<<

University of Nebraska - Lincoln DigitalCommons@University of Nebraska - Lincoln

Publications, Agencies and Staff of the U.S. Department of Commerce U.S. Department of Commerce

3-7-2001

Evolution of river

Healy Hamilton

Susana Caballero

Allen G. Collins

Robert L. Brownell Jr. NOAA, [email protected]

Follow this and additional works at: https://digitalcommons.unl.edu/usdeptcommercepub

Part of the Environmental Sciences Commons

Hamilton, Healy; Caballero, Susana; Collins, Allen G.; and Brownell, Robert L. Jr., "Evolution of river dolphins" (2001). Publications, Agencies and Staff of the U.S. Department of Commerce. 111. https://digitalcommons.unl.edu/usdeptcommercepub/111

This Article is brought to you for free and open access by the U.S. Department of Commerce at DigitalCommons@University of Nebraska - Lincoln. It has been accepted for inclusion in Publications, Agencies and Staff of the U.S. Department of Commerce by an authorized administrator of DigitalCommons@University of Nebraska - Lincoln. Proceedings: Biological Sciences, Vol. 268, No. 1466 (Mar. 7, 2001), pp. 549-556 B THE ROYAL doi 10.1098/rspb.2000.1385 92 SOCIETY

Evolution of river dolphins Healy E{amiltonltsSusana Caballero2,Allen G. Coll;nsl and Robert L. BrownellJr3 lMuseumof Paleontologyand Department of IntegrativeBiology, University of California,Berkeley, CA 94720, USA 2FundacionSubarta, Carrara 24Foeste, no 3-110,Cali, Colombia 3SouthwestFisheries Science Center, PO Box 271,La jrolla, CA 92038, USA The world's river dolphins (, Pontoporia,Lipotes and Platanista)are among the least known and most endangered of all cetaceans. The four extant genera inhabit geographically disjunct river systems and exhibit highly modified morphologies, leading many cetologists to regard river dolphins as an unnatural group. Numerous arrangements have been proposed for their phylogenetic relationships to one another and to other odontocete cetaceans. These alternative views strongly aXect the biogeographical and evolu- tionary implications raised by the important, although limited, fossil record of river dolphins. We present a hypothesis of river relationshipsbased on phylogenetic analysis of three mitochondrial genes for 29 cetacean species, concluding that the four genera representthree separate, ancient branches in odonto- cete evolution. Our molecular phylogeny correspondswell with the first fossil appearances of the primary lineages of modern odontocetes. Integrating relevant events inTertiary palaeoceanography,we develop a scenario for evolution during the globally high sea levels of the Middle . We suggest that ancestorsof the four extant river dolphin lineages colonized the shallow epicontinental seas that inun- dated the Amazon, Parana, and Indo-Gangetic river basins, subsequently remaining in these extensive waterways during their transition to freshwater with the Late Neogene trend of sea-level lowering. Keywords: ; fossil; phylogeny; Odontoceti; Miocene; epicontinental seas long, independent evolutionary histories. River dolphins 1. INTRODUCTION are highly modified taxa that have more autapomorphies Four genera of toothed cetaceans comprise the peculiar than shared characters useful for determining their and poorly known 'river dolphins'. Although several affiliations (Messenger 1994). Furthermore,river dolphin marine delphinids are commonly found in rivers quite far classifications have often assumed monophyly (Simpson upstream, river dolphins are morphologically and phylo- 1945; Kasuya 1973;Zhou 1982), although some characters genetically distinct from marine dolphins and most are used to unite river dolphins, such as an elongate rostrum restricted to freshwaterecosystems. Since the first sugges- and mandibular symphysis, may be primitive for odonto- tions of their affinities were advanced in the l9th century cete cetaceans. When exisiting taxa are few and so (Gray 1863; Flower 1867), the evolutionary relationship of distinctly modified that homologous characters are diffi- river dolphins to one another and to other odontocetes cult to detect, the fossil record of the group should play has remained controversial (Simpson 1945; Kasuya 1973; an important role in resolving taxonomic relationships Zhou 1982; Muizon 1984, 1988a; Fordyce & Barnes 1994; (Gauthier etal. 1988). Messenger 1994; Rice 1998). Despite diXering in detail, There are various fossil taxa related to extant genera, recent morphological systematic studies of modern and with the exception of Lipotes.Unfortunately, the record is fossil taxa ( Muizon 1988a,c, 1994; Heyning 1989; not yet complete enough to determine key character pola- Messenger & McGuire 1998) largely corroboratedearlier rities at intermediate stages. The fossil history of river views that each extant lineage is relatively ancient and dolphins has a long and confusing treatment in the that river dolphins comprise an unnatural group. Non- literature, with many fossils described as members of monophyly of river dolphins is consistent with their taxonomic groups no longer recognized; a comprehensive highly disjunct geographical distributions (figure 1): the re-examination is needed. A robust hypothesis of the dolphin, Inia geoffrensis,and the La Plata relationships among extant lineages is critical for river dolphin, Pontoporiablainvillei, are found in South exploring the biogeographical and evolutionary implica- America; the Yangtze river dolphin, Lipotesvexillifer, and tions of river dolphin fossils. Indian river dolphin, Platanistagangetica, inhabit rivers on Higher-level molecular phylogenetic studies of ceta- opposite sides of continental Asia. Placing the four river ceans have primarily focused on the relationship between dolphin lineages within the evolutionary tree of cetaceans cetaceans and artiodactyls (Graur & Higgins 1994; can help resolve the confused state of odontocete beta Montelgard et al. 1997) and on the hypothesis of odonto- (Heyning 1989; Fordyce et al. 1985; Fordyce & cete (Milinkovitch et al. 1993; Hasegawa et al. Barnes 1994; figure 2) and refine our understanding of 1997; Messenger & McGuire 1998). River dolphins were odontocete evolution. discussed in Arnason & Gullberg's (1996) cytochrome b The difficulties of confronting river dolphin systematics phylogeny of cetaceans, which provided additional using morphological analyses may relate directly to their evidence for a distinct, though unresolved, position for Platanista.Two recent studies have specifically addressed river dolphin phylogeny using DNA sequence analysis. *Authorfor correspondence (heals(socrates.berkeley.edu). Yang & Zhou (1999) were the first to include all four

Proc.R. Soc.Lond. B (2001)268, 54>556 :549 <) 2001 The Royal Society Received4 October 2000 Accepted24 October2000 8 broadlyapproachriver fromdolphinshas within beenin everytheto epicontinental sampleprimary bothlineage extensivelyseas of ofodontocete. the Middleand

550 H. Hamilton and others Evolutionofriverdolphins

With problematicphylogenies, for which odontocetes {mboldtlana certainlyqualify, it maybe moreuseful to add taxa rather than to add characters(Hillis 1996;Graybeal 1998). Our

:) Our objective is to reconstruct the evolutionary F history of river dolphins.We begin by presentinga I [ t hypothesis of the phylogeneticrelationships of extant < Inia river dolphinsbased on a multiplemitochondrial gene } geoffrensisJ phylogeny of 29 species of cetaceans. We consider l geoffrensif biogeographicaland stratigraphicalaspects of the fossil r ( z^/ recordof river dolphinsin relationto our phylogenetic ensis | hypothesis.Integrating the palaeontologicaldata with : knownevents inTertiary palaeoceanography, we conclude ,/- with a detailedscenario for the evolutionof the world's

/ Miocene.

/ 2. MATERIALAND METHODS Pontoporiablainvillei - Our data set is comprisedof the complete cytochromeb (1140bp),partial 12S (385bp), and partial 16S (530bp) mito- - - chondrialgenes, for 29 speciesbroadly representative of each primarylineage of odontocete.In additionto sequencesavail- able from previousstudies of cetacean molecularsystematics (Milinkovitchet al. 1994; Arnason & Gullberg1996; LeDuc et al. 1999),we sequencedeither the ribosomalgene fragments and/or the completecytochrome b for non-overlappingtaxa. In all, we generated 44 new sequences (GenBank accession numbersAF334482-AF334525). We analysedsequences of Inia of knownprovenance from Brazil,Peru and Bolivia,as well as Inia from GenBank (accessionnumber X92534; Arnason & Gullberg1996), in to evaluatethe suggestionthat the Bolivian form, Inia geoXrensisboliviensis, is distinct from Inia geoXrensisgeoXrensis (da Silva 1994; Pilleri & Gihr 1977).The partial12S sequence for Lipotesvexillifer was not availablefor this analysis.The mysticeteoutgroup consists of four speciesfrom three families.The taxa in this study, with tissue source, scientificand common names, are listed at the archivedweb pages of the Universityof CaliforniaMuseum of Paleontology (www.ucmp.berkeley.edu/archdata/ HamiltonetalOl/river. html), as are the primersequences, gene sequences,and data set align- ments. Figure1. Geographicaldistribution of extant riverdolphins. Samples were obtained either by biopsy darting, from (a) IniageoXrensis humboldtiana inhabits the OrinocoRiver museumspecimens, or fromthe GeneticsTissue Archive, South- system.I.g. geoXrensis is found throughout the mainstem west FisheriesScience Center,LaJolla, CA, USA. DNA was AmazonRiver and its tributaries.I.g. boliviensisoccurs in the Amazontributaries of easternBolivia, geographically isolated extractedby standardphenol-chloroform/ethanol precipitation by severalhundred kilometres of rapids.Pontoporia blainvillei is or with the QIAamp DNA extraction kit (Qiagen, Inc., restrictedto coastalSouth Atlantic waters. (b) Lipotesvexillifer Valencia,CA, USA). After an initial 2 min denaturationat is an extremelyendangered river dolphin that occursonly in 94 °C, PCRconsisted of 35 cycles,30 s at 94 °C, 45 s at 48-52 °C the lowerand middlereaches of the YangtzeRiver. Platanista and 90s at 72°C. The productswere visualized,cleaned and minorinhabits the IndusRiver system.P. gangeticais foundin directlysequenced in both directionson an ABI 377 automated the -BrahmaputraRiver system. DNA sequencer(Applied Biosystems, Foster City, CA, USA). Sequenceswere edited with Sequencherv. 3.0 sequenceanalysis river dolphintaxa in a molecularphylogenetic analysis, software(GeneCodes Corporation, Ann Arbor,MI, USA) and but their limited data set of only 307 base pairs (bp) of aligned manuallyin BioEdit 4.7.8 (Tom Hall). Four sites of the cytochromeb gene is insufficientto addressthe phylo- ambiguousalignment in the 16Sgene wereexcluded. geny of deeplydiverging taxa. In contrast,the molecular All phylogeneticanalyses were carried out usingPAUP 4.0b3a phylogenyof Cassenset al. (2000) analysesfive genes for (Swofford2000). Tree searches were conductedwith optimality 19 cetaceanspecies, both nuclearand mitochondrial,yet criteriaof parsimonyand maximumlikelihood. Twenty replicate even this largedata set resultsin low bootstrapvalues for searcheswere made for the maximum-likelihoodtree, assuming key nodes in riverdolphin phylogeny, particularly under the HKY85model of nucleotideevolution (Hasegawa et al. 1985) the maximum-likelihoodcriterion of molecularevolution. witha transitionto transversion(Ti:Tv) ratio of 6.0 anda gamma

Proc.R. Soc.Lond. B (2001) _ PontoporiidaeDelphinidae | _ | * Delphinidae MonodontidaePontoporiidae

Evolutionofriverdolphins H. Hamiltonand others 551

morphologywith fossil taxa morphologyof extanttaxa | molecularsequences

(a) (c) (e) - Physeteridae Physeteridae Physeteridae [ - Ziphiidae Ziphiidae Squalodontidaet r-{platanistid LX | Platanistidae | | Ziphiidae _ Squalodelphidaet .. Eurhinodelphidaet _ - r Imldae Lipotidae | Pontoporiidae _ _ Iniidae Pontoporiidae

_ -Monodontidae Phocoenidae Phocoenidae -Phocoenidae Delphinidae Delphinidae Delphinidae

.b} te (f) Physeteridae Physeteridae = Mysticeti

r SZiphudae l l Ziphiidae L Physeteridae

_ E Squalodelphidaet | '1 | Platanistidae Ziphiidae Platanistidae _ Lipotidae . . Pontoporiidae Iniidae Platanistidae Iniidae _ _ e _ Lipotidae Pontoporiidae u g Monodontidae _ g Monodontidae | H g Llpotldae

7 Kentnodontidaet l Phocoenidae 1 Iniidae Phocoenidae I -r

Figure2. Alternativehypotheses of odontocetephylogeny. Some endings have been emendedto standardizetaxonomic comparisons.(a) Muizon (1988a, 1991), (b) Barnes(1990); (c) Heyning (1989), (d) Messenger& McGuire(1998); (e) Arnason & Gullberg(1996), (G) Yang & Zhou (1999).

shapeparameter of 0.2.The assumedratio of Ti:Tvand the shape ,monodontids and moderndolphins, essentially of the distributionof substitutionrates were estimated under the Muizon'sconcept of the InfraorderDelphinida (Muizon criterionof likelihoodusing trees obtainedby both neighbour 1988a, 1991).In both analyses,beaked compose joining and unweightedparsimony. Parsimony searches (with the sistergroup to (Heyning 1989). The data 1000 replicates)were carriedout with a range of differential indicatethat non-platanistidriver dolphins are the extant weighting to assess the impact of these correctionson tree representativesof early lineages that divergedfrom the topology.Two bootstrap analyses were performed, one with trees stem leading to Delphinoidea(porpoises, monodontids found by neighbourjoining (with Jukes-Cantor corrected and dolphins), supporting their ranking as separate distances)and one with treesobtained using weighted parsimony families. Our analysis suggestsInia and Pontoporiaare (transversionscounting six timesas muchas transitions).Finally, monophyleticand together form the sister group of supportindices were calculatedfor each node presentin the Delphinoidea(Muizon 1984), and suggestsa distinction weightedparsimony analysis (Bremer 1988). betweenthe Bolivianand Amazonforms of Inia. The two analysesyield contradictinghypotheses for the relation- shipswithin Delphinoidea.The maximum-likelihoodtree 3. RESULTS indicates that porpoises and marine dolphins form a The maximum-likelihoodtree and the consensusof clade, while the weighted parsimony tree groups three most parsimonioustrees are largely congruent porpoiseswith monodontids,a view recentlyadvanced (figure3). The Physeteridae,represented by Physeterand (Waddellet al. 2000). Kogia,are basalodontocetes and do not forma cladewith Ziphiidae,the beakedwhales, contradicting some classifi- cations (Fordyce1994; Muizon 1991).The long-suspected 4. DISCUSSION polyphylyof riverdolphins is supportedby the mitochon- The phylogenetic relationships of river dolphins drial sequencedata. In both trees,Platanista gangetica and suggestedby ouranalysis allows for a refinedunderstanding Platanistaminor, representing Platanistidae, are sister to of odontocetesystematics and evolution,a long-elusive the remainingodontocetes, although bootstrap support goal. Just as the extensiveadaptations involved in the for this node is low. The remainingriver dolphin taxa transitionfrom land mammalto aquaticmammal have (Litotes,Inia and Pontoporia)are paraphyleticallyarranged obscured cetacean origins, each primary odontocete at the base of a well-supportedclade that also includes lineage exhibits a suite of highly derived characters

Proc.R. Soc.Lond. B (2001) 552 H. Hamiltonand others Evolutionofriverdolphins

(b) (a) Balaena mysticetus Eschrichtiusrobustus Balaenopteraphysalus Megapteranovaeangliae Physeter catodon Kogia breviceps Kogia simus Plal;anistagangetica Platanista minor Berardiusbairdii Tasmacetusshepardii Ziphiuscavirostris Mesoplodonbidens Mesoplodon europaeus Lipotes vexillifer Pontopona blvinvillei lnia geoffrensis boliviensis Inia geoffrensis-GenBanl Inia geoffrensis-Brazfl Inia geoffrensis-Peru Delphinapterusleucas Monodonmonoceros Neophocoenaphocoenoides phocoena Lagenorhynchusalbirostris Sousa chinensis Steno bredanensis Lagenorhynchusobscurus Lissodelphisborealis Orca orca Pseudorca crassidens

Figure3. Optimaltrees under the criteriaof (a) maximumlikelihood and (b) parsimony.The maximum-likelihoodtree was obtainedby carryingout 20 replicateheuristic searches, assuming the HKY85 modelof nucleotideevolution with a transitionto transversionratio of 6.0 and a gammashape parameter of 0.2. Bootstrapvalues (derivedfrom 1000replicates of neighbour- joining searchesusing JukesWantor corrected distances) are shownat the nodes.Values less than 50 are denotedby ' < '. The tree to the right is the consensusof threemost parsimonioustrees of length5416 foundwith 1000replicate heuristic searches. Transversionswere weighted six timesas heavilyas transitions.Above each node are parsimonybootstrap values ( 1000 replicates)and Bremersupport indices, separated by a verticalbar. The rangeof transitionto transversionweighting (from equal to ten times,as well as transversionsonly, denotedby an asterisk)that yieldseach clade is reportedbelow each corresponding node. The GenBankaccession number for 'Inia-GenBank'is X92534 (Arnason& Gullberg1996). without clear evidence of sequential forms. Thus alpha settings.During episodes of low sea level, nearshoresedi- taxonomic assignments are considerablyless controversial ments are eroded, abridgingthe record.Archaic forms than higher-level systematics. River dolphins provide an disappearand more advancedgroups emerge in succes- extreme example. Although the generic designations are sive waves with no clear origins. Many fossil cetaceans not disputed, their taxonomic ranks are undecided, and are known from single specimens,numerous taxa have many possible combinations of their interrelationship been erectedon the basisof undiagnostic,isolated or frag- have been proposed (figure 2). Similarly, the phylogenetic mentarybones, and the classificationhistory of extinct affinities of the remaining odontocete lineages are also cetaceansis long and bewildering.A confidentgrasp of unresolved (Heyning 1989; Rice 1998). The placement of modernphylogeny will help clarify the relationshipsof the river dolphins among these lineages, as indicated by past to presenttaxa. our molecular analysis, suggests a resolution that is Extincttaxa assignedto the Platanistidaeare well docu- notably concordant with the first appearance of these mented, particularlyZarhachis and Pomatodelphis,long- groups in the fossil record (figure 4). beaked Middle to Late Miocene cetaceans recovered primarilyfrom shallowepicontinental sea depositsof the (a) The fossil record oSriver dolphins Atlanticcoast of NorthAmerica (Kellogg 1959; Gottfried The fossil record of pelagic is understandably et al. 1994; Morgan 1994; table 1). Possibleplatanistid limited. Fossil cetaceans are primarily recovered from relativesare Squalodelphinidaeand at leastsome members rocks that formed in nearshore and continental-shelf of (Muizon 1994; Fordyce 1994), two well- depositional environments, and only rarely from deep-sea known, extinct families of archaic, medium-sized

Proc.R. Soc.Lond. B (2001) Evolutionofriverdolphins H. Hamilton and others 553

35 30 25 20 15 10 5 0 I ,1 111 111111111 111111111 1111111111 I y L l Miocene | Pliocene | g |

fL Early | Late | Early | Middle | Late |Early|Late | O |

______Mysticeti

______Physeteridae

______Platanistidae

______Ziphiidae

Lipotidae

,, ,,,, ______Iniidae

,,,,,,,______Pontoporiidae _ _ _ _ _ Delphiriidae ------4

I s it * s ______. nono ontloae ______Phocoenidae

Figure4. Generalcorrespondence between the hypothesizedphylogeny and fossilrecord of Odontoceti.Finer dotting indicates the uncertaindates for some earliestfossil occurrences. Lipotidae is the only clade for whichfossils are not yet definitivelyknown. heterodonts.Other fossil relatives of the Platanistidae (table 1).With the placement of most previously described include membersof the Dalpiaziniidae(Muizon 1994) 'iniids' in other extinct groups (Muizon 1988b; Cozzuol and Waipatiidae(Fordyce 1994, p. 147).If these lineages 1996), the family may be regarded as a freshwater South are monophyletic,then Platanista is the sole extant American endemic. The partial skull, rostral and memberof a once-abundantand diverseclade of archaic mandibular fragments known as Goniodelphis,from the odontocetes.The side-swimming,blind and highlyendan- Early Pliocene Palmetto Fauna of central Florida, are the gered Indian river dolphinhas long been recognizedas only fossil remains outside South America that can be 'the genus. . . presentingthe greatesttotal of modifications considered plausibly as Iniidae (Morgan 1994). However, known in any cetacean' (Miller 1923, p.41). Both fossil Muizon (1988b) regarded this material as too incomplete and extant platanistidswarrant further investigation for for a confident determination. Significantly, both fossil potentialinsights into cetaceanevolution. genera clearly assigned to Iniidae, Ischyrorhynchusand The assignmentof fossil taxa within non-platanistid Saurocetes,are found far south of Inids present range, river dolphins has been misdirected by inaccurate occurring only in the fluvial Late Miocene Ituzaingo conceptsof the systematicrelationship of extant taxa. In formation of the Parana basin, Argentina (with the most earlier classifications,Inia and Lipoteswere placed possible exception of fragmentary mandibular remains togetherin Iniidae,while Pontoporia(Stenodelphis in earlier reported from Brasil; Rancy et al. 1989). The Ponto- works)was sometimesclassified within Delphinidae,the poriidae have a broader geographical and geological marine dolphins (Miller 1923).For over a century,this range. Three species of Parapontoporiahave been described conceptof Iniidaewas a repositoryfor earlydolphin-like from nearshore shallow water deposits of California and fossilodontocetes (Kellogg 1944; Rensberger 1969; Wilson Baja California (Barnes 1985). The members in this 1935).With the descriptionof (Barnes 1984, Northern Hemisphere genus have been placed in their 1985),an extinct genus consideredintermediate between own subfamily, Parapontoporiinae, based on their asym- Lipotes and Pontoporia,subsequent classifications some- metrical cranial vertices. The subfamily Pontoporiinae, times placed Lipotes in the Pontoporiidae(Fordyce & identified by symmetrical cranial vertices, is restricted to Barnes 1994). Systematicrevision and more rigorous the Southern Hemisphere. Two fossil genera have been diagnosisof fossil taxa leave the majorityof generalized described from the of southern coastal small odontocetes outside of Lipotidae, Iniidae and Peru, the Pliocene Pliopontos,very similar to Pontoporia, Pontoporiidae.The Lipotidaehave essentiallyno fossil and the geologically youngest occurrence of the family, record. A single mandibularfragment from freshwater the Middle Miocene Brachydelphis(Muizon 1983, 1988c). sedimentsin southern China, known as Prolipotesand Another fossil, the Late Miocene Pontistes,is found in the tentativelydated as Miocene (Zhouet al. 1984),cannot be Parana formation, marine sediments of the Parana basin, confirmedas a Lipotid.Both Iniidae and Pontoporiidae Argentina, underlying and adjacent to those with fossil are representedby South American fossil relatives iniids (Cozzuol 1985).

Proc.R. Soc.Lond. B (2001) 554 H. Hamilton and others Evolutionofriverdolphins

Table 1. IdentiWicationand stratigraphy offossil riverdolphins taxon location stratigraphy:formation/age reference family Platanistidae Zarhachis Maryland Calvert Formation/Middle Miocene Kellogg (1924); Gottfried etal. (1994) Pomatodelphis Florida Agricola Fauna, Bone Valley/ Kellogg (1959); Morgan (1994) Middle Miocene family Lipotidae Prolipotes(?) Southern China Miocene (?) Zhou etal. (1984) family Pontoporiidae Brachydelphis coastal Peru Pisco Formation/Middle Miocene Muizon (1988c) Pliopontos coastal Peru Pisco Formation/Early Pliocene Muizon (1983), (1984) Pontistes Argentina Parana Formation/Late Miocene Cozzuol (1985), (1996) Parapontoporia California, Mexico SanOiego/LatePliocene;Almejas/ Barnes (1984), (1985) Late Miocene family Iniidae Ischyrhorhynchus Argentina Ituzaingo Formation/Late Miocene Cozzuol (1985), (1996) Saurocetes Argentina ItuzaingoFormation/LateMiocene Cozzuol (1988), (1996) Goniodelphis(?) Florida Palmetto Fauna, Bone Valley/ Morgan (1994) Late Miocene

(b) The evolution of river dolphins Miocene would have inundated large areas of the fore- The Middle Miocene was a time of globally high sea land basin, creating a shallow marine habitat. Fossils levels, with three significantmarine trangressive-regressive have not yet been recovered from these regions, but plata- cycles recorded worldwide (Haq et al. 1987). With the nistids are known to have inhabited Miocene epiconti- resulting large-scale marine transgressions on to low- nental seas in North America (table 1; Morgan 1994; lying regions of the continents, shallow epicontinental Gottfried et al. 1994). Platanistais the only surviving seas became prominent marine ecosystems. The Indo- descendant of an archaic odontocete that ventured into Gangetic plain of the Indian subcontinent, the Amazon the epicontinental seas of the Indo-Gangetic basin, and and Parana river basins of South America, and the remained through its transition to an extensive freshwater Yangtze river basin of China are vast geomorphic systems ecosystem during the Late Neogene trend of sea-level whose fluvio-deltaic regions were penetrated deeply by regression. Although the palaeogeography of the two marine waters during high sea-level stands. The shallow river systems would suggest a history of isolation, the estuarine regions created by the mixing of riverine and genetic distance we observed in our small sample of E marine waters probably supported diverse food resources, gangeticaand E minoris surprisinglylow (figure 3). particularly for aquatic animals able to tolerate osmotic Several lines of evidence suggest Miocene marine diffierences between fresh and saltwater systems. We incursions penetrated deeply into continental South propose that the ancestorsof the four extant river dolphin America (Hoorn et al. 1995; Lovejoy et al. 1998). To the taxa were inhabitants of Miocene epicontinental seas. north, incursions were along the course of the Amazon Draining of the epicontinental seas and reduction of the river palaeodrainage (Hoorn 1994), and to the south, into nearshore marine ecosystem occurred with a Late the Parana river basin ( Cozzuol 1996). During the Miocene trend of sea-level regression, which continued highest global stand of Miocene sea levels, the Parana throughout the Pliocene, interrupted by only moderate and Amazon river basins may have been connected, and relatively brief events of sea-level rise (Hallam 1992). forming an interior seaway that divided the continent, As sea levels fell, these archaic odontocetes survived termed the Paranense Sea (Von Ihering 1927).The largely in river systems, while their marine relatives were ignored hypothesis of the Paranense Sea is supported by superceeded by the radiation of Delphinoidea. Cassens sedimentological data (Rasanen et al. 1995) and bio- et al. (2000) also noted the persistence of river dolphins geographical data from foraminifera (Boltovsky l991) and during the radiation of delphinoids. They suggest that molluscs (Nuttall 1990). The existence of the Paranense extant river dolphin lineages 'escaped extinction' by adap- Sea is consistent with the distribution of both modern tation to their current riverine habitats. All extant organ- and fossil South American river dolphin taxa. isms have escaped extinction by being adequately adapted We hypothesize that the dolphins entered the seaway to their presentcircumstances. By integrating phylogenetic, from the north, diversified within its complex fluvial- palaeoceanographicand fossil data, we provide an explicit estuarine-marine system, and colonized its farthest hypothesis for the evolution and modern distribution of reaches, to the south-west Atlantic Ocean. Lowering of riverdolphins. global sea levels drained the inland sea, separating the The Indo-Gangetic foreland basin is a broad, flat plain northern and southern river basins, and isolating the of sediment delivered throughout the Cenozoic by an taxa. Iniid ancestors remained in the immense Amazon intricate network of migrating rivers descending from the basin, which was developing its modern transcontinental tectonically dynamic Himalayan mountains (Burbank aspect with the uplift of the Venezuelan Andes and clock- et al. 1996). The increased sea levels of the Middle wise rotation of its palaeodrainage (Hoorn et al. 1995).Inia

Proc.R. Soc.Lond. B (2001) Evolutionofriverdolphins H. Hamilton and others 555 evolved during the Amazon's transformation to a fresh- their guidanceregarding technical advice. We thank D. Lind- water system of extraordinary size, diversity and abun- berg andJ. Lipps for access to the facilitiesof the Molecular dance. The Parana river basin is a fraction of the size of its PhylogeneticsLaboratory, U.C. Berkeley.The detailed com- ments of two anonymousreviewers northern counterpart.The iniid fossil genera Ischyrorhynchus greatly improved the manuscript,which also benefitedfrom the thoughtfulcomments and Saurocetes)found along the banks of the Rio Parana, ofJ. W.Vatentine. This is UCMPpublication 1733. belong to genera that disappeared with the retreat of the continental sea ecosystem. Pontoporiafollowed the marine waters receding from the Parana basin to colonize the REFERENCES nearshore coastal zone north and south of the La Plata Arnason, U. & Gullberg,A. 1996 Cytochromeb nucleotide . sequencesand the identificationof five primarylineages of Parts of eastern and southern China are low-lying extantcetaceans. Mol. Biol. Evol. 13, 407-417. deltaic regions formed of sediments deposited by the Barnes,L. G. 1984 Fossilodontocetes from the Almejasforma- area's river systems, such as the Yangtze and the Yujiang. tion, Isla Cedros,Mexico. Paleobios 42, 1-46. Significant sea-level rise would transform these regions Barnes,L. G. 1985 Fossilpontoporiid dolphins from the Pacific into shallow waterways of mixed fluvial and marine Coastof NorthAmerica. Contrih Sci. Aat. Hist. Mus. LA County 363, 1-34. origin. Several fossil locales in nearbyJapan confirm the Barnes,L. G. 1990The fossilrecord and evolutionary relationship presence of odontocetes in the western Pacific during the of thegenus Sursiops. In Thebottlenose dolphin (ed. S. Leatherwood Miocene (Ichishima et al. 1995), potential colonizers of & R. Reeves),pp. 3-26. San Oiego,CA: AcademicPress. the Asian epicontinental seas. Our scenario is consistent Boltovsky,E. 1991Ihering's hypothesis in the light of foramini- with the geographical occurrence of the mandibular frag- ferologicaldata. Lethaia 24, 191-198. ment known as Prolipotes,inland of the Yujiang river delta Bremer,K. 1988 The limits of amino acid sequencedata in in southern China. If our phylogenetic interpretation is angiospermphylogenetic reconstruction. Evolution 42, 795-803. correct, then non-platanistid river dolphins are paraphy- Burbank,O. W., Beck,R. A. & Mulder,T. 1996The Himalayan letic, and Lipotes, like Platanista, is the sole surviving forelandbasin. In Thetectonic evolution of Asia (ed. A.Yin & taxon of a deeply divergent branch in cetacean evolution. M. Harrison),pp. 149188. CambridgeUniversity Press. The ancestry of non-platanistid river dolphins might be Cassens,I. (and 12 others) 2000 Independentadaptation to riverinehabitats allowed survivalof ancient lineages.Proc. found in the progenitors of one of two well-known groups AatlAcad. Sci. USA97, 11343-11 347. of fossil cetaceans. Eurhinodelphinids were long-beaked, Cozzuol,M. A. 1985The Odontocetiof the 'Mesopotamiense'of medium-sized odontocetes, sometimes encountered as the the ParanaRiver ravines. Invest. Cetacea 17, 39-54. dominant vertebrates in Miocene marine fossil forma- Cozzuol, M. A. 1988 Una nueva especie de Saurodelphis tions. In the TarkaroolooBasin of the Lake Frome region Burmeister,1891. Anaeghiniana 25, 39-45. of Southern Australia, eurhinodelphinid fossils from Cozzuol, M. A. 1996 The record of aquatic in several distinct horizons of the Middle Miocene Namba southern South America. Munchner-Geowissenschaftliche formation record the adaptation of at least one member of Abhandlungen30, 321-342. this group to a freshwater environment (Fordyce 1983). da Silva,V. M. F. 1994 Aspectsof the biologyof the Amazon Kentriodontids were small to medium-sized odontocetes dolphinsgenus Inia and SotaliaJ2uviatilis.Ph.O. dissertation, St John's that are probably basal delphinoids (Barnes 1990). Both College,Cambridge, UK. Flower,W. H. 1867Description of the skeletonof Iniageof)=rensis. groups were widespread, and both have a fossil record Wrans.zZool. Soc. Lond. 6, 87-116. extending from the late Oligocene to the Late Miocene. Fordyce,R. E. 1983 Rhabdosteiddolphins from the Middle Significantly, some fossil specimens now classified as Miocene, Lake FromeArea, South Australia.Alcheringa 7, either kentriodontids or eurhinodelphinids were first 27-40. described as iniids ( Kellogg 1955; Rensberger 1969). Fordyce,R. E. 1994 Waipatiamaerewhenua, new genus and new Neither eurhinodelphinidsnor kentriodontidsare likely to species,an archaiclate Oligocenedolphin. Proc. San Diego Soc. have given rise to non-platanistid river dolphins, as each Nat.Hist. 29,147-176. group is diagnosed based on their distinctivemorphologies. Fordyce,R. E. & Barnes,L. G. 1994The evolutionaryhistory of Nevertheless, a small, long-beaked, polydont Oligocene whalesand dolphins. A. Rev.Earth Planet. Sci.22, 419-455. ancestor of either extinct group is a plausible progenitor of Fordyce,R. E., Barnes,L. G. & Miyazaki,N. 1995 General aspectsof the evolutionaryhistory of whales and dolphins. extant I)elphinida SensuMuizon). A re-evaluation of both IslandArc 3, 373-391 (for1994). and Eurhinodelphinidae in light of our Gauthier,J., Kluge,A. G. & Rowe,T. 1988Amniote phylogeny revised understanding of river dolphin phylogeny should and the importanceof fossils.Cladistics 4, 105-209. provide further insights into the evolution of marine and Gottfried,M. O., Bohaska,O. J. & WhitmoreJr,F. C. 1994 freshwaterodontocetes. Miocenecetaceans of the Chesapeakegroup. Proc. San Diego Soc.;Nat. Hist. 29, 229-238. This researchwas supportedby the RemingtonKellogg Fund of Graur,O. & Higgins, O. G. 1994 Molecularevidence for the the Museumof Paleontology,tJniversity of California,Berkeley, inclusionof cetaceanswithin the order Artiodactyla.Mol. and the InternationalFund for AnimalWelfare (both to H.H.). Biol.Evol. 11, 357-364. A.G.C. is supportedby NSF grantEAR-9814845. For generous Gray,J. E. 1863On the arrangementof the cetaceans.Proc. zZool. facilitationof access to tissue samples, we are indebted to Soc.Lond. pp. 197-202. B Curryand K. Robertsonat the SouthwestFisheries Science Graybeal,A. 1998 Is it better to add taxa or charactersto a Center,LaJolla, California,to N. Bernaland L Villalbaof the difficult phylogenetic problem? Syst.Biol. 47, 9-17. National Museum of Natural History, La Paz, Bolivia, and Hallam, A. 1992 Phanerozoicsea-level changes. NewYork: Columbia to J. Mead at the SmithsonianInstitution, Washington, DC. University Press. R. LeDuc and E. Archerof the SouthwestFisheries Science Haq, B. H., Hardenbol,J. & Vail, P. R. 1987 Chronology of fluc- Centerare gratefullyacknowledged for cyt b PCR primersand tuating sea levels since theTriassic. Science235, 1156-1162.

Proc. R. SOG.Lond. B (2001) 556 H. Hamiltonand others Evolutionofriverdolphins

Hasegawa,M., Kishino, H. & Yano,T. 1985 Dating of the Muizon,C. de 1983Pliopontos littoralis, un noveauPlatanistidae human-apesplitting by a molecularclock of mitochondrial Cetaceadu Pliocenede la cote peruvienne.C. R. Acad.Sci. DNA. 7. Mol.Evol. 22, 160-174. Paris,296, 625-628. Hasegawa,M., Adachi,J. & Milinkovitch,M. C. 1997 Novel Muizon,C. de 1984Les vertebresfossiles de la FormationPisco phylogenyof whalessupported by total molecularevidence. 7 (Perou)II. Les odontocetesdu Pliocene Inferieurde Sud- Mol.Evol. 44(Suppl. 1),117-120. Sacaco.Edit. Rech. Sur les Civil. Mem. 50, 1-188. Heyning,J. E. 1989 Comparativefacial anatomyof beaked Muizon, C. de 1988a Les relations phylogenetiquesdes whales (Ziphiidae)and a systematicrevision among the Delphinida.Anns. Paleon. 74, 157-227. familiesof extant Odontoceti.Contrib. Sci. Aat. Hist. Mus. LA Muizon, C. de 1988b Le polyphyletismedes Acrodelphidae, County.405, 1-64. odontoceteslongirostres du Mioceneeuropeen. Bull. Mus. Aatl Hillis, D. M. 1996 Inferringcomplex phylogenies. Aature 383, Hist.Aat. C 4 Ser.10, 31-88. 130-131. Muizon,C. de 1988cLes vertebresfossiles de la FormationPisco Hoorn,C. 1994An environmentalreconstruction of the palaeo- (Perou).III. Les odontocetesdu Miocene.Edit. Rech. Sur les Amazon River system (Middle-Late Miocene, NW Civil.Mem. 78, 1-244. Amazonia).Palaeogeo. Palaeoclim. Palaeoecol. 112,187-238. Muizon,C. de 1991A new Ziphiidaefrom the EarlyMiocene of Hoorn, C., Guerrero,J., Sarmiento,G. & Lorente,M. 1995 Washingtonstate and a phylogeneticanalysis of the major Andeantectonics as a causefor changingdrainage patterns in groupsof odontocetes.Bull. Mus. Aatl Hist. Aat. Paris34, Miocenenorthern South America. Geology 23, 237-240. 279-326. Ichishima,H., Barnes,L. G., Fordyce,R. E., Kimura,M. & Muizon,C. de 1994Are the squalodontidsrelated to the plata- Bohaska,D. J. 1995 A review of kentriodontinedolphins: nistoids?Proc. San Diego Soc. Aat. Hist. 29, 135-146. systematicsand biogeography.Island Arc 3, 486-492 (for Nuttall, C. P. 1990 A review of the Tertiary non-marine 1994). molluscanfaunas of the Pebasianand other inlandbasins of Kasuya,T. 1973 Systematicconsideration of Recent toothed northwesternSouth America. Bull. Br. Mus. Aat. Hist. Geol. 45, whales based on the morphologyof the tympano-periotic 165-371. bone.Sci. Rep. Whales Res. Inst. 24, 87-108. Pilleri,G. & Gihr,M. 1977Observations on the Bolivian(Inia boli- Kellogg, A. R. 1924 A fossil from the Calvert viensisd'Orbigny, 1834) and the Amazonian buSeo (Inia Formationof Maryland.Proc. US. AatlMus. 63, 1-14. geofJrrensisde Blainville, 1817) with descriptionof a new sub- Kellogg,A. R. 1928The historyof whales theiradaptation to species(Iniageofjrrensishumboldtiana). Invest.Cetacea 8, 11-76. life in the water.Q; Rev. Biol. 3, 29-76, 174-208. Rancy,A., BoquetinVillanueva,J., Pereirade Souza Filho,J., Kellogg,A. R. 1944 Fossilcetaceans from the FloridaTeriary. Santos,J. c. R. & Negri, F. R. 1989 Lista preliminarda Bull.Mus. Comp. Zool. 94, 433-471. faunado Neogenoda regiaoriental. Ameghiniana 26, 249. Kellogg,A. R. 1955Three Mioceneporpoises from the Calvert Rasanen, M., Linna, A., Santos,J. & Negri, F. 1995 Late Cliffs,Maryland. Proc. US. AatlMus. 105, 101-154. Miocene tidal deposits in the Amazonianforeland basin. Kellogg,A. R. 1959Description of the skullof Pomatodelphisinae- Science269, 386-389. qualis.Bull. Mus. Comp. Z°°l 121,1-26. Rensberger,J. M. 1969A new iniid cetaceanfrom the Miocene LeDuc, R. G., Perrin,W. F. & Dizon, A. E. 1999 Phylogenetic of California.Univ. Cal.it: Publ. Geol. Sci. 82, 1-43. relationshipsamong the delphinidcetaceans based on full Rice, D. W. 1998Marine mammals of theworld. Special Publication cytochromeb sequences.Mar. Mamm. Sci. 15, 619-648. no.4. Lawrence,KS: MarineMammal Society. Lovejoy,N. R., Bermingham,E. & Martin,A. P. 1998 Marine Simpson,G. G. 1945The principlesof classificationand a classi- incursionsinto SouthAmerica. Aature 396, 421-422. ficationof mammals.Bull. Am. Mus. Aat. Hist. 85, 1-350. Messenger,S. 1994 Phylogeneticrelationships of platanistoid Swofford,D. L. 2000 PAUP*:Phylogenetic analysis using parsimony river dolphins:assessing the significanceof fossil taxa. Proc. ( *andother methods). Sunderland, MA: SinauerAssociates. SanDiego Soc. Aat. Hist. 29, 125-133. Von Ihering,H. 1927Die Geschichtedes Atlantischen Ozeans. Jena, Messenger,S. & McGuire,J. 1998 Morphology,molecules, and Germany:Gustav Fisher. the phylogeneticsof cetaceans.Syst. Biol. 47, 90-124. Waddell,V., Milinkovitch,M. C., Berube, M. & Stanhope, Milinkovitch,M. C., Orti, G. & Meyer,A. 1993Revised phylo- M. J. 2000 Molecular phylogeneticexamination of the geny of whales suggestedby mitochondrialribosomal DNA Delphinoideatrichotomy. Mol. Phylogenet. Evol. 15, 314-318. sequences.Aature 361, 346-348. Wilson, L. E. 1935 Miocene marine mammals from the Milinkovitch,M. C., Meyer, A. & Powell, J. R. 1994 BakersfieldRegion, California.Peabody Mus. Aat HistBull. 4, Phylogenyof all majorgroups of cetaceansbased on DNA 1-143. sequencesfrom 3 mitochondrialgenes. Mol. Biol. Evol. 11, Yang,G. & Zhou, K. 1999A studyon the molecularphylogeny 939-948. of riverdolphins. Acta Theriol. Sinica 19, 1-9. Miller,G. S. 1923The telescopingof the cetaceanskull. Smith. Zhou, K. 1982Classification and phylogenyof the Superfamily Misc.Coll. 76, 1-70. Platanistoidea,with notes on evidenceof the monophylyof Montgelard, C., Catzeflis, F. M. & Douzery, E. 1997 the Cetacea.Sci. Rep. Whales Res. Inst. fiokyo 34, 93-108. Phylogeneticrelationships of artiodactylsand cetaceansas Zhou, K., Zhou, M. & Zhao, Z. 1984 First discoveryof a deducedfrom the comparisonof cytochromeb and 12SrRNA Tertiaryplatanistoid fossil from Asia. Sci.Rep. Whales Res. Inst. mitochondrialsequences. Mol. Biol. Evol. 14, 550-559. 35, 173-181. Morgan,G. S. 1994 Miocene and Pliocene marine faunas from the Bone Valley formationof central Florida. As this paperexceeds the maximumlength normally permitted, Proc.San Diego Soc. Aat. Hist. 29, 239-268. the authorshave agreed to contributeto productioncosts.

Proc.R. Soc.Lond. B (2001)