Hydrobiologia (2005) 542:221–233 H. Segers & K. Martens (eds), Aquatic Biodiversity II DOI 10.1007/s10750-004-1389-x
Springer 2005 Phylogeography and speciation in the Pseudocrenilabrus philander species complex in Zambian Rivers Cyprian Katongo1, Stephan Koblmüller2, Nina Duftner2, Lawrence Makasa3 & Christian Sturmbauer2,* 1 Department of Biological Sciences, University of Zambia at Lusaka, Zambia Department of Zoology, University of Graz, Universitätsplatz 2, 8010 Graz, Austria 3 Department of Fisheries at Mpulungu, Ministry of Agriculture and Cooperatives, Zambia (*Author for correspondence: Tel.: +43-316-380-5595, Fax: +43-316-380-9875, E-mail: christian.sturmbauer@uni-graz.at) 2 Key words: mtDNA sequences, geographic speciation, phylogeography Abstract Haplochromine cichlids form the most species-rich lineage of cichlid fishes that both colonized almost all river systems in Africa and radiated to species flocks in several East African lakes. The enormous diversity of lakes is contrasted by a relatively poor albeit biogeographically clearly structured species diversity in rivers. The present study analyzed the genetic structure and phylogeographic history of species and populations of the genus Pseudocrenilabrus in Zambian rivers that span two major African drainage systems, the Congo- and the Zambezi-system. The mtDNA phylogeny identifies four major lineages, three of which occur in the Congo-system and one in the Zambezi system. Two of the Congo-clades (Lake Mweru and Lunzua River) comprise distinct albeit yet undescribed species, while the fish of the third Congo-drainage clade (Chambeshi River and Bangweulu swamps), together with the fish of the Zambezi clade (Zambezi and Kafue River) are assigned to Pseudocrenilabrus philander. Concerning the intraspecific genetic diversity observed in the sampled rivers, most populations are highly uniform in comparison to lacustrine haplochromines, suggesting severe founder effects and/or bottlenecking during their history. Two bursts of diversification are reflected in the structure of the linearized tree. The first locates at about 3.9% mean sequence divergence and points to an almost simultaneous colonization of the sampled river systems. Subsequent regional diversification (with about 1% mean sequence divergence) occurred contemporaneously within the Kafue River and the Zambezi River. The clear-cut genetic biogeographic structure points to the dominance of geographic speciation in this lineage of riverine cichlid fishes, contrasting the importance of in situ diversification observed in lake cichlids. Introduction With an estimated number of 2400 species (Snoeks, 2001) cichlid fishes represent the most species-rich family of vertebrates. More than 1500 species live in the Great East African Lakes alone (Turner et al., 2001). The cichlid species flocks of these lakes are the most spectacular examples of adaptive radiation and explosive speciation within a single group of organisms (Kosswig, 1947; Fryer & Iles, 1972; Greenwood, 1984). Numerous species of haplochromine cichlids occur in North–Central–East and Southern African rivers, of which many are still undescribed. The relationships of riverine cichlids are poorly known to date, but their diversity seems to follow a biogeographic pattern, in that distinct genera are restricted to particular biogeographic regions or river systems 222 (Bell-Cross, 1966, 1975; Roberts, 1975; Greenwood, 1984; Meyer et al., 1991; Stiassny, 1991; Skelton, 1994; De Vos & Seegers, 1998; Salzburger et al., 2002; Seehausen, 2002; Verheyen et al., 2003). Only a few genera (Tilapia, Oreochromis, Thoracochromis and Pseudocrenilabrus) are known to have distributions spanning many river systems and ichthyogeographic regions. The genus Pseudocrenilabrus occurs from the northernmost edge of Egypt down to South Africa. Morphologically, the genus Pseudocrenilabrus is characterized by its small body size, an upper pharyngeal apophysis of the Haplochromis-type, a moderately protractile upper jaw, bicuspid and/or conical teeth in the outer row and tricuspid teeth in the inner rows of both jaws, ctenoid scales, an interrupted single-pored lateral line and a rounded caudal fin as autapomorphy. The males lack egg spots in the anal fin, displaying a distinct red or orange spot on the distal tip of the anal fin. They are maternal mouthbrooders with up to 120 fry per clutch. The genus comprises three species currently considered as valid, even if the actual taxonomic situation is considerably more complex (Greenwood, 1989). One species, Pseudocrenilabrus multicolor, consists of two highly distinct subspecies, one colonizing the Nile system and the second the Lake Victoria region. Pseudocrenilabrus nicholsi is only found in the Republic of Congo, and the third nominal species, Pseudocrenilabrus philander comprises three sub-species plus a complex of distinct geographically separated populations pointing to a considerable sub-structuring of this species in southern Africa: P. philander luebberti in the area of Otavi in SW-Africa, P. philander dispersus from SW-Africa to Mozambique, and P. philander philander in SE and southern Africa (Skelton, 1991). In Zambia, the target region of this study, only P. philander philander has been reported to date. In terms of phylogenetic placement of the genus among other haplochromine lineages, a single molecular phylogenetic study exists that included P. multicolor (Salzburger et al., 2002). According to the study of Salzburger et al. (2002), the diversification of haplochromine cichlids was connected to the ‘‘primary lacustrine radiation’’ of Lake Tanganyika cichlids, comprising both riverine haplochromine lineages and endemic Tanganyikan lineages, summarized as the ‘H-lineage’ (Nishida 1991; 1997, re- defined by Salzburger et al., 2002). Interestingly, haplochromine cichlids are likely to be paraphyletic with respect to several endemic Tanganyikan lineages. Pseudocrenilabrus was placed as a member of the H-lineage in a clade containing both riverine taxa and the members of the endemic Tanganyikan tribe Tropheini. De Villiers et al. (1992) showed by means of analysis of restriction enzyme variation that allopatric populations of P. philander exhibit only a low degree of genetic variation despite their sometimes clear morphological distinctness. The present study aims to place the genus Pseudocrenilabrus in a representative taxonomic framework of African haplochromines and focuses on the phylogeographic structuring of P. philander philander in Zambian river systems. Therefore, we used a molecular approach to study the pathway of diversification of this lineage of cichlid fishes in the context of Southern Africa’s geological and hydrographic history. We sequenced three mitochondrial gene segments of nine populations in six Zambian rivers that span two major African drainage systems, to derive a mtDNA phylogeny. Material and methods In total we analyzed 43 individuals of the genus Pseudocrenilabrus plus 23 species from other African cichlid lineages. DNA sequences were obtained for three mitochondrial gene segments [1047 bp of the NADH2-gene (26 individuals), 402 bp of cytochrome b (26 individuals); and 358 bp of the control region (43 individuals)]. Detailed information about the sampled individuals is given in Table 1 and Figure 1. Voucher specimens are available from the authors. DNAextraction, PCR-amplification, and automatic DNA-sequencing were performed according to standard methods (see Salzburger et al., 2002; Koblmüller et al., 2004) using published primers for all gene segments (Kocher et al., 1989, 1995; Meyer et al., 1994). DNA sequences were aligned using Clustal X (Thompson et al., 1997) and alignments were further improved by eye in the case of the control region. Phylogenetic analyses were performed in two steps. The first step 223 Table 1. Characterization of the studied species, with sampling locality, drainage system, and GenBank accession numbers Species Locality Drainagea GeneBank accession number NADH 2 cytochrome b control region Eretmodus cyanostictus Tanganicodus irsacae Lake Tanganyika Lake Tanganyika LT LT AF398220b AF398219b AF428155b Z21778b – – Spathodus erythrodon Lake Tanganyika LT AF398218b AF428156b – b Altolamprologus compressiceps Lake Tanganyika LT AF398229 AF428163b – Neolamprologus brichardi Lake Tanganyika LT AF398227b Z29997b – Lamprologus mocquardi Congo River CO AF398225b Z29995b – Cyphotilapia frontosa Lake Tanganyika LT U07247b AF428169b – Triglachromis otostigma Lake Tanganyika LT AF398217b Z30004b – Limnochromis auritus Xenotilapia sima Lake Tanganyika Lake Tanganyika LT LT AF398216b U07270b Z21775b Z21772b – – Ophthalmotilapia ventralis Lake Tanganyika LT U07257b Z21771b – Callochromis macrops Lake Tanganyika LT U07242b Z21760b – Perissodus microlepis Lake Tanganyika LT AF398222b AF428168b – Plecodus straeleni Lake Tanganyika LT AF398221b Z21777b – Cyprichromis leptosoma Lake Tanganyika LT AF398224b AF428154b – Paracyprichromis brieni Lake Tanganyika LT AF398223b Z21776b – Astatotilapia burtoni Gnathochromis pfefferi Lake Tanganyika Lake Tanganyika LT LT AF317266b U07248b Z21773b AF428166b – – Tropheus moorii Lake Tanganyika LT U07267b Z12037b – b Lobochilotes labiatus Lake Tanganyika LT U07254 AF428170b – Astatoreochromis alluaudi ? LV AF398234b AF428157b – Orthochromis malagaraziensis Malagarazi River LT AF398232b AF428161b – Orthochromis polyacanthusc ? CO/LU AF398231b AF428159b – Pseudocrenilabrus multicolor ? (ornamental fish trade) NI AY602992 AY600141 AY602995 Pseudocrenilabrus nicholsi Pseudocrenilabrus philander ? (ornamental fish trade) Mukula Stream–Bangweulu CO LU AY602994 – AY600143 – AY602996 AY612141 AY612143 AY612145 AY612160 AY612161 Pseudocrenilabrus philander Mukula Stream–Bangweulu LU – – AY612142 AY612144 Pseudocrenilabrus philander Chambeshi–Chambeshi River LU – – AY612137 AY612138 AY612139 AY612155 AY612156 Pseudocrenilabrus Lunzua River LT – – AY612140 sp. nov. ‘yellow’ AY612146 AY612147 AY612148 AY612149 224 Table 1. (Continued) Species Pseudocrenilabrus sp. nov. ‘blue’ Locality Lunzua River Drainagea LT GeneBank accession number NADH 2 cytochrome b control region – – AY612150 AY612151 AY612152 AY612153 AY612154 AY615797 Pseudocrenilabrus Mwatishi River–Lake Mweru LU – – AY612133 sp. nov. ‘orange’ AY612134 AY612135 AY612136 Pseudocrenilabrus philander Lake Ithezi-Thezi KF – – Pseudocrenilabrus philander Lake Ithezi-Thezi KF – – AY612158 AY612159 Pseudocrenilabrus philander Lukanga Swamps KF – – AY612124 Pseudocrenilabrus philander Lukanga Swamps KF – – AY612125 Pseudocrenilabrus philander Pseudocrenilabrus philander Lukanga Swamps Lake Kariba KF MZR – – – – AY612126 AY612127 Pseudocrenilabrus philander Lake Kariba MZR – – AY612128 Pseudocrenilabrus philander Lake Kariba MZR – – AY612129 Pseudocrenilabrus philander Lake Kariba MZR – – AY612130 Pseudocrenilabrus philander near Marromeu LZR – – AY612157 Pseudocrenilabrus philander Kabala UZR AY602993 AY600142 AY612131 Pseudocrenilabrus philander Kabala UZR – – AY612132 Pseudocrenilabrus philander Kabala UZR – – AY612162 AY612163 Note: Species names were assigned according to fishbase (http://www.fishbase.org). For Pseudocrenilabrus only distinct haplotypes per locality are listed with separate accession numbers for each individual. a Drainage systems to which sampling sites belong to: CO, Congo basin; LT, Lake Tanganyika; LU, Luapula River; LV, Lake Victoria; KF, Kafue River; LZR, lower Zambezi River; MZR, middle Zambezi River; UZR, upper Zambezi River; NI, Nile basin. b GenBank accession numbers of sequences published elsewhere (Sturmbauer & Meyer, 1992; Sturmbauer & Meyer, 1993; Sturmbauer et al., 1994; Kocher et al., 1995; Salzburger et al., 2002). c There is a taxonomic problem with Orthochromis polyacanthus. This is reviewed by Greenwood & Kullander (1994), who state that the distribution of this species is limited to Lake Mweru. Due to the unknown origin of the sample and the earlier supposed synonymy of O. polyacanthus with O. stormsi, the sample used might be a different species of the genus Orthochromis. addressed the phylogenetic placement of the genus Pseudocrenilabrus among African cichlids and was based on a combined data set of cytochrome b and NADH 2, defining three representatives of the Tanganyikan cichlid tribe Eretmodini as the outgroup, justified by Salzburger et al. (2002). The second step analyzed the phylogenetic structure among nine Zambian populations of Pseudocrenilabrus using P. multicolor and P. nicholsi as outgroups. Phylogenetic trees were constructed using three alternative algorithms, maximum parsimony (MP), neighbor-joining (NJ) and maximum likelihood (ML) using the computerprogram PAUP* 4.0b2a (Swofford, 2000). In MP transversion mutations were weighted over transition mutations, according to the observed frequency in regions of the same degree of variation, as derived from a sliding-window 225 Figure 1. Map of the Zambian River systems, showing the sampling sites. analysis (Sturmbauer & Meyer, 1992). Specifically, weightings were set to 2:1 in regions of more than 10% of genetic variation, and 3:2 in regions of less than 10% genetic variation. For MP of the combined data set of cytochrome b and NADH 2 we applied the same weighting scheme as Salzburger et al. (2002). The appropriate model of molecular evolution for the maximum likelihood analysis was evaluated by the likelihood ratio test implemented in the computer program Modeltest 3.0 (Posada & Crandall, 1998) to be used for NJ and ML. This test justified the use of the TrN model of molecular evolution (Tamura & Nei, 1993) for the combined data set of cytochrome b and NADH 2 and the HKY model (Hasegawa et al., 1985) for the control region. For the combined data set we applied the estimated proportion of invariable sites (0.45), the gamma shape parameter (0.95), and the base frequencies (A: 0.27; C: 0.34; G: 0.11; T: 0.28). For the control region we applied the base frequencies (A: 0.36; C: 0.19; G: 0.14; T: 0.32) and the transition/ transversion ratio (2.24). We assessed the robustness of the resulting topologies by applying standard measures of confidence (bootstrap and quartet-puzzling frequencies) with 1000 pseudoreplicates (bootstrap) and 25 000 random quartets (quartet-puzzling), and in the case of maximum likelihood under the fast stepwise addition option in PAUP* 4.0b2a (Swofford, 2000). To visualize the relatedness and the degree of genetic diversity within and among populations of Pseudocrenilabrus we constructed a minimum 226 spanning tree, based upon the NJ- and an unweighted parsimony trees of step two of our analysis. To determine the sequence of the major cladogenic events in the Zambian lineages of the genus Pseudocrenilabrus, we constructed a linearized tree based on a 358 bp segment of the D-Loop performing the two-cluster test implemented in the computer program LINTRE (Takezaki et al., 1995). First, rate constancy was tested for all internal nodes in the topology based on the sequences of the control region using the branch length test of LINTRE. In our case, the test was performed applying the HKY model of molecular evolution. No rate heterogeneity was detected at a high significance level (pa < 0.01) so that none of the taxa had to be excluded from further analyses. Then, a tree for the given topology was constructed under the assumption of rate constancy, which is termed a linearized tree. Since no accurate geology-based dating for the rate of base substitution of the control region is available to date for African cichlids we refrain from absolute age estimates and compare average observed levels of genetic divergences to those observed in other lineages. Our pairwise calculations of mean sequence divergence were based upon the Kimura2-parameter model (K2P; Kimura, 1980), since this model was also used by Sturmbauer et al. (2001) and Baric et al. (2003). Results Phylogenetic placement of the genus Pseudocrenilabrus Maximum parsimony resulted in two most parsimonious trees [2637 steps; CI excluding uninformative sites, 0.46; retention index (RI), 0.60; and rescaled consistency index (RC), 0.36; trees not shown]. The ML-tree is depicted in Figure 2. The following relationships were consistently found by all three algorithms: the non-mouthbrooding Lamprologine cichlids occupied the most ancestral split followed by a series of short-branched splits of seven lineages, corresponding to the radiation of the H-lineage (Limnochromini, Ectodini, Cyphotilapiini, Cyprichromini, Perissodini, Haplochromini I and Haplochromini II/Tropheini; see also Salzburger et al., 2002). The branching order among these seven lineages varied with respect to the tree building algorithm. Pseudocrenilabrus was unambiguously placed as sister group to the Congo River species Orthochromis polyacanthus and as ancestral branch within the Haplochromini II/Tropheini-lineage. Phylogeographic structure among Zambian Pseudocrenilabrus populations Phylogenetic analysis identified four major lineages, three of which occur in the Congo-system and one in the Zambezi system. Maximum parsimony resulted in two most parsimonious trees [268 steps; CI excluding uninformative sites, 0.75; retention index (RI), 0.96; and rescaled consistency index (RC), 0.77; trees not shown]. The strict consensus tree of the two MP trees, the NJ- and the ML-tree are depicted in Figure 3. MP and ML resulted in almost identical topologies and placed the populations of Chambeshi River and Mukula Stream in the Bangweulu swamps (both Luapula System) as most ancestral split, followed by the Lunzua River population (Lake Tanganyika – Congo River drainage), followed by the population of the Mwatishi River estuary in Lake Mweru (Luapula System), followed by a more complex clade of populations of the Zambezi River drainage. In this clade the samples from Kafue River were placed ancestral to the population samples from three localities of the Zambezi River. In NJ, the clades of the Lunzua population and the Lake Mweru population changed their relative position. The minimum spanning tree (Fig. 4), derived from one of the two MP trees clearly reflects the four distinct lineages: the Chambeshi River and Lake Bangweulu individuals, the Lunzua River individuals, the Lake Mweru specimens, and finally the Kafue- and Zambezi-River samples. The Chambeshi-Bangweulu clade is separated from the Lunzua River population by 11–12 mutations, from the Lake Mweru population by 19–20 mutations, and from the Kafue- and ZambeziRiver samples by 14–24 mutations. The Lunzua River population is separated from the Lake Mweru population by 12 mutations and from the Kafue- and Zambezi-River samples by 8–16 mutations. Finally, the Lake Mweru population is separated from the Kafue- and Zambezi-River samples by 10–18 mutations. Concerning the 227 Figure 2. Phylogenetic placement of the genus Pseudocrenilabrus. Maximum likelihood tree comprising 26 species of East African cichlids, representing 9 distinct tribes [ERE, Eretmodini; LAM, Lamprologini; CYT, Cyphotilapiini; LIM, Limnochromini; CYP, Cyprichromini; HAP, Haplochromini; Per, Perissodini; ECT, Ectodini; TRO, Tropheini; assignment to tribes follows Salzburger et al. (2002) and Takahashi (2003)], based upon 1047 bp of the NADH 2 gene and 402 bp of the cytochrome b gene, using the substitution model TrN + I + C (Tamura & Nei, 1993). Eretmodus cyanostictus, Tanganicodus irsacae and Spathodus erythrodon were used as outgroup taxa. Quartet puzzling values (Strimmer & Von Haeseler, 1996) larger than 50 are shown above the branches. intra-lineage diversity, three of the four lineages are highly uniform and contain identical haplotypes only, except for the Bangweulu-population in which two individuals differ from the others by one mutation, albeit five of the seven sampled Bangweulu-individuals share the same haplotype with the specimens from the Chambeshi River that were sampled about 200 km apart. Only the lineage from the Kafue- and Zambezi-River shows genetic and geographic structure in that the indi- viduals from Kafue River are sister to those from the Zambezi River populations. Within these populations, a greater genetic diversity is observed, but no subdivision between the populations above and below the Victoria Falls is detected. The linearized tree (Fig. 5) points to two diversification events, the first (3.9% K2P-distance ± 1.2%) concerning an almost contemporary colonization of all sampled river systems, and the second (1.0% K2P-distance ± 0.6%) a further 228 Figure 3. Strict consensus tree of two MP trees [268 steps; CI excluding uninformative sites, 0.75; retention index (RI), 0.96; rescaled consistency index (RC), 0.77], the NJ and the ML tree [substitution model HKY (Hasegawa et al., 1985] of 41 taxa of the Pseudocrenilabrus philander species complex from nine distinct localities, plus the two outgroup taxa Pseudocrenilabrus multicolor and P. nicholsi, based upon 358 bp of the most variable part of the control region. Bootstrap values obtained from neighbor joining are shown above the branches, while numbers in the middle represent parsimony bootstrap values. Quartet puzzling values are depicted below the branches. Only bootstrap and quartet puzzling values larger than 40 are shown. Abbreviations in parentheses refer to major drainage systems (see Table 1). Roman numerals refer to the four distinct clades within the Zambian Pseudocrenilabrus: I, Chambeshi– Bangweulu clade; II, Lake Mweru clade; III, Lunzua clade; IV, Kafue–Zambezi clade. Pictures of the corresponding fishes are shown beside the clades. For the Lunzua clade both distinct color morphs, blue and yellow, are depicted (M, male; F, female). regional diversification within the Kafue–Zambezi clade. In fact, the conflicting branching order among the Lunzua- and Chambeshi populations among MP-ML versus NJ can be interpreted by the linearized tree as a contemporaneous diversification. Discussion Our analysis indicates a sister group relationship of the genus Pseudocrenilabrus to Orthochromis polyacanthus from the Congo River. Such a close relationship is corroborated morphologically by the lack of egg-spots on the anal fin in both genera. Concerning the intrageneric relationships of P. philander, P. nicholsi and P. multicolor, P. philander forms the most ancestral split (6.4% K2P-distance ± 1.3%), placing P. nicholsi and P. multicolor as sister taxa (4.7% K2P-distance), pointing to an origin of the genus in the Congo River, followed by a south- and northward range expansion. Also, the closest non-mouthbrooding sister group of the H-lineage, the lamprologine cichlids, are restricted to Lake Tanganyika, the Malagarazi and the Congo River. Salzburger et al. (2002) suggested a connection of the radiation of riverine haplochromines to the primary lacustrine radiation of the H-lineage, thus constraining the origin of haplochromines to the early CongoTanganyika basin (see also Lévêque, 1997). The level of genetic divergence between Pseudocrenila- 229 Figure 4. Minimum spanning tree of 14 haplotypes of the Pseudocrenilabrus philander species complex based on 358 bp of the most variable part of the mitochondrial control region (outgroup taxa: Pseudocrenilabrus multicolor, P. nicholsi). The topology corresponds to one of the most parsimonious trees that was most similar to the NJ tree. Each crossbar indicates one base substitution. •, branching points for which no intermediate haplotype was found. Abbreviations refer to major drainage systems (see Table 1). Numbers in parantheses refer to the number of identical haplotypes. brus philander and the other two species of Pseudocrenilabrus (6.4% K2P-distance ± 1.3%) is highly similar to that observed in the endemic Tanganyikan genus Tropheus (6.1% K2P-distance ± 1.4%, Baric et al., 2003), suggesting a similar evolutionary age for diversification and spread of these riverine haplochromines and this lineage of endemic lake cichlids, given that the rates of molecular evolution are roughly the same. The colonization of North-, Central-, East- and Southern Africa by transgression of present day water sheds must be interpreted in the light of the complex interactions of a series of geological and palaeoclimatological changes (Roberts, 1975; Skelton, 1994; reviewed in Lévêque, 1997). Phylogeographic structure among Zambian Pseudocrenilabrus populations The evolutionary relationships of the surveyed populations show a clear phylogeographic pattern in that all water bodies contain a genetically and sometimes also morphologically distinct entity. This is in contrast to the study of de Villiers et al. (1992) who found identical or almost identical restriction enzyme patterns in distant populations in Zimbabwe, Namibia and South Africa. The individuals sampled in Lunzua River are strikingly different from all remaining populations. Lunzua River drains into Lake Tanganyika, which in turn is connected to the Congo River via the Lukuga, but its upper reaches are separated from the lake by high water falls. Below the falls, only Astatotilapia burtoni was caught which was assigned to the ‘modern haplochromines’ (Salzburger et al., 2002). Interestingly, the Pseudocrenilabrus population in the upper Lunzua River comprises two distinct color morphs, one being bluish on the head and the body sides and the second one being yellowish (see photographs in Fig. 3). This distinction is also clearly visible in females: blue females do not have a yellowish-orange anal fin 230 Figure 5. Linearized tree based on a 358 bp segment of the most variable part of the mitochondrial control region. The linearized tree was compiled with the computer program LINTRE (Takezaki et al., 1995) after performing a branch length test (Takezaki et al., 1995) to test for differences in base substitution rates, using the substitution model HKY (Hasegawa et al., 1985). Gray bars refer to major diversification events. The observed mean sequence divergences using the substitution model K2P (Kimura, 1980) are depicted for the corresponding diversification events. Abbreviations in parentheses refer to major drainage systems (see Table 1). Roman numerals refer to the four distinct clades within the Zambian Pseudocrenilabrus: I, Chambeshi–Bangweulu clade; II, Lake Mweru clade; III, Lunzua clade; IV, Kafue–Zambezi clade. while yellow females do. However, in terms of mtDNA, they all share a single haplotype in the control region (n ¼ 11), so that nuclear data need to be carried out to further analyze their distinctness. Moreover, it became evident that the Lake Mweru population is clearly distinct from the Chambeshi–Bangweulu population, not only in terms of genetics, but also in terms of morphology, and is currently being described as a new species (Jos Snoeks, personal communication), which will make P. philander paraphyletic. This distinctness of the Lake Mweru population reflects the separation of the upper Luapula River system in two faunal sub-regions by the Johnston- and further upstream the Mumbatuta Falls, which are likely to represent important dispersal barriers (Jackson, 1961). In contrast to the specimens in Lake Mweru, the individuals from Chambeshi and Bangweulu are morphologically highly similar to Pseudocrenilabrus philander philander from the Zambezi drainage, even if they display similar genetic distances (about 4% K2P-distance) to the fish from the Zambezi system as the two other Congo-drainage clades, as is also evident from the linearized tree analysis (Fig. 5). Within the clade of the Zambezi drainage, a clear geographic substructure became evident: The samples from Kafue River occupy the most ancestral branches, while all individuals from the Zambezi River form a clade. Interestingly, no genetic substructure seems to exist among the Zambezi River individuals that were sampled at three localities above and below the Victoria Falls, representing the three distinct sections of the Zambezi, each having distinctive geomorphic characters (Wellington, 1955). Given that the Victoria Falls are a geologically ancient barrier, this finding is highly surprising. Since the upper reaches of the Kafue River system are in close geographical vicinity to the upper reaches of the Luapula River system (Fig. 1), we hypothesize 231 that the Kafue River was first colonized by Pseudocrenilabrus via an ancient connection to the Luapula system, according to the model of drainage evolution for southern Africa proposed by Skelton (1994) which suggests a series of opportunities for faunal exchange between the southwest, southeast and central African river systems during the tertiary. In general, the fish fauna of the upper Zambezi River and the Kafue River shows clear links to the Congo (Jackson, 1961). The occurrence of common fish species in the Kafue and Congo drainage requires an earlier link between these river systems. Moore & Larkin (2001) interpreted the abrupt change in course of the Chambeshi River, which at present forms a headwater of the Luapula system, as a capture elbow, implying that this river originally formed the headwaters of the Kafue River. Given the impossibility for Pseudocrenilabrus to get up the Victoria Falls, the upper reaches of the Zambezi River must have been colonized first, by an ancient connection to the Kafue River which might have been closed by a river capture event in the late Pleistocene (Thomas & Shaw, 1991; Moore & Larkin, 2001). Since the upper reaches of the Zambezi River and its tributaries are in close geographical vicinity to the Kafue River (Fig. 1), this scenario seems to be likely, also supported by the fact that the fish fauna of the Kafue River shows considerable similarity to the upper Zambezi River despite its present connection to the middle Zambezi River (Jackson, 1961; Skelton, 1994). This implies that the middle and lower Zambezi may have been colonized by Pseudocrenilabrus from two directions, via the Kafue and by individuals ‘‘jumping’’ down the Victoria Falls. Diversification and speciation in Pseudocrenilabrus The genetic diversity observed within the sampled populations is strikingly uniform in all except for the Zambezi drainage populations. These low levels of genetic diversity contrast all data of lacustrine cichlid species in the three Great East African Lakes (Moran & Kornfield, 1993; Moran et al., 1994; Verheyen et al., 1996; Sturmbauer et al., 1997; Nagl et al., 1998; Albertson et al., 1999; Rüber et al., 1999; Baric et al., 2003; Verheyen et al., 2003), suggesting severe founder effects during colonization. In addition, Pseudocrenilabrus seems to avoid water current of the main river beds and predominately occurs in quiet side arms and swamps. Many individuals are found in temporarily flooded areas during rainy season, so that seasonal bottlenecking might also contribute to the striking genetic uniformity of most sampled populations. The relatively greater degree of genetic variation in the Zambezi system (1.0% K2P-distance ± 0.6%) may be due to a larger and more stable effective population size in Kafue, and due to a bi-directional colonization of the middle and lower Zambezi, via the Kafue River and via the upper Zambezi. These hypotheses can be tested further when data from additional cichlid species of the Kafue–Zambezi drainage are available. The observed species diversity of the genus Pseudocrenilabrus is low in comparison to other riverine cichlid genera in Zambia (Skelton 1991). Each of the rivers seems to be inhabited by one species only. In the upper Lunzua River we only found Tilapia cf. sparrmanii in addition to Pseudocrenilabrus. In all other rivers, several other species of the genera Serranochromis, Sargochromis, Pharyngochromis, Tilapia and Oreochromis were found. Thus, Pseudocrenilabrus is likely to be well adapted to one particular niche but seems not competitive enough against other cichlids to undergo speciation by niche segregation. Speciation in Pseudocrenilabrus is rather likely to be driven by geographic separation. Acknowledgements We wish to thank C. Kapasa, H. Phiri and the team at the Department of Fisheries, Ministry of Agriculture and Cooperatives, and L. Mumba of the University of Zambia for their support during fieldwork. We are further indebted to O. Seehausen for providing tissue samples and K. Sefc for valuable comments on the manuscript. S. K., N. D. and C.S. were supported by the Austrian Science Foundation (grant P15239), C. K. by the OEAD, Austrian Ministry of Foreign Affairs and also by a fellowship from the Royal Museum for Central Africa, Belgium. S.K. and N.D. were also supported by the University of Graz. 232 References Albertson, R. C., J. A. Markert, P. D. Danley & T. D. Kocher, 1999. Phylogeny of a rapidly evolving clade: the cichlid fishes of lake Malawi, East Africa. Proceedings of the Natural Academy of Science, USA 96: 5107–5110. Baric, S., W. Salzburger & C. Sturmbauer, 2003. Phylogeography and evolution of the Tanganyikan cichlid genus Tropheus based upon mitochondrial DNA sequences. Journal of Molecular Evolution 56: 54–68. Bell-Cross, G., 1966. The distribution of fishes in Central Africa. Fish Research Bulletin of Zambia 4: 3–15. Bell-Cross, G., 1975. A revision of certain Haplochromis species of Central Africa. Occasional Papers of the National Museums and Monuments of Rhodesia 7: 405–464. De Villiers, D. L., E. H. Harley & A. J. Ribbink, 1992. Mitochondrial DNA restriction enzyme variation in allopatric populations of Pseudocrenilabrus philander. South African Journal of Science 88: 96–99. De Vos, L. & L. Seegers, 1998. Seven new Orthochromis species (Teleostei: Cichlidae) from the Malagarasi, Luiche and Rugufu basins (Lake Tanganyika drainage), with notes on their reproductive biology. Ichthyological Explorations of Freshwaters 9(4): 371–420. Fryer, G. & T. D. Iles, 1972. The Cichlid Fishes of the Great Lakes of Afrika: Their Biology and Evolution, TFH Publications, Neptune City. 641 pp. Greenwood, P. H., 1984. African cichlid fishes and evolutionary theories. In Echelle, A. A. & I. Kornfield (eds), Evolution of Fish Species Flocks. University of Maine at Orono Press, Orono: 141–154. Greenwood, P. H., 1989. The taxonomic status and phylogenetic relationships of Pseudocrenilabrus Fowler (Teleostei, Cichlidae). Ichthyological Bulletins of the J.L.B. Smith Institute of Ichthyology, 16 pp. Greenwood, P. H. & S. O. Kullander, 1994. A taxonomic review and redescription of Tilapia polyacanthus and T. stormsi (Teleostei: Cichlidae), with descriptions of two new Schwetzochromis species from the Upper Zaire River drainage. Ichthyological Explorations of Freshwaters 5(2): 161–180. Hasegawa, M., H. Kishino & T. Jano, 1985. Dating of the human-ape splitting by a molecular clock of mitochondrial DNA. Journal of Molecular Evolution 22: 160–174. Jackson, P. B. N., 1961. The Fishes of Northern Rhodesia. A Check List of Indigenous Species. Government Printer, Lusaka, Zambia. Kimura, M., 1980. A simple method for estimating evolutionary rate of base substitutions through comparative studies of nucleotide sequences. Journal of Molecular Evolution 16: 111–120. Koblmüller, S., W. Salzburger & C. Sturmbauer, 2004. Evolutionary relationships in the sand-dwelling cichlid lineage of Lake Tanganyika suggest multiple colonization of rocky habitats and convergent origin of biparental mouthbrooding. Journal of Molecular Evolution 58: 79–96. Kocher, T. D., W. K. Thomas, A. Meyer, S. V. Edwards, S. Pääbo, F. X. Villablanca & A. C. Wilson, 1989. Dynamics of mitochondrial DNA evolution in animals: amplification and sequencing with conserved primers. Proceedings of the National Academy of Sciences, USA 86: 6196–6200. Kocher, T. D., J. A. Conroy, K. R. McKaye, J. R. Stauffer & S. F. Lockwood, 1995. Evolution of NADH dehydrogenase subunit 2 in East African cichlid fishes. Molecular Phylogenetics and Evolution 4: 420–432. Kosswig, C., 1947. Selective mating as a factor for speciation in cichlid fish of East African lakes. Nature 159: 604. Lévêque, C., 1997. Biodiversity Dynamics and Conservation: The Freshwater Fish of Tropical Africa. Cambridge University Press, Cambridge, New York. Meyer, A., T. D. Kocher & A. C. Wilson, 1991. African fishes. Nature 350: 467–468. Meyer, A., J. M. Morrissey & M. Schartl, 1994. Recurrent origin of sexually selected trait in Xiphophorus fishes inferred from a molecular phylogeny. Nature 368: 539–541. Moore, A. E. & P. A. Larkin, 2001. Drainage evolution in south-central Africa since the breakup of Gondwana. South African Journal of Geology 104: 47–68. Moran, P. & I. Kornfield, 1993. Retention of an ancestral polymorphism in the mbuna species flock (Teleostei: Cichlidae) of Lake Malawi. Molecular Biology and Evolution 10(5): 1015–1029. Moran, P., I. Kornfield & P. N. Reinthal, 1994. Molecular systematics and radiation of haplochromine cichlids (Teleostei: Perciformes) of Lake Malawi. Copeia 1994(2): 274–288. Nagl, S., H. Tichy, W. E. Mayer, N. Takahata & J. Klein, 1998. Persitence of neutral polymorphisms in Lake Victoria cichlid fish. Proceedings of the Natural Academy of Science, USA 95: 14238–14243. Nishida, M., 1991. Lake Tanganyika as an evolutionary reservoir of old lineages of East African fishes: inferences from allozyme data. Experientia 47: 974–979. Nishida, M., 1997. Phylogenetic relationships and the evolution of Tanganyika cichlids: a molecular perspective. In Kawanabe, H., M. Hori & M. Nagoshi (eds), Fish communities in Lake Tanganyika. Kyoto University Press, Kyoto, Japan: 3–23. Posada, D. & K. A. Crandall, 1998. Modeltest: testing the model of DNA substitution. Bioinformatics 14: 817–818. Roberts, T. R., 1975. Geographical distribution of African freshwater fishes. Zoological Journal of the Linnean Society 57: 249–319. Rüber, L., E. Verheyen & A. Meyer, 1999. Replicated evolution of trophic specializations in an endemic cichlid lineage from Lake Tanganyika. Proceedings of the Natural Academy of Science, USA 96: 10230–10235. Salzburger, W., A. Meyer, S. Baric, E. Verheyen & C. Sturmbauer, 2002. Phylogeny of the Lake Tanganyika cichlid species flock and its relationships to the Central and East African haplochromine cichlis fish faunas. Systematic Biology 51: 113–135. Seehausen, O., 2002. Patterns in fish radiation are compatible with Pleistocene desiccation of Lake Victoria and 14 600 year history for its cichlid species flock. Proceedings of the Royal Society, London B 269: 491–497. Skelton, P. H., 1991. Pseudocrenilabrus. In Daget, J., J. P. Gosse, G. G. Teugels & D. F. E. T. van den Audenaerde (eds), Check-list of freshwater fishes of Africa (CLOFFA). 233 ISNB, Brussels; MRAC, Tervuren; and ORSTOM, Paris 4: 394–398. Skelton, P. H., 1994. Diversity and distribution of freshwater fishes in East and Southern Africa. Annales du Musée de la République de l’Afrique Centrale, Zoologique 275: 95– 131. Snoeks, J., 2001. Cichlid diversity, speciation and systematics: examples from the Great African lakes. Journal of Aquariculture & Aquatic Sciences 9: 150–166. Stiassny, M. L. J., 1991. Phylogenetic interrelationships of the family Cichlidae: an overview. In Keenleyside, M. H. A. (ed.), Cichlid Fishes. Behaviour, Ecology and Evolution. Chapman & Hall, London: 1–35. Strimmer, K. & A. Von Haeseler, 1996. Quartet puzzling: a quartet maximum-likelihood method for reconstructing tree topologies. Molecular Biology and Evolution 7: 964– 969. Sturmbauer, C. & A. Meyer, 1992. Genetic divergence, speciation, and morphological stasis in a lineage of African cichlid fishes. Nature 358: 578–581. Sturmbauer, C. & A. Meyer, 1993. Mitochondrial phylogeny of the endemic mouthbrooding lineages of cichlid fishes of Lake Tanganyika, East Africa. Molecular Biology and Evolution 10: 751–768. Sturmbauer, C., E. Verheyen & A. Meyer, 1994. Mitochondrial phylogeny of the Lamprologini, the major substrate spawning lineage of cichlid fishes from Lake Tanganyika in eastern Africa. Molecular Biology and Evolution 11: 691– 703. Sturmbauer, C., E. Verheyen, L. Rüber & A. Meyer, 1997. Phylogeographic patterns in populations of cichlid fishes from rocky habitats in Lake Tanganyika. In Stepien, C. A. & T. D. Kocher (eds), Molecular Systematics of Fishes. Academic Press, San Diego: 97–111. Sturmbauer, C., S. Baric, W. Salzburger, L. Rüber & E. Verheyen, 2001. Lake level fluctuations synchronize genetic divergence of cichlid fishes in African lakes. Molecular Biology and Evolution 18: 144–154. Swofford, D. L., 2000. PAUP*: Phylogenetic Analysis Using Parsimony (and other methods), Beta Version 4.0. Sinauer, Sunderland, MA. Takahashi, T., 2003. Systematics of Tanganyikan cichlid fishes (Teleostei: Perciformes). Ichthyological Research 50: 367–382. Takezaki, N., A. Rzhetsky & M. Nei, 1995. Phylogenetic test of the molecular clock and linearized trees. Molecular Biology and Evolution 12: 823–833. Tamura, K. & M. Nei, 1993. Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Molecular Biology and Evolution 10(3): 512–526. Thomas, D. S. G. & P. A. Shaw, 1991. The Kalahari Environment. Cambridge University Press, UK. 284 pp. Thompson, J. D., D. G. Higgins & T. J. Gibson, 1997. CLUSTAL X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Research 24: 4876–4882. Turner, G. F., O. Seehausen, K. E. Knight, C. J. Allender & R. L. Robinson, 2001. How many species of cichlid fishes are there in African lakes? Molecular Ecology 10: 793–806. Verheyen, E., L. Rüber, J. Snoeks & A. Meyer, 1996. Mitochondrial phylogeny of rockdwelling cichlid fishes reveals evolutionary influence of historical lake level fluctuations of Lake Tanganyika, Africa. Philosophical Transactions of the Royal Society, London B 351: 797–805. Verheyen, E., W. Salzburger, J. Snoeks & A. Meyer, 2003. Origin of the cichlid fishes from Lake Victoria, East Africa. Science 300: 325–329. Wellington, J. H., 1955. Southern Africa – A Geographic Study. Vol. 1, Physical Geography. Cambridge University Press, UK, 528 pp.