Molecular Phylogenetics and Evolution 60 (2011) 385–397 Contents lists available at ScienceDirect Molecular Phylogenetics and Evolution journal homepage: www.elsevier.com/locate/ympev Phylogenetic relationships and the temporal context for the diversification of African characins of the family Alestidae (Ostariophysi: Characiformes): Evidence from DNA sequence data Jairo Arroyave a,b,⇑, Melanie L.J. Stiassny a a b Department of Ichthyology, Division of Vertebrate Zoology, American Museum of Natural History, Central Park West at 79th St., New York, NY 10024, USA Department of Biology, The Graduate School and University Center, The City University of New York, 365 Fifth Avenue, New York, NY 10016, USA a r t i c l e i n f o Article history: Received 14 December 2010 Revised 15 April 2011 Accepted 20 April 2011 Available online 30 April 2011 Keywords: Alestidae Phylogeny Systematics Divergence time estimation Biogeography a b s t r a c t Phylogenetic relationships within the family Alestidae were investigated using parsimony, maximum likelihood, and Bayesian approaches based on a molecular dataset that included both nuclear and mitochondrial markers. Multiple representatives of all but two of the recognized alestid genera were included, which allowed for testing previous hypotheses of intergeneric relationships and the monophyly of several genera. The phylogenetic position of the Neotropical genus Chalceus with respect to the family Alestidae was also examined. In order to understand the temporal context of alestid diversification, Bayesian methods of divergence time estimation using fossil data in the form of calibration priors were used to date the nodes of the phylogenetic tree. Our results rejected the monophyly of the family as currently recognized (Alestidae sensu lato) and revealed several instances of poly- and paraphyly among genera. The genus Chalceus was recovered well nested within Neotropical characiforms, thus rejecting the hypothesis that this taxon is the most basal alestid. The estimated mean divergence time for the alestid clade (Alestidae sensu stricto) was 54 Mya with a 95% credibility interval of 63–49 Mya. These results are incongruent with the hypothesis that the origin of the family Alestidae predates the African-South American DriftVicariance event. Ó 2011 Elsevier Inc. All rights reserved. 1. Introduction Fishes of the order Characiformes are among the most diverse and abundant components of the Neotropical and African freshwater ichthyofaunas. Although by far the greatest diversity of the order occurs in the Neotropics, African freshwaters harbor more than 200 species currently arrayed in four families. With a total of 121 recognized species in 21 genera, and including well-known forms such as the giant African goliath tigerfish (Hydrocynus goliath) or the Congo tetra (Phenocogrammus interruptus) popular with aquarists, Alestidae is the most speciose of the African characiform families (Eschmeyer and Fong, 2010). While a few alestids are found in the lower Nile basin and in scattered localities in the horn of Africa, the family is primarily sub-Saharan where it occurs predominantly in lowland rivers, reaching highest diversity in the Congo River basin, Lower Guinea, and the coastal rivers of West Africa (Roberts, 1975; Zanata and Vari, 2005; Eschmeyer, 2010). Alestidae was erected by Géry (1977) to include the African members of the family Characidae sensu Greenwood et al. (1966), originally classified into the subfamilies Hydrocyninae and Alesti⇑ Corresponding author. Fax: +1 212 769 5642. E-mail address: jarroyave@amnh.org (J. Arroyave). 1055-7903/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.ympev.2011.04.016 nae (Roberts, 1969). Alestid subfamilial taxonomy is widely recognized to be artificial, based on combinations of often broadly overlapping features, many of which are variable within and among species (Schaefer, 2007). The subfamily Hydrocyninae, for instance, was created to accommodate the exceptionally largebodied tigerfishes of the genus Hydrocynus. The subfamily Alestinae (which includes all remaining alestid genera) was divided into Alestini and Petersiini to accommodate mid-sized and dwarf species, respectively (Roberts, 1969; Géry, 1995). Despite previous claims by Vari (1979, p. 342) suggesting that recognition of Hydrocyninae would render Alestinae non-monphyletic, Géry (1995) maintained Roberts’ (1969) subfamilial classification. In a study focusing on the higher-level relationships among characiform fishes, Ortí and Meyer (1997) presented for the first time preliminary evidence for the monophyly of the family. Their results, however, were based on a very limited taxon sampling that included only three alestid genera. Although Vari (1998) argued that previous studies (Roberts, 1969; Vari, 1979, 1995) provided morphological characters indicative of alestid monophyly, his conclusions were based on inductive generalizations rather than a list of synapomorphies resulting from a comprehensive cladistic analysis. Relationships among members of Alestidae were investigated further using phylogenetic methods by Murray and Stewart (2002), 386 J. Arroyave, M.L.J. Stiassny / Molecular Phylogenetics and Evolution 60 (2011) 385–397 followed by the works of Hubert et al. (2005), Zanata and Vari (2005), and Calcagnotto et al. (2005). Although Murray and Stewart (2002) provided several synapomorphies supporting the monophyly of the family, their study focused on the relationships of the genera Alestes and Brycinus, and included limited sampling of alestid generic diversity. Likewise, the sampling of alestid taxa in the study of Hubert et al. (2005) consisting of only six genera. Such limited taxon sampling necessarily decreases phylogenetic accuracy (Hillis, 1998), rendering the results of those studies suspect at best. By contrast, the monumental contribution of Zanata and Vari (2005) included representatives of 19 alestid genera and surveyed a large number of characters from a wide variety of morphological systems, resulting in the most comprehensive compendium of alestid comparative anatomy and the ensuing phylogenetic hypothesis derived from that dataset. Similarly, the molecular phylogeny of Calcagnotto et al. (2005) – although primarily focused on suprafamilial relationships within the order Characiformes – included representatives of 14 alestid genera, and comparative data from two mitochondrial (16S and cyt-b) and four nuclear (RAG2, sia, fkh, and trop) genes. Both studies corroborated the monophyly of Alestidae, yet the recovered topologies were incongruous with each other (Fig. 1). Based on both the age of the oldest fossil assignable to Alestidae and the time of the African/South American drift-vicariance event, Zanata and Vari (2005, p. 120, Fig. 44) proposed age estimates for higher-level clades within the family. Such estimates, however, are problematic for several reasons. First, use of palaeontological data alone very likely underestimates the ages of lineages, as the appearance of the oldest fossil is expected to postdate the origin of the clade it belongs to (Marshall, 1990). Second, use of a biogeographic event to estimate the age of the alestid clade is critically dependent on the phylogenetic accuracy of the topology, yet the position of the South American genus Chalceus with respect to alestids as hypothesized by Zanata and Vari (2005) might be an artifact of incomplete taxon sampling. This is because several lin- eages of Neotropical characins were not represented in their set of outgroup taxa. Third, even if the status of the Neotropical genus Chalceus as the most basal alestid is correct, it is not the only instance of an African/South American sister-group relationship across the phylogeny of the order (Ortí and Meyer, 1997; Buckup, 1998; Calcagnotto et al., 2005; Malabarba and Malabarba, 2010). Thus, there is no compelling reason to propose the breakup of Gondwana as the cause of the split between Chalceus and the alestid clade. As a result, the temporal context of alestid diversification as proposed by Zanata and Vari (2005) remains in need of critical evaluation. Reliably inferring the pattern and timing of cladogenetic events in the alestid phylogeny is therefore essential for an improved understanding of the evolutionary history of the family and its implications for the historical biogeography of characiform fishes. Molecular-dating techniques, although still far from perfect, allow for estimation of the ages of clades by taking into account many of the uncertainties associated with converting genetic distances into time units (Rutschmann, 2006). In theory, dating the nodes of a phylogenetic tree from DNA sequence data requires a constant rate of substitution among lineages (i.e., a strict molecular clock) and available fossil information (often treated as fixed) to calibrate the clock. Recently, however, so-called ‘‘relaxed-clock’’ methods (e.g., Sanderson, 1997; Rambaut and Bromham, 1998; Thorne et al., 1998; Drummond et al., 2006) allow departures from clock-like behavior. These present an enticing approach to divergence time estimation because molecular datasets rarely conform to a strict clock and such a lack of rate constancy has been regarded as the main obstacle for an accurate molecular dating of phylogenies (Smith and Peterson, 2002). In particular, the Bayesian relaxed-clock method of Drummond et al. (2006) excels in allowing users to employ probability distribution-based calibrations instead of fixed-age nodes, thus modeling some of the uncertainties associated with calibrating substitution rates from palaeontological data. Fig. 1. Previous hypotheses of alestid interrelationships based on molecular (Calcagnotto et al., 2005) and morphological data (Zanata and Vari, 2005). J. Arroyave, M.L.J. Stiassny / Molecular Phylogenetics and Evolution 60 (2011) 385–397 Although the works of Zanata and Vari (2005) and Calcagnotto et al. (2005) are hitherto the best studies addressing the monophyly and the phylogenetic relationships of Alestidae, they disagree on the placement of the Neotropical genus Chalceus, and derived discordant topologies. Therefore, a phylogenetic analysis of Alestidae based on independent evidence seems desirable, especially when a focused molecular phylogeny of the family has yet to be proposed. Furthermore, improved estimates for the absolute ages of clades in the alestid tree can be generated from fossil and DNA sequence data by means of molecular-dating techniques. Thus, the main objectives of the present study are to present the most comprehensive phylogeny of Alestidae based on molecular data and to date the nodes of the resultant phylogenetic tree. Accordingly, previous hypotheses of alestid interrelationships, the monophyletic status of the family and its genera, and the phylogenetic position of Chalceus are tested. By establishing the temporal context of alestid diversification, alternative biogeographic hypotheses explaining the Gondwanan distribution of the order Characiformes are explored. 2. Materials and methods 2.1. Taxon sampling Representatives of all valid genera, except Petersius and Virilia (for which tissue vouchers were unavailable) were sampled for phylogenetic analyses. Whenever possible, multiple species of each genus were included. Although multiple individuals were sequenced for most species, the analyses presented herein are based on a reduced dataset that includes DNA sequence data for only one individual per species. This was partially due to avoid analyses of an unnecessarily redundant and larger dataset, but mostly because it was a requirement for the Bayesian method of phylogeny and divergence time estimation employed in this study. Sequencing of multiple individuals per species, however, allowed for an improved control of sequence quality and contamination issues. Tissues were primarily obtained from specimens collected during field expeditions of the ongoing NSF-funded Biotic Surveys and Inventories Congo Project (http://research.amnh.org/vz/ichthyology/congo/ index.html). Additional tissues were obtained from samples deposited in the Ambrose Monell Cryo Collection (AMCC) at the American Museum of Natural History (AMNH), augmented by donations from colleagues at the Cornell Museum of Vertebrates (USA), the Zoologische Staatssammlung München (Germany), and the South African Institute for Aquatic Biodiversity. Material examined (voucher specimens) and GenBank accession numbers for the gene sequences generated and included in this study are listed in Table 1. A total of 64 terminals (53 ingroup and 11 outgroup species) encompassed the taxonomic sampling for phylogenetic analyses. Outgroup choice was informed by previous higher-level phylogenetic hypotheses of characiform relationships (e.g., Buckup, 1998; Calcagnotto et al., 2005). Hence, outgroup taxa included representatives of 7 characiform families: the remaining three African families (i.e., Citharinidae, Distichodontidae, and Hepsetidae), and four of the Neotropical families historically regarded as closely related to alestids (i.e., Erythrinidae, Ctenoluciidae, Acestrorhynchidae, and Characidae). 2.2. Marker selection and character sampling Aiming to recover both deeper and more recent divergences, four protein-coding genes with markedly different rates of substitution were chosen for phylogenetic analyses. These consist of two nuclear (SH3PX3 and myh6) and two mitochondrial (COI and cyt-b) markers, totaling more than 3000 bp. SH3PX3 and myh6 were pro- 387 posed by Li et al. (2007) as promising markers with potential use in molecular systematics of actinopterygian fishes. These have been employed in empirical phylogenetic studies of Cypriniformes (Chen et al., 2008), Gasterosteiformes (Kawahara et al., 2009), Stomiiformes (DeVaney, 2008), Cyprinodontiformes (Meredith et al., 2010), and Perciformes (Li et al., 2010), among others. On the other hand, the mitochondrial markers COI and cyt-b have been consistently used in evolutionary studies across most animal phyla (e.g., Folmer et al., 1994; Kocher et al., 1989), and have proven useful in resolving phylogenetic relationships of characiform fishes (e.g., Calcagnotto et al., 2005; Javonillo et al., 2010). 2.3. DNA extraction, amplification and sequencing Total genomic DNA was extracted from tissues (muscle and fin clips), preserved in 95% EtOH and stored frozen, using DNEasy Tissue Extraction Kit (Qiagen) following the manufacturer’s protocol. For each sample and gene, DNA amplification via Polymerase Chain Reaction (PCR) was preformed in a 25-lL volume containing one Ready-To-Go PCR bead (GE Healthcare), 21 lL of PCR-grade water, 1 lL of each primer (10 lM), and 2 lL of genomic DNA. Doublestranded PCR products were purified using AMPure (Agencourt). Primer sequences and PCR profiles are listed in Table 2. Sequencing of each strand of amplified product was performed in a 5-lL volume containing 1 lL of primer (3.2 lM), 0.75 lL of BigDyeÒ Ready Reaction Mix, 1 lL of BigDyeÒ buffer, and 2.25 lL of PCR-grade water. Sequencing reactions consisted of a 2-min initial denaturation at 95 °C, followed by 35 cycles of denaturation at 95 °C for 30 s, annealing at 50 °C for 60 s, and extension at 72 °C for 4 min, followed by a 3-min final extension at 72 °C. For sequencing COI, however, annealing temperature was set at 45 °C. All sequencing reactions were purified using CleanSEQ (Agencourt) and electrophoresed on an Applied Biosystems 3700 automated DNA sequencer at the AMNH Molecular Systematics Laboratories. Contig assemblage and sequence editing was performed using the software Geneious Pro v4.6.2. 2.4. Phylogenetic analyses Phylogenetic analyses were carried out using both parsimony and model-based approaches. Prior to phylogeny estimation, nucleotide base correspondences (‘‘primary homologies’’ sensu de Pinna, 1991) were identified by means of multiple sequence alignment (MSA). An alignment for each gene was performed using MUSCLE (Edgar, 2004) under default parameters, followed by concatenation of individual alignments. Calculation of the number of variable and parsimony informative sites was conducted using MEGA v4. (Tamura et al., 2007). When using parsimony as an optimality criterion, both non-additive optimization (Fitch, 1971) and direct optimization (Wheeler, 1996) were implemented in the programs TNT [Willi Hennig Society edition] (Goloboff et al., 2003; Goloboff et al., 2008) and POY v4.1. (Varón et al., 2008), respectively. A dynamic homology approach (i.e., direct optimization) was applied with the expectation of finding more parsimonious phylogenetic hypotheses because in addition to minimizing substitutions and insertion–deletion events, nucleotide correspondences (i.e., putative homologies) themselves are chosen to minimize tree length (Wheeler, 2001; Wheeler et al., 2006). In all parsimony analyses gaps were treated as a fifth state and no cost for gap opening was specified. An indel/substitution cost ratio of 1:1 was set as the transformation cost matrix to be used in calculating the length of the tree. Tree searches were performed by building a set of Wagner trees (Farris, 1970) by random addition sequence of taxa (RAS), followed by tree bisection and reconnection (TBR) branch swapping (Swofford and Olsen, 1990), perturbation using the Parsimony Ratchet (Nixon, 1999), and tree fusing (Goloboff, 1999). Branches 388 J. Arroyave, M.L.J. Stiassny / Molecular Phylogenetics and Evolution 60 (2011) 385–397 Table 1 Taxa, voucher specimens and GenBank accession numbers for the gene sequences included in the analyses. Taxon Outgroup Citharinidae Citharinus gibbosus Distichodontidae Distichodus fasciolatus Distichodus kolleri Hepsetidae Hepsetus odoe Ctenoluciidae Ctenolucius hujeta Erythrinidae Hoplias malabaricus Acestrorhynchidae Acestrorhynchus sp. Characidae Brycon sp. Hemigrammus erythrozonus Chalceus erythrurus Chalceus macrolepidotus Ingroup Alestidae Alestes baremoze Alestes inferus Alestes liebrechtsii Alestes macrophtalmus Alestopetersius caudalis Alestopetersius compressus Alestopetersius hilgendorfi Alestopetersius nigropterus Arnoldichthys spilopterus Bathyaethiops caudomaculatus Bathyaethiops greeni Brachypetersius altus Brycinus bimaculatus Brycinus comptus Brycinus fwaensis Brycinus grandisquamis Brycinus imberi Brycinus kingsleyae Brycinus lateralis Brycinus macrolepidotus Brycinus nurse Brycinus opisthotaenia Brycinus poptae Brycinus sp. ‘‘Yaekama’’ Brycinus cf. batesii Bryconaethiops boulengeri Bryconaethiops microstoma Bryconaethiops sp.’’Mpozo’’ Bryconaethiops yseuxi Bryconalestes longipinnis Clupeocharax schoutedeni Duboisialestes tumbensis Hemigrammopetersius barnardi Hydrocynus brevis Hydrocynus forskahlii Hydrocynus goliath Hydrocynus vittatus Ladigesia roloffi Lepidarchus adonis Micralestes acutidens Micralestes humilis Micralestes lualubae Micralestes occidentalis Micralestes stormsi Nannopetersius lamberti Nannopetersius mutambuei Phenacogrammus aurantiacus Phenacogrammus interruptus Phenacogrammus sp.’’Tshimbi’’ Voucher GenBank accession number COI cyt-b myh6 SH3PX3 AMNH 241039 JF800930 JF800989 JF801096 JF801161 AMNH 240041 AMNH 249814 JF800987 JF800986 JF801034 JF801033 JF801093 JF801092 JF801156 JF801155 AMNH 243489 JF800936 JF800995 JF801041 JF801108 AMCC 102198 JF800932 JF800991 JF801102 JF801104 AMCC 110702 – JF801035 JF801077 JF801158 AMCC 116689 JF800935 JF800994 JF801097 JF801107 AMCC 116693 AMCC 116707 AMNH 233413 AMCC 102150 JF800933 JF800988 – JF800931 JF800992 – JF801036 JF800990 JF801087 JF801086 JF801078 JF801039 JF801105 JF801157 JF801159 JF801103 AMCC 102186 AMNH 242137 AMNH 246418 CU 91702 AMNH 242193 AMNH 242455 AMNH 242460 AMNH 246321 AMCC 102151 CU 93170 CU 93147 AMNH 247407 AMNH 242467 AMNH 244215 ZSM 37775 ZSM 38246 AMNH 245523 AMNH 247482 AMCC 102120 AMNH 246621 AMCC 116712 AMNH 236470 CU 93137 ZSM 39400 AMNH 249531 AMNH 243601 AMNH 246417 AMNH 242188 AMNH 239455 AMNH 249561 AMNH 242486 AMNH 246659 CU 93792 AMCC 102185 AMNH 249806 AMNH 239463 AMNH 245529 ZSM 39625 ZSM 39626 AMNH 239476 AMNH 246569 AMNH 239489 ZSM 36182 AMNH 239483 ZSM 35562 AMNH 246579 ZSM 39495 AMNH 233442 ZSM 39307 JF800979 JF800982 JF800981 JF800980 JF800968 JF800970 JF800965 JF800966 JF800934 JF800972 JF800973 JF800971 JF800962 JF800959 JF800960 JF800951 JF800956 JF800961 JF800958 JF800950 JF800957 – JF800948 JF800947 JF800949 JF800955 JF800954 JF800953 JF800952 JF800940 JF800937 JF800969 JF800939 JF800984 JF800983 JF800978 JF800985 JF800938 – JF800942 – JF800941 JF800944 JF800943 JF800964 JF800963 JF800974 JF800977 JF800975 JF801026 JF801029 JF801028 JF801027 JF801017 JF801019 JF801015 JF801016 JF800993 JF801021 JF801022 JF801020 JF801013 JF801012 – JF801004 JF801009 – JF801011 JF801003 JF801010 JF801037 JF801002 – – JF801008 JF801007 JF801006 JF801005 JF800998 – JF801018 JF800997 JF801031 JF801030 JF801025 JF801032 JF800996 – JF801000 JF801038 JF800999 JF801001 – – JF801014 JF801023 JF801024 – JF801070 JF801073 JF801072 JF801071 JF801062 JF801064 JF801060 JF801061 JF801040 JF801098 JF801099 JF801088 JF801084 JF801057 JF801083 JF801082 JF801054 JF801058 JF801056 JF801049 JF801055 JF801085 JF801048 JF801047 JF801081 JF801053 JF801052 JF801051 JF801050 JF801080 JF801042 JF801063 JF801091 JF801075 JF801074 JF801069 JF801076 JF801101 JF801100 JF801044 JF801079 JF801043 JF801046 JF801045 JF801059 JF801089 JF801065 JF801068 JF801066 JF801148 JF801151 JF801150 JF801149 JF801137 JF801139 JF801134 JF801135 JF801106 JF801141 JF801142 JF801140 JF801131 JF801129 JF801130 JF801121 JF801126 – JF801128 JF801120 JF801127 – JF801118 JF801117 JF801119 JF801125 JF801124 JF801123 JF801122 JF801111 JF801109 JF801138 JF801110 JF801153 JF801152 JF801147 JF801154 – JF801162 JF801113 JF801160 JF801112 – JF801114 JF801133 JF801132 JF801143 JF801146 JF801144 389 J. Arroyave, M.L.J. Stiassny / Molecular Phylogenetics and Evolution 60 (2011) 385–397 Table 1 (continued) Taxon Voucher Phenacogrammus taeniatus Rhabdalestes maunensis Rhabdalestes septentrionalis Tricuspidalestes sp.’’Tshimbi’’ GenBank accession number ZSM 39279 SAIAB 71928 TM-503 ZSM 39324 COI cyt-b myh6 SH3PX3 JF800976 JF800946 JF800945 JF800967 – – – – JF801067 JF801095 JF801094 JF801090 JF801145 JF801116 JF801115 JF801136 Institutional abbreviations: AMCC (Ambrose Monell CryoCollection, AMNH), AMNH (American Museum of Natural History), CU (Cornell University Museum of Vertebrates), SAIAB (South African Institute for Aquatic Biodiversity), ZSM (Zoologische Staatssammlung München), TM (personal collection of Timo Mortiz). Table 2 Primers and PCR profiles for amplification of the genes used in this study. Gene Source Primer Primer Sequencea PCR thermal profileb COI Folmer et al. (1994) Kocher (1989) Irwin et al. (1991) Li et al. (2007) GGTCAACAAATCATAAAGATATTGG TAAACTTCAGGGTGACCAAAAAATCA AAAAAGCTTCCATCCAACATCTCAGCATGATGAAA AACTGCCAGTCATCTCCGGTTTACAAGAC CATMTTYTCCATCTCAGATAATGC ATTCTCACCACCATCCAGTTGAA GGAGAATCARTCKGTGCTCATCA CTCACCACCATCCAGTTGAACAT GTATGGTSGGCAGGAACYTGAA CAAACAKCTCYCCGATGTTCTC GACGTTCCCATGATGGCWAAAAT CATCTCYCCGATGTTCTCGTA (95 °C/60s, 42 °C/60s, 72 °C/90s)  35 cyt-b LCO1490 HCO2198 L14841 H15915 myh6_F459 myh6_R1325 myh6_F507c myh6_R1322c SH3PX3_F461 SH3PX3_R1303 SH3PX3_F532c SH3PX3_R1299c myh6 SH3PX3 Li et al. (2007) (95 °C/60s, 50 °C/60s, 72 °C/120s)  35 (98 °C/10s, (98 °C/10s, (98 °C/10s, (98 °C/10s, (98 °C/10s, (98 °C/10s, (98 °C/10s, (98 °C/10s, 53 °C/30s, 51 °C/30s, 62 °C/30s, 60 °C/30s, 55 °C/30s, 53 °C/30s, 62 °C/30s, 60 °C/30s, 72 °C/45s) 72 °C/45s) 72 °C/45s) 72 °C/45s) 72 °C/45s) 72 °C/45s) 72 °C/45s) 72 °C/45s)         15 15 15 15 15 15 15 15 + + + + a Listed from 50 to 30 . Conditions for denaturation, annealing and extension steps for each cycle are listed in parenthesis, followed by the number of cycles. All reactions included a 5-min initial denaturation at 95 °C and a 7-min final extension at 72 °C. c Primers used during a second (nested) PCR required for successful amplification. b for which the minimum possible length was zero were collapsed. TNT analyses included 100 RAS. Tree searches in POY were executed using the search command, which performs as many builds, swaps, ratchets, and tree fusings as possible within the specified time (A. Varón, pers. comm.). POY analyses were carried out on the AMNH Parallel Computing Cluster over a period of 5 days, using 32 processors. Tree statistics for parsimony analyses included tree length (L), ensemble consistency index (CI; Kluge and Farris, 1969) and ensemble retention index (RI; Farris, 1989). Maximum likelihood [ML] (Felsenstein, 1981) and Bayesian (Rannala and Yang, 1996) analyses comprise the model-based approaches to phylogenetic inference used in this study. In order to accommodate the potential process heterogeneity that might occur among the different gene regions of our dataset, both ML and Bayesian analyses were conducted on the concatenated alignment partitioned into gene regions with parameters unlinked. Nucleotide substitution model selection for each gene partition was accomplished by means of both hierarchical likelihood ratio tests (hLRTs) and the Akaike Information Criterion (AIC) using the program jModelTest (Posada, 2008). Acceptance/rejection of the null hypothesis in the hLRTs was based on a significance level of 0.05. Both indices determined that, for all gene partitions, the model that best described the evolution of the sequence data over the phylogeny was the GTR + I + C (see Section 3). ML analyses were conducted with RAxML v7.0.4. (Stamatakis, 2006) through the Cyberinfrastructure for Phylogenetic Research (CIPRES) project web server (Miller et al., 2009). Bayesian analyses included the co-estimation of phylogeny and divergence dates, and were carried out in BEAST v1.4.8. (Drummond and Rambaut, 2007). Clade support was estimated by means of the bootstrap character resampling method (Felsenstein, 1985) for TNT and RAxML (Stamatakis et al., 2008) analyses, the Bremer Index (Goodman et al., 1982; Bremer, 1988) for POY analyses, and clade posteriors for BEAST analyses. Bootstrap support values were calculated based on 1000 pseudoreplicates. 2.5. Divergence time estimation Prior to estimation of divergence times for the alestid phylogeny, a likelihood ratio test (LRT) was performed using the program PAUP⁄ v4.0. (Swofford, 2003) in order to determine whether rates of nucleotide substitution in our dataset departed significantly from expectations under the assumption of a strict molecular clock. Results of the LRT indicated that a strict molecular clock was inappropriate to estimate dates of divergence in our dataset, so we relied on the Bayesian relaxed-clock method of Drummond et al. (2006) under the uncorrelated lognormal (UCLN) rate variation model with default parameters as implemented in the software BEAST. In the UCLN relaxed-clock model there is no correlation of rates on adjacent branches of the tree: the rate on each branch is drawn independently and identically from a lognormal distribution described by mean and variance parameters (Drummond et al., 2006). Selection of this model was based on the idea that autocorrelated relaxed-clock models might be inappropriate under several conditions, for instance, when disparity in life-history traits among taxa contributes a substantial amount of variation in the inherited determinants of rates (Smith et al., 2010). In addition, empirical and simulation studies have shown that the UCLN model performs well both in terms of accuracy and power, even when the data were generated under alternative models (Drummond et al., 2006). We assumed uniform priors for both the GTR + I + C and UCLN model parameters, and a Yule process prior for topology and divergence times. The Yule process implies a prior distribution under a pure-birth stochastic branching process model (i.e., speciation as random lineage splitting) and its use seems appropriate if priors on trees are expected to reflect the underlying physical process by which trees are generated (Velasco, 2008). Moreover, various authors have claimed that uniform priors on topologies introduce bias in favor of smaller and larger clades and against medium sized 390 J. Arroyave, M.L.J. Stiassny / Molecular Phylogenetics and Evolution 60 (2011) 385–397 ones (Pickett and Randle, 2005; Goloboff and Pol, 2005; Yang, 2006), as well as a skewed distribution on tree shapes (Velasco, 2008). Likewise, labeled histories (i.e., topologies with a temporal ordering of the nodes) are not equally probable if all topologies are allowed the same prior probability (Velasco, 2008). A Yule process, instead, produces each labeled history with equal probability (Edwards, 1970). When implemented in BEAST, this prior assumes a constant speciation rate per lineage and has a single parameter (yule.birthRate) that represents the average net rate of lineage birth. Under this prior, branch lengths are expected to be exponentially distributed with a mean of yule.birthRate 1 (Drummond et al., 2007). Because fossils correctly placed in a phylogeny can only provide minimum age estimates for a particular lineage (Marshall 1990), estimates of substitution rates were calibrated using fossil information in the form of probabilistic priors rather than as fixed values. Specifically, a lognormally distributed prior (with a rigid lower bound, mean and standard deviation parameters) was chosen to accommodate some of the uncertainties associated with the use of fossil data to specify the ages of the internal nodes used as calibration points. This prior is generally regarded as the most appropriate for modeling palaeontological information (Hedges and Kumar 2004; Drummond et al., 2006) and its use implies that the actual cladogenetic event is most likely to have occurred at some time prior to the earliest appearance of the fossil (Ho, 2007). Absolute estimates of divergence times for the alestid phylogeny were calculated using the two oldest fossils unambiguously assignable to the family Alestidae as calibration points. These included isolated teeth from the Late Miocene (5–7 Mya) of Kenya and Zaire assignable to the most recent common ancestor (MRCA) of the genus Hydrocynus (Van Neer, 1992; Stewart, 1994, 2003), and from the Early Eocene (49–54.8 Mya) of the Iberian Peninsula assignable to the MRCA of the genera Alestes, Brycinus and Bryconaethiops (De la Peña Zarzuelo, 1996; Zanata and Vari, 2005, p. 123). The mean and standard deviation parameters of the lognormal prior probability distribution associated with these calibrations were chosen so that most of the probability lies within the interval representing the age of the fossil, yet allowing the true age of divergence to extend much further back in time. The rigid minimum bounds were set as the lower limit of the fossil age interval (i.e., 5 and 49 Mya, respectively). BEAST analyses were implemented using the Markov Chain Monte Carlo (MCMC) algorithm run for 20  106 generations, with a sampling frequency of 1000 generations and default proposal mechanisms. Convergence of the MCMC algorithm to a stationary distribution, and thus the number of generations to be discarded as burn-in, was determined by examining trace plots of posterior probability vs. number of generations using Tracer v1.5 (Rambaut and Drummond, 2009). The maximum a posteriori (MAP) tree (Rannala and Yang, 1996), a chronogram indicating the mean ages of all nodes with their associated 95% highest posterior density (HPD) intervals, and the posterior probabilities of nodes were cal- culated from the set of post burn-in trees using TreeAnnotator v1.5.3 (Drummond and Rambaut, 2007). 3. Results 3.1. Nucleotide substitution model selection Summary of the results from the statistical selection of best-fit models performed in jModelTest is presented in Table 3. Both hLRTs and AIC established that the model that fit our data best for all gene partitions was the general time reversible (GTR; Rodriguez et al., 1990) with a proportion of invariable sites (I) and a gamma-distributed rate heterogeneity among sites (C). 3.2. Nucleotide homologies A matrix corresponding to the hypotheses of primary homology as determined by MSA is presented in Appendix A. The concatenated alignment included a total of 3271 characters, from which 1323 were variable, 9 indels, and 1145 parsimony informative. None of the indels was phylogenetically informative. Individual alignments for COI, cyt-b, myh6 and SH3PX3 consisted of 658, 994, 820, and 799 nucleotides, respectively. 3.3. Alestid phylogeny Two equally most parsimonious trees of length 8344 were found during POY analyses. In contrast, cladogram search in TNT resulted in three equally most parsimonious trees of length 8358. The strict consensus trees of POY and TNT were identical in topology, except for the position of the least inclusive clade containing Ctenolucius hujeta and Arnoldichthhys spilopterus, which was recovered more basally by POY (Fig. 2). Similarly, the most optimal trees recovered by ML and Bayesian analyses were topologically equivalent (Fig. 3). Bootstrap values in the TNT topology were notably low for several nodes. By contrast, most nodes in the POY topology were well supported, having Bremer values above 10. Similarly, most clades in the ML and Bayesian topologies were well supported (bootstraps > 75% and clade posteriors > 0.9) (Fig. 3). In general, the parsimony topology was concordant with that obtained by model-based methods. Both approaches recovered the African genera Arnoldichthhys and Lepidarchus nested well within the Neotropical members of the outgroup, thus rejecting the monophyly of the family as currently recognized (Alestidae sensu lato). Moreover, contrary to the hypothesis of Zanata and Vari (2005), the Neotropical genus Chalceus was resolved as more closely related to a subset of South American characins than to African alestids. All analyses recovered the remaining members of the ingroup (hereafter referred to as Alestidae sensu stricto [s.s.]) resolved into four major lineages (Clades A–D in Fig. 4). Resolution within each lineage, however, differed slightly between methods. 3.4. Monophyly of alestid genera Table 3 AIC values, likelihood function ( lnL), and likelihood ratio (LR) statistics with associated P-values for the null and alternative models for each gene partition. Gene Model AIC COI GTR+I + C HKY + I + C GTR + I + C HKY + I + C GTR + I + C GTR + C GTR + I + C HKY + I + C 25451.3778 25455.2096 34722.1206 34782.2185 22416.9149 22419.9031 11416.9149 11435.4361 cyt+b myh6 SH3PX3 lnL 12421.6889 12427.6048 17057.0603 17093.5782 5211.5098 5183.8185 5404.4574 5417.718 LR P-value 11.8318 0.018647 73.036001 0 4.617401 0.03165 26.5212 2.50E 05 Due to taxon sampling limitations (i.e., availability of tissues), it was only possible to test monophyly for 10 of the 15 polytypic alestid genera. Overall, our results corroborated the monophyly of the genera Bryconaethiops, Hydrocynus, Alestes, Nannopeterius, Micralestes, and Phenacogrammus. The last-named, however, is not monophyletic according to the results of the parsimony analyses. This is because P. aurantiacus was found more closely related to Clupeocharax schoutedeni than to the other Phenacogrammus species included in this study (Fig. 2). On the other hand, we found no support for the monophyletic status of four currently recognized genera. Brycinus was recovered as polyphyletic with its members J. Arroyave, M.L.J. Stiassny / Molecular Phylogenetics and Evolution 60 (2011) 385–397 391 Fig. 2. Phylogenetic relationships of Alestidae as inferred by parsimony analyses and represented by the strict consensus tree obtained with direct optimization (L = 8344; CI = 0.26 RI = 0.52). Bremer and bootstrap (>50%) support values are shown above and below branches, respectively. The position of the clade Ctenolucius hujeta + Arnoldichthhys spilopterus in the TNT topology (i.e., non-additive optimization) is represented by the dotted branch. Outgroup taxa are in gray font. Neotropical members of the outgroup are labeled with an asterisk. distributed in two separate clades: one at the base of the alestid tree and sister to all other alestids; the other recovered as the sister group of Bryconaethiops. Similarly, the genera Bathyaethiops, Alestopetersius, and Rhabdalestes were recovered as paraphyletic given their placement with respect to members of the genera Brachypetersius, Tricuspidalestes, and Hemigrammopetersius respectively (Figs. 2 and 3). 3.5. Divergence time estimates Estimates of divergence times for the alestid phylogeny (Fig. 5) indicate that the origin of the family Alestidae s.s. dates back to the Early Tertiary (54 Mya; 95% HPD interval = 63–49) and that most diversification occurred during the mid-Tertiary within a period of just 30 My (40–10 Mya). Furthermore, our analyses indicate that the split between the lineage leading to the MRCA of alestids and the lineage that includes all other characiforms – except citharinoids – occurred during the Late Cretaceous (78 Mya; 95% HPD interval = 99–59). The estimated age of the nodes used as calibration points did not exactly match the age of the fossils assigned to those clades, especially for the node representing the MRCA of the genus Hydrocynus, in which the estimated age (21 My, 95% HPD interval = 30–13) is considerably older than the 5-My minimum age based on fossil data. On the other hand, the estimated age of the MRCA of the genera Alestes, Brycinus and Bryconaethiops (54 My; 95% HPD interval = 63–49) did not differ significantly from 392 J. Arroyave, M.L.J. Stiassny / Molecular Phylogenetics and Evolution 60 (2011) 385–397 Fig. 3. Phylogenetic relationships of Alestidae as inferred by ML and Bayesian analyses and represented by the most optimal tree recovered by RAxML. Bootstrap values (>50%) and clade posteriors (>0.5) are shown above and below branches, respectively. Outgroup taxa are in gray font. Neotropical members of the outgroup are labeled with an asterisk. that suggested by the fossil age (49 My). In addition to the timing of cladogenetic events within the Alestidae, our results indicate that the split between citharinoids and the lineage leading to the remaining characiforms must have occurred sometime between the mid- and Late Cretaceous (87 Mya; 95% HPD interval = 119–59). This latter result, however, should be viewed with caution given that our sampling of Neotropical characiform lineages is far from comprehensive and only two fossil calibrations – restricted to the alestid clade – were included in the analyses. 4. Discussion 4.1. Monophyly and intergeneric relationships of Alestidae All optimality criteria resulted in a similar pattern of relationships, suggesting that the alestid phylogeny as inferred from our data is robust and not significantly affected by analytical method. However, in contrast to model-based approaches, parsimony analyses do not support a sister-group relationship between Hydrocynus and Alestes nor do they support a monophyletic Phenacogrammus (exclusive of Clupeocharax). Additionally, results from different methods revealed a few subtle differences in the branching pattern among outgroup taxa (Fig. 4). Although our results indicate that continued inclusion of the genera Arnoldichthhys and Lepidarchus in Alestidae would render the family non-monophyletic, we cannot discount the possibility that the phylogenetic position of the diminutive West African Lepidarchus as revealed by our analyses is an artifact of missing data, given that none of the mitochondrial genes was successfully sequenced for this particular taxon. However, as most of the phylogenetic signal resolving intergeneric relationships (i.e., deeper divergences) is provided by less variable nuclear markers, we could anticipate that failure to include mitochondrial data would be problematic mostly for resolving divergences at less inclusive levels. Interestingly, this is the first time that such a hypothesis regarding the position of Lepidarchus has been proposed. Of the few studies that have investigated generic interrelationships of Alestidae, only the morphology-based analysis of Zanata and Vari (2005) included comparative data for this genus. In contrast to our results, they hypothesized a sister-group relationship between Lepidarchus and Ladigesia (Fig. 1), a relationship supported by eleven morphological synapomorphies (Zanata and Vari, 2005, p. 113). Given the morphological support for this species pair, our findings regarding the position of Lepidarchus should perhaps be viewed with caution because of the missing data issue. We note however, that most of the characters optimized as synapomorphic for the Lepidarchus + Ladigesia sister-pair by Zanata and Vari (2005) are reversals potentially homoplastically associated with developmental truncation at small size (true also for Tricuspidalestes which is also a diminutive species, and the proposed sister group to the Lepidarchus + Ladigesia pair). Hopefully, future analyses without missing data, and ultimately a comprehensive reevaluation of both molecular and morphological data in a total evidence context, will elucidate the actual pattern and taxonomic changes will be proposed if necessary. J. Arroyave, M.L.J. Stiassny / Molecular Phylogenetics and Evolution 60 (2011) 385–397 393 Fig. 4. Summary trees representing intergeneric relationships within Alestidae as inferred by parsimony (right) and model-based approaches (left). Major subfamilial clades are indicated by the letters A–D. Fig. 5. BEAST chronogram showing divergence time estimates for the alestid phylogeny represented by the mean ages of clades (nodes heights). Gray bars correspond to the 95% highest posterior density (HPD) intervals and are shown for intergeneric and higher-level divergences only. Timescale is in millions of years before present. Nodes used as calibration points are indicated by a dagger ( ). African taxa shown in black font; Neotropical taxa shown in gray font. 394 J. Arroyave, M.L.J. Stiassny / Molecular Phylogenetics and Evolution 60 (2011) 385–397 Contrary to the unexpected placement of Lepidarchus in our study, the phylogenetic position of the genus Arnoldichthhys is perhaps not as surprising. Even though both Zanata and Vari (2005) and Calcagnotto et al. (2005) recovered this genus as basal to the Alestidae, the latter acknowledged that this result was indeed weakly supported. Despite attributing the lack of support to a long-branch attraction artifact, Calcagnotto et al. (2005, p. 144) did not discount the possibility that the monophyly of Alestidae may be compromised by the problematic placement of Arnoldichthys. Zanata and Toledo-Piza (2004) and Zanata and Vari (2005) have noted the striking morphological resemblance between Arnoldichthys (a monotypic genus, endemic to the Niger Delta region of west Africa) and members of the Neotropical genus Chalceus (with five species distributed widely in South America) (Zanata and Toledo-Piza, 2004), and there is certainly no question that these two genera do bear a striking phenotypic similarity that is not shared with other African alestids. However, our analyses consistently recover the Neotropical taxon Ctenolucius hujeta as the sister group of Arnoldichthhys, and place Chalceus in a clade with the genera Acestrorhynchus, Brycon, and Hemigrammus. So despite the morphological characters enumerated by Zanata and Vari (2005) supporting the placement of Chaleus and Arnoldichthys as sequential sister taxa to the remaining alestids, our analyses place Chalceus and Arnoldichthys well separated from the members of that clade. While our sampling of South American characins is far from comprehensive enough to allow us unequivocally to state that Chalceus is indeed more closely related to an assemblage that includes Acestrorhynchus, Brycon, and Hemigrammus or that Arnoldichthys is more closely related to the Ctenoluciidae than to any other characiforms, we find no support for their alignment with the remaining African taxa. In view of this conflict, we suggest it prudent to exclude both genera from the family in order to maintain a monophyletic Alestidae well supported by congruent molecular and morphological character data. The phylogenetic analyses presented here recovered an Alestidae s.s. consisting of four major suprageneric assemblages represented by well-supported clades (A–D in Fig. 4). Except for the position of Bryconalestes and the genera not included in their study, Calcagnotto et al. (2005) recovered the same four major subfamilial clades, yet the branching pattern between, and resolution within each clade were not exactly as in the phylogeny proposed herein. Both studies strongly reject the monophyly of the subfamily Alestinae, as well as the monophyly of the tribes Alestini and Petersiini, and thus do not conform to the existing subfamilial and tribal classification (Roberts, 1969; Géry, 1995). Accordingly, we concur with Calcagnotto et al.’s (2005) claim that ‘‘continued recognition of subfamilies and tribes within Alestidae must be reconsidered’’. In the interest of promoting a phylogeny-based taxonomy, we suggest that the aforementioned suprageneric clades form a useful basis for future revisional studies. Our results further corroborate some previous findings regarding intergeneric relationships at less inclusive levels within the alestid phylogeny. Namely, they support a close relationship between Hydrocynus and Alestes (Brewster, 1986; Murray and Stewart, 2002; Hubert et al., 2005; Calcagnotto et al., 2005), a close relationship between Ladigesia and Micralestes (Géry, 1968; Calcagnotto et al., 2005), and a sister-group relationship between a subset of Brycinus s.l. and the genus Bryconaethiops (Murray and Stewart, 2002; Calcagnotto et al., 2005). These results are, however, in conflict with the morphology-based hypothesis proposed by Zanata and Vari (2005), suggesting that ultimate resolution of these discrepancies lie with an augmentation and reevaluation of both molecular and morphological data in the context of a total evidence analysis. Concern may be raised about not having combined our molecular data set with the morphological matrix of Zanata and Vari (2005) under a total evidence approach. There are, however, two main reasons behind our choice. First, the fact that our taxon sampling was substantially non-overlapping (especially at the species-level) with that of Zanata and Vari (2005), which would inevitably lead to considerable amounts of missing data. Second, and probably most importantly, the fact that at present we do not have access to all the comparative material necessary for a meaningful re-examination of characters and homology statements in order to understand and resolve potential conflicts. The focus of this study, as previously stated, was to reconstruct the most comprehensive phylogeny of the family based on independent evidence. Despite relying on DNA sequence data only, we believe that the molecular characters used in the present study proved to be quite informative, as indicated by the overall degree of phylogenetic resolution and high clade support values (Figs. 2 and 3). Although a combined analysis represents a major priority for future research (given the extensive disparity between the most comprehensive morphology- and DNA-based phylogenies to date), we consider that the results presented herein are a useful contribution to help direct revisionary studies in order to facilitate a much-needed improvement of the alpha-taxonomy of the family (Stiassny and Schaefer, 2005; Schaefer, 2007). 4.2. Monophyly of alestid genera Attempts at a phylogeny-based generic-level classification of Alestidae are generally recent, yet they remain conflicting and poorly supported by apomorphy-based diagnoses (Schaefer, 2007). Although a provisional scheme of alestid interrelationships was provided by Zanata and Vari (2005), the monophyletic status of most genera remains to be fully assessed. Overall, our results rejected the monophyly of Brycinus, Rhabdalestes, Alestopetersius, and Bathyaethiops. The polyphyly of Brycinus, a speciose and commercially important Pan-African genus, has been repeatedly suggested (Murray and Stewart, 2002; Hubert et al., 2005; Calcagnotto et al., 2005), all of who recognize the monophyly of the ‘‘macrolepidotus’’ species group of Paugy (1986); a grouping of eight large-bodies species mainly distributed in Central Africa and which includes the type species of the genus (B. macrolepidotus). By corroborating this hypothesis, our results underscore the desirability of a future revisionary study of this economically important genus. Certainly, continued recognition of Brycinus as presently conceived should be reconsidered and efforts focused on revisional and phylogenetic studies to resolve the composition and relationships of the rump Brycinus s.l. Similarly, previous studies did not support the monophyly of Rhabdalestes (Hubert et al., 2005; Zanata and Vari, 2005) or Alestopetersius (Zanata and Vari, 2005), and a more encompassing study involving all nominal species within both genera must precede any change in the generic-level classification of these taxa. Likewise, although our results rejected the monophyly of the genus Bathyaethiops based on the phylogenetic position of Brachypetersius altus, proposing taxonomic changes at this point is premature given the restricted sampling of Bathyaethiops and Brachypetersius species in our study. In contrast to previous hypotheses (Hubert et al., 2005; Zanata and Vari, 2005), and despite the claim of Stiassny and Mamonekene (2007: p 20) that ‘‘the genus Micralestes lacks a rigorous phylogenetic diagnosis and as currently conceived encompasses a wide range of external morphological diversity and considerable anatomical variability’’, our results strongly support the monophyly of this genus. Nevertheless, our sampling of Micralestes species includes only about a third of those currently recognized, mostly from the lower Congo River. Conversely, the monophyly of the genera Bryconaethiops, Hydrocynus, Nannopetersius, and Alestes appears to be well established based on the results of this and previous studies (Zanata and Vari, 2005; Calcagnotto et al., 2005). J. Arroyave, M.L.J. Stiassny / Molecular Phylogenetics and Evolution 60 (2011) 385–397 395 4.3. The phylogenetic position of the genus Chalceus 4.5. Characiform biogeography The novel hypothesis of a sister-group relationship between the Neotropical genus Chalceus and the family Alestidae served as the basis for expanding the limits of the family (Zanata and Vari, 2005). However, according to the phylogenetic hypothesis presented herein, the genus Chalceus is more closely related to other Neotropical characins than to the members of the family Alestidae s.s. This finding is in agreement with previous hypotheses of higher-level characiform relationships (Lucena, 1993; Ortí and Meyer, 1997; Calcagnotto et al., 2005). Interestingly, if Chalceus is in fact a member of a strictly Neotropical clade, the remarkable similarities between this taxon and the genus Arnoldichthys (Zanata and Toledo-Piza, 2004) might be an indication that the latter is indeed more closely related to Neotropical characins than to alestids, as suggested by our data. While the primary goal of this study was not to resolve the pattern and timing of characiform diversification, by including representatives of the most basal lineages (i.e., citharinoids) and several other families across the diversity of the order, the inferred chronogram provides new insights into the temporal context of characiform evolution and its biogeographic implications. The prevailing hypotheses explaining characiform biogeographic patterns (Fink and Fink, 1981; Lundberg, 1993; Buckup 1998; Malabarba and Malabarba, 2010) suggest that the disjunct distribution of the order be attributed to the African/South American drift-vicariance event and therefore the origin of characins necessarily precedes the mid-Cretaceous fragmentation of Gondwana. In contrast, our results indicate that the divergence between citharinoids and the clade that includes the remaining characiform taxa occurred during the Late Cretaceous (87 Mya; 95% HPD interval = 119–59). Thereby suggesting that the order Characiformes most likely originated after Africa and South America had separated [c. 110 Mya] (McLoughlin, 2001). Considering that a barrier to intercontinental migration of freshwater fishes might have established much earlier than 115 Mya (Briggs, 2005), vicariance hypotheses seem even less likely. Our results are consistent with the existing fossil record, since the earliest characiform fossils known to date come from the Cenomanian (c. 95 Myr) of Sudan and Morocco (Werner, 1994; Dutheil, 1999) and otophysan fossils only extend back to the Albian stage of the mid-Cretaceous [c. 110 Mya] (Gayet and Meunier, 2003). Thus, based on current data, we suggest that diversification of the order might actually have commenced in Africa, shortly after the Gondwanan split. Under this scenario, the opening of the Atlantic Ocean does not provide an adequate explanation for the modern distribution of characiform fishes and therefore biogeographic hypotheses must recourse to dispersalist arguments. Hypotheses involving marine dispersal of early characiform lineages have been proposed (Gayet, 1982; Filleul and Maisey, 2004; Otero et al., 2008), and although contentious in most cases (Malabarba and Malabarba, 2010), if a vicariance model does not readily explain current distributional patterns, dispersal scenarios should not be discounted simply because recent members of the order are intolerant of saltwater (Calcagnotto et al., 2005, p. 147). In light of our limited sampling of Neotropical taxa, and the recurrent recognition of multiple African/South American sister-group relationships within the order (Ortí and Meyer, 1997; Buckup 1998; Calcagnotto et al., 2005), hypothesizing possible scenarios involving marine dispersal is premature. Even if the general conclusions of this study hold, it would require a comprehensive and robust phylogeny of the order – ideally with estimates divergence times – to properly explain the contemporary Gondwanan distribution of characiform fishes. We emphasize however, a few caveats associated with the temporal context of characiform diversification as inferred from our analyses. First, although we included representatives of all four African families, our sampling of Neotropical lineages is far from comprehensive. In addition, after careful consideration of the available characiform fossil record we selected only two fossil calibration points, confined within the family Alestidae. This likely has implications for the accuracy of the estimated ages at deeper nodes in the phylogeny of the order. It is clear that inclusion of multiple calibration points has a strong impact on the overall accuracy of divergence time estimates (Near and Sanderson, 2004; Fulton and Strobeck, 2010), and that this is especially true for relaxed-clock methods, where multiple calibrations can act as landmark points detecting rate variation at multiple levels throughout the phylogeny (Benton and Donoghue, 2007). Hopefully, future studies investigating the timing of characiform diversification will use much larger datasets (in terms of both taxon and character sampling) 4.4. Timescale of alestid diversification Although the use of DNA sequence data to estimate the timing of evolutionary events is increasingly popular, this is the first study applying molecular-dating techniques to the estimation of absolute dates of divergence for a group of characiform fishes. Previous studies exploring the timing of origin and diversification of the order Characiformes (e.g., Lundberg, 1993; Lundberg et al., 1998; Malabarba and Malabarba, 2010) and the family Alestidae (Zanata and Vari, 2005) did so based solely on palaeontological evidence. Despite not conforming to the general conclusions of these previous studies, our results are nonetheless consistent with the existing characiform fossil record (Malabarba and Malabarba, 2010). Our hypothesis of an Early Tertiary origin of the family Alestidae s.s. – certainly much younger than suggested by Zanata and Vari (2005) – implies that the alestid radiation long postdates the Mesozoic fragmentation of Gondwana. Our age estimates for the origin of genera and higher-level clades within Alestidae are likewise far more recent than previously proposed and lead to the unanticipated conclusion that the origin and diversification of alestids took place in African waters. Most contemporary river basins of Sub-Saharan Africa were formed only after the Late Cretaceous, when Central Africa emerged above sea level (Stankiewicz and de Wit, 2006). Interestingly, the timing of origin of the Alestidae s.s., as inferred from our analysis, broadly coincides with the development of the modern African river network and related ecosystems of the region. Similarly, while most diversification in the family took place throughout the mid-Tertiary (40–10 Mya), the highest rates of cladogenesis occurred during the Early Miocene (25–15 Mya), after the Congo was first captured by a western coastal river draining to the Atlantic following the uplift of the East African Highlands (Stankiewicz and de Wit, 2006). Although we have no evidence of a causal relationship between such geologic events and the diversification of alestids, temporal correspondences like these are noteworthy given the current distribution patterns of the family. The estimated ages of the nodes constrained by our fossil-based calibrations (nodes in Fig. 5) are older than the ages of the fossils themselves. Such a discrepancy was particularly manifest for the node representing the MRCA of Hydrocynus, for which the estimated mean age (21 My; 95% HPD interval = 30–13) is about four times older than the minimum age suggested by the fossil itself (5 My). This mismatch between constrained and estimated node ages, demonstrates that fixed-age fossil calibrations may indeed be biased and that divergence times based solely on palaeontological data are often likely to underestimate true ages. Hence, the importance of using more refined methods that rely on probabilistic priors to calibrate molecular clocks. 396 J. Arroyave, M.L.J. Stiassny / Molecular Phylogenetics and Evolution 60 (2011) 385–397 and multiple, well-characterized, fossil-based calibration points across the phylogeny of the Characiformes. Regardless of the accuracy of our estimates, the inclusion of temporal information is crucial to formulating and testing biogeographic hypotheses. Traditionally, vicariance biogeographers have explained the distribution of taxa based on phylogenetic patterns in conjunction with Earth history; specifically, by ascribing sister-group relationships to the emergence of a geographic barrier. Although vicariance hypotheses make sense in light of the congruence between phylogenetic and palaeogeographic patterns (Nelson and Platnick, 1981; Nelson and Rosen, 1981; Humphries and Parenti, 1999), they hold only if geological and cladogenetic events indeed match in time (Donoghue and Moore, 2003). The ultimate test for vicariance biogeographic hypotheses, as shown in this study, is provided by temporal information in the form of the absolute ages of clades. 5. Conclusions The molecular phylogeny presented herein did not corroborate the family Alestidae as monophyletic, with putative members scattered throughout the phylogeny of the order. Our results rejected the hypothesis that the genus Chalceus is the most basal alestid, for it was recovered well nested within a clade of Neotropical characins. Likewise, the genera Arnoldichthhys and Lepidarchus were found to be more closely related to members of strictly Neotropical lineages. Moreover, the resulting phylogeny revealed several instances of poly- and paraphyly among alestid genera, highlighting the necessity for future revisionary studies to better understand the alpha-taxonomy of the family. In accordance with one of the primary goals of phylogenetic systematics, which is to maintain a classification scheme were only monophyletic units are recognized, we propose a redefined Alestidae (i.e., Alestidae sensu stricto) that does not include the genera Arnoldichthhys, Lepidarchus, and Chalceus. The origin of the Alestidae sensu stricto dates back to the Early Tertiary (63–49 Mya), implying that the diversification of the family took place in African waters long after the Mesozoic fragmentation of Gondwana. Acknowledgments This research was supported by an AMNH Graduate Student Fellowship and a CUNY Science Fellowship to J.A., and by the NSF-funded Congo project (Grant DEB-0542540 to M.L.J.S.). We thank Bob Schelly (AMNH) for his invaluable assistance and helpful guidance in the lab. We also thank Scott Schaefer, Dawn Roje, and Nelson Salinas for helpful comments on the manuscript. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.ympev.2011.04.016. References Benton, M.J., Donoghue, P.C., 2007. Paleontological evidence to date the tree of life. Mol. Biol. Evolut. 24, 26–53. Bremer, K., 1988. The limits of amino acid sequence data in angiosperm phylogenetic reconstructions. Evolution 42, 795–803. Brewster, B., 1986. A review of the genus Hydrocynus Cuvier 1819 (Teleostei: Characiformes). Bull. Brit. Mus. (Nat. Hist.) Zool. 50, 163–206. Briggs, J.C., 2005. The biogeography of otophysan fishes (Ostariophysi: Otophysi): a new appraisal. J. Biogeogr. 32, 287–294. Buckup, P.A., 1998. Relationships of the Characidiinae and phylogeny of characiform fishes (Teleostei: Characiformes). In: Malabarba, L.R., Reis, R.E., Vari, R.P., Lucena, Z.M.S., Lucena, C.A.S. (Eds.), Phylogeny and Classification of Neotropical Fishes. EDIPUCRS, Porto Alegre, Brazil, pp. 193–234. Calcagnotto, D., Schaefer, S.A., DeSalle, R., 2005. Relationships among characiform fishes inferred from analysis of nuclear and mitochondrial gene sequences. Mol. Phyl. Evolut. 36, 135–153. Chen, W., Miya, M., Saitoh, K., Mayden, R.L., 2008. Phylogenetic utility of two existing and four novel nuclear gene loci in reconstructing Tree of Life of rayfinned fishes: the order Cypriniformes (Ostariophysi) as a case study. Gene 423, 125–134. De La Peña Zarzuelo, A., 1996. Characid teeth from the Lower Eocene of the Ager Basin (Lérida, Spain): paleobiogeographical comments. Copeia 3, 746–750. de Pinna, M.C.C., 1991. Concepts and tests of homology in the cladistic paradigm. Cladistics 7, 367–394. DeVaney, S.C., 2008. Interrelationships of Fishes of the Order Stomiiformes. Unpublished Ph.D. Dissertation. University of Kansas, USA. 240 pp. Donoghue, M.J., Moore, B.R., 2003. Toward an integrative historical biogeography. Integr. Comp. Biol. 43, 261–270. Drummond, A.J., Rambaut, A., 2007. BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evolut. Biol. 7, 214. Drummond, A.J., Ho, S.Y.W., Phillips, M.J., Rambaut, A., 2006. Relaxed phylogenetics and dating with confidence. PLoS Biol. 4 (5), e88. Drummond, A.J., Ho, S.Y.W., Rawlence, N., Rambaut, A., 2007. A Rough Guide to BEAST 1.4. University of Auckland, New Zealand. Dutheil, D.B., 1999. An overview of the freshwater fish fauna from the Kem Kem beds (Late Cretaceous: Cenomanian) of southeastern Morocco. In: Arratia, G., Schultze, H.P. (Eds.), Mesozoic Fishes 2—Systematics and Fossil Record. Verlag Dr. Friedrich Pfeil, München, Germany, pp. 553–563. Edgar, R.C., 2004. MUSCLE: a multiple sequence alignment method with reduced time and space complexity. BMC Bioinform. 5, 113. Edwards, A.W.F., 1970. Estimation of the branch points of a branching diffusion process. J.R. Stat. Soc. Series B 32 (2), 155–174. Eschmeyer, W.N. (Ed.), 2010. Catalog of Fishes Electronic Version (12 July 2010). <http://research.calacademy.org/ichthyology/catalog/fishcatmain.asp>. Eschmeyer, W.N., Fong, J.D. 2010. Species of Fishes by family/subfamily. On-line Version Dated 07-12-2010. <http://research.calacademy.org/research/ ichthyology/catalog/SpeciesByFamily.asp>. Farris, J.S., 1970. Methods for computing Wagner Trees. Syst. Zool. 19, 83–92. Farris, J.S., 1989. The retention index and the rescaled consistency index. Cladistics 5, 417–419. Felsenstein, J., 1981. Evolutionary trees from DNA sequences: a maximum likelihood approach. J. Mol. Evolut. 17, 368–376. Felsenstein, J., 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39, 783–791. Filleul, A., Maisey, J.G., 2004. Redescription of Santanichthys diasii (Otophysi, Characiformes) from the Albian of the Santana Formation and comments on its implications for otophysan relationships. Am. Mus. Novit. 3455, 1–21. Fink, S.V., Fink, W.L., 1981. Interrelationships of the ostariophysan fishes (Teleostei). Zool. J. Linnean Soc. 72, 297–353. Fitch, W.M., 1971. Toward defining the course of evolution: minimum change for a specific tree topology. Syst. Zool. 20, 406–416. Folmer, O., Black, M., Hoeh, W., Lutz, R., Vrijenhoek, R., 1994. DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Mol. Marine Biol. Biotechnol. 3, 294–299. Fulton, T.L., Strobeck, C., 2010. Multiple fossil calibrations, nuclear loci and mitochondrial genomes provide new insight into biogeography and divergence timing for true seals (Phocidae, Pinnipedia). J. Biogeogr. 37, 814– 829. Gayet, M., 1982. Considération sur la phylogénie et la paléobiogéographie des Ostariophysaires. Geobios (numéro special) 6, 39–52. Gayet, M., Meunier, F.J., 2003. Paleontology and palaeobiogeography of catfishes. In: Arratia, G., Kapoor, B.G., Chardon, M., Diogo, R. (Eds.), Catfishes, vol. 2. Science Publishers, Enfield, NH, pp. 491–522. Géry, J., 1968. Ladigesia roloffi, a new genus and species of African characoid fishes. Trop. Fish Hobbyist 16 (8), 78–87. Géry, J., 1977. Characoids of the World. Tropical Fish Hobbyist Publications, Neptune City, NJ, 672 pp. Géry, J., 1995. Description of new or poorly known Alestinae (Teleostei: Characiformes: Alestidae) from Africa, with a note on the generic concept in the Alestinae. Aquat. J. Ichthyol. Aquat. Biol. 1 (4), 37–64. Goloboff, P.A., 1999. Analyzing large data sets in reasonable times: solutions for composite optima. Cladistics 15, 415–428. Goloboff, P.A., Pol, D., 2005. Parsimony and Bayesian phylogenetics. In: Victor, A.A. (Ed.), Parsimony, Phylogeny and Genomics. Oxford University Press, Oxford, pp. 148–159. Goloboff, P.A., Farris, J.S., Nixon, K., 2003. TNT: Tree Analysis using New Technology. Version 1.0, version Beta test v. 0.2. Program and Documentation. <http:// www.zmuc.dk/public/phylogeny/TNT/>. Goloboff, P.A., Farris, J.S., Nixon, K., 2008. TNT, a free program for phylogenetic analysis. Cladistics 24, 774–786. Goodman, M., Olson, C.B., Beeber, J.E., Czelusniak, J., 1982. New perspectives in the molecular biological analysis of mammalian phylogeny. Acta Zool. Fennica 169, 19–35. Greenwood, P.H., Rosen, D.E., Weitzman, S.H., Myers, G.S., 1966. Phyletic studies of teleostean fishes with a provisional classification of living forms. Bull. Am. Mus. Nat. Hist. 131, 339–354. Hedges, S.B., Kumar, S., 2004. Precision of molecular time estimates. Trends Genet. 202, 242–247. J. Arroyave, M.L.J. Stiassny / Molecular Phylogenetics and Evolution 60 (2011) 385–397 Hillis, D.M., 1998. Taxonomic sampling, phylogenetic accuracy, and investigator bias. Syst. Biol. 47, 3–8. Ho, S.Y.M., 2007. Calibrating molecular estimates of substitution rates and divergence times in birds. J. Avian Biol. 38, 409–414. Hubert, N., Bonillo, C., Paugy, D., 2005. Early divergence among the Alestidae (Teleostei, Ostariophyses, Characiformes): mitochondrial evidences and congruence with morphological data. Comptes Rendus Biol. 328, 477–491. Humphries, C.J., Parenti, L.R., 1999. Cladistic Biogeography: Interpreting Patterns of Plant and Animal Distributions, second ed. Oxford University Press, 187 pp. Irwin, D.M., Kocher, T.D., Wilson, A.C., 1991. Evolution of the cytochrome b gene of mammals. J. Mol. Evol. 32, 128–144. Javonillo, R., Malabarba, L.R., Weitzman, S.H., Burns, J.R., 2010. Relationships among major lineages of characid fishes (Teleostei: Ostariophysi: Characiformes), based on molecular sequence data. Mol. Phyl. Evolut. 54, 498–511. Kawahara, R., Miya, M., Mabuchi, K., Near, T.J., Nishida, M., 2009. Stickleback phylogenies resolved: evidence from mitochondrial genomes and 11 nuclear genes. Mol. Phyl. Evolut. 50, 401–404. Kluge, A.G., Farris, J.S., 1969. Quantitative phyletics and the evolution of anurans. Syst. Zool. 18, 1–32. Kocher, T.D., Thomas, W.K., Meyer, A., Edwards, S.V., Pääbo, S., Villablanca, F.X., Wilson, A.C., 1989. Dynamics of mitochondrial DNA evolution in animals: amplification and sequencing with conserved primers. Proc. Natl. Acad. Sci. USA. 86, 6196–6200. Li, C., Ortí, G., Zhang, G., Lu, G., 2007. A practical approach to phylogenomics: the phylogeny of ray-finned fish (Actinopterygii) as a case study. BMC Evolut. Biol. 7, 44. Li, C., Ortí, G., Zhang, G., 2010. The phylogenetic placement of sinipercid fishes (‘‘Perciformes’’) revealed by 11 nuclear loci. Mol. Phyl. Evolut. 56, 1096–1104. Lucena, C.A.S., 1993. Estudo filogenético da família Characidae com uma discussão dos grupos naturais propostos (Teleostei, Ostariophysi, Characiformes). Unpublished Ph.D. Dissertation, Universidade de São Paulo, Brazil. Lundberg, J.G., 1993. African-South American freshwater fish clades and continental drift: problems with a paradigm. In: Goldblatt, P. (Ed.), The Biotic Relationships between Africa and South America. Yale University Press, New Haven, pp. 156– 199. Lundberg, J.G., Marshall, L.G., Guerrero, J., Horton, B., Malabarba, M.C., Wesselingh, F., 1998. The stage for Neotropical fish diversification: a history of tropical South American Rivers. In: Malabarba, L.R., Reis, R.E., Vari, R.P., Lucena, Z.M.S., Lucena, C.A.S. (Eds.), Phylogeny and Classification of Neotropical Fishes. EDIPUCRS, Porto Alegre, Brazil, pp. 13–48. Malabarba, M.C., Malabarba, L.R., 2010. Biogeography of Characiformes: an evaluation of the available information of fossil and extant taxa. In: Nelson, J.S., Schultze, H.P., Wilson, M.V.H. (Eds.), Origin and Phylogenetic Interrelationships of Teleosts. Verlag Dr. Friedrich Pfeil, München, Germany, pp. 317–336. Marshall, C.R., 1990. The fossil record and estimating divergence times between lineages: maximum divergence times and the importance of reliable phylogenies. J. Mol. Evolut. 30, 400–408. McLoughlin, S., 2001. The breakup history of Gondwana and its impact on preCenozoic floristic provincialism. Aust. J. Bot. 49, 271–300. Meredith, R.W., Pires, M.N., Reznick, D.N., Springer, M.S., 2010. Molecular phylogenetic relationships and the evolution of the placenta in Poecilia (Micropoecilia) (Poeciliidae: Cyprinodontiformes). Mol. Phyl. Evolut. 55, 631– 639. Miller, M.A., Holder, M.T., Vos, R., Midford, P.E., Liebowitz, T., Chan, L., Hoover, P., Warnow, T., 2009. The CIPRES Portals. <www.webcitation.org/5imQlJeQa>. Murray, A.M., Stewart, K.M., 2002. Phylogenetic relationships of the African genera Alestes and Brycinus (Teleostei, Characiformes, Alestidae). Can. J. Zool. 80, 1887– 1899. Near, T.J., Sanderson, M.J., 2004. Assessing the quality of molecular divergence time estimates by fossil calibrations and fossil-based model selection. Philos. Trans. Roy. Soc. B. 359, 1477–1483. Nelson, G.J., Platnick, N.I., 1981. Systematics and biogeography: cladistics and Vicariance. Columbia University Press, New York, USA. Nelson, G., Rosen, D.E., 1981. Vicariance Biogeography: A Critique. Columbia University Press, New York, USA. Nixon, K.C., 1999. The Parsimony Ratchet, a new method for rapid parsimony analysis. Cladistics 15, 407–414. Ortí, G., Meyer, A., 1997. The radiation of characiform fishes and the limits of resolution of mitochondrial ribosomal DNA sequences. Syst. Biol. 46, 75–100. Otero, O., Valentin, X., Garcia, G., 2008. Cretaceous characiform fishes (Teleostei: Ostariophysi) from Northern Tethys: description of new material from the Maastrichtian of Provence (Southern France) and palaeobiogeographical implications. Geol. Soc. Lond. Spec. Publ. 295, 155–164. Paugy, D., 1986. Révision systématique des Alestes et Brycinus Africains (Pisces, Characidae). Éditions de L’ORSTOM, Collection Études et Théses, 295 pp. Pickett, K.M., Randle, C.P., 2005. Strange Bayes indeed: uniform topological priors imply non-uniform clade priors. Mol. Phyl. Evolut. 34 (1), 203–211. Posada, D., 2008. JModelTest: phylogenetic model averaging. Mol. Biol. Evolut. 25, 1253–1256. Rambaut, A., Bromham, L., 1998. Estimating divergence dates from molecular sequences. Mol. Biol. Evolut. 15, 442–448. Rambaut, A., Drummond, A.J., 2009. Tracer v1.5. <http://beast.bio.ed.ac.uk/Tracer> (accessed May 2010). 397 Rannala, B., Yang, Z., 1996. Probability distribution of molecular evolutionary trees: a new method of phylogenetic inference. J. Mol. Evolut. 43, 304–311. Roberts, T.R., 1969. Osteology and relationships of characoid fishes, particularly the genera Hepsetus, Salminus, Hoplias, Ctenolucius, and Acestrorhynchus. Proc. Cal. Acad. Sci. 36 (15), 391–500. Roberts, T.R., 1975. Geographical distribution of African freshwater fishes. Zool. J. Linnean Soc. 57, 249–319. Rodriguez, F., Oliver, J.L., Marin, A., Medina, J.R., 1990. The general stochastic model of nucleotide substitution. J. Theor. Biol. 142, 485–501. Rutschmann, F., 2006. Molecular dating of phylogenetic trees: a brief review of current methods that estimate divergence times. Divers. Distribut. 12, 35–48. Sanderson, M.J., 1997. A nonparametric approach to estimating divergence times in the absence of rate constancy. Mol. Biol. Evolut. 14, 1218–1231. Schaefer, S.A., 2007. Petersiini. In: Stiassny, M.L.J., Teugels, G.G., Hopkins, C.D. (Eds.), The Fresh and Brackish Water Fishes of Lower Guinea, West-Central Africa, vol. 1. MNHN, Paris, MRAC Tervuren, pp. 380–409, IRD Éditions Paris. Smith, A.B., Peterson, K.J., 2002. Dating the time of origin of major clades: molecular clocks and the fossil record. Annu. Rev. Earth Planet. Sci. 30, 65–88. Smith, S.A., Beaulieu, M., Donoghue, M.J., 2010. An uncorrelated relaxed-clock analysis suggests an earlier origin for flowering plants. Proc. Natl. Acad. Sci. USA 107, 5897–5902. Stamatakis, A., 2006. RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 22 (21), 2688–2690. Stamatakis, A., Hoover, P., Rougemont, J., 2008. A fast bootstrappping algorithm for the RAxML web-servers. Syst. Biol. 57, 758–771. Stankiewicz, J., de Wit, M.J., 2006. A proposed drainage evolution model for Central Africa–Did the Congo flow east? J. Afri. Earth Sci. 44, 75–84. Stewart, K.M., 1994. A Late Miocene fish fauna from Lothagam, Kenya. J. Vert. Paleontol. 14 (4), 592–594. Stewart, K.M., 2003. Fossil fish remains from Mio-Pliocene deposits at Lothagam, Kenya. In: Leakey, M.G., Harris, J.M. (Eds.), Lothagam: the Dawn of Humanity in Eastern Africa. Columbia University Press, New York, pp. 75–111. Stiassny, M.L.J., Mamonekene, V., 2007. Micralestes (Teleostei, Alestidae) of the Lower Congo River, with a description of a new species endemic to the Lower Congo River rapids in the Democratic Republic of Congo. Zootaxa 1614, 17–29. Stiassny, M.L.J., Schaefer, S.A., 2005. Rhabdalestes aeratis, new species (Characiformes: Alestidae): first occurrence of the genus from the Middle Congo River basin. Ichthyol. Explor. Freshwaters 16 (3), 271–278. Swofford, D.L., 2003. PAUP⁄. Phylogenetic Analysis Using Parsimony (⁄and Other Methods). Version 4. Sinauer Associates, Sunderland, Massachusetts, USA. Swofford, D.L., Olsen, G.J., 1990. Phylogeny reconstruction. In: Willis, D.M., Moritz, C. (Eds.), Molecular Systematics. Sinauer Associates, Sunderland, Massachusetts, USA, pp. 411–501. Tamura, K., Dudley, J., Nei, M., Kumar, S., 2007. MEGA4: molecular evolutionary genetics analysis (MEGA) software version 4.0.. Mol. Biol. Evolut. 24, 1596– 1599. Thorne, J.L., Kishino, H., Painter, I.S., 1998. Estimating the rate of evolution of the rate of molecular evolution. Mol. Biol. Evolut. 15, 1647–1657. Van Neer, W., 1992. New Late Tertiary fish fossils from the Sinda Region, eastern Zaire. African Study Monogr. 17, 27–47, Supplementary Issue. Vari, R.P., 1979. Anatomy, relationships and classification of the families Citharinidae and Distichodontidae (Pisces: Caracoidei). Bull. Brit. Mus. Nat. Hist. Zool. 36 (5), 261–344. Vari, R.P., 1995. The Neotropical fish family Ctenoluciidae (Teleostei: Ostariophysi: Characiformes): supra and intrafamilial phylogenetic relationships, with a revisionary study. Smithson. Contrib. Zool. 564, 1–97. Vari, R.P., 1998. Higher level phylogenetic concepts within Characiformes (Ostariophysii), a historical review. In: Malabarba, L.R., Reis, R.E., Vari, R.P., Lucena, Z.M.S., Lucena, C.A.S. (Eds.), Phylogeny and Classification of Neotropical Fishes. EDIPUCRS, Porto Alegre, Brazil, pp. 111–122. Varón, A., Vinh, L.S., Bomash, I., Wheeler, W.C. 2008. POY 4.1. American Museum of Natural History. <http://research.amnh.org/scicomp/projects/poy.php>. Velasco, J., 2008. The prior probabilities of phylogenetic trees. Biol. Philos. 23, 455– 473. Werner, C., 1994. Die kontinentale Wirbeltierfauna aus der unteren Oberkreide des Sudan (Wadi Milk Formation). Berliner Geowissenschaftliche Abhandlungen, Reihe E, Palaeobiologie 13, 221–249. Wheeler, W.C., 1996. Optimization alignment: the end of multiple sequence alignment in phylogenetics? Cladistics 12, 1–9. Wheeler, W.C., 2001. Homology and the optimization of DNA sequence data. Cladistics 17, S3–S11. Wheeler, W.C., Aagesen L., Arango, C.P., Faivovich, J., Grant, T., D’Haese, C., Janies, D., Smith, W.L., Varón, A., Giribet, G. 2006. Dynamic Homology and Phylogenetic Systematics: A Unified Approach Using POY. American Museum of Natural History, New York. Yang, Z., 2006. Computational Molecular Evolution. Oxford University Press, Oxford, USA. Zanata, A.M., Toledo-Piza, M., 2004. Taxonomic revision of the South American fish genus Chalceus Cuvier (Teleostei: Ostariophysi: Characiformes) with the description of three new species. Zool. J. Linnean Soc. 140, 103–135. Zanata, A.M., Vari, R.P., 2005. The family Alestidae (Ostariophysi, Characiformes): a phylogenetic analysis of a trans-Atlantic clade. Zool. J. Linnean Soc. 145, 1–144.