same district would be closer on cytochrome b phylogeny for these freshwater fish). We used the heuristic search option. The significance of the observed fit between district area and cytochrome b (gene) trees can be evaluated by comparing the distribution of the same measure of fit for the random trees (Page 1994a,b). One null hypothesis is that the district area tree is independent of the gene tree. We tested this hypothesis by generating 10,000 random trees, using a proportional-to-distinguishable option, with the same number of district and the same number of taxa as the actual phylogeny, and we measured how well these fit the observed gene tree in comparison with the district tree. The proportion of random gene trees that have the same (or greater) number of speciation-separation events as the observed trees is the probability of obtaining the observed value by chance alone.
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Phylogenetic Reconstruction
The most parsimonious tree displayed a length equal to 1,473 steps with the unweighted parsimony (CI of 0.445 and an overall retention index (RI) of 0.633; tree not shown). 454 characters are variables, of which 345 characters are informative for the parsimony. The first divergence concerns species located in the Iberian Peninsula (C. macrolepidotus and C. arcasii). Surprisingly, the different species of Telestes are the sister group of C. genei. This topology was not supported by the bootstrapped tree. The Chondrostoma genus was monophyletic, with a low bootstrap proportion (52%). The two topologies (bootstrapped and not) were statistically different, because of the position of the Telestes species.
With the weighting parsimony transversion-transition ratio equal to 2:1 at the third position, we found three most parsimonious trees (1,720 steps) that were not significantly different from the bootstrapped tree (polyotomy, Table 2). 454 characters are variables, of which 347 characters are informative for the parsimony. This tree displayed two different clusters (Figure 2). The first group comprised the different species of Telestes; the second one, the different species of Chondrostoma. The first divergence concerned species located in the Iberian Peninsula (southern Iberia and central Iberia), then species of southern France (with a close affinity to the Ebro-Cantabric area), and then species of the Padano Venetian area, with no distinction for those of the Danubian, Albanian, and Aegeo Macedo Anatolian areas. This topology was not significantly different from the most parsimonious tree (with the unweighted parsimony criterion) with the Templeton and Kishino Hasegawa tests (alpha = 0.05; Figure 3).
On the other hand, the trees in NJ, bootstrapped NJ (for the Juke and Cantor distance and for the Tamura-Nei distance with gamma shape equal to 0.21) yielded the same topology because Telestes was included within the Chondrostoma genus in the latter model, but this topology was not supported. This result was not surprising because gamma distances are generally more realistic than nongamma distances, but they have larger variances than the latter. For this reason, they do not necessarily produce better results in phylogenetic inference unless the number of nucleotides used is very large (Nei and Kumar, 2000).
Lack of Resolution for Different Nodes: Soft Versus Hard Polytomies
When we removed each species involved in one of the 527 completely unresolved quartets we found only one species that induced a lack of resolution for the tree topology (i.e., by removing only the species C. soetta, representing one of the two species belonging to the Padano Venetian district, we obtained a weak bootstrap support (58%) for one basal node (node A as described previously in the neighbor-joining tree). When we conserved only four species representing the four groups involved in the unresolved branches (i.e., C. soetta–C. genei, C. toxostoma, C. lusitanicum–C. lemmingii, and C. polylepis–C. duriensis–C. p. willkommii), none of the branches was statistically supported (except node A, Figure 2), so these subsamples of species indicated a real lack of resolution in the tree. Removing the Telestes genus and C. soetta species induced a strong bootstrap support for node A (85% in NJ only) and a weak bootstrap support for node B (53% in NJ only).
On the 17 domains tested by jackknifing, none was responsible for the lack of bootstrap support for the four branches (A, B, C, and D). There are 2m–2 branches in a rooted tree (on a completely resolved tree), so for 42 taxa we have 82 branches. Thirty-five sites (7.15%; 7:A, 10:B, 6:C, and 12:D) were assigned to the four branches (4.88%), and 454 sites were assigned to the other branches.
The distribution of the substitution pattern for each site belonging to the four branches yielded a high rate of substitution in comparison to the other sites (Figure 4). The mean and SE were, respectively, for branch A, B, C, D equal to 6.88 ± 1.77, 7.34 ± 1.99, 7.82 ± 5.80, and 6.34 ± 4.38. The distribution of the other sites (belonging to the other branches) corresponded to a gamma distribution (Figure 4A). Two regression lines explained the correlation between the number of substitutions and the CI value for the apomorphic position belonging to the four branches (Figure 5). The first one yielded a r2 equal to.8237; the second one, a r2 equal to.7935.
|
|
Genetic Distances for the Different Species of Chondrostoma
We found that the species level (based on morphology) in the Chondrostoma genus could be separated in two mtDNA levels of divergence. The distribution of these two genetic divergence levels was binomial, whatever the distance chosen (Jukes and Cantor distance or Tamura-Nei with gamma shape = 0.96). The distribution of pairwise comparison for the Jukes and Cantor distance (the first one is the oldest) fit with a classical normal distribution (µ = 0.083;
± 0.010), gave a nonsignificant 2 test (P =.37). The second one (the earliest) fit with a classical normal distribution (µ = 0.033; ± 0.009), gave a nonsignificant 2 test (P =.80). Using the 1% of substitution per million years, we obtained a mean of time divergence (in millions of years) of 4.17 ± 0.53 for the first one (T1) and a mean of time divergence of 1.65 ± 0.47 for the second one (T2). The results were similar for Tamura-Nei with gamma shape = 0.21 with a higher variance (µ = 0.132; ± 0.024, and T1: 6.57 ± 1.19; µ = 0.043; ± 0.015, and T2: 2.13 ± 0.73). We tested the possible artifacts due to morphological bias, such as cryptic species, hybridization and introgression, by using all the taxa with no morphological criteria (528 pairwise comparisons). The distribution of sequence divergence displayed the same pattern as the previous distribution (Figure 6). This result indicated clearly that for the same level of taxonomic rank (i.e., the different Chondrostoma species), there are two levels of genetic differentiation.
|
Cytochrome b Tree Versus District Area Tree
Because only completely resolved topology can be tested (Page 1994), we used the most parsimonious tree with the weighting parsimony transversion-transition ratio equal to 2:1 at the third position for the gene tree, and two completely resolved topologies for the district trees.
We tested the most parsimonious tree (molecular data) against the classical hypothesis of Banarescu (1992) as described in the introduction (Figure 7A). The heuristic search found 8 speciation-separation events, 25 speciation events, 0 migratory events, and 22 sorting events (corresponding to sister species present in the area in the past but absent nowadays). We found no statistical support between the cytochrome b gene tree and the district area tree; the association between these two trees may be due to chance alone (with P = 0.8227).
|
Furthermore, we tested the most parsimonious tree (molecular data), with the district tree corresponding to the closest dichotomous topology to the molecular tree (Figure 7B). The heuristic search found 10 speciation-separation events, 23 speciation events, 0 migratory events, and 10 sorting events. We found no statistical support between the cytochrome b gene tree and the district area tree; this association between the two trees may be due to chance alone (with P =.4291).
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Chondrostoma Radiation or Molecular Pitfall?
Molecular phylogenies are providing major insights into our understanding of evolution. However, phylogenetic analyses of molecular data often recover polytomies (multifurcation rather than bifurcation relationships). The polytomy including the C. toxostoma, C. soetta, C. genei, and C. polylepis cluster and the eastern Mediterranean cluster, with short branches connecting the taxa, is the signature of a rapid cladogenesis event. Evolutionary theory suggests that in certain cases, such as after colonization of ecologically open territory, lineages should experience a burst of rapid evolutionary diversification (Jackman et al. 1999). If cladogenesis is truly simultaneous, the failure to resolve dichotomous relationships among the taxa in question is called a "hard polytomy" (Walsh et al. 1999). Multiple branching events that are extremely rapid in succession may be empirically indistinguishable from simultaneous branching and are called a "soft polytomy."
If organisms rapidly speciated from a common ancestor, molecular characters may not recover phylogenetic history. Internodes separating species are too short to contain more than a small number of shared derived characters, and these characters will be difficult to distinguish from homoplasious ones and therefore render the recovery of phylogenetic patterns very difficult (Fiala and Sokal 1985; Kraus and Miyamoto 1991; Lanyon 1988). In addition, the taxa resulting from such a radiation will contain a number of randomly shared ancestral polymorphisms, which may be contradictory and thus further complicate the recovery of the true phylogenetic history (Simon et al. 1994).
The biological reality of polytomies is a topic of active debate (DeSalle et al. 1994; Hoelzer 1997). Considerable work has centered on increasing the size of data sets, employing different types of data, and using alternative methods of phylogenetic reconstruction to resolve presumably soft polytomies. Although polytomies often can be resolved into sequential bifurcations by an increase in data, some polytomies remain uncertain even after analysis of very large data sets (Jackman et al. 1999, Walsh et al. 1999). Several authors have attempted to determine the amount of sequence data required to resolve a soft polytomy, using simulation. As demonstrated by Walsh et al. (1999) for the Auklet mtDNA, 215 to 1,237 bp should be sufficient (when a phylogenetic signal has been detected) to resolve sequential lineage splits that occurred at least 100,000 years apart, 2.6 million years ago. These authors analyzed 1,044 bp of cytochrome b, 522 bp of ND6, 507 bp of 12S rDNA, and 842 bp of ATPase 6 and 8, and demonstrated that it was not judicious to sequence more mtDNA.
Several researchers argue that an ancestral lineage sometimes can generate multiple descendent lineages at one time, resulting in multiple simultaneous speciation events represented by a hard polytomy. Biogeographic hypotheses, fossil data, and geologic evidence all suggests that cladogenesis need not occur strictly by dichotomous branching events. Hoelzer and Melnick (1994) have argued that geologic changes such as glaciations and changes in sea level could have isolated several populations of a widespread species, initiating their divergence and ultimate speciation through a peripatric process. Simultaneous isolation of multiple populations of an ancestral species by a single vicariant event is indicated in a number of species, such as the cichlid fishes in Africa (Meyer et al. 1990).
This radiation pattern in the phylogeny shows concomitant short internodes joining the mtDNA clades with no bootstrap support. The classical pitfall in obtaining this pattern is due to a high number of taxa with respect to the number of variable characters (in distance method) that are informative characters (in parsimony methods). This is not the case in our study, for which 454 characters are variables and 345 characters are parsimony informative for 42 taxa. Without the outgroup species (three of them) and Telestes species (six taxa)—in essence, 33 taxa—we obtain 324 variables characters and 237 characters informative for the parsimony. Another problem is using partial cytochrome b gene (Nei et al. 1998), but this was not the case in our study (we used the complete gene). It is difficult to prove a radiation (corresponding to a polytomy in the tree), because the lack of bootstrap support could be due to any number of pitfalls; however, the different tests that we did permit us to discriminate two kinds of unresolved branches in the data set. Phylogenetic analyses incorporating subsampling of species can identify cases in which including all species in the phylogenetic analysis produces the appearance of simultaneous branching of lineages when the branching is actually sequential. Using the analytical procedure recommended by Jackman et al. (1999) for using aligned molecular sequences to detect significantly supported relationships, we provide evidence for sequential branching for branches A and B and a simultaneous branching for the branches C and D.
Morphological Species Versus Molecular Species
The first general result is the observation of two levels of divergence in the distribution of pairwise distance. For a same level of morphological differentiation two levels of molecular differentiation exist. This result shows that morphology and cytochrome b are decoupled in the Chondrostoma genus. Two morphologically distinct sister species may have a high genetic distance (it is the case for C. genei and C. soetta) or low genetic distance (C. macrolepidotus and C. arcasii, or C. vardarensis and C. prespensis). Two subspecies (low morphological differences) may have a high genetic distance (C. polylepis duriensis and C. polylepis willkommii) or low (C. holmwoodii holmwoodii and C. holmwoodii meanderes). Two alternative explanations may clarify this observation. First, as already mentioned in some molecular systematic studies (Agnèse et al. 1990; Berrebi et al. 1990; Durand et al. 2002) the morphometric characters used by authors to infer the phylogenetic relationships within the Cyprinidae family are sometimes irrelevant. This problem is especially pronounced for low systematic levels such as in species or subspecies, where usually few morphometric traits can be used to determined these systematic levels.
The second explanation may be possible introgression mtDNA of endemic species by recent Danubian invaders. Durand et al. (2000) highlighted in the Squalius complex strong discrepancies between phylogenetic links in two related species, depending on whether molecular markers were used to infer the phylogeny (mtDNA or enzymatic loci). According to these authors, theses differences may be the consequence of genetic introgression that followed secondary contacts. This was probably also the case for the Chondrostoma genus, which may explain the lack of congruence between the taxonomic status and the phylogenetic distance of several species, such as C. prespensis, C. vardarensis, C. nasus, C. regium, C. holmwoodii, and C. oxyrhynchum. Thus, genetic introgression may explain their close link even if this assumption has to be demonstrated by the use of several independent markers.
The Chondrostoma Phylogeography: An Insight Into Mediterranean Biogeography?
According to our results, our tree topology is not in agreement with a gradual colonization of the Mediterranean areas as suggested by the Banarescu assumption. A dichotomic tree topology would be in accordance with a long and gradual colonization of Mediterranean rivers. Whereas, as mentioned above, a rapid and large colonization event as suggested in Bianco's theory would produce a burst-like tree topology that is closer to the observed tree. However, is the Lago Mare theory suggested by Bianco (1990) a rational alternative? The Messinian salinity crisis (MSC) and the theory of the Mediterranean Sea desiccation (Hsü et al. 1973, 1977) have been extensively debated for more than 25 years. This controversy recently took issue with a new model (Clauzon et al. 1996), which highlights a chronological distinction between marginal and deep evaporation, based on the results of new approaches (Butler et al. 1995; Clauzon et al. 1996). In this two-step model the MSC began 5.8 million years ago with a first sea-level fall, which was moderate (150 m) because the Mediterranean basin was not isolated to the Atlantic basin. However, in a second interval, after a brief flooding event, the Mediterranean basin became fully isolated (5.6 to 5.32 million years), leading to a sea-level drop of 2,000 m (Clauzon et al. 1997).
At the same time, the Paratethys also experienced an important sea-level drop that drained the Pannonian basin (Serbia) and isolated the Dacian (Roumania) and the Euxinic basins (Balck Sea; Clauzon et al., in press). Water in the Euxinic basin was limited to the deepest level of the basin (as with the Mediterranean basin), whereas the Dacian was suspended (and formed a lake) as with the Panonian basin (in the Pô valley; Clauzon et al., in press). During this period, the European hydrographic networks were very active, because fluvial incisions were made in Mediterranean margins (Clauzon et al. 1996). Paratethys basins were freshwater lakes (Marinescu 1992), whereas the Mediterranean basins were hypersaline. However, the Lago Mare fauna from the Paratethys colonized Mediterranean margins, which suggests that dispersion for freshwater or brackish organisms was possible on the coastal border of the Mediterranean basin then (Clauzon et al., submitted). The end of the MSC took place when the Atlantic reconnected to the Mediterranean basin through the Strait of Gibraltar or the Baltic Corridor (Krijgsman et al. 1999).
Lastly, if the MSC is still not falsified and seems to have played an important role in some freshwater or brackish organisms' dispersion, are our results consistent with a Mediterranean rivers colonization during this geological event? If the first Chondrostoma radiation (under the assumption that the molecular evolution of the cytochrome b followed a molecular clock rate of 1% per million years) occurred during the MSC (5.5 million years; Krijgsman et al. 1999), is it the only element in agreement with the Lago Mare theory? As indicated in the introduction, the MSC was not only a dispersion event but also a strong vicariant process. The main prediction of the historical biogeography is that concordant geographic distribution implies a common history (Croizat et al. 1974). Under a pure dispersalist scenario, similar distribution patterns can occur only by chance. Thus, it is only when strong vicariant events are responsible for concordant distribution that a common history for all taxa is expected. The amount of the phylogeography (Avise et al. 1987) to apply to the historical biogeography depends on not only the geographic distribution of independent taxa but also their genealogical partition. Thus, the genealogical concordance across multiple codistributed taxa in highlighting area relationships allows powerful inference concerning historical biogeography.
Taking into account previous results obtained in the phylogeography of the Leuciscus subgenus Squalius (Durand et al. 2000), a genus with similar ecological habitat preferences and with a similar geographic range distribution in Europe, our Chondrostoma phylogenetic tree displays a close phylogenetic-branching pattern, suggesting a close dispersion history. The first general result is the observation of two levels of divergence, suggesting two radiation events. Both radiations in both phylogenies show concomitant short internodes joining the mtDNA clades (especially in the Chondrostoma phylogeny). The first divergence concerns species located in the Iberian Peninsula, southern France (for Chondrostoma, Figure 8), Italy, Greece (for Squalius), and Turkey and central Europe. A second (earlier) divergence includes the Turkish, Danubian, Greek, and Italian (for Squalius) species-populations. Within the Squalius phylogeny the basal position of L. borysthenicus, which occurs around the Black Sea and in the northeastern part of Greece, suggests a Paratethytian origin for the Squalius genus. This first radiation occurred 5.11 to 3.87 million years ago. Thus, the Lago Mare theory seems to be the most parsimonious explanation for the high radiation in both these genera. During this event a Paratethyan ancestor of the Chondrostoma and Leuciscus subgenus Squalius may have reached all the main North Mediterranean rivers located in the Iberian Peninsula, southern France, Italy, and Greece. Mesopotamia also might have been colonized during this broad dispersal event.
|
The second wave of colonization probably took place approximately 2.23 to 1.07 million years ago. During this event, populations located in Mesopotamia may have reached the Danube through the Anatolian inland lake (Demirsoy 1996; Gorur et al. 1995) and the oligohaline Black and Caspian Seas. This assumption of a Mesopotamian origin is based on the high genetic divergence of Turkish Chondrostoma and L. cephalus within the second clade. From the Danube some population crossed the Alpine chain (probably in the Kosovar plain and in the Vardar valley), explaining the Danubian affinities for Greek Chondrostoma and the presence of L. cephalus around the Adriatic and the Aegean Seas (Durand et al. 1999). During this second colonization event, many secondary contacts probably occurred between the Messinian relicts (endemic to Mediterranean rivers) and the Mesopotamian invaders, explaining discrepancies between morphology and genetic distance (Durand et al. 2000).
|
|
Acknowledgments |
|---|
Durand is currently at Centre IRD de Bel Air, Dakar, Senegal; Laroche is currently at IUEM, Ressources halieutiques—Poissons Marins, Plouzané, France. We thank Erhan Ünlü, who kindly provided the Turkish samples used in this study. We are grateful to J.-P. Suc for his suggestions and information about the geological background of the Mediterranean area during the Messinian Salinity Crisis, and to Maria Villanueva, Bob Britton, and one reviewer for helpful comments.
|
Footnotes |
|---|
Corresponding Editor: Martin Tracey
Received September 11, 2002
Accepted April 23, 2003
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
-
Agnèse JF, Berrebi P, Lévêque C, Guéguan JF, 1990. Two lineages, diploid and tetraploid demonstrated in African species of Barbus (Osteichthyes, Cyprinidae). Aquat Living Resour. 3:305-311.
Alves MJ, Coelho MM, Collares-Pereira MJ, Dowling TE, 1997. Maternal ancestry of the Rutilus alburnoides complex (Teleostei, Cyprinidae) as determined by analysis of cytochrome b sequences. Evolution. 51:1584-1592.[CrossRef][Web of Science]
Apostolidis AP, Triantaphyllidis C, Kouvatsi A, Economidis PS, 1997. Mitochondrial DNA sequence variation and phylogeography among Salmo trutta L. (Greek brown trout) populations. Mol Ecol. 6:531-542.[CrossRef][Medline]
Avise JC, Arnold J, Ball RM, Bermingham E, Lamb T, Neigel JE, Reeb CA, Saunders NC, 1987. Intraspecific phylogeography: the mitochondrial DNA bridge between population genetics and systematics. Ann Rev Ecol Syst. 18:489-522.[Web of Science]
Banarescu PM, 1960. Fauna republicii populare Romîne Pisces-Osteichthyes. Bucaresti: Ed. Acad. Repub. Pop.
Banarescu PM, 1992. Zoogeography of fresh waters, vol. 2: distribution and dispersal of freshwater animals in North America and Eurasia. Wiesbaden: Aula Verlag.
Berrebi P, Lévêque C, Cattaneo-Berrebi G, Agnèse JF, Guégan JF, Machordom A, 1990. Diploid and tetraploid African Barbus (Osteichthyes, Cyprinidae): on the coding of differential gene expression. Aquat Living Resour. 3:313-323.
Bianco PG, 1990. Potential role of the palaeohistory of the Mediterranean and Paratethis basins on the early dispersal of Euro-Mediterranean freshwater fishes. Ichthyol Explor Freshwaters. 1:167-184.
Briolay J, Galtier N, Brito MR, Bouvet Y, 1998. Molecular phylogeny of Cyprinidae inferred from cytochrome b DNA sequences. Mol Phylogenet Evol. 9:100-108.[CrossRef][Web of Science][Medline]
Brito MR, Briolay J, Galtier N, Bouvet Y, Coelho MM, 1997. Phylogenetic relationships within genus Leuciscus (Pisces, Cyprinidae) in Portuguese freshwaters, based on mitochondrial DNA cytochrome b sequences. Mol Phylogenet Evol. 8:435-442.[CrossRef][Web of Science][Medline]
Butler RWH, Lickorish WH, Grasso M, Pedley HM, Ramberti L, 1995. Tectonics and sequence stratigraphy in Messinian basins, Sicily: constraints on the initiation and termination of the Mediterranean salinity crisis. Geol Soc Am Bull. 107:425-439.
Chang YS, Huang FL, Lo TB, 1994. The complete nucleotide sequence and gene organization of carp (Cyprinus carpio) mitochondrial genome. J Mol Evol. 38:138-155.[CrossRef][Web of Science][Medline]
Clauzon G, Rubino JL, Casero P, 1997. Regional modalities of the Messinian Salinity Crisis in the framework of two phases model. In: Neogene basins of the Mediterranean region: controls and correlation in space and time, R.C.M.N.S. Inter.-Coll., Catania, Program and Abstracts; 44–46.
Clauzon G, Suc JP, Gautier F, Berger A, Loutre MF, 1996. Alternate interpretation of the Messinian salinity crisis: controversy resolved? Geology. 24:363-366.
Clauzon G, Suc JP, Popescu SM, Marunteanu M, Rubino JL, Marinescu F, Jipa D, in press. Influence of the Mediterranean sea eustasy over the Dacic basin (eastern Paratethys) in the late Neogene. Basin Res.
Coad BW, 1996. Zoogeography of the fishes of the Tigris-Euphrates basin. Zool Mid East. 13:51-70.
Collares-Pereira MJ, 1978. Chondrostoma toxostoma arrigonis (Steindachner, 1866), a new Cyprinid to Portugal. Arq Mus Boc. 6:(12) 219-239.
Collares-Pereira MJ, 1980. Contribution to the knowledge of Iberian Cyprinid Chondrostoma lemmingi (Stein., 1866) and its affinities with Chondrostoma arrigonis (Stein., 7(12): 1866). Arq Mus Boc. 7(12):151-178.
Croizat L, Nelson G, Rosen DE, 1974. Centers of origin and related concepts. Syst Zool. 23:265-287.[CrossRef][Web of Science]
Cunha C, Mesquita N, Dowling TE, Gilles A, Coelho MM, 2002. Phylogenetic relationships of Eurasian and American cyprinids using cytochrome b sequences. J Fish Biol. 61:929-944.[CrossRef]
Demirsoy A, 1996. General and Turkish zoogeography. Ankara, Turkey: Meteksan A. S.
DeSalle R, Absher R, Amato G, 1994. Speciation and phylogenetic resolution. Trends Ecol Evol. 9:297-298.[CrossRef]
Dowling TE, Tibbets CA, Minckley WL, Smith GR, 2002. Evolutionary relationships of the plagopterins (Teleostei: Cyprinidae) from cytochrome b sequences. Copeia 3:655–678.
Durand JD, Templeton AR, Guinand B, Imsiridou A, Bouvet Y, 1999. Nested clade and phylogeographic analyses of the chub, Leuciscus cephalus (Teleostei, Cyprinidae) in Greece: implications for Balkan Peninsula biogeography. Mol Phylogenet Evol. 13:566-580.[CrossRef][Web of Science][Medline]
Durand JD, Tsigenopoulos CS, Unlü E, Berrebi P, 2002. Phylogeny and Biogeography of the family Cyprinidae in the Middle East inferred from cytochrome b DNA—evolutionary significance of this region. Mol Phylogenet Evol. 22:91-100.[CrossRef][Web of Science][Medline]
Durand JD, Ünlü E, Doadrio I, Pipoyan S, Templeton AR, 2000. Origin, radiation, dispersion and allopatric hybridization in the chub, Leuciscus cephalus. Proc R Soc Lond B. 267:1687-1697.[Medline]
Elvira B, 1987. Taxonomic revision of the genus Chondrostoma agassizi, 1835 (Pisces, Cyprinidae). Cybium. 11:11-140.
Elvira B, 1990. Iberian endemic freshwater fishes and their conservation status in Spain. J Fish Biol. 37:231-232.[CrossRef]
Elvira B, 1991. Further studies on the taxonomy of the genus Chondrostoma (Osteichthyes, Cyprinidae): species from eastern Europe. Cybium. 15:147-150.
Felsenstein J, 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution. 39:783-791.[CrossRef][Web of Science]
Fiala KL, Sokal RR, 1985. Factors determining the accuracy of cladogram estimation: evaluation using computer simulation. Evolution. 39:609-622.[CrossRef][Web of Science]
Gilles A, Chappaz R, Cavalli L, Lörtscher M, Faure E, 1998a. Genetic differentiation and introgression between putative subspecies of L. soufia (Teleostei: Cyprinidae) of the region of the Mediterranean Alps. Can J Fish Aquat Sci. 55:2341-2354.[CrossRef]
Gilles A, Lecointre G, Faure E, Chappaz R, Brun G, 1998b. Mitochondrial phylogeny of the European cyprinids: implications for their systematics, reticulate evolution and colonization time. Mol Phylogenet Evol. 10:132-143.[CrossRef][Web of Science][Medline]
Gorur N, Sakinc M, Barka A, Akkök R, 1995. Miocene to Pliocene palaeogeographic evolution of Turkey and its surroundings. J Hum Evol. 28:309-324.[CrossRef]
Hedges S, 1992. The number of replications needed for accurate estimation of the bootstrap P value in phylogenetic studies. Mol Biol Evol. 9:366-369.[Web of Science][Medline]
Hillis D, Bull J, 1993. An empirical test of bootstrapping as a method for assessing confidence in phylogenetic analysis. Syst Biol. 42:182-192.[CrossRef][Web of Science]
Hoelzer GA, 1997. Inferring phylogenies from mtDNA variation: mitochondrial-gene trees versus nuclear-gene trees. Evolution. 51:622-626.[CrossRef][Web of Science]
Hoelzer GA, Melnick DJ, 1994. Patterns of speciation and limits to phylogenetic resolution. Trends Ecol Evol. 9:104-107.[CrossRef]
Howes GJ, 1991. Systematics and biogeography: An overview. In: Cyprinid fishes—systematics, biology and exploitation, vol. 3 (Winfield IJ and Nelson JS, eds). London: Chapman and Hall; 1–33.
Hsü KJ, Montadert L, Bernoulli D, Cita MB, Erickson A, Garrison RE, Kidd RB, Melieres F, Müller C, Wright R, 1977. History of the Mediterranean salinity crisis. Nature. 267:399-403.[CrossRef][Web of Science]
Hsü KJ, Ryan WBF, Cita MB, 1973. Late Miocene desiccation of the Mediterranean. Nature. 242:240-244.[CrossRef]
Jackman TR, Larson A, De Queiroz K, Losos JB, 1999. Phylogenetic relationships and tempo of early diversification in anolis lizards. Syst Biol. 48:254-285.[CrossRef][Web of Science]
Kishino H, Hasegawa M, 1989. Evaluation of the maximum likelihood estimate of the evolutionary tree topologies from sequence data, and the branching order in Hominoidea. J Mol Evol. 29:170-179.[CrossRef][Web of Science][Medline]
Kocher TD, Stepien CA, 1997. Molecular systematics of fishes. San Diego, California. Academic Press.
Kocher TD, Thomas WK, Meyer A, Edwards SV, Paabo S, Villablanca FX, Wilson AC, 1989. Dynamics of mitochondrial DNA evolution in animals: Amplification and sequencing with conserved primers. Proc Natl Acad Sci USA. 86:6196-6200.
Kosswig C, 1955. Zoogeography of the Near East. Syst Zool. 4:49-73.
Kraus F, Miyamoto MM, 1991. Rapid cladogenesis among the pecoran ruminants: Evidence from mitochondrial DNA sequences. Syst Zool. 40:117-130.[CrossRef]
Krijgsman W, Hilgen FJ, Raffi I, Sierro FJ, Wilson DS, 1999. Chronology, causes and progression of the Messinian salinity crisis. Nature. 400:652-655.[CrossRef]
Kumar S, 1996. PHYLTEST: phylogeny hypothesis testing software. Pennsylvania State University, University Park.
Lanyon SM, 1988. The stochastic mode of molecular evolution: what consequences for systematic investigations? Auk. 105:565-573.[Web of Science]
Lecointre G, Philippe H, Lê H, Le Guyader H, 1993. Species sampling has a major impact on phylogeny inference. Mol Phylogenet Evol. 2:205-224.[CrossRef][Medline]
Machordom A, Doadrio I, 2001. Evidence of a Cenozoic Betic-Kabilian connection based on freshwater fish phylogeography (Luciobarbus, Cyprinidae). Mol Phylogenet Evol. 18:252-263.[CrossRef][Web of Science][Medline]
Marinescu F, 1992. Les bioprovinces de la paratéthys et leurs relations. Paleontologia i Evolució. 24–25:445-453.
Mathias P, 1921. Etude du genre Chondrostoma dans l'Europe occidentale et la région circumméditerranéenne. Mém Soc Zool France. 28:(1–2) 1-52.
Mesquita N, Carvalho G, Shaw P, Crespo E, Coelho MM, 2001. River basin–related genetic structuring in an endangered fish species, Chondrostoma lusitanicum, based on mtDNA sequencing and RFLP analysis. Heredity. 86:253-264.[CrossRef][Web of Science][Medline]
Meyer A, Kocher TD, Basassibwaki P, Wilson AC, 1990. Monophyletic origin of Lake Victoria cichlid fishes suggested by mitochondrial DNA sequences. Nature. 347:550-553.[CrossRef][Medline]
Nelva A, 1988. Origine et biogéographie des deux Chondrostomes français: Chondrostoma nasus et Chondrostoma toxostoma (Pisces, Cyprinidae). Cybium. 12:287-299.
Nei M, Kumar S, 2000. Molecular evolution and phylogenetics. New York: Oxford University Press.
Nei M, Kumar S, Takahashi K, 1998. The optimization principle in phylogenetic analysis tends to give incorrect topologies when the number of nucleotides or amino acids used is small. Proc Natl Acad Sci USA. 13:12390-12397.
Page RDM, 1994a. Maps between trees and cladistic analysis of historical associations among genes, organisms, and areas. Syst Biol. 43:58-77.[CrossRef]
Page RDM, 1994b. Parallel phylogenies: reconstructing the history of host-parasite assemblages. Cladistics. 10:155-173.[CrossRef][Web of Science]
Penzo EG, Gandolfi G, Bargelloni L, Colombo L, Patarnello T, 1998. Messinian salinity crisis and the origin of freshwater lifestyle in western Mediterranean gobies. Mol Biol Evol. 15:1472-1480.
Philippe H, 1993. MUST: a computer package of management utilities for sequences and trees. Nucleic Acids Res. 21:5264-5272.
Por FD, Dimentman C, 1985. Continuity of Messinian biota in the Mediterranean basin. In Geological evolution of the Mediterranean basin (Stanley DJ and Wezel FC, eds). Berlin: Springer; 545–557.
Posada D, Crandall KA, 1998. MODELTEST: testing the model of DNA substitution. Bioinformatics. 14:817-818.
Saiki RK, Gelfand DH, Stoffel S, Scharf SJ, Higuchi R, Horn GT, Mullis KB, Erlich HA, 1998. Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science. 239:487-491.
Saitou N, Nei M, 1987. The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol Biol Evol. 4:406-425.[Abstract]
Simon C, Frati F, Beckenbach A, Crespi B, Liu H, Flook P, 1994. Evolution, weighting, and phylogenetic utility of mitochondrial gene sequences and a compilation of conserved polymerase chain reaction primers. Entomol Soc of Am. 87:651-686.
Smith GR, Dowling TE, Gobalet K, Lugaski T, Shiozawa D, Evans P, 2002. Biogeography and rates of evolution of Great Basin fishes. In Great Basin Aquatic Systems History (Hershler R, Madsden DB, and Curry DR, eds). Contributions to the Earth Sciences, no. 33. Washington, DC: Smithsonian Institution 175-234.
Strimmer K, Von Haeseler A, 1996. Quartet puzzling: a quartet maximum likelihood method for reconstructing tree topologies. Mol Biol Evol. 13:964-969.[Web of Science]
Templeton AR, 1983. Phylogenetic inference from restriction endonuclease cleavage site maps with particular reference to the evolution of humans and the apes. Evolution. 37:221-244.[CrossRef][Web of Science]
Walsh HE, Kidd MG, Moum T, Friesen VL, 1999. Polytomies and the power of phylogenetic inference. Evolution. 53:932-937.[CrossRef][Web of Science]
Zardoya R, Doadrio I, 1999. Molecular evidence on the evolutionary and biogeographical patterns of European cyprinids. J Mol Evol. 49:227-237.[CrossRef][Web of Science][Medline]
Zardoya R, Economidis PS, Doadrio I, 1999. Phylogenetic relationships of Greek Cyprinidae: molecular evidence for at least two origins of the Greek Cyprinid fauna. Mol Phylogenet Evol. 13:122-131.[CrossRef][Web of Science][Medline]
Zharkikh A, Li WH, 1992. Statistical properties of bootstrap estimation of phylogenetic variability from nucleotide sequences. I. Four taxa with a molecular clock. Mol Biol Evol. 9:1119-1147.[Abstract]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. I. Robalo, C. S. Santos, V. C. Almada, and I. Doadrio Paleobiogeography of Two Iberian Endemic Cyprinid Fishes (Chondrostoma arcasii-Chondrostoma macrolepidotus) Inferred from Mitochondrial DNA Sequence Data J. Hered., March 1, 2006; 97(2): 143 - 149. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||