TWENTY-FIVE Molecular signatures of Neogene biogeographical events in the Amazon fi sh fauna

Nathan R. Lovejoy1, Stuart C. Willis2 and James S. Albert3 1University of Toronto Scarborough, Toronto, Ontario, Canada 2University of Nebraska-Lincoln, Lincoln, Nebraska, USA 3University of Louisiana Lafayette, Lafayette, Louisiana, USA

Abstract

Molecular genetics can contribute to biogeography by clarifying species limits (and thereby distributions) and phylogenetic relationships. Molecular data also offer the tantalizing prospect of dating the ages of lineages by adding a timescale to phylogenetic reconstructions. Thus, molecular analyses signifi cantly enhance our ability to test models and hypotheses that address the complex relationship between biological evolution and palaeo- geographic history in the Amazon drainage basin. Here, we review the use of molecular data for understand- ing Amazonian fi sh biogeography, with particular emphasis on Neogene palaeogeographic events. We provide an overview of previous work in the fi eld, and briefl y mention the possibilities and pitfalls of molecular bio- geographical approaches. Challenges for molecular investigations of the Amazon fi sh fauna include taxon selection and sampling, molecular clock assumptions and calibration issues, and identifying clearly testable hypotheses. We provide recommendations for future investigations and methodological improvements.

Introduction understood in the context of the larger-scale climatic and geo- logical histories of South and Middle American river basins. With more than 3000 species, Amazonian fi shes constitute the most species-rich aquatic continental fauna on Earth (Lundberg et al. 2000; Reis et al. 2003). Although the evolutionary and Amazonia as the core of Neotropical fresh waters ecological forces underlying the formation of this spectacular diversity are incompletely understood, advances are continually In terms of taxonomic composition, the Amazonian ichthyo- being made in the fi elds of historical biogeography and phylo- fauna comprises the core of the Neotropical freshwater region, geography. In this chapter we consider the role of molecular data a vast (~17 million km2) and ecologically heterogeneous assem- sets for biogeography, at both the species and population levels. blage of geographical areas extending from the Isthmus of We briefl y review major patterns in the geographic distribu- Tehuantepec in southern Mexico (16°N) to the La Plata drainage tions of Amazonian freshwater fi shes, summarize recent work basin in northern Argentina (34°S), and including all of Central using new molecular data sets and new methods for the analysis America and northern South America east of the Andes. The of biogeographical data. A central theme of this chapter is that Neotropical ichthyofaunal region is therefore restricted to the understanding the diversifi cation of Amazonian freshwater fi shes humid tropical portions of the Neotropical realm as originally requires detailed species-level or even population-level know ledge defi ned (Wallace 1876), excluding the cooler and/or more arid ofUNCORRECTED geographical distributions and phylogenetic interrelation- areas of the Southern Cone,PROOF the Andean Altiplano and the Pacifi c ships. In turn, this biological information can only be completely slopes of Chile and Peru. The Neotropical ichthyofauna is easy to recognize. Fishes from throughout this very wide area belong to relatively few higher-level taxa, each of which is characterized by a distinct suite of morpho- Amazonia, Landscape and Species Evolution, 1st edition. Edited by C. Hoorn and F.P. Wesselingh. © 2010 Blackwell Publishing, logical traits. According to current classifi cation, the Neotropical ISBN 978-1-4051-8113-6 ichthyofauna includes 17 orders of fi shes; this compares with

HHoorn_ch25_FPP.inddoorn_ch25_FPP.indd 440505 77/27/2009/27/2009 3:48:083:48:08 PMPM 406 N.R. Lovejoy et al.

26 orders in the Mississippi River basin and adjacent drainages. freshwater fl uvial system on the planet, discharging about 16% However, despite its relatively poor diversity at higher taxonomic of the world’s fl owing freshwater into the Atlantic (Goulding levels, the Neotropical ichthyofauna is extremely diverse at lower et al. 2003). taxonomic levels. Estimates of total Neotropical fi sh richness range The largest geological feature of the Neotropics is the South between 5000 and 8000 species, compared to about 1000 species American Platform, an ancient (Precambrian-Paleozoic >250 Ma) in North America (Lundberg et al. 2000). This disproportionate block of continental crust underlying all of Amazonia and adja- distribution of taxonomic categories, with many lower taxa and cent regions (De Almeida et al. 2000; see also Chapters 2 &3). few higher taxa, has resulted from a lengthy history of geographi- The role of the South American Platform on the diversifi cation cal isolation and in situ diversifi cation (Lundberg 1998). of Amazonian fi shes cannot be overemphasized. This platform As in most of the Earth’s fresh waters, the Neotropical ich- lies very low in the Earth’s mantle, with more than half its total thyofauna is dominated by ostariophysan fi shes, which include area less than 100 m elevation (P. Petry & J. Albert, unpublished the tetras, catfi shes and electric knifefi shes. Ten ostariophysan observation). As a result, several areas in South America have clades contain approximately 75% of all Neotropical species; the been exposed to marine transgressions and regressions over the most diverse by far being the Loricarioidea (armoured catfi shes course of the past c. 110 million years. Documenting the exact and relatives, with around 1500 species currently described), extents of these marine transgressions is an active area of research and the Characoidea (tetras and relatives, with about 1750 spe- (e.g. Hoorn 1993; Hoorn et al. 1995; Monsch 1998; Roddaz et al. cies currently described). Highly diverse non-ostariophysan 2005; Rebata et al. 2006; Westaway 2006; Hovikoski et al. 2007a, clades include the Neotropical (500+ species) and 2007b; see also Chapter 8), yet regardless of the exact positions of the Cyprinodontiformes (killifi shes, 600+ species). The great palaeocoastlines, it is clear that episodes of marine transgression majority (>97%) of Neotropical freshwater fi sh species are mem- drastically affected the extent and distribution of habitat available bers of these primary and secondary freshwater fi sh taxa, which to obligate freshwater species. have little or no tolerance for salt water and very poor capaci- Several large-scale (~103–104 km2) geological features defi ne ties for dispersal over marine barriers. These fi shes trace their the contours of the major Amazonian watershed (see Chapter 4). origins to before the Early Cretaceous separation of Africa and The two principal highland structures of the South American South America (c. 110 Ma; Myers 1949, 1966). These clades can Platform are the Guiana in the northeast and Brazilian therefore be considered the ecosystem incumbents (sensu Wilson Shield in the southeast, which together comprise about half the 2003). Including a handful of recent anthropogenic transplants, area of Amazonia. These shields are of Proterozoic (>600 Ma) there are few examples of fi shes from other continents that have origins and vastly pre-date the radiations of fi shes in the naturally established themselves in cis-Andean South American Late Cretaceous and Paleogene (120–23 Ma). The shields attain waters (Hrbek et al. 2007; Pérez et al. 2007). Indeed, the only only modest altitudes (mostly up to c. 1000 m), and have long fi sh taxa that appear to have successfully joined Amazon com- since lost their easily eroded sediments; they are drained by rivers munities during the whole of the Cenozoic are certain groups of of low-sediment clear waters (e.g. Xingu, Tocantins, Trombetas, marine origin (Lovejoy et al. 1998, 2006; Boeger & Kritsky 2002). Rio Branco). Along the entire western margin of South America, Most of these marine- derived clades are represented by relatively the Andes constitute another substantial infl uence on the few species, with only the potamotrygonid stingrays attaining a distribution of Amazonian freshwater fi sh taxa. Rising to almost moderate level of diversity (>25 spp.). 7000 m, the Andes are of Late Cretaceous to Cenozoic age, and By the standards of biogeography in a global context, the therefore much younger than the shields (see Chapter 4) The margins of the Neotropical region are remarkably sharp (Miller sediment-rich waters draining the Andes are referred to as white 1966; Myers 1966; Lomolino et al. 2006). Only a few Neotropical waters (Marañon, Napo, Madeira, Meta, etc.) lineages, including several characins, catfi shes and cichlids, have Whereas the current watersheds of the Amazon and Orinoco dispersed as far north as central Mexico. There are only about a drainage basins are of Neogene age, the phenotypes, behaviours dozen or so Neotropical freshwater fi shes known from the north- and ecologies of many Amazonian fi shes appear to trace their ori- ern pampas of Argentina (Casciotta et al. 1989; Menni & Gomez gins to the Paleogene or even Late Cretaceous (Lundberg 1998; 1995; López et al. 2002), and the Patagonian fauna is quite dis- see Chapter 17). The earliest palaeo-ichthyofaunas with a modern tinct. Similarly, very few fi sh taxa from other regions of the world Amazonian taxonomic composition are from the Paleocene Santa are present in the Neotropics – a handful of North American Lucia and Maiz Gordo Formations of eastern Bolivia and northern derivatives, an African catfi sh clade (Lundberg et al. 2007), and Argentina, respectively, and the Eocene Lumbrera Formation of a few marine-derived taxa occur in southern Mexico and nuclear northern Argentina (Malabarba et al. 2006). These faunas are Middle America (Guatemala and Honduras). from some of the earliest known sediments of a large river basin that drained northwards from the area of the modern Pantanal to the area of the modern western Amazon, and which is presumed GeologicalUNCORRECTED and palaeogeographic context to have been an area of lowland PROOF freshwater habitats following a major marine regression c. 59–55 Ma (Lundberg et al. 1998; see also The modern Amazon drainage basin drains about 7 × 106 km2 of Chapter 26). From that time up until the Late Miocene assembly northern South America east of the Andes and between the Guiana of the modern Amazonian watershed c. 11–12 Ma, this north- and Brazilian Shields. This region is largely covered by moist fl owing river basin must have been the epicentre of freshwater fi sh tropical lowland forests, with some areas of seasonally fl ooded diversifi cations. The onset of the transcontinental wetlands and savannas. Amazonia is the greatest interconnected was preceded by a mega-wetland that had two successive stages

HHoorn_ch25_FPP.inddoorn_ch25_FPP.indd 440606 77/27/2009/27/2009 3:48:083:48:08 PMPM Molecular signatures in Amazon fi sh fauna 407

(see Chapter 8). The Pebas mega-wetland (24–11 Ma) was a series challenging task. In the tropics, taxonomic work on fi shes is com- of lakes, swamps, marginal marine embayments and lowland fl u- plicated both by the vast diversity of the fauna and by the diffi cul- vial habitats, while the subsequent Acre mega-wetland (11–7 Ma) ties of achieving geographically expansive sampling across terrain was a fl uvio-tidal system (Wesselingh et al., 2002; see also that is often remote and inaccessible. The sheer size of Neotropical Chapter 8). Presumably, these dramatic ecosystem changes must river systems hampers collecting projects that seek adequately to have had correspondingly signifi cant effects on fi sh diversity and sample and assess species ranges. Taxonomic progress to date, biogeography. Eastern Amazonia, including the Amazonas and based mostly on morphology, has been a hard-won multinational Marajó sedimentary basins, was formed during the Paleozoic effort (Reis et al. 2003). and went through renewed episiodes of subsidence during the Molecular data offer welcome assistance for species identifi ca- Late Cretaceous (c. 100–65 Ma) and Cenozoic (65–0.5 Ma) (Costa tion and taxonomic analysis. Genetic approaches are expected et al. 2001; see also Chapter 3). The modern Amazon drainage to be particularly valuable when morphology-based taxon- system was assembled through a series of Andean tectonic shifts omy is obscured by phenotypic conservatism, as in the case of during the Miocene (Hoorn 1993; Hoorn et al. 1995; see also potentially cryptic species (e.g. Lovejoy & de Araújo 2000), or Chapter 4). These events isolated the trans-Andean Magdalena by extreme phenotypic variability or plasticity (e.g. Albert et al. and Maracaibo drainage basins and resulted in distinct Amazon 1999; Fernandes et al. 2002; Albert & Crampton 2003). In these and Orinoco drainages with essentially modern drainage confi gu- situations, molecular data can provide valuable insights that can rations (Cooper et al. 1995; Hoorn et al. 1995). The eastern stretch assist and direct morphological efforts. In several recent examples, of the Amazon River captured the Western Amazon to form the molecular data were able to help establish species ranges, distin- modern Amazon watershed at about 11 Ma (Dobson et al. 2001; guish cryptic species, and elucidate confusing cases of intraspe- Mapes et al. 2006; Figueiredo et al. unpublished). cifi c polyphenism in cichlids (Pérez et al. 2007; Willis et al. 2007), characins (Sivasundar et al. 2001; Dergam et al. 2002; Hubert et al. 2006), potamotrygonid freshwater stingrays (Toffoli et al. 2008) Molecular data and biogeography: considerations and and freshwater needlefi shes (Lovejoy & de Araújo 2000). opportunities While molecular data represent a nearly unlimited source of information for species investigation, their use raises a number The biogeography of Amazonian fi shes has previously been of methodological and practical issues. One concern is that most addressed with morphology- and -based studies. These studies to date have essentially been based on a single molecu- usually take the form of cladistic analyses of individual clades, lar locus: mitochondrial DNA (mtDNA). Mitochondrial DNA where conclusions are drawn about the biogeographical implica- is readily amplifi ed from ethanol-preserved tissue, but because tions of reconstructed phylogenetic trees (e.g. Vari 1988; Albert the mitochondrial genome is non-recombining and maternally et al. 2005; Hulen et al. 2005). Other studies include syntheses inherited, it may not necessarily share the same genealogical his- of phylogenetic (e.g. Schaefer 1997) or taxonomic data (Hubert tory as loci from the nuclear genome. Indeed, mtDNA lineages & Renno 2006) that use historical biogeographical methods to have been observed to cross species boundaries in freshwater interpret patterns derived from multiple clades of fi shes. To date, fi shes as a result of hybridization (Bermingham & Avise 1986; these studies have made progress in discerning the effects of pal- Smith 1992; Bernatchez & Wilson 1998), and this phenomenon aeogeographic events, but often in areas that are peripheral (albeit has been observed in Neotropical fi shes (Willis et al. 2007). Until related) to Amazonia, such as eastern Brazil and northwestern the prevalence of such confounding factors is assessed, reliance South America (Vari & Malabarba 1998). Biogeographical stud- on a single genetic locus such as mtDNA is ill advised. This warn- ies of Amazon fi shes have resulted in patterns that are ‘often con- ing is relevant to the proposed ‘barcoding’ of the fauna of tropical tradictory or at best only partially congruent’, as a result of the regions (e.g. Neigel et al. 2007), an approach to species identifi ca- highly complex geological history of the region, and the great age tion and discovery based on a fragment of a single mitochondrial of many constituent clades (Vari & Malabarba 1998). gene. In general, more robust molecular approaches will require Given such diffi culties, could molecular data sets and approaches multiple genetic loci, and several authors have proposed methods enhance our ability to discern and interpret biogeographical pat- to combine morphological and molecular data to estimate species terns? In this section, we consider the role of molecular data in boundaries (Wiens & Servedio 2000; Wiens & Penkrot 2002; Sites the biogeography of South American fi shes. We summarize & Marshall 2003, 2004). approaches and perspectives that molecular data make possible, In the foreseeable future, it is not likely that suffi cient molecular and briefl y mention potential challenges. Our discussion centers data will be on hand to allow the accurate identifi cation of all spe- on four topics: (i) the use of molecular data for species identifi ca- cies of Neotropical fi shes. This is primarily because multi-locus tion; (ii) phylogeography; (iii) molecular , and (iv) samples from across the ranges of species, many of which inhabit molecular-based age estimation. multiple drainages, would be a requirement. This means that mor- UNCORRECTEDphology will continue toPROOF guide much species discovery and iden- tifi cation, including the ongoing work of evolutionary biologists, Molecular data and species identity ecologists, conservation biologists and others. However, molecu- lar investigations can play a special role in ‘ground-truthing’ mor- Species are fundamental units of evolution and ecology, and are phology-based taxonomy by allowing detailed genetic dissections the basic components of most biogeographical analyses. However, of model clades. Such studies will help us to answer questions such identifying independently evolving species lineages can be a as: How common are sympatric complexes of cryptic species?

HHoorn_ch25_FPP.inddoorn_ch25_FPP.indd 440707 77/27/2009/27/2009 3:48:083:48:08 PMPM 408 N.R. Lovejoy et al.

How often are populations of widespread species genetically dis- investigation. As in the case of other types of molecular studies, tinct (perhaps to the point of separate species status)? Since a clear the large geographical ranges of some species, the logistical diffi - understanding of species identity and distribution is the lynchpin culties associated with making collections, and the extremely high for accurate investigation, such genetic investigations will surely species richness of the fauna as a whole, all conspire to make phy- contribute to Amazonian biogeography. logeographic studies of Amazonian fi shes extremely challenging.

Phylogeography Molecular phylogeny

The fi eld of historical biogeography has traditionally focused on Historical biogeography relies on robust phylogenetic hypotheses. the effects of Earth history events on the distributions of clades Thus, the development of novel phylogenetic methods and data and regional biotas (e.g. Morrone & Crisci 1995). By contrast, the sets is central to the advance of biogeographical understanding. fi eld of phylogeography has emerged as a distinctive discipline The advantages of molecular phylogenetics that are of particular that focuses on the analysis of intraspecifi c (population-level) data importance for the study of Amazonian fi sh clades include: (Avise et al. 1987; Bermingham & Martin 1998). Here we main- tain the distinction between these terms to emphasize the different 1 The potential to resolve phylogenies for species-rich but methods and concepts employed in these two related but distinct osteologically conservative groups (e.g. some tetras and areas of investigation. While historical biogeography and phylo- cichlids); geography both investigate the role of Earth history in shaping 2 the ability to test morphological phylogenies with alternative diversity, phylogeography explicitly incorporates population-level data sources; processes, such as gene fl ow and range expansion. 3 the ability to infer branch lengths for phylogenetic age Because of its emphasis at the species and population level, estimation (see below); phylogeography should be useful for reconstructing Amazonian 4 the ability quickly and effi ciently to collect vast amounts of fi sh biogeography at relatively shallow time horizons, on the order data for large numbers of samples. of tens of thousands to perhaps a few million years. We expect the approach to record relatively recent population events and proc- Perhaps the biggest challenge of applying molecular approaches esses associated with climate changes and hydrological evolution to Amazonian fi shes is obtaining appropriate taxon sampling of the landscape (e.g. stream capture events). Phylogeography from such a diverse and widespread fauna. Thus, the vast diversity should tell us about recent gene fl ow along major river systems of Amazonian fi shes both hinders and beckons the application of (Strange & Burr 1997), and very recent alterations of the aquatic molecular methods. landscape, including shifts in drainages and expansions/contrac- tions of aquatic habitats (e.g. Burridge et al. 2007). By emphasizing the boundaries between populations and species, phylogeog- Molecular estimates of clade ages raphy should also illuminate the patterns and processes associ- ated with speciation. The pioneering work of Bermingham et al. A great potential advantage of incorporating molecular sequence (e.g. Bermingham & Martin 1998; Reeves & Bermingham 2006) data in biogeographical analysis is the estimation of clade ages. on Central American freshwater fi shes demonstrates how the The capacity to examine phylogenies within an accurate tem- phylogeographic approach can provide important new insights. poral context is vital for tying cladogenetic patterns to specifi c However, to date, similar analyses of Amazonian fi sh taxa, espe- palaeogeographic events. Understanding of Amazonian clades, cially investigations of geographical phenomena, have been which exhibit a confusing array of temporally overlapping limited. A recent attempt to investigate Pleistocene refugia and biogeographical patterns, would certainly benefi t from time- piranha population structure (Hubert et al. 2007b) is described in calibrated phylogenies. more detail in a later section. Phylogenetic analysis of molecular data provides two distinct Concerns regarding the phylogeographic approach, especially kinds of information: branching order (tree topology) and branch its traditional reliance on single mitochondrial genes, have been lengths. Branching order provides the history of lineage splitting voiced (e.g. Degnan 1993; Hare 2001; Choat 2006). Because a single or speciation events, and can be used in conjunction with palaeo- gene tree can be consistent with different scenarios of population geographic or fossil data to provide estimates of relative ages of history, it can be diffi cult to falsify phylogeographic hypotheses. clades (Lundberg 1998). For example, a clade with representatives Also, a mitochondrial gene may provide a view of population on either side of an impermeable geographical barrier, such as the events that may or may not be mirrored in nuclear genes (Hudson Eastern Cordillera of the Colombian Andes, may be presumed to & Turelli 2003). Finally, the assumption that mitochondrial DNA be at least as old as the origin of that barrier. Because these types evolves inUNCORRECTED a neutral manner, incorporated into some population of age estimates depend only PROOFon the availability of phylogenies genetic models, may not be general (Meiklejohn et al. 2007) – this in conjunction with geological data, they are also derivable may call into question conclusions based on neutral models. A from morphology-based analyses. Methods for age estimation practical issue in phylogeographic analysis of Neotropical fi shes is using branch lengths, however, are currently only available for appropriate designation of species boundaries. Because relatively molecular data sets. Molecular sequences differ from qualitative few genetic surveys have been undertaken, understanding of spe- morphological data in that the constituent units (e.g. nucleotide cies limits is limited and likely to change during the course of an bases in the case of DNA) are thought to evolve in a manner

HHoorn_ch25_FPP.inddoorn_ch25_FPP.indd 440808 77/27/2009/27/2009 3:48:083:48:08 PMPM Molecular signatures in Amazon fi sh fauna 409

that is amenable to statistical analysis (Sanderson 1997; Yoder & preservation and recovery of fossils in fl uvial systems. Low-energy Yang 2000). lacustrine depositional environments, from which most fresh- The logic behind molecular estimates of branch lengths and water fossils are known, are rare in the present-day Amazonian clade ages is relatively straightforward. Branch lengths may hydrological setting. The high current fl ow and low pH of many be estimated using a variety of phylogenetic approaches and tropical rivers combined with high rates of biogenic decomposi- optimization procedures, and are then converted to ages using tion also reduce the probability of fossil formation. Further, the appropriate calibrations of rates of molecular evolution (e.g. discovery of sedimentary outcrops in the Amazon drainage basin Near et al. 2005). In practice, however, nearly every step involved is hindered by thickly vegetated landscapes and low topographic in this approach has been criticized and debated (e.g. Sanderson relief. To date, fossil fi shes from within the watershed of the mod- 1997; Yoder & Yang 2000; Pulquerio & Nichols 2006). Much of ern Amazon drainage basin are restricted to the Neogene of west- this discussion involves methodological issues that apply broadly ern Amazonia. However, fossil Amazonian fi sh faunas are also across taxa and geographical areas outside the Neotropics, and known from areas currently outside Amazonia, such as Andean are not discussed here. However, studies that attempt to derive basins to the west and the north. molecular-based ages for Neotropical fi shes face a number of Fossils of Neotropical freshwater fi shes are therefore rare, and additional challenges and opportunities, and these are mostly most groups are either poorly represented or entirely absent related to calibration issues (Hulsey et al. 2004). from the palaeontological record. As a result there is a dearth of Translating branch lengths (often measured as amount of fossils with phenotypes intermediate between the major groups sequence divergence) to absolute ages (in years) requires an that dominated Mesozoic and Cenozoic ichthyofaunas. In other estimate of evolutionary rate. Such rates can be assumed a priori words, most fi sh taxa are fully modern by the time of their fi rst (e.g. 2% divergence per million years is a rate that is sometimes appearance in the stratigraphic record, often being ascribed to used for mitochondrial genes of fi shes), but this approach has its modern genera. For example, the fi sh fauna of the Maastrichtian problems (see, e.g., Ho 2007). A preferable technique is to calculate (c. 71–66 Ma) El Molino Formation of Bolivia is dominated by a rate specifi cally for the gene(s) and clade in question, using one or non-teleost groups (e.g. dipnoans, pycnodonts, polypertiforms, more calibration points. There are sophisticated methods that relax lepisosteids) characteristic of the Cretaceous, and also some the assumption of the molecular clock, and allow rates of evolution archaic (an extinct siluriform Andinichthys, an osteo- to vary across the branches of a phylogenetic tree (e.g. Sanderson glossid) (Gayet et al. 2001, 2003). By contrast, the overlying 1997; Drummond et al. 2006; Yang & Rannala 2006). However, Paleocene (c. 60–58 Ma) Santa Lucia Formation is dominated by even for these methods, at least one calibration point is required, teleosts, especially characiform and siluriform taxa that character- and multiple calibrations are preferred for cross- validation (Hulsey ize modern faunas (DeCelles & Horton 2003; Gayet et al. 2003). et al. 2004; Near et al. 2005). Two types of calibrations are typically Such sudden faunal transformations indicate great gaps in the used, employing: (i) stratigraphic information from fossils, and/or preservational sequence, extremely rapid diversifi cation, or both. (ii) radiometric dating of palaeogeographic events. We elaborate More recent fossil formations provide some useful materials on these approaches below. for estimating minimum ages of certain Neotropical freshwater fi sh clades. Perhaps the best known is the Middle Miocene (c. 12 Ma) La Venta fauna in the Villavieja Formation of what is Using fossils to calibrate molecular estimates now the Magdalena valley of Colombia (Lundberg & Chernoff of clade ages 1992). Fishes of the trans-Andean (west of the eastern Andean cordilleras) La Venta fauna include many living forms that are The relative or absolute age of fossils is widely used in evolution- now known only in the cis-Andean (east of the Andean cor- ary and biogeographical studies to estimate minimum clade ages dillera) Amazon and Orinoco Basins. Many of these species (e.g. Jablonski et al. 1985; Bemis et al. 1997; Arratia 1999; Albert are indistinguishable from living species (e.g. gigas, & Fink 2007). The presence of a fossil with traits diagnostic of the pacu Colossoma macropomum), or are closely related to a particular taxon is direct evidence for its stratigraphic range Amazonian species (e.g. the catfi sh genera Brachyplatystoma and (Lundberg et al. 1998; Murray 2001; Malabarba et al. 2006). Of Hoplosternum). The geological isolation of the Magdalena from course a lineage may be older than the age of the oldest known the Amazon drainage basin began with the rise of the Eastern fossil, either because of sampling errors, or because the timing of Cordillera of Colombia about 12 Ma, suggesting a minimum age the lineage divergence pre-dated the acquisition of morphologi- for the divergence of lineages in cis- and trans-Andean basins cal features by which that taxon is recognized. Thus, fossil ages (Albert et al. 2006). The Late Miocene Urumaco Formation, in are sometimes used as minimum calibration points for molecular what is now the modern Falcon Basin of northern Venezuela, rate estimates (Near et al. 2005). preserves fossilized remains of fi shes from the estuary of the pal- The utility of the palaeontological record for calibration points aeo-Orinoco (Sanchez-Villagra & Aguilera 2006). Fishes of the dependsUNCORRECTED on the abundance, quality and taxonomic breadth of Urumaco Formation werePROOF isolated from the rest of the Orinoco fossils for the clade of interest. Unfortunately, the record of fossil drainage basin by the rise of the Western Merida Andes, between fi shes in the Amazon drainage basin is relatively sparse, especially 10 and 8 Ma (Hardman & Lundberg 2006). considering the very high diversity of this region (see Chapter 17). Aside from the general circumstances surrounding fossil preser- Freshwater fi shes are poorly represented as fossils worldwide, vation in Amazonia, taphonomic bias is expected and observed to compared with near-shore marine fi shes or many terrestrial impact certain clades more than others. Taphonomic biases include vertebrate groups. This is due to unfavourable conditions for the differential effects of body size, phenotype, ecology and behaviour

HHoorn_ch25_FPP.inddoorn_ch25_FPP.indd 440909 77/27/2009/27/2009 3:48:083:48:08 PMPM 410 N.R. Lovejoy et al.

on the environment of deposition, such as the physical and chem- that separate river basins and aquatic habitats, such as tectonic ical properties of water and sediments that affect disarticulation, uplifts or marine incursions, act as important barriers to gene fl ow transport, interment and mineralization. Fishes with larger size, and dispersal. These barriers are expected to lead to allopatric spe- robust spines or teeth, or other morphologically diagnostic bony ciation and ultimately may lead to differentiated clades isolated elements, those inhabiting shallow marine or lacustrine habitats on either side of the barrier. Observed amounts of sequence diver- or that occur in large schools, are all more likely to be preserved, gence between these separated taxa, divided by time since separa- and when discovered, correctly identifi ed. Among Neotropical tion, provides a rate of molecular evolution that can be used to freshwater fi shes the most abundant and readily identifi able fi sh obtain age estimates for other parts of the clade in question. fossils are the distinctly ornamented fi n spines and skull bones It is obvious that obtaining accurate estimates for the geological of pimelodid and doradid catfi shes, stingray caudal barbs, osteo- dates of palaeogeographic events is central to the biogeographi- glossid skull roofi ng bones, and the characteristic tooth plates of cal age calibration of molecular divergences. Information from lungfi shes, stingrays and some characins, groups whose remains dating methods like fi ssion track analysis and radiometric decay are common in the Villavieja and Urumaco Formations. Fossils provides the foundation for current understanding of the abso- of whole, articulated skeletons are very rare, probably due to the lute timing of tectonic events and the ages of sedimentary forma- high-energy nature of most Amazonian fl uvial environments, as tions. Biostratigraphy using indicator fossils (e.g. land mammal well as predation, scavenging, bacterial degradation and dislodge- stages) allows relative dating of geographically disparate faunas ment by burrowers. Taphonomic processes therefore result in an when absolute radiometric ages are scarce, as they are in the South under-representation of clades that include many small species, American record. such as tetras (Characinae). Empirical evidence that many impor- Ideally, a geological vicariance event for age calibration tant clades are under-represented as fossils comes from the gym- would be: notiform electric fi shes. Fragments of articulated mid-body and posterior body sections of fossilized gymnotiform electric fi shes 1 Rapid, separating taxa almost instantly, and also affecting all are known from the Late Miocene Yecua Formation of Bolivia, members of a regional fauna almost simultaneously; providing a minimum date c. 12 Ma for the presence of this taxon 2 spatially extensive, affecting a broad geographical area and in Amazonian waters. However, biogeographical and phylogenetic multiple phylogenetically independent taxa; data suggest that the gymnotiform lineage originated in the Lower 3 long-lived, of suffi cient geological duration that the lineages Cretaceous (>110 Ma), leaving a c. 100 million year gap in the fos- are still distributed in allopatry; sil record for this clade. 4 impermeable to all members of the fauna (semipermeable To date, few molecular biogeographical studies of Amazonian barriers are more diffi cult to perceive after the fact); fi shes have incorporated fossil-based calibrations, due in large 5 accompanied by a volcanism so that the date can be known part to the diffi culties outlined above. One successful example is with great precision by radiometric decay analysis. Hardman & Lundberg’s (2006) study of phractocephaline catfi sh diversifi cation. These authors used fossil representatives of the The most useful geological dating events are therefore tectonic extant Steindachneridion to calibrate a molecular timescale orogenies with plutonic activity, and from this perspective, the for the rest of the clade, and compared a calibration point derived dating of events is expected to be much more reliable in the west- from the orogeny of the Merida Andes. Aspects of this study ern than in the eastern Amazonia. highlight the characteristics that will make certain taxa amenable Although of great utility for estimating the age of taxa or whole to this type of analysis. Phractocephalines are relatively large fi shes faunas, dates obtained from palaeogeographic-based molecu- (>1 m adult total length), with robust bony elements that preserve lar calibrations may be subject to several sources of error. Some well as fossils. They are a relatively depauperate clade, with only a of these arise from inaccuracies in the methods for obtaining few extant species in four genera, of which two (Phractocephalus geological dates, and others from uncertainties in the effects of and Steindachneridion) are represented in the fossil record. They geological events on individual taxa. For example, the effects of are also widespread, with a distribution that includes the Maracaibo the Isthmus of Panama on geminate marine lineages are more and Orinoco drainages – this allows for comparisons between pal- complicated than expected – with gene fl ow being sundered in aeogeography and fossil calibration points. different lineages at different times (Bermingham et al. 1997). In summary, fossil-based calibrations of molecular rates hold Biases may also arise from incomplete phylogenetic resolu- some promise for Amazonian fi shes, as evidenced by Hardman & tion, incomplete sampling, or an actual history of widespread Lundberg (2006). However, the paucity of the fossil record, as well extinctions, all of which serve to overestimate the true divergence as the uneven taxonomic distribution of fossils, will necessarily time by reducing information on sister taxon relationships. All of limit the usefulness of the approach. these sources of error in fact hinder biogeographical age calibra- tion across the Eastern Cordillera of Colombia (Albert et al. 2006) UNCORRECTEDand the Western Merida Andes PROOF (Hardman & Lundberg 2006). In Using palaeogeographic events to calibrate molecular both of these situations, the trans-Andean fauna is known from estimates of clade ages palaeontological sources to have been formerly much richer, and these vicariance events were accompanied by regional extinc- This calibration method uses dates of palaeogeographic events tions. These extinctions eliminate relevant taxa and interfere provided by geological data as minimum ages for the divergence with accurate phylogenetic reconstruction and branch length of sister taxa. In the case of freshwater fi shes, geographical events estimation.

HHoorn_ch25_FPP.inddoorn_ch25_FPP.indd 441010 77/27/2009/27/2009 3:48:083:48:08 PMPM Molecular signatures in Amazon fi sh fauna 411

The timing of many key Neotropical palaeogeographic events • the separation of the Orinoco and Amazon Rivers into their remains the subject of active investigation. While the palaeo- modern drainage basin by the rise of the Vaupes Arch (Late geography of the Neogene is much better understood than that Miocene-Pliocene) (Hoorn 1993; Hoorn et al. 1995; Diaz de of earlier periods, the nature and timing of many major Neogene Gamero 1996; see also Chapter 4); events are still hotly debated (Hoorn et al. 1995; Hoorn 1996; • the separation of the Paraná and Amazon drainage basins Diaz de Gamero 1996; Campbell et al. 2001, 2006; Rossetti 2001; by the rise of the Michicola Arch (c. 42–35 Ma) and/or Chapare Campbell 2005; Hoorn & Vonhof 2006; see also Chapter 26). Buttress (c. 28–15 Ma) (Butler et al. 1995; DeCelles & Horton Nevertheless, the Cenozoic history of several major geographical 2003). features in Amazonia is suffi ciently well established to allow their use in biogeographical age calibrations. Also, important palaeo- Based on the criteria discussed above, orogenic dating of geographic events that did not directly involve Amazonia, such as lineages from the Maracaibo and Magdalena drainage basins the closure of the Central American Isthmus, can also be used as appears to represent the most reliable option for palaeogeo- calibration points for widely distributed fi sh clades. graphic calibrations. Although these vicariance events have To date, a relatively small number of events have been used to been associated with widespread extinctions – a possible source calibrate molecular rates for Amazonian fi sh clades (Fig. 25.1). of error – these events were spatially extensive, long-lived, These include: relatively impermeable to fi shes, and associated with well- accepted geological dates. Dating the division of Orinoquian • the orogeny of the Merida Andes and isolation of the Maracaibo and Amazonian faunas, in contrast, is much less straightfor- drainage basin (c. 10–8 Ma) (Duque-Caro 1990; Diaz de ward. Although separation of these two drainages is thought to Gamero 1996; Colletta et al. 1997; Lundberg et al. 1998); have occurred about 10 to 8 Ma, after the reconfi guration of an • the orogeny of the Eastern Cordillera of Colombia and isolation ancient Orinoco river system, they are in fact currently connected of the Magdalena drainage basin (c. 12–11 Ma) (Hoorn et al. via the Casiquiare river, and exchanges of headwater tributar- 1995; Guerrero 1997); ies between these two drainages has probably been relatively

Impermeable

Semipermeable 8

12

10–0 >30–10?

9–3

42–0

0–100 m 101–200 m UNCORRECTED PROOF201–300 m 301–400 m >400

Fig. 25.1 Map of South America showing locations of hydrogeographic barriers discussed in the text. Numbers refer to minimum divergence times in Ma. Base map generated by Paulo Petry using HydroSHEDS high-resolution elevation data.

HHoorn_ch25_FPP.inddoorn_ch25_FPP.indd 441111 77/27/2009/27/2009 33:48:09:48:09 PPMM 412 N.R. Lovejoy et al.

common throughout the intervening period. Thus, distributions freshwater plume may be an ongoing phenomenon in some taxa of fi sh lineages across the Amazonian and Orinoquian divide (e.g. Poecilia reticulata) (D. Phillips, personal communication). could be the result of more recent events rather than a singular 10–8 Ma vicariance (e.g. Willis et al. 2007), and calibrations using this event (e.g. Hubert et al. 2007a) should be viewed with Progress to date: molecular signatures of Neogene caution. Similarly, the hypothesized ‘separation’ of the Amazon palaeogeographic events and Paraná drainage basins at about 10 Ma (used by Musilová et al. 2008), is perhaps an oversimplifi cation of a complex his- Because fi shes are strictly confi ned to aquatic habitats, their torical process (Aguilera & De Aguilera 2003). Lundberg et al. biogeographical histories are expected to record a detectable (1998) describe multiple stream capture events over the past signature of river history, including drainage modifi cation and 40 million years that may have served to facilitate or disrupt gene capture events, as well as effects of marine incursions. However, fl ow between these two rivers. detection of these events depends on the scale of the analysis Unfortunately, not all fi sh clades have living representatives (taxonomic and temporal) and the response of individual clades in the Magdalena or Maracaibo drainages. The use of palaeo- (mediated by ecology) to particular palaeogeographic alterations. geographic events to date Amazonian genetic divergences in still In the case of Neogene Amazonia, large-scale riverine events of in its infancy, yet there are several promising areas for future particular importance include: work. Several palaeogeographic events have been proposed for dating fi shes in the drainages of the southeastern Brazilian 1 The break-up of the northward fl owing palaeo-Orinoco Shield, including the headwaters of the Upper Paraná, Tiete, Sao into separate Maracaibo/Magdalena, Orinoco and Amazon Francisco and Parnaíba drainage basins (Ribeiro 2006; Ingenito & drainage basins; Buckup 2007). Headwater stream capture in the western Guianas 2 the establishment of the east-west transcontinental axis of the (e.g. Caura-Uraricoera, Branco-Essequibo) is currently being modern Amazon River; used to date cichlids (Lopez-Fernandez et al. 2005a, 2005b, 2006) 3 connections between the Amazon/Madeira and upper Paraná and loricariid catfi shes (Armbruster 2004; Reis et al. 2006). The Rivers; Late Miocene-Pliocene (c. 9–3 Ma) rise of the Fitzcarrald Arch in 4 the existence of the vast Pebas mega-wetland system in what is southwestern Amazonia (Espurt et al. 2007; see also Chapter 6) now the area of the western Amazonia; is being used to help constrain divergence dates in some electric 5 periodic marine incursions into the lowland basins of various fi shes (J. Albert, unpublished observation). The hydrological river systems. separation of the trans-Andean San Juan (Pacifi c) and Atrato (Caribbean) drainage basins is undated but could in principle These events are of such signifi cant scale (geographically aid in understanding the timing of vicariance events within the and temporally) that they could reasonably be expected to Choco (Albert et al. 2006). leave detectable signatures on the biogeographical patterns of The formation of rapids and waterfalls in large rivers partially Amazonian taxa. At the level of populations, fi shes might be isolates lowland faunas from those of tributary headwaters, and expected to record the putative effects of climatic oscillations, may therefore be used to calibrate more recent (Plio-Pleistocene) recent drainage capture and vicariance events, as well as signifi cant speciation events. For example, the rapids at Porto Velho and habitat alterations (caused, e.g., by marine introgressions). the Iguaçu Falls both mark the lower limits for many fi sh species endemic to the Upper Madeira (Bolivian) and Upper Paraná drainage basins respectively (Chernoff et al. 2000; Castro Results from species- and clade-level analyses et al. 2005; Kullander & Ferreira 2006; Ingenito & Buckup 2007). Indeed, rapids may serve as semipermeable fi lters for dispersal on To date, molecular analyses have been applied to Amazonian all of the large Amazon tributaries. Several studies have attempted fi shes representing a broad range of sizes, life-history to correlate the locations of so-called structural arches (Roddaz strategies and habitat preferences, including: migratory detri- et al. 2005) with biogeographical distributions, hypothesized tivorous Prochilodus Characiformes (Sivasundar et al. 2001; to serve as the geophysical underpinnings for rapids in western Turner et al. 2004; Moyer et al. 2005), large-bodied riverine Amazonia (Da Silva & Patton 1998). However, the accuracy of predatory phractocephaline catfi shes (Pimelodidae) (Hardman geologically derived dates for the origins of rapids is clouded to & Lundberg 2006), small-bodied, stream-dwelling algivo- some extent by uncertainties about the mechanisms of their for- rous catfi shes (Hypostomus) (Montoya-Burgos 2003), ben- mation, resulting from a combination of regional tectonics and thic potamotrygonid stingrays (Lovejoy et al. 1998; Toffoli erosion under the infl uence of eustatic sea-level changes. Further, et al. 2008), as well as cichlids (Willis et al. 2007; Musilová et rapids are not generally a fi xed landscape feature as they continu- al. 2008), needlefi shes (Lovejoy & de Araújo 2000), piranhas ally erodeUNCORRECTED their basement sediments and therefore move upstream (Hubert et al. 2007a; Freeman PROOF et al. 2007; Orti et al. 2008) and on geological timescales. Continental islands, such as Trinidad, killifi shes (Hrbek & Larson 1999). All of these studies are based offer another potential source of more recent palaeogeographic on molecular phylogenetic analyses; some include age estimates dating events. However, precise dating of lineage-splitting can based on molecular data (e.g. Hardman & Lundberg 2006) and be hampered by a complex history of iterative isolations and some incorporate additional types of biogeographic analy- connections caused by long-term fl uctuations in sea level. In the ses (e.g. Dispersal Vicariance Analysis in Hubert et al. 2007a). case of Trinidad, transoceanic dispersal by means of the Orinoco The taxonomic scale ranges from comprehensive analyses of

HHoorn_ch25_FPP.inddoorn_ch25_FPP.indd 441212 77/27/2009/27/2009 3:48:103:48:10 PMPM Molecular signatures in Amazon fi sh fauna 413

small genera (e.g. three species of Potamorrhaphis needlefi shes; Freshwater Potamorrhaphis needlefi shes show a similar pattern, Lovejoy & de Araújo 2000) to investigations of larger clades (e.g. where widespread haplotype clades in the lower Orinoco and 60+ species of the killifi sh family Rivulidae; Hrbek & Larson Amazon Rivers are recently diverged relative to lineages in the 1999). upper Orinoco (Lovejoy & de Araújo 2000). Hubert et al. (2007a) Because of the ecological diversity of the taxa considered suggest that their data from Serrasalmus and Pygocentrus piranhas above, and the variety of proposed hypotheses and analyses, a provide additional evidence of marine effects on lowland taxa. simple summary of biogeographical fi ndings is elusive. A general Finally, Lovejoy et al. (1998, 2006) and Lovejoy & Collette (2001) conclusion, with signifi cant ramifi cations for the testability discussed the possibility that marine transgressions also played a of geological events, is that biogeographical patterns are often role in the evolutionary transitions of several marine fi sh lineages very complex. Even in relatively young groups (<10 Ma) with to freshwater habitats (including stingrays, needlefi shes, ancho- relatively few (<10) species, observed patterns suggest that vies and others). dispersal, extinction and hybridization have taken place (e.g. Moyer et al. 2005; Hubert et al. 2007a; Willis et al. 2007) – these Results from intraspecifi c (phylogeographic) analyses processes can obscure and complicate biogeographical interpre- tation. A related issue is that many historical or current barriers A wide variety of intraspecifi c genetic patterns have been reported affecting Amazonian fi shes are neither permanent nor impassi- for Amazonian fi sh groups. In studies of large-bodied fi sh species ble, as evidenced by multiple reconstructions of dispersal across inhabiting Amazonian fl oodplains, including Brachyplatystoma them. For example, although the fracture of the palaeo-Orinoco catfi shes (Batista & Alves-Gomes 2006), (Hrbek resulted in separate Amazon and Orinoco drainages approxi- et al. 2005) and Colossoma macropomum (Santos et al. 2007), little mately 8 Ma, since that time multiple lineages have moved between population genetic structure was observed, suggesting extensive these basins using the Casiquiare or the Rupununi portals, or gene fl ow across hundreds of kilometres. In contrast, the needle- other dispersal pathways (e.g. Lovejoy & de Araújo 2000; Hubert fi sh Potamorrhaphis guianensis, a smaller-bodied species (30 cm) et al. 2007a; Willis et al. 2007). Dispersal is problematic because also found on Amazonian fl oodplains, shows substantial genetic it can erode the signature of earlier palaeogeographic events, and differences among geographically separated populations (Lovejoy complicate attempts to use palaeogeographic barriers to calibrate & de Araújo 2000). Cooke & Beheregaray (2007) found extremely age estimates. high levels of genetic variability in a nuclear intron for cardinal Despite these impediments, there are clear signals of some tetras (Paracheirodon axelrodi) sampled across the Negro drainage palaeogeographic events. The Neogene orogeny of the north- basin. In contrast, investigations of some prochilodontid species ern Andes is well recorded by endemic Magdalena/Maracaibo from a number of river drainages have reported comparably low lineages in several clades, and as expected for a permanent amounts of intraspecifi c genetic variation (Sivasundar et al. 2001; barrier, observed biogeographical patterns are consistent with Turner et al. 2004; Moyer et al. 2005). However, none of these a single vicariant event, and no post-barrier dispersal (Love- studies explicitly tested hypotheses concerning the possible effects joy et al. 1998; Sivasundar et al. 2001; Montoya-Burgos 2003; of Neogene palaeogeographic events on observed intraspecifi c Turner et al. 2004; Albert et al. 2006; Hardman & Lundberg 2006). patterns. Several authors have hypothesized that observed sister clades In a fi rst attempt to link intraspecifi c patterns to palaeogeo- in the Orinoco and Amazon Rivers can be explained by the graphic events, Hubert et al. (2007b) proposed to test the effects of conversion of the palaeo-Orinoco into separate Amazon and Pleistocene climatic fl uctuations by examining population genetic Orinoco drainage basins. For example, Hubert et al. (2007a), variation within the piranha Serrasalmus rhombeus in the Madeira assumed that Orinoco and Amazon sister lineages of piranhas River. These authors hypothesized a role for forest refuges in pre- were established by this palaeogeographic event. Relationships serving genetic diversity during reductions in the species range, between Paraná and Amazon clades have been attributed to and presented results that they interpreted as evidence for a popu- headwater capture or boundary displacement events in several lation expansion during the last 800,000 years. However, the refu- studies (Sivasundar et al. 2001; Montoya-Burgos 2003; Musilová gia theory in general has been widely criticized on both theoretical et al. 2008). Montoya-Burgos (2003) and Hubert et al. (2007a) and empirical grounds (Endler 1982; Weitzman & Weitzman present additional hypotheses associating biogeographical 1982; Lundberg 1998; Colinvaux & De Oliveira 2000; Colinvaux patterns with river changes involving the Madeira, Tocantins, et al., 2000; see also Chapter 20), and Hubert et al. (2007b) do Negro and Ucayali. not present palaeobotanical evidence for the existence of forest Although debate continues about the extent of Neogene habitat, particularly in the regions hypothesized to function as ref- marine transgressions in the Amazon region, there seems little ugia. Indeed, a recent palaeovegetation study of Amazonia during doubt that such events strongly affected the distribution of the Pleistocene (Anhuf et al. 2007) does not indicate forest refugia freshwater fi shes. A simple prediction is that marine inundated in the areas hypothesized by Hubert et al. (2007b) (i.e. Aripuanã lowlandUNCORRECTED regions would have been uninhabitable by freshwater and Beni regions). Further, PROOF S. rhombeus inhabits rapids and large fi shes; hence, we expect such regions to be occupied by lineages river channels, and the distribution of this species is not tightly that are recently diverged compared to more upland relatives. linked to forested habitats. The role of forest refugia in limiting Hrbek & Larson (1999) documented such a pattern in rivulid kil- population sizes and dispersal therefore remains unclear. Thus, lifi shes, where taxa on upland shield formations were determined although the study is a valuable initial attempt, the conclusions to have diverged earlier than taxa in lowland regions, including of Hubert et al. (2007b) are best viewed with caution until more the Orinoco llanos, Amazonian savanna, and coastal regions. detailed investigations can be carried out.

HHoorn_ch25_FPP.inddoorn_ch25_FPP.indd 441313 77/27/2009/27/2009 3:48:113:48:11 PMPM 414 N.R. Lovejoy et al.

Conclusions and future challenges Albert, J.S., Fernandes-Matioli, F.M., de Almeida-Toledo, L.F. (1999) A new species of Gymnotus (Gymnotiformes, Teleostei) from The use of molecular genetics in biogeography is currently in a southeastern Brazil: towards the deconstruction of Gymnotus pioneering stage of development. Many studies continue to suf- carapo. Copeia 1999, 410–421. Albert, J.S., Crampton, W.G.R., Thorsen, D.H., Lovejoy, N.R. (2005) fer from a reliance on analyses of single clades, the use of single Phylogenetic systematics and historical biogeography of the genes (usually mitochondrial), and the assumption that gene trees Neotropical electric fi sh Gymnotus (Teleostei: Gymnotiformes). accurately represent species trees. Studies on the biogeography Syst Biodivers 2, 375–417. of Amazonian fi shes offer additional challenges, including the Albert, J.S., Lovejoy, N.R., Crampton, W.G.R. (2006) Miocene extraordinary taxonomic complexity of the fauna and the associ- tectonism and the separation of cis- and trans-Andean river basins: ated logistical requirements of sample collection and experimen- Evidence from Neotropical fi shes. J S Am Earth Sci 21, 14–27. tal design. Yet even a cursory review of the list of references cited Anhuf, D., Ledru, M.P., Behling, H., Da Cruz, J.F.W., Cordeiro, R.C., for this chapter demonstrates the advances now being made in all Van der Hammen, T. et al. (2006) Paleo-environmental change in these areas. Amazonian and African rainforest during the LGM. Palaeogeogr The fi elds of phylogenetics and biogeography face an even Palaeocl 239, 510–527. Armbruster, J.W. (2004) Phylogenetic relationships of the greater challenge: to generate and test useful alternative hypoth- suckermouth armoured catfi shes () with emphasis on eses. As in the other historical sciences (e.g. , astronomy), the Hypostominae and the Ancistrinae. Zool J Linn Soc 141, 1–80. progress is made by new and more accurate methods of meas- Arratia, G. (1999) The monophyly of the Teleostei and stem-group urement, new sources of information, new hypotheses with teleosts. In: Arratia, G., Schultz, H.-P. (eds) Mesozoic Fishes 2 – increasingly restrictive predictions, and new models for matching Systematics and Fossil Record. Munchen: Verlag Dr Friedrich Pfeil, observations to predictions. The use of molecular data in histori- pp. 265–334. cal biogeography and phylogeography offers great potential in all Avise, J.C., Arnold, J., Ball, R.M., Bermingham, E., Lamb, T., Neigel, these regards. J.E. (1987) Intraspecifi c phylogeography: The mitochondrial DNA Interactions between biology, palaeogeography and geology are bridge between and systematics. Ann Rev Ecol increasingly proving to be mutually benefi cial. Each of these fi elds Syst 18, 489–522. can and should be used to test hypotheses proposed by the others. Batista, J.S., Alves-Gomes, J.A. (2006) Phylogeography of Brachyplatystoma rousseauxii (Siluriformes – Pimelodidae) in the For biologists, the delineation of testable biogeographical hypoth- offers preliminary evidence for the fi rst case of eses is often diffi cult, in part because Neotropical palaeogeography “homing” for an Amazonian migratory catfi sh. Genet Mol Res 5, is such a dynamic fi eld, and also because the relevant geological 723–740. literature can be diffi cult to parse. In this regard, it is important Bemis, W.E., Findeis, E.K., Grande, L. (1997) An overview of for Neotropical biogeographers to set up explicit hypotheses with Acipenseriformes. Environ Biol Fish 48, 25–72. care, and to consider multiple alternatives. Bermingham, E., Avise, J.C. (1986) Molecular zoogeography of freshwater fi shes in the southeastern United States. Genetics 113, 939–965. Acknowledgements Bermingham, E., Martin, A.P. (1998) Comparative mtDNA phylogeography of neotropical freshwater fi shes: testing shared history to infer the evolutionary landscape of lower Central We thank G. Arratia, J. Armbruster, J. Baskin, E. Bermingham, America. Mol Ecol 7, 499–517. D. Bloom, P. Buckup, P. Chakrabarty, W. Crampton, W. Fink, M. Bermingham, E., McCafferty, S.S., Martin, A.P. (1997) Fish Goulding, S. Kullander, G. Orti, H. Lopez-Fernandez, J. Lundberg, biogeography and molecular clocks: Perspectives from the L. Malabarba, T. Near, J. Neigel, P. Petry, R. Reis, S. Schaefer, G. Panamanian Isthmus. In: Kocher, T., Stepien, C. (eds) Molecular Smith and K. Winemiller for thoughtful discussions of the ideas Systematics of Fishes. San Diego, CA: Academic Press, pp. 113–128. presented in this chapter. We thank two anonymous reviewers for Bernatchez, L., Wilson, C.C. (1998) Comparative phylogeography of their comments on an earlier version of the manuscript. N.R.L. Nearctic and Palearctic fi shes. Mol Ecol 7, 431–452. acknowledges support from an NSERC Discovery Grant. J.S.A. Boeger, W.A., Kritsky, D.C. (2002) Parasites, fossils and geologic acknowledges support from US National Science Foundation history: Historical biogeography of the South American freshwater croakers, Plagioscion spp. (Teleostei, Sciaenidae). Zool Scr 32, 3–11. DEB 0138633, 0215388 and 0614334. Burridge, C.P., Craw, D., Waters, J.M. (2007) An empirical test of freshwater vicariance via river capture. Mol Ecol 16, 1883–1895. Butler, R.F., Richards, D.R., Sempere, T., Marshall, L.G. (1995) References Paleomagnetic determinations of vertical-axis tectonic rotations from late Cretaceous and Paleocene strata of Bolivia. Geology 23, Aguilera, O., De Aguilera, D.R. (2003) Two new otolith-based sciaenid 799–802. species of the genus Plagioscion from South American Neogene Campbell, A.E. (2005) Shelf-geometry response to changes in relative marineUNCORRECTED sediments. J Paleontol 77, 1133–1138. sea level on a mixed carbonate-siliciclastic PROOF shelf in the Guyana Albert, J.S., Crampton, W.G.R. (2003) Seven new species of the Basin. Sediment Geol 175, 259–275. Neotropical electric fi sh Gymnotus (Teleostei, Gymnotiformes) Campbell, K.E., Heizler, M., Frailey, C.D., Romero-Pittman, L., with a redescription of G. carapo (Linnaeus). Zootaxa 28, 1–54. Prothero, D.R. (2001) Upper Cenozoic chronostratigraphy of the Albert, J.S., Fink, W.L. (2007) Phylogenetic relationships of fossil southwestern Amazon Basin. Geology 29, 595–598. Neotropical electric fi shes (Osteichthyes: Gymnotiformes) from Campbell, K.E. Jr, Frailey, CD., Romero-Pittman, L. (2006) The the Upper Miocene of Bolivia. J Vertebr Paleontol 27, 17–25. Pan-Amazonian Ucayali , late Neogene sedimentation

HHoorn_ch25_FPP.inddoorn_ch25_FPP.indd 441414 77/27/2009/27/2009 3:48:113:48:11 PMPM Molecular signatures in Amazon fi sh fauna 415

in Amazonia, and the birth of the modern Amazon River system. Endler, J.A. (1982) Pleistocene Forest Refuges: Fact or Fancy? In: Palaeogeogr Palaeocl 239, 166–219. Prance, G.T. (ed.) Biological Diversifi cation in the Tropics. New York: Casciotta, J.R., López, H.L., Menni, R.C., Aiquelarena, A.M. (1989) Columbia University Press, pp. 641–657. The fi rst fi sh fauna from the Salado River (Central Argentina, Espurt, N., Baby, P., Brusset, S., Roddaz, M., Hermoza, W., Regard, V. South America) with additions to the Dulce River and limnological et al. (2007) How does the Nazca Ridge infl uence the comments. Archiv für Hydrobiologie115, 603–661. modern Amazonian foreland basin? Geology 35, 515–518. Castro, R.M.C., Casatti, L., Santos, H.E., Vari, R.P., Melo, A.L.A., Fernandes, C.C., Lundberg, J.G., Riginos, C. (2002) Largest of Martins, L.S.E. et al. (2005) Structure and composition of the all electric-fi sh snouts: Hypermorphic facial growth in male stream ichthyofauna of four tributary rivers of the upper Rio Apteronotus hasemani and the identity of Apteronotus anas Parana basin, Brazil. Ichthyol Explor Fres 16, 193–214. (Gymnotiformes: Apteronotidae). Copeia 2002, 52–61. Chernoff, B., Machado-Allison, A., Willink, P., Sarmiento, J., Barrera, Freeman, B., Nico, L.G., Osentoski, M., Jelks, H.L., Collins, T.M. S., Menezes, N., Ortega, H. (2000) Fishes of three Bolivian rivers: (2007) Molecular systematics of Serrasalmidae: Deciphering the Diversity, distribution and conservation. Interciencia 25, 273–283. identities of piranha species and unraveling their evolutionary Choat, J.H. (2006) Phylogeography and reef fi shes: Bringing ecology histories. Zootaxa 1484, 1–38. back into the argument. J Biogeogr 33, 967–968. Gayet, M., Marshall, L.G., Sempere, T., Meunier, F.J., Cappetta, H., Colinvaux, P.A., De Oliveira, P.E. (2000) Palaeoecology and climate Rage, J.C. (2001) Middle Maastrichtian vertebrates (fi shes, of the Amazon basin during the last glacial cycle. J Quaternary Sci amphibians, dinosaurs and other reptiles, mammals) from Pajcha 15, 347–356. Pata (Bolivia). Biostratigraphic, palaeoecologic and palaeo- Colinvaux, P.A., De Oliveira, P.E., Bush, M.B. (2000) Amazonian and biogeographic implications. Palaeogeogr Palaeocl 169, 39–68. neotropical plant communities on glacial time-scales: The failure of Gayet, M., Jegu, M., Bocquentin, J., Negri, F.R. (2003) New characoids the aridity and refuge hypotheses. Quaternary Sci Rev 19, 141–169. from the Upper Cretaceous and Paleocene of Bolivia and the Mio- Colletta, B., Roure, F., DeToni, B., Loureiro, D., Passalacqua, H., Pliocene of Brazil: Phylogenetic position and paleobiogeographic Gou, Y. (1997) Tectonic inheritance, crystal architecture, and implications. J Vertebr Paleontol 23, 28–46. contrasting structural styles in the Venezuela Andes. Tectonics 16, Goulding, M.J., Barthem, R., Ferreira, E.J.G. (2003) The Smithsonian 777–794. Atlas of the Amazon. Washington, DC: Smithsonian Books. Cooke, G.M., Beheregaray, L.B. (2007) Extremely high variability in Guerrero, J. (1997) Stratigraphy, sedimentary environments and the S72 intron of the Amazonian cardinal tetra (Paracheirodon Miocene uplift of the Colombian Andes. In: Kay, R.F. et al. (eds) axelrodi). J Fish Biol 71, 132–140. Vertebrate Paleontology in the Neotropics: The Miocene Fauna of La Cooper, M.A., Addison, F.T., Alvarez, R., Coral, M., Graham, R.H., Venta, Colombia. Washington DC: Smithsonian Institution Press, Hayward, A. et al. (1995) Basin development and tectonic history pp. 15–43. of the Llanos basin, Eastern Cordillera, and Middle Magdalena Hardman, M., Lundberg, J.G. (2006) Molecular phylogeny and a valley, Colombia. AAPG Bulletin 79, 1421–1443. chronology of diversifi cation for “phractocephaline” catfi shes Costa, C.H., Murillo, M.V., Sagripanti, G.L., Gardini, C.E. (2001) (Siluriformes: Pimelodidae) based on mitochondrial DNA and Quaternary intraplate deformation in the southeastern Sierras nuclear recombination activating gene 2 sequences. Mol Phylogenet Pampeanas, Argentina. J Seismol 5, 399–409. Evol 40, 410–418. Da Silva, M.N.F., Patton, J.L. (1998) Molecular phylogeography and Hare, M.P. (2001) Prospects for nuclear gene phylogeography. Trends the evolution and conservation of Amazonian mammals. Mol Ecol Ecol Evol 16, 700–706. 7, 475–486. Ho, S.Y.W. (2007) Calibrating molecular estimates of substitution De Almeida, F.F.M., Brito Neves, B.B., Carneiro, C.D.R. (2000) The rates and divergence times in birds. J Avian Biol 38, 409–414. origin and evolution of the South American Platform. Earth Sci Hoorn, C. (1993) Marine incursions and the infl uence of Andean Rev 50, 77–111. tectonics on the Miocene depositional history of northwestern DeCelles, P.G., Horton, B.K. (2003) Early to middle Tertiary foreland Amazonia: results of a palynological study. Palaeogeogr Palaeocl basin development and the history of Andean crustal shortening in 105, 267–309. Bolivia. Geol Soc Am Bull 115, 58–77. Hoorn, C. (1996) Miocene deposits in the Amazonian foreland basin. Degnan, S.M. (1993) The perils of single gene trees – Mitochondrial Science 273, 122–123. versus single-copy nuclear DNA variation in white-eyes (Aves: Hoorn, C., Vonhof, H. (2006) Neogene Amazonia: Introduction to the Zosteropidae). Mol Ecol Notes 2, 219–225. special issue. J S Am Earth Sci 21, 1–4. Dergam, J.A., Paiva, S.R., Schaeffer, C.E., Godinho, A.L., Vieira, F. Hoorn, C., Guerrero, J., Sarmiento, G.A., Lorente, M.A. (1995) Andean (2002) Phylogeography and RAPD-PCR variation in Hoplias tectonics as a cause for changing drainage patterns in Miocene malabaricus (Bloch, 1794) (Pisces, Teleostei) in southeastern Brazil. northern South-America. Geology 23, 237–240. Genet Mol Biol 25, 379–387. Hovikoski, J., Gingras, M., Räsänen, M., Rebata, L.A., Guerrero, Diaz de Gamero, M.L. (1996) The changing course of the Orinoco J., Ranzi, A. et al. (2007a) The nature of Miocene Amazonian River during the Neogene: a review. Palaeogeogr Palaeocl 123, epicontinental embayment: High-frequency shifts of the low- 385–402. gradient coastline. GSA Bulletin 119, 1506–1520. Dobson, D.M., Dickens, G.R., Rea, D.K. (2001) Terrigenous sediment Hovikoski, J., Räsänen, M., Gingras, M., Lopéz, S., Romero, on Ceara Rise: a Cenozoic record of South American orogeny and L.A.R., Melo, J. (2007b) Palaeogeographical implications of the UNCORRECTEDerosion. Palaeogeogr Palaeocl 165, 215–229. Miocene Quendeque FormationPROOF (Bolivia) and tidally-infl uenced Drummond, A.J., Ho, S.Y.W., Phillips, M.J., Rambaut, A. (2006) Relaxed strata in southwestern Amazonia. Palaeogeogr Palaeocl 243, phylogenetics and dating with confi dence. PLoS Biol 4, e88. 23–41. Duque-Caro, H. (1990) Major Neogene events in panamaic South Hrbek, T., Larson, A. (1999) The evolution of diapause in the killifi sh America. In: Tsuchi, R. (ed.) Pacifi c Neogene Events, their Timing, family Rivulidae (Atherinomorpha, Cyprinodontiformes): A Nature and Interrelationships. Tokyo: Tokyo University Press, molecular phylogenetic and biogeographic perspective. Evolution pp. 101–114. 53, 1200–1216.

HHoorn_ch25_FPP.inddoorn_ch25_FPP.indd 441515 77/27/2009/27/2009 3:48:113:48:11 PMPM 416 N.R. Lovejoy et al.

Hrbek, T., Taphorn, D.C., Thomerson, J.E. (2005) Molecular phylogeny Lovejoy, N.R., Bermingham, E., Martin, A.P. (1998) Marine incursion of Austrofundulus Myers (Cyprinodontiformes: Rivulidae), with into South America. Nature 396, 421–422. a revision of the genus and the description of four new species. Lovejoy, N.R., Albert, J.S., Crampton, W.G.R. (2006) Miocene marine Zootaxa 825, 1–39. incursions and marine/freshwater transitions: Evidence from Hrbek, T., Seekinger, J., Meyer, A. (2007) A phylogenetic and Neotropical fi shes. J S Am Earth Sci 21, 5–13. biogeographic perspective on the evolution of poeciliid fi shes. Mol Lundberg, J.G. (1998) The temporal context for the diversifi cation of Phylogenet Evol 43, 986–998. Neotropical fi shes In: Malabarba, L., Vari, R.P., de Lucena, C.A.S., Hubert, N., Renno, J.F. (2006) Historical biogeography of South de Lucena, Z.M.S. (eds) Phylogeny and Classifi cation of Neotropical American freshwater fi shes. J Biogeogr 33, 1414–1436. Fishes. Porto Alegre: Museu de Ciências e Tecnologia, pp. 49–68. Hubert, N., Duponchelle, F., Nuñez, J., Rivera, R., Renno, J.-F. (2006) Lundberg, J.G., Akama, A. (2005) Brachyplatystoma capapretum: Evidence of reproductive isolation among closely related sympatric a new species of goliath catfi sh from the Amazon Basin, with a species of Serrasalmus (Ostariophysi, Characidae) from the Upper reclassifi cation of allied catfi shes (Siluriformes: Pimelodidae). Madeira River, Amazon, Bolivia. J Fish Biol 69, 31–51. Copeia 2005, 492–516. Hubert, N., Duponchelle, F., Nunez, J., Garcia-Davila, C., Paugy, Lundberg, J.G., Chernoff, B. (1992) A Miocene fossil of the Amazonian D., Renno, J.F. (2007a) Phylogeography of the piranha genera fi sh Arapaima (Teleostei, ) from the Magdalena river Serrasalmus and Pygocentrus: implications for the diversifi cation of region of Colombia – biogeographic and evolutionary implications. the Neotropical ichthyofauna. Mol Ecol 16, 2115–2136. Biotropica 24, 2–14. Hubert, N., Duponchelle, F., Nunez, J., Rivera, R., Bonhomme, F., Lundberg, J.G., Marshall, L.C., Guerrero, J., Horton, B., Malabarba, Renno, J.F. (2007b) Isolation by distance and Pleistocene expansion M.C.S.L., Wesselingh, F. (1998) The temporal context for the of the lowland populations of the white piranha Serrasalmus diversifi cation of Neotropical fi shes. In: Malabarba, L., Vari, R.P., de rhombeus. Mol Ecol 16, 2488–2503. Lucena, C.A.S., de Lucena, Z.M.S. (eds) Phylogeny and Classifi cation Hudson R.R., Turelli. M. (2003) Stochasticity overrules the ‘three- of Neotropical Fishes. Porto Alegre: Museu de Ciências e Tecnologia, times rule’: genetic drift, genetic draft, and coalescent times for pp. 13–48. nuclear loci versus mitochondrial data. Evolution 57, 182–190. Lundberg, J.G., Kottelat, M., Smith, G.R., Stiassny, M.L.J., Gill, Hulen, K.G., Crampton, W.G.R., Albert, J.S. (2005) Phylogenetic A.C. (2000) So many fi shes, so little time: An overview of recent systematics and historical biogeography of the Neotropical electric ichthyological discovery in continental waters. Ann MO Bot Gard fi sh Sternopygus (Teleostei: Gymnotiformes). Syst Biodivers 3, 87, 26–62. 407–432. Lundberg, J.G., Sullivan, J.P., Rodiles-Hernández, R., Hendrickson, Hulsey, C.D., de Leon, F.J.G., Johnson, Y.S., Hendrickson, D.A., Near, D.A. (2007) Discovery of African roots for the Mesoamerican T.J. (2004) Temporal diversifi cation of Mesoamerican fi shes Chiapas catfi sh, Lacantunia enigmatica, requires an ancient across a major biogeographic boundary. Mol Phylogenet Evol 31, intercontinental passage. P Acad Nat Sci Phila 156, 39–53. 754–764. Malabarba, M.C., Zuleta, O., Del Papa, C. (2006) Proterocara argentina, Ingenito, L.F.S., Buckup, P.A. (2007) The Serra da Mantiqueira, south- a new fossil cichlid from the Lumbrera Formation, Eocene of eastern Brazil, as a biogeographical barrier for fi shes. J Biogeogr 34, Argentina. J Vertebr Paleontol 26, 267–275. 1173–1182. Mapes, R.W., Nogueira, A.C.R., Coleman, D.S., Leguizamon-Vega, Jablonski, D., Flessa, K.W., Valentine, J.W. (1985) Biogeography and A.M. (2006) Evidence for a continent scale drainage inversion in paleobiology. Paleobiology 11, 75–90. the Amazon Basin since the Late Cretaceous. Abstract, Philadelphia Kullander, S.O., Ferreira, E.J.G. (2006) A review of the South American Annual Meeting, Geological Society of America, Abstracts with cichlid genus Cichla, with descriptions of nine new species programs 38(7), 518. (Teleostei: Cichlidae). Ichthyol Explor Fres 17, 289–398. Meiklejohn, C.D., Montooth, K.L., Rand, D.M. (2007) Positive and Lomolino, M.V., Riddle, B.R., Brown, J.H. (2006) Biogeography, negative selection on the mitochondrial genome. Trends Genet 23, 3rd edn. New York: Sinauer. 259–263. López, H.L., Morgan, C.C., Montenegro, M.J. (2002) Ichthyological Menni, R.C., Gomez, S.E. (1995) On the habitat and isolation of Ecoregions of Argentina. La Plata: División Zoología Vertebrados, Gymnocharacinus bergi (Osteichthyes, Characidae). Environ Biol Museo de La Plata. Fish 42, 15–23. Lopez-Fernandez, H., Honeycutt, R.L., Stiassny, M.L.J., Winemiller, Miller, R.R. (1966) Geographical distribution of Central American K.O. (2005a) Morphology, molecules, and character congruence in freshwater fi shes. Copeia 1966, 773–801. the phylogeny of South American geophagine cichlids (Perciformes, Monsch, K.A. (1998) Miocene fi sh faunas from the northwestern Labroidei). Zool Scr 34, 627–651. Amazonia basin (Colombia, Peru, Brazil) with evidence of marine Lopez-Fernandez, H., Honeycutt, R.L., Winemiller, K.O. (2005b) incursions. Palaeogeogr Palaeocl 143, 31–50. Molecular phylogeny and evidence for an adaptive radiation of Montoya-Burgos, J.I. (2003) Historical biogeography of the catfi sh geophagine cichlids from South America (Perciformes: Labroidei). genus Hypostomus (Siluriformes: Loricariidae), with implications Mol Phylogenet Evol 34, 227–244. on the diversifi cation of Neotropical ichthyofauna. Mol Ecol 12, Lopez-Fernandez, H., Baechle, D.C.T., Kullander, S.O. (2006) Two new 1855–1867. species of Guianacara from the of eastern Venezuela Morrone, J.J., Crisci, J.V. (1995) Historical biogeography – Introduction (Perciformes : Cichlidae). Copeia 2006, 384–395. to methods. Ann Rev Ecol Syst 26, 373–401. Lovejoy, UNCORRECTEDN.R., de Araújo, M.L.G. (2000) Molecular systematics, Moyer, G.R., Winemeller, K.O., PROOF McPhee, M.V., Turner, T.F. (2005) biogeography, and population structure of Neotropical freshwater Historical demography, selection, and coalescence of mitochondrial needlefi shes of the genus Potamorrhaphis. Mol Ecol 9, 259–268. and nuclear genes in Prochilodus species of northern South America. Lovejoy, N.R., Collette, B.B. (2001) Phylogenetic relationships of new Evolution 59, 599–610. world needlefi shes (Teleostei: Belonidae) and the biogeography of Murray, A.M. (2001) The fossil record and biogeography of transitions between marine and freshwater habitats. Copeia 2001, the Cichlidae (: Labroidei). Biol J Linn Soc 74, 324–338. 517–532.

HHoorn_ch25_FPP.inddoorn_ch25_FPP.indd 441616 77/27/2009/27/2009 3:48:113:48:11 PMPM Molecular signatures in Amazon fi sh fauna 417

Musilová, Z., Rícan, O., Janko, K., Novák, J. (2008) Molecular Sivasundar, A., Bermingham, E., Orti, G. (2001) Population structure phylogeny and biogeography of the Neotropical cichlid fi sh tribe and biogeography of migratory freshwater fi shes (Prochilodus: Cichlasomatini (Teleostei: Cichlidae: Cichlasomatinae). Mol Characiformes) in major South American rivers. Mol Ecol 10, Phylogenet Evol 46, 659–672. 407–417. Myers, G.S. (1949) Salt-tolerance of fresh-water fi sh groups in Smith, G.R. (1992) Introgression in fi shes: Signifi cance for relation to zoogeographical problems. Bijdragen tot de Dierkunde paleontology, cladistics, and evolutionary rates. Syst Biol 41, 28, 315–322. 41–57. Myers, G.S. (1966) Derivation of the freshwater fi sh fauna of Central Strange, R.M., Burr, B.M. (1997) Intraspecifi c phylogeography of America. Copeia 1966, 766–773. North American highland fi shes: A test of the Pleistocene vicariance Near, T.J., Bolnick, D.I., Wainwright, P.C. (2005) Fossil calibrations hypothesis. Evolution 51, 885–897. and molecular divergence time estimates in centrarchid fi shes Toffoli, D., Hrbek, T., de Araújo, M.L.G., de Almeida, M.P., Charvet- (Teleostei: Centrarchidae). Evolution 59, 1768–1782. Almeida, P., Farias, I.P. (2008) A test of the utility of DNA barcoding Neigel, J., Domingo, A., Stake, J. (2007) DNA barcoding as a tool for in the radiation of the freshwater stingray genus Potamotrygon coral reef conservation. Coral Reefs 26, 487–499. (Potamotrygonidae: Myliobatiformes). Genet Mol Biol 31, Ortí, G., Sivasundar, A., Dietz, K., Jégu, M. (2008) Phylogeny of the 324–336. Serrasalmidae (Characiformes) based on mitochondrial DNA Turner, T.F., McPhee, M.V., Campbell, P., Winemiller, K.O. sequences. Genet Mol Biol 31, 1(Suppl.), 343–351. (2004) Phylogeography and intraspecifi c genetic variation of Pérez, G.A.C., Rican, O., Orti, G., Bermingham, E., Doadrio, I., prochilodontid fi shes endemic to rivers of northern South America. Zardoya, R. (2007) Phylogeny and biogeography of 91 species of J Fish Biol 64, 186–201. heroine cichlids (Teleostei: Cichlidae) based on sequences of the Vari, R.P. (1988) The Curimatidae, a lowland Neotropical fi sh family cytochrome b gene. Mol Phylogenet Evol 43, 91–110. (Pisces: Characiformes); distribution, endemism, and phylogenetic Pulquerio, M.J.F., Nichols, R.A. (2006) Dates from the molecular biogeography. In: Heyer, E.V.W.R. (ed.) Neotropical Distribution clock: how wrong can we be? Trends Ecol Evol 22, 180–184. Patterns. Rio de Janeiro: Academia Brasiliera de Ciências, pp. 313– Rebata, L.A., Gingras, M.K., Räsänen, M.E., Barberi, M. (2006) Tidal- 348. channel deposits on a delta plain from the Upper Miocene Nauta Vari, R.P., Malabarba, L. (1998) Neotropical ichthyology: An overview. Formation, Maranon Foreland Sub-basin, Peru. Sedimentology 53, In: Malabarba, L., Vari, R.P., de Lucena, C.A.S., de Lucena, Z.M.S. 971–1013. (eds) The Temporal Context for the Diversifi cation of Neotropical Reeves, R.G., Bermingham, E. (2006) Colonization, population fi shes. Phylogeny and Classifi cation of Neotropical Fishes. Porto expansion, and lineage turnover: Phylogeography of Mesoamerican Alegre: Museu de Ciências e Tecnologia, pp. 1–12. characiform fi sh. Biol J Linn Soc 88, 235–255. Wallace, A.R. (1876) The Geographical Distribution of With a Reis, R.E., Kullander, S.O., Ferraris, C.J. (2003) Checklist of the Study of the Relations of Living and Extinct Faunas as Elucidating the Freshwater Fishes of South and Central America. Porto Alegre: Past Changes of the Earth’s Surface. London: Macmillan & Co. Edipucrs. Weitzman, S.H., Weitzman, M. (1982) Biogeography and evolutionary Reis, R.E., Pereira, E.H.L., Armbruster, J.W. (2006) , a new diversifi cation in neotropical freshwater fi shes, with comments on loricariid catfi sh subfamily (Teleostei, Siluriformes), with revisions the refugia theory. In: Prance, G.T. (ed.) Biological Diversifi cation in of and . Zool J Linn Soc 147, 277–299. the Tropics. New York: Columbia University Press, pp. 403–422. Ribeiro, A.C. (2006) Tectonic history and the biogeography of the Wesselingh, F.P., Räsänen, M.E., Irion, G., Vonhof, H.B., Kaandorp, freshwater fi shes from the coastal drainages of eastern Brazil: an R., Renema,W. et al. (2002) Lake Pebas: a palaeoecological example of faunal evolution associated with a divergent continental reconstruction of a Miocene, long-lived lake complex in western margin. Neotrop Ichthyol 4, 225–246. Amazonia. Cainozoic Res 1, 35–81. Roddaz, M., Viers, J., Brusset, S., Baby, P., Herail, G. (2005) Sediment Westaway, R. (2006) Late Cenozoic sedimentary sequences in Acre provenances and drainage evolution of the Neogene Amazonian state, southwestern Amazonia: Fluvial or tidal? Deductions from foreland basin. Earth Planet Sci Lett 239, 57–78. the IGCP 449 fi eldtrip. J S Am Earth Sci 21, 120–134. Rossetti, D.d.F. (2001) Late Cenozoic sedimentary evolution in Wiens, J.J., Penkrot, T.L. (2002) Delimiting species based on DNA northeastern Pará, Brazil, within the context of sea level changes. and morphological variation and discordant species limits in spiny J S Am Earth Sci 14, 77–89. lizards (Sceloporus). Syst Biol 51, 69–91. Sanchez-Villagra, M.R., Aguilera, O.A. (2006) Neogene vertebrates Wiens, J.J., Servedio, M.R. (2000) Species delimitation in systematics: from Urumaco, Falcon State, Venezuela: Diversity and signifi cance. inferring diagnostic differences between species. Proc Roy Soc Lond J Syst Palaeontol 4, 213–220. B 267, 631–636. Sanderson, M.J. (1997) A nonparametric approach to estimating Willis, S.C., Nunes, M.S., Montana, C.G., Farias, I.P., Lovejoy, N.R. divergence times in the absence of rate constancy. Mol Biol Evol (2007) Systematics, biogeography, and evolution of the neotropical 14, 1218–1231. peacock basses Cichla (Perciformes: Cichlidae). Mol Phylogenet Santos, M.CF., Ruffi no, M.L., Farias, L.P. (2007) High levels of genetic Evol 44, 291–307. variability and panmixia of the tambaqui Colossoma macropomum Wilson, E.O. (2003) The origins of hyperdiversity. In: Wilson, E.O. (Cuvier, 1816) in the main channel of the Amazon River. J Fish Biol (ed.) Pheidole in the New World: A Dominant, Hyperdiverse Ant 71, 33–44. Genus. Cambridge, MA: Harvard University Press, pp. 13–18. Schaefer,UNCORRECTED S.A. (1997) The neotropical cascudinhos: Systematics Yang, Z., Rannala, B. (2006) PROOF Bayesian estimation of species divergence and biogeography of the Otocinclus catfi shes (Siluriformes: times under a molecular clock using multiple fossil calibrations Loricariidae). Proc Acad Nat Sci Phila 148, 1–120. with soft bounds. Mol Biol Evol 23, 212–226. Sites, J.W. Jr, Marshall, J.C. (2003) Species delimitation: A renaissance Yoder, A.D., Yang, Z.H. (2000) Estimation of primate speciation dates issue in systematic biology. Trends Ecol Evol 18, 462–470. using local molecular clocks. Mol Biol Evol 17, 1081–1090. Sites, J.W. Jr, Marshall, J.C. (2004) Operational criteria for delimiting species. Ann Rev Ecol Evol Syst 35, 199–229.

HHoorn_ch25_FPP.inddoorn_ch25_FPP.indd 441717 77/27/2009/27/2009 3:48:113:48:11 PMPM UNCORRECTED PROOF

HHoorn_ch25_FPP.inddoorn_ch25_FPP.indd 441818 77/27/2009/27/2009 3:48:113:48:11 PMPM