discovery and phylogenetic analysis of a riverine …

597

q 2002 The Society for the Study of Evolution. All rights reserved.

Evolution, 56(3), 2002, pp. 597–616

DISCOVERY AND PHYLOGENETIC ANALYSIS OF A RIVERINE SPECIES FLOCK OFAFRICAN ELECTRIC FISHES (MORMYRIDAE: TELEOSTEI)

JOHN P. SULLIVAN,1,2 SEBASTIEN LAVOUE,3,4 AND CARL D. HOPKINS1,5

1Department of Neurobiology and Behavior, Cornell University, Ithaca, New York 148532E-mail: [email protected]

3Museum National d’Histoire Naturelle, Ichtyologie Generale et Appliquee; 43 rue Cuvier 75005, Paris, France4E-mail: [email protected]

5E-mail: [email protected]

Abstract. The evolution of species-specific mate recognition signals is of particular interest within speciose mono-phyletic groups with restricted distributions (known as ‘‘species flocks’’). However, the explosive nature of speciationin these clades makes difficult the reconstruction of their phylogenetic history. Here we describe a species flock ofriverine mormyrid fishes from west-central Africa in which electric signals may play a role in the reproductive isolationof sympatric species. In our recent field collections, totaling more than 1400 specimens from many localities, werecognize 38 forms that are distinct in their morphologies and electric organ discharge (EOD) characteristics. Of these38, only four clearly correspond to described species. Here we treat these forms as operational taxonomic units (OTUs)in a phylogenetic analysis of cytochrome b sequence data from a sample of 86 specimens. We examined support inthe molecular data for the monophyly of these 38 OTUs considered together, the monophyly of each phenotypicallydelimited OTU considered individually, and for relationships among OTUs congruent with those inferred from thedistribution of morphological and EOD character states. Trees obtained by both maximum-parsimony and maximum-likelihood analyses, rooted with sequence data from outgroup taxa, provide evidence for the monophyly of these 38OTUs with respect to other mormyrid fishes. The small genetic distances between many distinct forms suggest theirrecent divergence. However, in many instances the cytochrome b tree topology fails to support the monophyly ofindividual OTUs and close relationships between OTUs that are similar in morphology and EOD characteristics. Inother cases, individuals from distinct OTUs share identical or nearly identical haplotypes. Close examination of thesecases suggests that unnatural OTU definition is not the sole cause of this pattern, and we infer an incongruence betweenthe mitochondrial gene tree and the organismal phylogeny caused by incomplete mitochondrial lineage sorting and/or introgression across forms. The apparently rapid diversification in this clade of riverine electric fishes and theproblems associated with recovering a meaningful species-level phylogeny from mitochondrial data parallel findingsin other species flocks. Selection on EOD waveforms as mate recognition signals may be involved in the radiationof these fishes. This is the first description of a freshwater fish species flock from a riverine, as opposed to a lacustrine,environment.

Key words. Cytochrome b, electric fish, phylogeny, speciation, species flock.

Received November 27, 2000. Accepted December 7, 2001.

Species-specific signals often mediate prezygotic repro-ductive isolation between sympatric, closely related species.The evolution of such signals and their role in the origin andmaintenance of species boundaries is frequently investigatedin cases of ‘‘species flocks’’ in which several to many specieshave diversified from a single common ancestor in a geo-graphically restricted area, often over an evolutionarily shortperiod of time. Examples include the haplochromine cichlidfish species flocks of the African Great Lakes in which visualcolor patterns and displays maintain reproductive isolationamong species (Lowe 1953; Fryer and Iles 1972; Seehausenet al. 1997; van Oppen et al. 1998; Knight and Turner 1999)and the Galapagos finch, Hawaiian picture-wing Drosophila,and Hawaiian cricket species flocks in which sounds largelyserve this function (Hoy et al. 1988; Otte 1994; Grant andGrant 1996; Shaw 1999). Species flocks are attractive to stu-dents of signal evolution and speciation because speciesboundaries are of recent origin (or in some cases still incom-plete) and because rapid diversification has produced a largenumber of parallel cases from which to observe general pat-terns. Knowledge of species relationships is critical for signalevolution studies, and yet the very features that make speciesflocks compelling can complicate the analysis of their internalphylogenetic structure. In this paper we describe the discov-ery and a preliminary phylogenetic analysis of a previously

unrecognized species flock of freshwater electric fishes fromAfrica in which simple electric signals may play an importantrole in the reproductive isolation of sympatric species.

The African weakly electric fish superfamily Mormyroidea(families Mormyridae 1 Gymnarchidae) is a remarkablemodern radiation from within the superorder Osteoglosso-morpha, one of the oldest and phylogenetically most basalgroups of extant teleosts (Greenwood 1973; Patterson andRosen 1977; Lauder and Liem 1983; Arratia 1997). Mor-myroids have diversified in African freshwater riverine en-vironments, whereas other osteoglossomorph lineages haveundergone a general reduction in their global diversity anddistribution since the Early Tertiary (Li 1997). Although re-stricted to the African continent, more than 200 living speciesof mormyroids are recognized (Daget et al. 1984); the re-maining living osteoglossomorphs, distributed on all conti-nents except Europe and Antarctica, comprise just 18 species(Nelson 1994). Uniquely among osteoglossomorphs and like-ly related to their evolutionary success, all mormyroids elec-trolocate and communicate by means of specialized electricorgans and receptors, enabling them to effectively exploit thenocturnal environment.

In this paper we present evidence based on recent collec-tions from Gabon in west-central Africa that this mormyroidradiation is continuing explosively within a monophyletic

598 JOHN P. SULLIVAN ET AL.

group we refer to as the ‘‘Gabon-clade Brienomyrus.’’ Thisgroup includes Paramormyrops gabonensis Taverne 1971,several described species currently placed in the genus Brien-omyrus, and many taxonomically unrecognized forms pos-sessing distinct electric organ discharge (EOD) waveformsand morphologies. We are interested in the evolutionary his-tory of species-specific EOD and electric organ character-istics in these fishes and the role of electric communication,if any, in the origin and maintenance of species diversity.Comparative and experimental investigations into these is-sues require better knowledge of the number of Gabon-cladeBrienomyrus species, their distributions, and phylogenetic re-lationships. Here we undertake a preliminary study of mor-phological and EOD variation to estimate species-level di-versity and a complimentary study of genetic variation amongthe identified groups. We assess congruence between groupsthat we recognize by phenotype and groups diagnosed byDNA sequence characters. Specifically, we wish to know ifsequence data will (1) demonstrate that all phenotypes thatwe presume to belong to the Gabon-clade identified in ourearlier studies represent a monophyletic group; (2) supportthe monophyly and thereby the potential species status ofeach phenotype considered individually (when sequences ofmultiple individuals are available); and (3) support relation-ships among identified groups congruent with the distributionof several presumed phylogenetically informative phenotypiccharacters. If agreement is good between the phenotypic andgenetic data, we will use the molecular phylogeny to lookfor patterns of EOD and electric organ character evolution.

Taxonomy and Previous Phylogenetic Studies of the GeneraBrienomyrus and Paramormyrops

Taverne (1971) established the genus Brienomyrus as part ofhis osteology-based taxonomic revision of the Mormyridae, butwithout explicit reference to any uniquely shared derived char-acters (synapomorphies). Within the genus, he recognized twosubgenera: B. (Brienomyrus) and B. (Brevimyrus). Currently,there are nine valid species within the first subgenus (B. bra-chyistius, B. longianalis, B. sphekodes, B. kingsleyae, B. curvif-rons, B. longicaudatus, B. batesii, B. tavernei, and B. hopkinsi)and a single species (B. niger) in the latter (Alves-Gomes andHopkins 1997; Teugels and Hopkins 1998).

Taverne et al. (1977a) later established the genus Para-mormyrops, describing the type species, P. gabonensis, fromthe Ivindo River of Gabon and referring to this genus Mar-cusenius jacksoni Poll, a species from the upper ZambesiRiver basin of Angola. Taverne did not include P. gabonensiswithin Brienomyrus apparently because of the presence of anossified lateral ethmoid in this species, a bone absent in theBrienomyrus specimens he examined, but which we find isintermittently present as a small ossified or cartilaginous el-ement in some Brienomyrus specimens (pers. obs.).

Recent studies have demonstrated the polyphyly of thegenus Brienomyrus with several molecular datasets (Alves-Gomes and Hopkins 1997; Lavoue et al. 2000; Sullivan etal. 2000). S. Lavoue, J. P. Sullivan, and C. D. Hopkins (un-publ. ms.) performed an unweighted parsimony analysis onthe combined data used in these previous studies (from themitochondrial 12S, 16S, cytochrome b, and nuclear RAG2

genes) with sequence data from two introns in the nuclearS7 gene for 38 species belonging to 17 nominal genera withinthe subfamily Mormyrinae. The single most parsimonioustree that results (Fig. 1) indicates that Brienomyrus (Brien-omyrus) brachyistius, the type species of the genus, is notclosely related to any other included Brienomyrus species,but is instead the sister group to Isichthys henryi and that thispair together is the sister group to species of Mormyrus. Quiteseparate from this clade, Brienomyrus (Brevimyrus) niger isweakly supported as the sister species to Hyperopisus bebeat the base of a large clade containing species of Marcusenius,Hippopotamyrus, Gnathonemus, and Campylomormyrus.

A third clade containing nominal Brienomyrus species (B.hopkinsi, B. longicaudatus) in addition to P. gabonensis andan undescribed species (VAD in this study) appears as thesister group to Marcusenius ntemensis. These species (in-cluding M. ntemensis) are endemic to a particular region oflower Guinea: the Ogooue, Ntem, and Woleu/Mbini Riverbasins of Gabon, southern Cameroon, and Equatorial Guinea.The sister group of these species—Boulengeromyrus knoepf-fleri, Ivindomyrus opdenboschi, and Pollimyrus marchei—areall endemic to the same region (Daget et al. 1984; KamdemToham 1998). The monophyly of this third clade of Brien-omyrus species plus P. gabonensis is supported by high boot-strap and decay index values in this analysis and by a unique22-bp inversion in the first intron of the S7 gene (S. Lavoue,J. P. Sullivan, and C. D. Hopkins, unpubl. ms.).

In this study, we are concerned with members of this third,taxonomically unrecognized clade. Despite their demonstrat-ed remote relationship to the type species of Brienomyrus,the inclusion of P. gabonensis, and their wide distribution inthe lower Guinea and Congo ichthyofaunal regions, we in-formally refer to them as the ‘‘Gabon-clade Brienomyrus’’because they dominate the mormyrid fauna in our collectionsfrom this country. We have begun morphological studies tosearch for phenotypic synapomorphies of this group, and spe-cies descriptions and a revision of their taxonomy are un-derway (C. D. Hopkins, G. G. Teugels, R. J. Rundell, and J.P. Sullivan, unpubl. ms.).

The Gabon-Clade Brienomyrus

Gabon-clade Brienomyrus species range in adult size fromabout 100 mm to 250 mm standard length. They have relativelyelongate, moderately compressed bodies; rounded nontubularsnouts; and small terminal to somewhat subterminal mouthsbearing five to nine (upper jaw) and six to nine (lower jaw)bicuspid, pincerlike teeth. Dorsal and anal fins are set far backon the body and are roughly symmetrical above and belowthe midline. They have fleshy, somewhat bulbous chins; smallbut functional eyes; and are light gray or light brown to nearblack, with little patterning of pigmentation on the body orfins. Like other mormyrids, they are nocturnally active and areprobably benthic foragers of insect larvae (Winemiller andAdite 1997). General uniformity of jaw structure and dentitionsuggests little trophic divergence among them, although stom-ach contents have not been compared. In the forested regionsof Gabon, they are found in a wide variety of flowing waterhabitats, from the shallowest stream headwaters to rocky sub-strates in the deepest portion of large rivers. Species com-

599SPECIES FLOCK OF AFRICAN ELECTRIC FISHES

FIG. 1. Phylogenetic relationships among 36 species belonging to 17 nominal genera of the Mormyrinae as reflected by the mostparsimonious tree recovered from an unweighted analysis of the combined molecular sequence datasets from 12S, 16S, cytochrome b,RAG2, and S7 introns 1 and 2 (4256 aligned bp, 939 informative characters, CI 5 0.51, RI 5 0.62). The genus Brienomyrus is shownto be polyphyletic. The clade of species from Gabon is distinct from Brienomyrus (Brienomyrus) brachyistius, the type species ofBrienomyrus Taverne, and from Brienomyrus (Brevimyrus) niger. The numbers above nodes are bootstrap values (shown if above 50%);those below nodes are Bremer decay indices (adapted from S. Lavoue, J. P. Sullivan, and C. D. Hopkins, unpubl. ms.). The monophylyof the Mormyrinae and the sister group relationship between Myomyrus macrops and remaining Mormyrinae was established by Lavoueet al. (2000) and by Sullivan et al. (2000).


position differs among these habitat types and from region toregion. Gabon-clade Brienomyrus are frequently the numeri-cally most abundant fish in stream inventories. In some lo-calities, five or six distinct forms, presumably different species,are found together with species of other mormyrid genera(pers. obs.). Little is known regarding the life history, ecologyand mating system of these fishes.

The Mormyrid Electric Organ and Electric OrganDischarge

Mormyrids generate electric discharges using four parallelcolumns of electrocytes in the caudal peduncle (Schlichter1906; Lissmann 1951; Bennett and Grundfest 1961; Bennett1970, 1971). Each electrocyte of the electric organ generatesa pulsating waveform identical to the overall EOD waveformrecorded externally to the fish. Among species, EOD pulsesvary in duration from a fraction of a millisecond to 10 msecand have an amplitude of 100 mV to several hundred mil-livolts. These EODs are repeated at irregular intervals of 10to 100 times per second. The waveform of the discharge iscontrolled by the structure and physiology of the electrocytesmaking up the electric organ (Bennett 1971; Bass 1986b),whereas the rhythm of the discharges is regulated by activityof a pacemaker region in the medulla.

Several studies of mormyrid species assemblages havedocumented intraspecific stereotypy and interspecific differ-ences in EOD waveforms (Hopkins 1980, 1981, 1986a; Bass1986a,b; Bass et al. 1986). Among species, EODs differ induration, number, and amplitude of peaks, polarity, and waveshape. In many species, male and female EODs differ duringthe breeding season (Hopkins 1980; Hopkins and Bass 1981;Bass and Hopkins 1983; Kramer 1997). Adult male EODsare typically longer in duration than those of females. TheseEOD differences have behavioral relevance. In a study onone Brienomyrus species from Gabon, it was demonstratedthat males discriminate between the EODs of conspecific andheterospecific females and between those of conspecificmales and females based on temporal cues in the EOD, thatis, the duration and shape of the EOD waveform (Hopkinsand Bass 1981; Hopkins 1986b).

The relationship of certain species-specific EOD waveformcharacteristics to electrocyte anatomy is well understood.Electrocytes in mormyrids are flattened disk-shaped cells thatare electrically active on both anterior and posterior faces.A complex network of electrically active tube-shaped stalks,which is an outgrowth of each electrocyte, receives neuralinnervation and transmits electrical excitation from the syn-apse to the disk faces. Innervation is on the thickest portionof the stalks near the center of the electrocyte disk. The stalksystem then branches repeatedly to form ever finer stalkletsthat eventually merge with the posterior face of the electro-cyte. In some species the stalks penetrate through the disksurface of the electrocyte before fusing with the noninner-vated face on the opposite side of the cell. In other species,stalks are nonpenetrating, and the stalklets fuse on the in-nervated face of the cell. Species with nonpenetrating stalkelectrocytes produce EODs with only two phases, or peaks(Fig. 2B, left), whereas species with penetrating stalk elec-trocytes have EODs with an additional initial phase (Fig. 2B,

right). The complex anatomy of these electrocytes and thefunctional relationship between the stalks and the electrocytefaces in generating EOD waveforms have been reviewed byBennett (1971), Bass (1986a), Alves-Gomes and Hopkins(1997), and Sullivan et al. (2000).

In a previous study (Sullivan et al. 2000), we generated aphylogenetic hypothesis from mitochondrial and nuclear se-quence data for all major groups of mormyrids. Unweightedparsimony optimization of this electrocyte character on thetree, in conjunction with observations of electrocyte ontog-eny, led us to hypothesize that nonpenetrating electrocytestalks are the primitive condition for the family Mormyridae,and that penetrating stalks may have evolved only once inthe common ancestor of the large subfamily Mormyrinae,followed by multiple independent reversals to the nonpene-trating condition. Whereas both forms of electrocyte arefound within the Gabon-clade Brienomyrus, this analysis in-dicated that penetrating stalk electrocytes are the primitivecondition for the Gabon-clade Brienomyrus as well as for theoutgroup taxa used in this study.

MATERIALS AND METHODS

Specimens Used in This Study

We collected and recorded the electric signals of more than1400 specimens that we identified as Gabon-clade Brienomyrusduring field trips to Gabon and the Central African Republicbetween 1998 and 2000. Field sites are listed in Table 1 andmapped in Figure 3. Methods of capture included funnel trapsbaited with earthworms lowered to river and stream bottoms,hook and line, cast netting, rotenone exposure followed by re-suscitation, and dip netting following localization of fishes’EODs with handheld electrode-amplifier units.

To record a fish’s EOD, we placed each specimen indi-vidually into a plastic container with water from the fieldsite. An electrode was positioned at least 10 cm from eitherend of the fish and oriented parallel to its body axis. EODswere amplified (CWE BMA-831/XR Bioamplifier, CWE Inc.,Ardmore, PA; DC settings, 0–50 kHz bandwidth), capturedby a DaqBook (IoTech, Inc., Cleveland, OH) analog to digitalconverter (16 bits at 100 kHz) and recorded onto the harddisk of a laptop computer.

We euthanized the fishes by overdose of the anestheticMS222. Tissue samples were subsequently removed from thedorsal musculature with a clean scalpel blade and preservedin 90% ethanol or a saturated NaCl solution containingDMSO and EDTA (Seutin et al. 1991). All specimens werefixed in phosphate-buffered 10% formalin, later transferredto 75% ethanol, and catalogued at the Cornell UniversityMuseum of Vertebrates (CU) or the American Museum ofNatural History (AMNH). We attached permanent tags to allspecimens bearing a unique number by which specimen, tis-sue sample, and EOD recording are linked in our records.All animal procedures followed National Institutes of Healthguidelines under a protocol approved by the Cornell Uni-versity Institutional Animal Care and Use Committee.

Diagnosis of Operational Taxonomic Units and Choice ofSpecimens for Sequencing

Each specimen was assigned to an operational taxonomicunit (OTU) after examination of its external morphology and


FIG. 2. Operational taxonomic units (OTUs) are diagnosed using both external morphology and electric organ discharge (EOD) char-acteristics. (A) Tracings of the dorsal profiles of the heads of four Gabon-clade Brienomyrus OTUs illustrate the difference betweenblunt snouts (top) and sharp snouts (bottom). The electrocytes in the electric organ have nonpenetrating stalks with posterior innervation(type NPp, left) or penetrating stalks with anterior innervation (type Pa), as illustrated by a schematic of a single electrocyte in sagittalview. (B) Four types of EODs recorded from each of the OTUs in (A). Individuals with type NPp electrocytes produce EODs with onlytwo phases: peak P1 and peak P2. OTUs with penetrating stalks (type Pa) have EODs with an initial head-negative phase (P0) in additionto P1 and P2. For each OTU, multiple EODs have been normalized to the same peak-peak amplitude and superimposed with head positivityupward. (C) Within each category of snout and electrocyte type, we recognize distinct OTUs on the basis of more subtle, but consistent,differences in morphology and EODs. Illustrated are three allopatrically distributed sharp-snouted forms with Pa electrocytes. The first,BEN, from the Ivindo River, has a relatively short EOD compared to the others. SP4 from the Louetsi River has a similar morphologyto the first, but slightly longer EOD duration; SP7 from the Ntem River differs from the other two in head shape and in EOD duration.

EOD waveform, recorded in the field. OTUs that did notcorrespond to described species were given unique, three-character alphanumeric codes. We base some of our OTUson the natural groups derived from a principal componentanalyses (PCA) of morphometric, meristic, and EOD data.We measured 29 linear morphometric characters and 12 me-

ristic counts of fin rays, scales, and teeth on 415 preservedspecimens. These measures were then subjected to separatePCAs to look for natural clusters of individuals. We alsomade quantitative measures of their signals by determiningthe time of occurrence, voltage, and slope at as many as 10different landmark points on each EOD. For each individual,


TABLE 1. List of all 89 specimens sequenced in this study, organized by operational taxonomic unit (OTU). Included is a brief explanationof the OTU abbreviation, the total number of voucher specimens in the field samples, the number of individuals sequenced, the collectionlocalities, and the GenBank, museum, and individual specimen numbers. NPp, electrocyte with nonpenetrating stalk, posterior innervation; Pa,electrocyte with penetrating stalk, anterior innervation.

OTU OTU description No. Seq. Locality Genbank no. Museum no.Specimen

no.

BEN

Brienomyrus curvifronsB. hopkinsiB. longicaudatus

manuscript abbr.

described speciesdescribed speciesdescribed species

20

5524

5

2

111

Ivindo basinIvindo River

AF477450AF477455AF477469AF201575AF201576

CU78338CU78344CU81661CU78352CU78355

20182295205022852289

BN1

BN2

blunt, NPp, 1

blunt, NPp, 2

87

53

4

1

Louetsi/Ngounie

Upper Ogooue basin

AF477458AF477459AF477432AF477457AF477431

CU84579CU84566CU80521CU84579CU80474

26822718312426813542

Boulengeromyrus knoepffleriBON

described speciesmanuscript abbr.

744

14

Ivindo RiverLouetsi/Ngounie

AF201573AF477470AF477471AF477472AF477473

CU79692CU84649CU84649CU80310CU84567

22482656265929792711

BP1 blunt, Pa, 1 210 9 Coastal/MayumbaLouetsi/Ngounie

Lower Ogooue basinMouvanga/Ngounie

AF477478AF477460AF477400AF477403AF477444

CU80520CU84659CU80341CU80320CU84583

33582684319528182530

BP4 blunt, Pa, 4 1 1

Okano basinWoleu basinUpper Ogooue basin

AF477443AF477445AF477441AF477477AF477442

CU80355not cataloguedCU81308CU80892CU80511

30162704163837713389

BP5BP6BP7BX1CAB

blunt, Pa, 5blunt, Pa, 6blunt, Pa, 7blunt, mixed, 1manuscript abbr.

4617

43073

11113

Okano basinCoastal/Pointe MbiniIvindo River

AF477468AF477430AF477424AF477479AF477466

CU80862CU80476CU81309CU81264CU80816

23223547163440182116

IN1IP1Ivindomyrus opdenboschi

intermediate, NPp, 1intermediate, Pa, 1described species

15

28

111

Ntem basin

Lower Ogooue basinNtem basinIvindo River

AF477423AF477422AF477439AF477421AF477480

CU80586CU80893CU80591CU80586CU81668

17893848184416922184

LIBLISMAG

from Librevillemanuscript abbr.manuscript abbr.

152569

114

Coastal/LibrevilleUpper Ogooue basinIvindo River

Ntem basin

AF477446AF477427AF477452AF477451AF477415

CU80867CU81090CU78326CU78323CU80904

24273366229721683945

Marcusenius ntemensis

NGO

described species

from Ngounie River

21

11

2

2

Ntem RiverIvindo RiverNtem RiverLouetsi/Ngounie

AF477416AF201593AF477418AF477463AF477464

CU80902CU79706CU80723CU84665CU84665

39992186161627082710

NZO

OFF

from Nzoundou

from Offoue River

4

16

2

2

Ngounie River

Louetsi/NgounieUpper Ogooue River

AF477462AF477461AF477447AF477402

CU84661CU84661CU84667CU80526

2553255226433394

PARParamormyrops gabonensis

SANSN2

manuscript abbr.described species

from Sangha Riversharp, NPp, 2

4231

852

12

13

Upper Ogooue basinIvindo RiverNtem RiverSangha River basinLouetsi/Ngounie

AF477419AF201603AF477425AF477467AF477437

CU80934CU79702CU80713AMNH231046CU80299

34612048398024852969

SN3 sharp, NPp, 3 11 3Upper Ogooue RiverMouvanga/Ngounie

AF477438AF477435AF477465AF477456AF477436

CU84664CU80458CU84603CU84603CU80356

25953415260626193027

SN4

SN7SN8SP2

sharp, NPp, 4

sharp, NPp, 7sharp, NPp, 8sharp, Pa, 2

76

21

40

2

116

Upper Ogooue River

Ntem RiverLouetsi/Ngounie

AF477434AF477433AF477428AF477420AF477398

CU80458CU80463CU80496CU80928CU81312

33963465366616113203


TABLE 1. Continued

OTU OTU description No. Seq. Locality Genbank no. Museum no.Specimen

no.

Mouvanga/NgounieAF477399AF477405AF477404AF477406AF477407

CU84571CU84582CU80231CU84578CU84602

26722638301325422585

SP4 sharp, Pa, 4 72 7 Louetsi/Ngounie

Mouvanga/Ngounie

AF477408AF477449AF477448AF477412AF477409

CU80305CU84602CU84602CU84658CU80358

29952673267125923014

SP6SP7SP8

sharp, Pa, 6sharp, Pa, 7sharp, Pa, 8

1932

8

111

Upper Ogooue RiverNtem basinUpper Ogooue River

AF477411AF477410AF477413AF477414AF477417

CU80357CU80337CU80485CU80877CU80488

30263342365839663657

SZA

TEN

manuscript abbr.

manuscript abbr.

80

61

4

3

Ivindo RiverNtem basinOkano basinWoleu RiverIvindo River

AF477475AF477440AF477429AF477401AF477453

CU80848CU80881CU80929CU80884CU80809

20083839163337882011

VAD

TOTALS: OTUs: 41

manuscript abbr.

Vouchers:

42

1457

3

89

Ntem basinIvindo River

Woleu basin

AF477454AF477426AF477474AF201578AF477476

CU80807CU81311CU79704CU79740CU80888Sequences:

21913850210524253814

89

FIG. 3. Map of west-central Africa with the collection localities of the specimens of the Gabon-clade Brienomyrus indicated as closedcircles. All field sites are in the either Gabon or the Central African Republic. Locality names indicated here identify the collection basinfor each specimen listed in Table 1.


→

FIG. 4. Photographs of live or preserved specimens from all 38 OTUs recognized in this study and their electric organ discharge (EOD)waveforms, organized by general snout shape and electrocyte type. Also shown are photographs of the holotypes of three describedspecies in this group that we could not confidently associate with those in our collections. All EODs are plotted on the same time basewith head positivity upward. For OTUs with pronounced sex differences in EODs, both male and female waveforms are shown.

we subjected the separate measures to PCA. Of the 16 OTUsderived from this analysis, four are described species ofBrienomyrus and 12 are being prepared for species descrip-tions in separate publications (C. D. Hopkins, G. G. Teugels,R. J. Rundell, and J. P. Sullivan, unpubl. ms.; G. G. Teugelsand C. D. Hopkins, unpubl. ms.).

Additional OTUs were recognized without PCA and withless formal morphological analysis than would be requiredfor the preparation of species descriptions. This consisted ofcareful side-by-side visual comparison of the external mor-phology of all specimens in the laboratory (including thosepreviously treated by PCA). We also examined and comparedthe EOD waveforms of each specimen assigned to an OTUby overlaying them on a computer screen with similar EODsin our database. Few of these OTUs are diagnosed by single,unique phenotypic character states. However, within eachbroad class of snout shape and electrocyte type (Fig. 2A),covariance of variable phenotypic features among these spec-imens such as relative head width, forehead slope, eye size,degree of lower jaw protrusion, caudal peduncle width, scalecounts, and EOD waveform characteristics is not random, butusually delineates discrete and recognizable forms betweenwhich we do not observe intermediates. New OTUs werecreated when we judged specimens to differ consistently inthe combination of these characteristics from all those pre-viously examined. Because in most cases both male and fe-male specimens were available (sex can be determined easilyin adults from the appearance of the anal fin), expected sexualdimorphisms in morphology and EOD waveforms could beaccounted for in these comparisons.

We suspect that many, if not most, of these additionalOTUs will merit species status upon more detailed and formalstudy. For the purposes of this study, however, the avoidanceof recognizing polyphyletic assemblages as OTUs was ofmore importance than determining placement of the specieslevel, which would be required in a work of taxonomy. Forexample, the relatively subtle differences we used to distin-guish the three allopatric OTUs in Figure 2C could be in-terpreted as evidence for three separate species, or alterna-tively as geographic variation among disjunct populations ofa single widespread species. When differences were consis-tent, as in this case, we recognized separate OTUs. Our ex-pectation was that if conspecific populations were dividedinto multiple OTUs, sequences from these would cluster to-gether in the phylogenetic analysis. In these cases, subsequenttaxonomic studies could determine placement of the specieslevel. We recognized that such potential splitting would beless problematic for the phylogenetic analysis than wouldmistaken lumping of nonsister lineages. Our procedure forrecognizing smallest diagnosable clusters conforms in prin-ciple to the phylogenetic species concept of Cracraft (1983),although it is not our opinion that species as they are rec-ognized in taxonomy must always be such.

From a total of 1450 voucher specimens (Table 1), weselected 85 individuals for sequencing representing all iden-tified OTUs. When possible, we sequenced two or more rep-resentatives of each OTU from different populations. In somecases, we sequenced additional specimens of an OTU if thefirst two sequences failed to form a monophyletic group ina preliminary parsimony analysis.

Choice of Molecular Marker

We chose to sequence the complete mitochondrial cyto-chrome b gene because two earlier studies using this markerin mormyrids (Lavoue et al. 2000; Sullivan et al. 2000) in-dicated its utility and provided outgroup sequences.

In an effort to obtain a complementary but independentdataset from nuclear markers, we assayed a number of nuclearintron and spacer regions, most of which have been usedsuccessfully in species-level and population-level fish stud-ies. We were successful in obtaining sequences of 400–800bp for a subsample of several Gabon-clade BrienomyrusOTUs for five such markers: (1) RAG 1, second intron, prim-ers designed from data in Willet et al. (1997); (2) S7 introns1 and 2, primers from Chow and Hazama (1998); (3) cal-modulin intron, primers from Chow (1998); (4) GABA intron,primers from J. M. Quattro (pers. comm.); and (5) an LDH-A intron, primers from T. Dowling (pers. comm.). In eachcase, variable sites among the five or six taxa sequenced weretoo few (maximum pairwise P-distance , 1%) to warrantsequencing all individuals used in this study. We likewiseassayed the internal transcribed spacer 1 (ITS1) region of thenuclear ribosomal RNA genes with a number of differentprimers from the literature. Although we were successful inobtaining clean ITS1 sequences from mormyrid species out-side the study group, all sequence chromatograms from Ga-bon-clade Brienomyrus appeared to contain signal from mul-tiple polymorphic ITS1 copies. For this reason, we limitedour dataset to mitochondrial cytochrome b sequences.

Choice of Outgroups

Outgroup cytochrome b sequences were available from ourprevious study (Sullivan et al. 2000) in which we had se-quenced the complete cytochrome b gene from 41 speciesbelonging to all 19 recognized genera of mormyroid fishes.For this analysis, we chose outgroup sequences from the twonearest, consecutive outgroups to the clade formed by thefour Gabon-clade Brienomyrus taxa included in the Sullivanet al. (2000) and Lavoue, J. P. Sullivan, and C. D. Hopkins,(unpubl. ms.) studies (see Fig. 1): M. ntemensis, B. knoepffleri,and I. opdenboschi. Sequences from two individuals of out-group taxon M. ntemensis were used: one from the IvindoRiver and another from the Ntem River.


→

FIG. 5. A strict consensus tree of 680 equally parsimonious trees, each of 636 steps, shown as a phylogram using ACCTRAN characteroptimization, produced from an unweighted parsimony analysis of the complete cytochrome b gene from Gabon-clade Brienomyrus andoutgroup taxa. For all trees, CI 5 0.49, RI 5 0.84, RC 5 0.41 with uninformative sites excluded. Table in inset lists the different OTUs,according to the origin of the names. OTUs NGO, NZO, OFF, and SAN all are abbreviations for one of the collecting localities wherethese fish were captured. Some of the OTU names refer to manuscript names of two manuscripts currently in preparation; others referto the shape of the snout (B, blunt; I, intermediate; S, sharp) and to the type of electrocytes in the electric organ (NPp, nonpenetratingstalk, posterior innervation; Pa, penetrating stalk, anterior innervation). The remaining OTUs are described species. The letters in thisand the following figure refer to nodes that are discussed in the text.

Polymerase Chain Reaction and Sequencing

We extracted DNA from tissue samples with the QIAamptissue kit (Quiagen, Inc., Valencia, CA). Sequences of theprimers used to amplify the mitochondrial cytochrome b genewere taken from Palumbi (1996) and are: 59-TGA TAT GAAAAA CCA TCG TTG-39 (L14724) and 59-CTT CGA TCTTCG rTT TAC AAG-39 (H15930). Polymerase chain reaction(PCR) volumes of 50 ml consisted of approximately 100–500ng genomic DNA, 1X GeneAmp Gold PCR buffer (PE Ap-plied Biosystems, Foster City, CA), 1.25 units of AmpliTaqGold (PE Applied Biosystems), concentrations of 0.2 mM ofeach primer, 200 mM of each dNTP, and 3 mM of MgCl.We carried out PCR on a Hybaid TouchDown thermocycler(Hybaid Limited, Teddington, Middlesex, England) using aninitial 958C denaturation step for 10 min followed by 35repetitions of a three-step cycle consisting of 948C for 1 min,annealing for 1 min at 428C, and extension at 728C for 1.5min, followed by a final extension at 728C for 7 min. Wepurified our PCR products with the Wizard PCR Preps DNApurification kit (Promega, Madison, WI). We sequenced thedouble-stranded PCR products directly in both directionswith the primers used for amplification by automated dye-terminator cycle chemistry on a PE Applied Biosystems 377automated sequencer. We edited the sequences from the gelchromatograms with the Sequencher software package (GeneCodes Corp., Ann Arbor, MI).

Phylogenetic Analysis Methods

We reconstructed cytochrome b phylogeny using bothmaximum-parsimony (MP) and maximum-likelihood (ML)analyses in PAUP* version 4.0b3 (Sinauer, Inc., Sunderland,MA) on an Apple PowerMac G4 computer. For the MP anal-ysis, we performed a heuristic search on the cytochrome bdataset in which all characters and classes of substitutionwere equally weighted. Starting trees for tree bisection-re-connection (TBR) branch swapping were obtained by 100iterations of the random stepwise addition sequence.PAUP*’s default settings for a heuristic MP search were usedin all other cases. Relative support for the internal nodes ofthe MP trees was estimated by bootstrap analysis (Felsenstein1985) consisting of 1000 pseudoreplicates in PAUP* (startingtrees obtained by single iteration of random stepwise addi-tion; MAXTREES set to 2000; otherwise parameter settingswere identical to MP heuristic search).

For the ML analysis, we performed an initial parameter-rich heuristic tree search using the general-time-reversiblemodel with rate heterogeneity (Yang 1994) in which the six-way substitution rate matrix was estimated from the datasetby ML, as were site specific rates for each of the three codon

positions. Assumed nucleotide frequencies were those de-termined empirically from the dataset, a molecular clock wasnot enforced, and starting branch lengths were obtained usingthe Rogers-Swofford approximation method (Rogers andSwofford 1998). TBR branch swapping began from a startingtree obtained by neighbor joining. The option to collapsebranches of insignificantly different length was used. Becauseof the impractical amount of computation time required tocomplete branch swapping in a parameter-rich analysis of adataset this large, we halted the analysis after it had founda best tree that remained unchanged through 2000 subsequentbranch swappings. The substitution rate matrix parametersand codon-specific substitution rates estimated for this treewere then entered as fixed parameters in a subsequent heu-ristic ML search in which branch swapping was allowed tocontinue to completion.

Mutational saturation in the dataset was estimated by look-ing for nonlinearity in a plot of pairwise observed raw dis-tance (PAUP*’s adjusted character distance) against tree-cor-rected distance (PAUP*’s patristic distance) for each codonposition and each class of nucleotide substitution on one ofthe most parsimonious trees resulting from the MP analysis,as described by Hassanin et al. (1998).

RESULTS

Operational Taxonomic Unit Diagnosis

We recognized 38 OTUs among our 1450 collected spec-imens (Fig. 4, Table 1). Four of these we identified as Brien-omyrus curvifrons (Taverne et al. 1977b), B. longicaudatus(Taverne et al. 1977b), B. hopkinsi (Taverne and Thys VanDen Audenaerde 1985), and Paramormyrops gabonensis(Taverne et al. 1977a) based on comparison to type materialand the original descriptions. We could not confidently as-sociate three other previously described Gabon-clade Brien-omyrus species from this region, B. sphekodes (Sauvage1880), B. kingsleyae (Gunther 1896), and B. batesii (Boulen-ger 1906) to any of our remaining 34 OTUs. However, wecannot rule out the possibility that these three species are infact represented among them. In the case of B. sphekodes andB. kingsleyae, for which no illustrations were provided in theoriginal descriptions, the poor condition of the type speci-mens made comparisons to our recently collected specimensdifficult. Although the specimens we assigned to the OTUVAD in this study resemble the types of B. batesii, we hes-itated to make the identification based on a consistent dif-ference in the number of circumpeduncular scales.

In three unusual cases (MAG and BEN, SP2 and SP4, andBP1 and BN1) we encountered cases of sympatric, morpho-logically indistinguishable forms within which we observe


two different EOD waveforms (taking sexual dimorphismsinto account). We chose to recognize them provisionally asseparate OTUs in the analysis, although we consider it pos-sible that each pair may represent a single species with anEOD polymorphism. We sequenced more of these forms forthe analysis than we did for most other OTUs with the hopeof resolving their relationships.

Dataset Characteristics

Sixty-five unique 1140 bp cytochrome b haplotypes wererecovered from the 85 Gabon-clade Brienomyrus specimenssequenced. These made up the ingroup dataset for the par-simony analysis. We observed no indels, nonsense mutations,or ambiguous sites in any of the ABI chromatograms thatmight indicate amplification of extramitochondrial copies ofthe gene. Maximum uncorrected p-distance between the out-group and ingroup sequences is 10%. We observed no evi-dence of mutational saturation for transitions or transversionsin any codon position in our plots of pairwise adjusted char-acter distance against patristic distance as calculated inPAUP*.

Of 287 variable sites in the sequences, 232 were parsi-mony-informative characters: 188 (81%) in the third-codonposition, 33 (14%) in the first-codon position, and 11 (5%)in the second-codon position. Average base frequenciesacross all sites are C 5 32%, A 5 29%, T 5 25%, G 5 14%,but the informative sites of the third-codon position showincreased high C, low G bias (C 5 43.5%, A 5 29.5%, T 522%, G 5 5%) as reported previously for mormyroid fishes(Lavoue et al. 2000; Sullivan et al. 2000), for other fishes(Meyer 1993; Lydeard and Roe 1997), and for other verte-brate groups (Irwin et al. 1991; Kornegay et al. 1993). A chi-squared test failed to reveal any significant heterogeneity inbase frequencies among the sequences.

Phylogenetic Analysis Results

An unweighted MP analysis in PAUP* yielded 680 trees,each of 636 steps (CI 5 0.49, RI 5 0.84, RC 5 0.41, un-informative sites excluded). A strict consensus of these trees,depicted as a phylogram (using ACCTRAN character opti-mization) is shown in Figure 5. A consensus tree of 1000parsimony bootstrap pseudoreplicates in which nodes re-ceiving less than 50% bootstrap level support are collapsedis shown in Figure 6. Well-supported nodes defining majorclades are labeled A–J in Figures 5 and 6.

The ML analysis yielded a tree with an ln-likelihood scoreof 4958.3. Estimated relative substitution rates were 0.336for position 1, 0.08 for position 2, and 2.58 for position 3.Because branches of insignificantly different length were col-lapsed during the ML search, some nodes are not resolved.The topology of this ML tree (not shown) is extremely similarto that of the MP consensus tree: The only topological in-congruence between them concerns the pattern of interrela-tionships of clades H, I, and J (Figs. 5, 6), none of whichhave significant character support.

Both analyses indicate that M. ntemensis is the sister groupto all ingroup specimens sequenced (node A; Figs. 5, 6),supporting the monophyly of the additional, putative Gabon-clade Brienomyrus with those included in previous studies.

Mean p-distance between M. ntemensis and the Gabon-cladeBrienomyrus taxa is 8.3%.

Within the Gabon-clade Brienomyrus, nodes B, C, D, E,and F (Figs. 5, 6) are well supported by long internal branchesand high parsimony bootstrap values. Nodes B and E definethe most basal division within the Gabon-clade Brienomyrus.The OTUs in these two clades are separated from each otherby 7% p-distance. Clade B includes the OTUs VAD and SZA,both of which belong to separate, well-defined clades (D andC, respectively, separated from each other by an average 6.5%p-distance. The two VAD specimens sequenced, one fromthe Woleu River basin and the other from the Ivindo, appearas a monophyletic group which is sister to OTU BP7 1634from the Okano River basin. Sister to these taxa is clade Cin which two distinctly different OTUs, IP1 and LIS, bothknown from only single populations, are nested within theSZA OTU, a widespread form found in the Ivindo, Ntem,Woleu, and Okano Rivers in northern Gabon. All membersof clade B have penetrating stalk–type electric organs (Fig.7D).

The most basal division in the remaining taxa separatesthe OTU BON, an OTU only known from the Louetsi Riverin southern Gabon, from clade F. Clade F, which is supportedby an exceedingly long branch relative to others on the tree,is in turn divided into a large clade (G) whose sister groupconsists of two individuals of the OTU BP1: one from theOkano River, the other from the Woleu River Basin. Thesequences of these two BP1s differ from those in clade G byan average of 3.0% p-distance.

Outside of clade G, all ingroup OTUs have penetratingstalk type electric organs, the apparent primitive conditionfor the entire subfamily Mormyrinae (Sullivan et al. 2000).Within clade G, OTUs have a mix of penetrating and non-penetrating stalk–type electric organs (Fig. 7D).

Within clade G, the remainder of the Gabon-clade Brien-omyrus (32 OTUs) segregate into three relatively well-sup-ported clades—H, I, and J—that are of uncertain relationshipto each other and that are poorly resolved internally. Withinthese clades, sequences of different OTUs often differ by1.0% p-distance or less, whereas in some cases specimensassigned to distinct OTUs share identical cytochrome b hap-lotypes (e.g., BP6 3547/BN2 3542; BP1 3016, 2530, 2704/SN3 2619).

Furthermore, only two OTUs represented by sequencesfrom multiple populations in this large clade appear as mono-phyletic groups on the tree (OFF in Figs. 5, 6; P. gabonensisin Fig. 5). Other OTUs (SP2, SP4, MAG, TEN, SN2, SN3,CAB, NZO, and BP1) represented by multiple sequences,either from the same or from different populations, appearnonmonophyletic on the tree. The remaining OTUs abovenode G are represented by single individuals; therefore, theirmonophyly cannot be tested.

In summary, the basal clades in the tree (below clade G)are characterized by long internal branch lengths and OTUsequence monophyly or at least coherence (although SZAsequences are paraphyletic with respect to IP1 and LIS, theyare at least restricted to a single, well-defined clade C), where-as the large clade G is characterized by short internal branchesand general OTU sequence polyphyly. This pattern in cladeG is not in general a consequence of lack of resolution of


FIG. 6. Consensus MP bootstrap tree based on 1000 pseudoreplicates on the same dataset used for Figure 5. Nodes supported by bootstrapproportions greater than or equal to 50% are shown with bootstrap values indicated.


FIG. 7. The cytochrome b tree fails to recover higher level groups of Gabon-clade Brienomyrus suggested by shared morphological andelectric organ character states, all presumably derived within the group based on outgroup comparison. The phylogenetic tree from Figure5 is reproduced showing the presence (in black) or absence (gray) of each character state for each OTU. (A) OTUs MAG, BEN, SP2,SP4, SP6, SP7, and SP8, B.curvifrons, and B.hopkinsi all share very sharp snouts with terminal mouths and a jutting lower jaw (inset).Yet these taxa do not form a monophyletic group on the cytochrome b haplotype tree. (B) OTUs with 16 instead of 12 circumpeduncularscales (inset) do not form a monophyletic group and OFF and B. longicaudatus that additionally share an elongate caudal peduncle, adistinctive sloping head shape, and large adult size do not appear as sister taxa. (C) OTUs TEN and BN2 that share monophasic EODwaveforms (inset) similarly do not form a monophyletic group. (D) OTUs SP6, SP8, MAG, and SP2 all possess reversed polarity EODwaveforms (inset) in which the initial head negative P0 derived from current flowing through penetrating stalks, becomes a major phaseof the EOD and P2 is reduced. These OTUs do not form a monophyletic group on the tree. (E) OTUs possessing nonpenetrating stalkelectrocytes with posterior innervation (type NPp, inset) do not form a monophyletic group.


the data. In many cases, the nonmonophyly of same-OTUsequences is supported by high bootstrap values (e.g., forTEN, SP4, SN2, SN3, BP1, and BN1). Furthermore, haplo-types of allopatric OTUs that we presumed to be closelyrelated due to relatively minor differentiation among them(e.g., BEN, SP4, SP7 in Fig. 2C; B. longicaudatus and OFFand B. hopkinsi and SN8) do not appear as nearest relativeson the tree.

The status of the sympatric OTU pairs that are morpho-logically indistinguishable but that have different EODs(MAG and BEN, SP2 and SP4, and BP1 and BN1) remainsunclear. In no case do the haplotypes of these OTUs (eithersingly, or with its pair) appear monophyletic on the tree,although there are cases of haplotype identity (or near iden-tity) within each pair.

DISCUSSION

Explaining Operational Taxonomic Unit and HaplotypeTree Incongruence

The departure of the tree topology in clade G from ex-pectations of OTU sequence monophyly and of close rela-tionships between sequences of similar OTUs may indicateeither real incongruence between the mitochondrial haplotypetree and the overall organismal phylogeny or poor corre-spondence between our OTUs and monophyletic groups ofthese fishes.

Lineage Sorting

Gene tree–species tree incongruence caused by incompletegene lineage sorting has been implicated in a number ofstudies of other fish species flocks (Moran and Kornfield1993; McMillan and Palumbi 1995; Strecker et al. 1996;Kornfield and Parker 1997; Parker and Kornfield 1997) andin Darwin’s finches (Sato et al. 1999). This situation resultswhen the rate of lineage splitting or speciation exceeds therate of stochastic sorting of allelic polymorphisms within anorganismal lineage. The phylogeny of alleles sampled inthese cases may differ from larger organismal phylogeny(Pamilo and Nei 1988; Harrison 1991).

Incomplete mitochondrial lineage sorting is suggested bythe presence of divergent haplotypes within single popula-tions (Moran and Kornfield 1993). There are several exam-ples of this in our dataset. For instance, the haplotypes ofSN3 3027 and 2619 are nested within clade J, whereas SN32606 from the same population is nested within clade I. Hap-lotypes of SP4 2671 and 2673 are nested together in cladeH, whereas that of SP4 2995 from the same population isnested within clade J. These results suggest that it may befutile to interpret the pattern of haplotype relationships withinclade G in terms of overall organismal relationships.

Introgression

The second explanation for gene tree–species tree incon-gruence is introgression by hybridization (Smith 1992). Thisis possible among interfertile species in which reproductiveisolation is maintained by extrinsic barriers or by behaviorsthat break down under some conditions (Arnold 1997; Dowl-ing and Secor 1997). Introgression of the mitochondrial ge-

nome between populations of closely related species is notuncommon (Harrison 1989). Cases in our data particularlysuggestive of introgression are those in which a specimen’shaplotype is a nearer relative to the haplotype of a different,but sympatric OTU than it is to haplotypes of the same OTUfrom another population. An example of this is observed onthe tree with the sequence data from the OTU TEN. ThisOTU is among the most recognizable forms in our collectionsdue to its elongate body, small eyes, protruding lower jaw,and monophasic EOD. Yet cytochrome b haplotypes of twoindividuals from the Ivindo River population (2011, 2191)cluster with those of other Ivindo OTUs (clade H in Figs. 5,6) rather than with the haplotype of the TEN specimen 3850from the Ntem River, about 200 km distant. In other cases,specimens of distinct, but sympatric OTUs share identicalhaplotypes (e.g., in clade I: BP6 3547 and BN2 3542 andSN7 3666 & SN2 3415; in clade J: BP1 3016, 2530, 2704,and SN3 2619. Strangely, throughout much of clade G, geo-graphical proximity seems to be a more consistent predictorof clade membership than is OTU identity. For instance, allthe haplotypes appearing in the large clade J are from a singleregion of southern Gabon, although other haplotypes of thesesame OTUs and from the same localities appear again inclades H and I. It is difficult to conceive how any processother than introgression across multiple OTU boundariescould produce this pattern.

Evaluating Possible Problems with Operational TaxonomicUnit Diagnosis

The cytochrome b haplotype tree is consistent with patternsthat would be produced by incomplete lineage sorting andintrogression in clade G, although without additional data,assessing the relative importance of each is difficult. Alter-natively, if our OTUs fail to represent natural groups, thenonmonophyly of their cytochrome b haplotypes would beexpected. We could have made three kinds of errors in thesediagnoses. If errors of any one of these kinds were madesystematically, we should be able to detect them post hoc,because each results in different predictions.

In the first case, we may have been deceived by convergentevolution into inadvertently lumping a number of differentnatural groups into single OTUs (i.e., we designated fewerOTUs than there are natural groups or species). We acceptthat convergence of individual characteristics is possible andperhaps likely in some cases. However, for convergent evo-lution to explain the examples outlined above, both mor-phology and EOD features would have had to coevolve iden-tically and repeatedly in different localities. We know of noreason to believe that selection should favor the associationof a particular EOD waveform with a particular external mor-phology. For some particularly distinctive OTUs, such as theOTU TEN mentioned above, we find it impossible to believethat the similarity between populations is due to convergenceand not to common ancestry.

In the second case, the mistaken use of intraspecific poly-morphisms or plastic characters in our OTU diagnoses couldexplain OTU/cytochrome b haplotype tree incongruence (i.e.,we designated more OTUs than there are natural groups orspecies). If many of our OTUs are indeed unnatural groupings


below the species level, one would predict that the distri-bution of those derived phenotypic character states commonto two or more OTUs should more often correspond to mono-phyletic groups of cytochrome b haplotypes than those weused to diagnose single OTUs, and that examples of haplo-type identity between individuals placed in different OTUsmay often be due to their conspecific status.

We consider the first of these predictions in Figure 7. Ac-cepting the monophyly of clade G, the tree topology belownode G and the outgroup relationships shown in Figure 1,we identified five characters states, each of which appearsuniquely within several OTUs of clade G (shown in black inFig. 7A–E). These character states are: (1) the very sharpsnout combined with jutting lower jaw morphology of nineOTUs (Fig. 7A); (2) 16 circumpeduncular scales found infour OTUs versus 12 in all others (Fig. 7B); (3) monophasicEOD waveforms (vs. typical biphasic or triphasic waveforms;Fig. 7C); (4) EOD waveforms with polarity reversed withrespect to typical EODs (Fig. 7D); and (5) nonpenetratingstalk electrocytes (Fig. 7E). These character states all appearto be derived within the Gabon-clade Brienomyrus basedupon outgroup comparisons. However, none of these char-acter states (or their alternate states) correspond to mono-phyletic groups of cytochrome b haplotypes.

The second prediction of the hypothesis that many of ourOTUs demarcate unnatural groups below the species level isthe conspecific status of those sympatric specimens from dif-ferent OTUs which share cytochrome b haplotypes. As al-ready stated, we recognize that this could be the case forthree morphologically cryptic OTU pairs (MAG and BEN,SP2 and SP4, and BP1 and BN1) within which OTUs weredistinguished solely by EOD characteristics. However, in oth-er cases of haplotype identity or near identity, such as thatof SN3 2619 and SN3 3027 to haplotypes of three BP1 spec-imens (Fig. 5, clade J), the substantial and consistent mor-phological and EOD differentiation between the forms coun-terindicates their conspecificity. We anticipate a more thor-ough resolution of this issue from microsatellite studies cur-rently underway in communities of sympatric Gabon-cladeBrienomyrus (M. A. Arnegard, unpubl. ms.).

The third type of error that would have produced a patternof nonmonophyly of OTU cytochrome b haplotypes is thediagnosis of OTUs by evolutionarily primitive (plesiomorph-ic) characteristics for the group. OTUs diagnosed by prim-itive characteristics would appear paraphyletic on the treewith respect to other OTUs and would produce a patternsimilar to that observed for the OTU SZA with respect toLIS and IP1 in clade C (Figs. 5, 6). The use of additional,independent markers may confirm that the widespread OTUSZA is truly paraphyletic with respect to the allopatric OTUsLIS and IP1 that are known only from single populations.However, as in this case, OTU paraphyly should often beinterpretable as such, because some OTU coherence on thetree will be maintained. Because this pattern is not clearlyrepeated for OTUs in clade G, we think it unlikely that useof primitive character states in OTU definition explains allthe cases of OTU nonmonophyly.

Some of the OTUs we recognized may not represent naturalspecies. However, these considerations suggest that no singletype of error was repeatedly made in their diagnosis and

strengthens the case for the role of incomplete lineage sortingand introgression in producing the general pattern of non-monophyly of same-OTU haplotypes in clade G, as we sug-gest above.

This failure of cytochrome b sequences to resolve mono-phyletic OTUs in clade G (which contains 32 of the 38 Ga-bon-clade Brienomyrus we recognized) renders the topologyof clade G useless as a framework upon which to map electricorgan and EOD characteristics with a view to inferring evo-lutionary patterns. However, we do note with interest thatonly in this clade, characterized by great phenotypic diversityand apparently little concomitant genetic divergence, the ex-clusive appearance in some OTUs of presumably derivedNPp-type electrocytes, in addition to the presumably prim-itive Pa-type electrocytes in others. Sullivan et al. (2000)noted that species with the derived NPp-type electrocytesusually have longer duration EODs than do those with Pa-type. The general characteristics of clade G are consistentwith our hypothesis in Sullivan et al. (2000) that selectionmay favor reversal to nonpenetrating stalk electrocytes inthose primitively penetrating stalk lineages in which contactbetween many recently speciated forms necessitates novelEOD characteristics, such as increased duration, which canenhance species recognition (Hopkins and Bass 1981). Un-fortunately, our inability to posit OTU interrelationships inclade G from the cytochrome b data precludes closer ex-amination of this character displacement hypothesis.

Estimating the Age of the Gabon-Clade Brienomyrus

Similar terminal branch lengths in Figure 5 and in the MLanalysis phylogram (not shown) suggest that base substitu-tions are accumulating at relatively equal rates in differentcytochrome b lineages of Gabon-clade Brienomyrus. Usingthe software RRTree (Robinson et al. 1998), we conductedrelative rates tests on the sequences when grouped into theseven major clades recovered in the parsimony analysis (la-beled clades H, I, J, C, D and unlabelled clades BP1 1638/BP1 3771, BON in Figs. 5, 6). They revealed no significantdifferences in evolutionary rates in any pairwise comparisonbetween groups, relative to the outgroup sequences from M.ntemensis.

Alves-Gomes (1999) has proposed a molecular clock ratein mormyroid fishes for the mitochondrial 12S and 16S rRNAgenes of 0.23% per million years. Cytochrome b distancesare on average 3.0 times greater than 12S and 16S distancesin corresponding pairwise comparisons of mormyrid se-quences in Sullivan et al. (2000). Thus, accepting Alves-Gomes’s suggested 12S/16S clock rate implies a cytochromeb clock rate in the neighborhood of 0.7% per million years.

To calculate this clock, Alves-Gomes (1999) used meanuncorrected 12S and 16S p-distance between 12 mormyroidsand the taxon Chitala chitala from the sister group to theMormyroidea, the notopteroids. He used the 65 million-year-old fossil Ostariostoma, putatively the sister taxon to thesetwo groups (Li and Wilson 1996), as a calibration point.Recently, however, a fossil notopterid from the Middle Cre-taceous of Morocco, Paleonotopterus greenwoodi, has beendescribed and phylogenetically placed as the sister group tothe extant notopteroids (Forey 1997; Taverne and Maisey


1999; Taverne 2000), pushing back the minimum age of thenotopteroid lineage and thus the mormyroid/notopteroid di-vergence to 100 million years ago. By itself, this findingimplies that the Alves-Gomes clock rate is overestimated byat least a factor of 1.5. However, pairwise mormyroid andnotopteroid 12S and 16S p-distances plotted against the cor-responding p-distances from the much more slowly evolvingRAG2 nuclear gene (in fig. 1 of Sullivan et al. 2000) suggestthat 12S/16S p-distance underestimates the actual divergencein these genes between mormyroids and notopteroids by atleast a factor of three, due to significant substitutional sat-uration at this phylogenetic distance. Applying these correc-tions doubles the implied (maximum) clock rates to 0.46%per million years for 12S/16S and to about 1.4% per millionyears for cytochrome b. This conjectural cytochrome b clockrate is roughly in the middle range of mitochondrial codinggene clock rates (0.9–2.5% per million years) estimated forfishes or applied to them in other studies (McCune 1997).

The confidence interval on this estimate for a cytochromeb clock rate in mormyroids must be thought to be exceedinglywide. Nevertheless, applying this rate to the data implies thatthe Gabon-clade Brienomyrus stem group arose roughly 6million years ago (based on an average pairwise uncorrecteddistance of 8.3% between M. ntemensis vs. taxa in clade Ain Figs. 5, 6). Likewise, this rate implies an age of about 2million years for the large clade G (based on the mean pair-wise uncorrected distance of 3.0% between BP1 1638/BP13771 and members of clade G). Thus, it is possible that thebulk of the Gabon-clade Brienomyrus diversity originatedwithin the Quaternary, and much of it within only the past500,000 years.

Endemism and Phylogeographic Patterns

The precise distributional boundaries of the Gabon-cladeBrienomyrus in African freshwaters remain poorly known,but available data suggest that they may be restricted to riverbasins of lower Guinea (from the Sanaga River of Cameroonsouth to the Kouilou-Niari River of the Republic of Congo)and to the Congo River basin (Teugels and Hopkins 1998).Their center of diversity appears to be within the Ogooueand Ntem River basins of Gabon and Cameroon. To date,most of our collection effort has been concentrated in thisregion, although this conclusion is supported by study of theexisting literature and all available museum collections. Thediversity of this mormyrid clade appears to be lower in theCongo basin. In our collections from forest streams in theSangha River basin of the Central African Republic, a CongoRiver tributary, we found only a single species from this clade(SAN in our analysis). Similar habitats in the Ogooue basintypically host four to six distinct forms. This pattern is re-versed for the mormyrid genera Marcusenius, Campylomor-myrus, Stomatorhinus, Mormyrops, and Mormyrus, which arespecies-rich in most Congo basin localities, but are repre-sented in low diversity, or not at all, in the rivers of lowerGuinea.

In addition to being the center of Gabon-clade Brienomyrusdiversity, the Ogooue and Ntem Rivers are also the group’sprobable center of origin, because the two nearest, sequentialoutgroups (M. ntemensis and the clade containing B. knoepf-

fleri, I. opdenboschi, and P. marchei) are themselves Ogooue/Ntem endemics. In fact, all but one of these outgroup taxa(P. marchei) are further restricted to the Ntem and an adjacentOgooue tributary, the Ivindo River, the upper portion ofwhich is believed to have been captured from the Ntem atsome time in the past (Thys van den Audenaerde 1966; Olivry1986). Fish dispersal between the adjacent Ntem and theIvindo headwaters may continue to occur during periods ofhigh water. Assuming an origin in the Ogooue/Ntem region,(perhaps in an ancient Ntem/Ivindo), members of the Gabon-clade Brienomyrus would have subsequently dispersed intothe Congo basin and into neighboring lower Guinea drain-ages.

The diversity and endemism of Gabon-clade Brienomyrusspecies in the Ogooue and Ntem basins is paralleled in thesimilarly forest-dependent killifish genus Aphyosemion forwhich Wildekamp (1993) listed 39 species and subspeciesfrom this region. The Ntem and part of the Ogooue basin arepositioned within Hamilton’s (1982) ‘‘Cameroon/Gabon corearea’’ of species richness for forest-associated groups of ter-restrial plants and animals. These contemporary patterns oforganismal distribution, in conjunction with pollen-based re-constructions of floral change, indicate the persistence ofseveral lowland forest refugia in portions of the Ogooue andNtem basins throughout the most arid periods of the Pleis-tocene (Maley 1987, 1991, 1996). During these periods mostof the modern equatorial forest belt of central Africa wasdominated by savanna. Combined, the phylogeographic, pa-leoecological, and molecular clock estimations point to an insitu diversification of Gabon-clade Brienomyrus in theOgooue/Ntem region during the past 2 million years.

The Gabon-Clade Brienomyrus as a Riverine Species Flock

Within fishes, the term ‘‘species flock’’ has largely beenused in intralacustrine contexts; the most famous examplesare the haplochromine cichlid radiations of the East AfricanGreat Lakes (Brooks 1950; Echelle and Kornfield 1984;Greenwood 1984; Meyer et al. 1990; Goldschmidt 1996;Kornfield and Smith 2000). The term has also been appliedto tilapiine cichlids of the crater lakes of Cameroon (Schliew-en et al. 1994); sculpins of Lake Baikal (Taliev 1955; Berg1965); cyprinids in both Lake Lanao, Philippines (Kornfieldand Carpenter 1984), and in Lake Tana, Ethiopia (Nagelkerkeet al. 1994); killifishes in Lake Titicaca (Parenti 1984; Parkerand Kornfield 1995); pupfishes in Lake Chichancanab, Mex-ico (Humphries 1984); and to a Mesozoic radiation of se-mionotid fishes in North American lakes (McCune et al. 1984;McCune 1996). Arguing that the term need not be restrictedto lacustrine fishes, Johns and Avise (1998) have applied itto radiations of northeastern Pacific Sebastes rockfishes andto Antarctic nototheniod icefishes (Ritchie et al. 1996) forwhich there is evidence of explosive speciation in the past.Although authors differ on the definition of a fish speciesflock (see Greenwood 1984; Ribbink 1984), we use the termto mean a monophyletic assemblage of species, at least large-ly restricted to the geographical area of their origin (i.e.,autochthonous), exhibiting a high level of sympatry, and rap-id, or explosive, speciation relative to their nearest relatives


in neighboring regions. We have shown how all of thesecriteria seem to apply to the Gabon-clade Brienomyrus.

Because river basins, like lakes and islands, are habitatscircumscribed by a boundary inside of which gene flow ispossible, but across which dispersal and invasion are rareevents, diversification of species flocks can take place withinthem. However, unlike many of the lakes hosting fish speciesflocks that began as biologically depauperate environments,the rivers of west-central Africa have undoubtedly alwaysharbored fish communities, although their courses and inter-connections have changed dynamically through time. Perhapsas a consequence of long-term occupation of their riverineenvironments by numerous other groups of fishes, the Gabon-clade Brienomyrus flock seems not to represent an adaptiveradiation in which species have diversified greatly in ecologyand morphology.

Conclusions

We have identified a species flock of mormyrid fishes inwest-central Africa. This is the first freshwater fish speciesflock described within a group of weakly electric fishes andthe first wholly within a riverine, as opposed to a lacustrine,environment. This species flock may have arisen within theprecursor of the modern Ogooue and Ntem River systems,and much of its diversification may have taken place duringthe past 2 million years. The species-specificity of EODwaveforms within this group and their demonstrated use inmate recognition suggest the possibility that selection onEODs may play an important role in the origin and/or main-tenance of species boundaries.

Study of species flocks holds out the promise of revealingdetails of the interrelated processes of speciation and signalevolution. However, progress in estimating their internal phy-logenies may depend on the development of methodologieswith the potential to overcome the limitations encounteredwith mitochondrial sequence data. Our analysis of cyto-chrome b sequences provides evidence for the monophyly ofthis flock and for a recent origin of much of its diversity, butfails to resolve much of its internal phylogenic structure. Weattribute this to the effects of incomplete mitochondrial lin-eage sorting and to introgression. These phenomena havebeen similarly problematic in the phylogenetic analysis ofother species flocks with mitochondrial datasets. Unfortu-nately, alternatives are limited in such cases. Apart from thedifficulty of finding a nuclear marker of appropriate evolu-tionary rate, the four-times-larger effective population sizeof nuclear markers relative to mitochondrial predicts theirgreater susceptibility to lineage sorting problems (Pamilo andNei 1988; Palumbi and Cipriano 1998). Phylogenetic analysisof species flocks may prove to be more successful with da-tasets consisting of many unlinked molecular characters, suchas SINE insertions (Shedlock and Okada 2000) or AFLPfragments (Albertson et al. 1999).

ACKNOWLEDGMENTS

For help in Gabon, we thank J. D. Mbega and J. H. Mveof IRAF, P. Posso and B. Bouroubou of IRET, O. Langrandand A. Kamdem Toham of WWF-CARPO, the Sisters of theImmaculate Conception in Lambarene, the Christian and Mis-

sionary Alliance in Bongolo, CIRMF in Franceville, J. Beckand C. Ella from the U.S. Peace Corps, and C. Aveling ofECOFAC. For help in the Central African Republic, we thankJ. B. Kindi-Moungo and the WWF office in Bangui. FromCornell, M. Arnegard helped collect fishes, EODs, and tissuesin Gabon, as did J. Friel, who additionally oversaw the cur-ation of specimens at the Cornell Museum of Vertebrates. G.Harned prepared histological slides of electric organs. Wethank M. L. J. Stiassny and J. Cracraft of the American Mu-seum of Natural History for making possible JPS’s collectiontrip to the Central African Republic. M. Arnegard, J. G. Lund-berg, A. R. McCune, G. Teugels, B. Turner, and K. Zamudioprovided helpful comments on an early draft of the manu-script. Funding for this work came from the following grantsto C.D. Hopkins: National Science Foundation InternationalProgram Grant INT-9605176, National Geographic Society5801-96, and the National Institute of Mental HealthMH37972.

LITERATURE CITED

Albertson, R. C., J. A. Markert, P. D. Danley, and T. D. Kocher.1999. Phylogeny of a rapidly evolving clade: the cichlid fishesof Lake Malawi, East Africa. Proc. Natl. Acad. Sci. USA 96:5107–5110.

Alves-Gomes, J. A. 1999. Systematic biology of gymnotiform andmormyriform electric fishes: phylogenetic relationships, molec-ular clocks and rates of evolution in the mitochondrial rRNAgenes. J. Exp. Biol. 202:1167–1183.

Alves-Gomes, J., and C. D. Hopkins. 1997. Molecular insights intothe phylogeny of mormyriform fishes and the evolution of theirelectric organs. Brain Behav. Evol. 49:324–351.

Arnold, M. L. 1997. Natural hybridization and evolution. OxfordUniv. Press, Oxford, U.K.

Arratia, G. 1997. Basal teleosts and teleostean phylogeny. F. Pfeil,Munich.

Bass, A. H. 1986a. Electric organs revisited: evolution of a ver-tebrate communication and orientation organ. Pp. 13–70 in T.H. Bullock and W. Heiligenberg, eds. Electroreception. JohnWiley and Sons, New York.

———. 1986b. Species differences in electric organs of mormyrids:substrates for species-typical electric organ discharge wave-forms. J. Comp. Neurol. 244:313–330.

Bass, A. H., and C. D. Hopkins. 1983. Hormonal control of sexualdifferentiation changes in electric organ discharge waveform.Science 220:971–974.

Bass, A. H., J.-P. Denizot, and M. A. Marchaterre. 1986. Ultra-structural features and hormone-dependent sex differences ofmormyrid electric organs. J. Comp. Neurol. 254:511–528.

Bennett, M. V. L. 1970. Comparative physiology: electric organs.Annu. Rev. Physiol. 32:471–528.

———. 1971. Electric organs. Pp. 347–491 in W. S. Hoar and D.J. Randall, eds. Fish physiology. Academic Press, New York.

Bennett, M. V. L., and H. Grundfest. 1961. Studies on the mor-phology and electrophysiology of electric organs. III. Electro-physiology of the electric organs in mormyrids. Pp. 113–135 inC. Chagas and A. P. d. Carvalho, eds. Bioelectrogenesis. El-sevier, New York.

Berg, L. S. 1965. Freshwater fishes of the USSR and adjacent coun-tries. Israel Program for Scientific Translation, Jerusalem.

Boulenger, G. A. 1906. Description of new mormyrid fish fromsouth Cameroon. Ann. Mag. Nat. Hist. 18:36–37.

Brooks, J. L. 1950. Speciation in ancient lakes. Q. Rev. Biol. 25:30–176.

Chow, S. 1998. Universal PCR primer for calmodulin gene intronin fish. Fish. Sci. 64:999–1000.

Chow, S., and K. Hazama. 1998. Universal PCR primers for S7ribosomal protein gene introns in fish. Mol. Ecol. 7:1247–1263.


Cracraft, J. 1983. Species concepts and speciation analysis. Curr.Ornithol. 1:159–187.

Daget, J., J.-P. Gosse, and D. F. E. Thys van den Audenaerde. 1984.Check-list of the freshwater fishes of Africa. ORSTOM/MRAC,Paris.

Dowling, T. E., and C. L. Secor. 1997. The role of hybridizationand introgression in the diversification of animals. Annu. Rev.Ecol. Syst. 28:593–619.

Echelle, A. A., and I. Kornfield. 1984. Evolution of fish speciesflocks. Univ. of Maine Press, Orono, ME.

Felsenstein, J. 1985. Confidence limits on phylogenies: an approachusing the bootstrap. Evolution 39:783–791.

Forey, P. L. 1997. A Cretaceous notopterid (Pisces: Osteoglosso-morpha) from Morocco. S. Afr. J. Sci. 93:564–569.

Fryer, G., and T. D. Iles. 1972. The cichlid fishes of the Great Lakesof Africa. Oliver and Boyd, Edinburgh.

Goldschmidt, T. 1996. Darwin’s dreampond: drama in Lake Vic-toria. MIT Press, Boston.

Grant, B. R., and P. R. Grant. 1996. Cultural inheritence of songand its role in the evolution of Darwin’s finches. Evolution 50:2471–2487.

Greenwood, P. H. 1973. Interrelationships of osteoglossomorphs.Pp. 307–332 in P. H. Greenwood, R. S. Miles, and C. Patterson,eds. Interrelationships of fishes. Academic Press, London.

———. 1984. What is a species flock? Pp. 13–20 in A. A. Echelleand I. Kornfield, eds. Evolution of fish species flocks. Univ. ofMaine Press, Orono, ME.

Gunther, A. 1896. Report on a collection of reptiles and fishes madeby Miss M. H. Kingsley during her travels on the Ogowe Riverand in Old Calabar. Ann. Mag. Nat. Hist. 17:261–285.

Hamilton, A. C. 1982. Environmental history of East Africa: a studyof the Quaternary. Academic Press, London.

Harrison, R. G. 1989. Animal mitochondrial DNA as a geneticmarker in population and evolutionary biology. Trends Ecol.Evol. 4:6–11.

———. 1991. Molecular changes at speciation. Annu. Rev. Ecol.Syst. 22:281–308.

Hassanin, A., G. Lecointre, and S. Tillier. 1998. The ‘evolutionarysignal’ of homoplasy in protein-coding gene sequences and itsconsequence for a priori weighting in phylogeny. C. R. Acad.Sci. Ser. III 321:611–620.

Hopkins, C. D. 1980. Evolution of electric communication channelsof mormyrids. Behav. Ecol. Sociobiol. 7:1–13.

———. 1981. On the diversity of electric signals in a communityof mormyrid electric fish in West Africa. Am. Zool. 21:211–222.

———. 1986a. Behavior of Mormyridae. Pp. 527–576 in T. H.Bullock and W. Heiligenberg, eds. Electroreception. John Wileyand Sons, New York.

———. 1986b. Temporal structure of non-propagated electric com-munication signals. Brain Behav. Evol. 28:43–59.

Hopkins, C. D., and A. H. Bass. 1981. Temporal coding of speciesrecognition signals in an electric fish. Science 212:85–87.

Hoy, R., A. Hoikkala, and K. Kaneshiro. 1988. Hawaiian courtshipsongs: evolutionary innovation in communication signals ofDrosophila. Science 240:217–219.

Humphries, J. M. 1984. Genetics and speciation in pupfishes fromLaguna Chichancanab, Mexico. Pp. 129–140 in A. A. Echelleand I. Kornfield, eds. Evolution of fish species flocks. Univ. ofMaine Press, Orono, ME.

Irwin, D. M., T. D. Kocher, and A. C. Wilson. 1991. Evolution ofthe cytochrome b gene of mammals. J. Mol. Evol. 32:128–144.

Johns, G. C., and J. C. Avise. 1998. Tests for ancient species flocksbased on molecular phylogenetic appraisals of Sebastes rock-fishes and other marine fishes. Evolution 52:1135–1146.

Kamdem Toham, A. 1998. Fish biodiversity of the Ntem RiverBasin (Cameroon): taxonomy, ecology and conservation. Ph.D.diss. Laboratory of Ecology and Aquaculture, Dept. of Biology,Katholieke Universiteit Leuven, Leuven, Belgium.

Knight, M. E., and G. F. Turner. 1999. Reproductive isolationamong closely related Lake Malawi cichlids: Can males rec-ognize conspecific females by visual cues? Anim. Behav. 58:761–768.

Kornegay, J. R., T. D. Kocher, L. A. Williams, and A. C. Wilson.

1993. Pathways of lysozyme evolution inferred from the se-quences of cytochrome b in birds. J. Mol. Evol. 37:367–379.

Kornfield, I., and K. E. Carpenter. 1984. Cyprinids of Lake Lanao,Philippines: taxonomic validity, evolutionary rates, and speci-ation scenarios. Pp. 69–84 in A. A. Echelle and I. Kornfield,eds. Evolution of fish species flocks. Univ. of Maine Press,Orono, ME.

Kornfield, I., and A. Parker. 1997. Molecular systematics of a rap-idly evolving species flock: the mbuna of Lake Malawi and thesearch for phylogenetic signal. Pp. 25–37 in T. D. Kocher andC. A. Stepien, eds. Molecular systematics of fishes. AcademicPress, New York.

Kornfield, I., and P. F. Smith. 2000. African cichlid fishes: modelsystems for evolutionary biology. Annu. Rev. Ecol. Syst. 31:163–196.

Kramer, B. 1997. Electric organ discharges and their relation to sexin mormyrid fishes. Naturwissenschaften 84:119–121.

Lauder, G. V., and K. Liem. 1983. The evolution and interrela-tionships of the actinopterygian fishes. Bull. Mus. Comp. Zool.150:95–197.

Lavoue, S., R. Bigorne, G. Lecointre, and J.-F. Agnese. 2000. Phy-logenetic relationships of mormyrid electric fishes (Mormyridae:Teleostei) inferred from cytochrome b sequences. Mol. Phylo-genet. Evol. 14:1–10.

Li, G.-Q. 1997. Notes on the historical biogeography of the osteo-glossomorpha (Teleostei). Pp. 54–66 in Proceedings of the 30thinternational geological congress. VSP International SciencePublishers, Zeist, The Netherlands.

Li, G.-Q., and M. V. H. Wilson. 1996. Phylogeny of the Osteo-glossomorpha. Pp. 163–174 in M. L. J. Stiassny, L. R. Parenti,and G. D. Johnson, eds. Interrelationships of fishes. AcademicPress, New York.

Lissmann, H. W. 1951. Continuous electric signals from the tail ofa fish, Gymnarchus niloticus Cuv. Nature 167:201–202.

Lowe, R. H. 1953. Notes on the ecology and evolution of Nyasafishes of the genus Tilapia with a description of T. saka Lowe.Proc. Zool. Soc. Lond. 122:1035–1041.

Lydeard, C., and K. J. Roe. 1997. The phylogenetic utility of themitochondrial cytochrome b gene for inferring relationshipsamong actinopterygian fishes. Pp. 285–303 in T. D. Kocher andC. A. Stepien, eds. Molecular systematics of fishes. AcademicPress, New York.

Maley, J. 1987. Fragmentation de la foret dense humide africaineet extension des biotopes montagnards au Quaternaire recent:nouvelles donnees polliniques et chronologiques. Implicationspaleoclimatiques et biogeographiques. Paleoecol. Afr. 18:307–334.

———. 1991. The African rainforest vegetation and paleoenviron-ments during late Quaternary. Clim. Change 19:79–98.

———. 1996. The African rainforest: main characteristics of chang-es in vegetation and climate from the Upper Cretaceous to theQuaternary. Proc. R. Soc. Edinb. 104B:31–73.

McCune, A. R. 1996. Biogeographic and stratigraphic evidence forrapid speciation in semionotid fishes. Paleobiology 22:34–48.

———. 1997. How fast is speciation? Molecular, geological, andphylogenetic evidence from adaptive radiations of fishes. Pp.585–609 in T. J. Givnish and K. J. Sytsma, eds. Molecular evo-lution and adaptive radiation. Cambridge Univ. Press, Cam-bridge, U.K.

McCune, A. R., K. S. Thomson, and P. E. Olsen. 1984. Semionotidfishes from the Mesozoic great lakes of North America. Pp. 585–610 in A. A. Echelle and I. Kornfield, eds. Evolution of fishspecies flocks. Univ. of Maine Press, Orono, ME.

McMillan, W. O., and S. R. Palumbi. 1995. Concordant evolution-ary patterns among Indo-West Pacific butterflyfishes. Proc. R.Soc. Lond. B 260:229–239.

Meyer, A. 1993. Evolution of mitochondrial DNA in fishes. Pp. 1–38 in P. W. Hochachka and T. P. Mommsen, eds. Biochemistryand molecular biology of fishes. Elsevier, Amsterdam.

Meyer, A., T. D. Kocher, P. Basasibwaki, and A. C. Wilson. 1990.Monophyletic origin of Lake Victoria cichlid fishes suggestedby mitochondrial DNA sequences. Nature 347:550–553.

Moran, P., and I. Kornfield. 1993. Retention of an ancestral poly-


morphism in mbuna species flock (Pisces: Cichlidae) of LakeMalawi. Mol. Biol. Evol. 10:1015–1029.

Nagelkerke, L. A. J., F. A. Sibbing, J. G. M. Van-Den-Boogaart,E. H. Lammens, and J. W. M. Osse. 1994. The barbs (Barbusspp.) of Lake Tana: a forgotten species flock? Environ. Biol.Fishes 39:1–22.

Nelson, J. S. 1994. Fishes of the world. John Wiley and Sons, NewYork.

Olivry, J. C. 1986. Fleuves et Rivieres du Cameroun. MESRES-ORSTOM, Paris.

Otte, D. 1994. The crickets of Hawaii: origin, systematics, andevolution. The Orthopterist’s Society, Philadelphia, PA.

Palumbi, S. R. 1996. Nucleic acids. II. The polymerase chain re-action. Pp. 205–248 in D. M. Hillis, C. Moritz, and B. K. Mable,eds. Molecular systematics. Sinauer Associates, Sunderland,MA.

Palumbi, S. R., and F. Cipriano. 1998. Species identification usinggenetic tools: the value of nuclear and mitochondrial gene se-quences in whale conservation. Heredity 89:459–464.

Pamilo, P., and M. Nei. 1988. Relationships between gene treesand species trees. Mol. Biol. Evol. 5:568–583.

Parenti, L. R. 1984. Biogeography of the Andean killifish genusOrestias with comments on the species flock concept. Pp. 85–92 in A. A. Echelle and I. Kornfield, eds. Evolution of fishspecies flocks. Univ. of Maine Press, Orono, ME.

Parker, A., and I. Kornfield. 1995. A molecular perspective onevaluation and zoogeography of cyprinodontid killifishes. Cop-eia 1995:8–21.

———. 1997. Evolution of the mitochondrial DNA control regionof the mbuna (Cichlidae) species flock of Lake Malawi, EastAfrica. J. Mol. Evol. 45:70–83.

Patterson, C., and D. E. Rosen. 1977. Review of the ichthyodec-tiform and other mesozoic teleost fishes and the theory and prac-tice of classifying fossils. Bull. Am. Mus. Nat. Hist. 158:81–172.

Ribbink, A. J. 1984. Is the species flock concept tenable? Pp. 21–25 in A. A. Echelle and I. Kornfield, eds. Evolution of fishspecies flocks. Univ. of Maine Press, Orono, ME.

Ritchie, P. A., L. Bargelloni, A. Meyer, J. A. Taylor, J. A. Mac-Donald, and D. M. Lambert. 1996. Mitochondrial phylogeny oftrematomid fishes (Nototheniidae, Perciformes) and the evolu-tion of Antarctic fish. Mol. Phylogenet. Evol. 5:383–390.

Robinson, M., M. Gouy, C. Gautier, and D. Mouchiroud. 1998.Sensitivity of the relative rate test to taxonomic sampling. Mol.Biol. Evol. 15:1091–1098.

Rogers, J. S., and Swofford, D. L. 1998. A fast method for ap-proximating maximum likelihoods of phylogenetic trees fromnucleotide sequences. Syst. Biol. 47:77–89.

Sato, A., C. O’Huigin, F. Figueroa, P. Grant, B. R. Grant, H. Tichy,and J. Klein. 1999. Phylogeny of Darwin’s finches as revealedby mtDNA sequences. Proc. Natl. Acad. Sci USA 96:5101–5106.

Sauvage, H. E. 1880. Etude sur la faune ichthyologique del’Ogooue. Arch. Mus. Natl. Hist. Nat. 2:5–56.

Schlichter, H. 1906. Uber den feineren Bau des schwach-elektrisch-en Organs von Mormyrus oxyrhynchus Geoff. Z. Wiss. Zool. 84:479–525.

Schliewen, U. K., D. Tautz, and S. Paabo. 1994. Sympatric spe-ciation suggested by monophyly of crater lake cichlids. Nature368:629–632.

Seehausen, O., J. J. M. van Alpen, and F. Witte. 1997. Cichlid fishdiversity threatened by eutrophication that curbs sexual selec-tion. Science 277:1808–1811.

Seutin, G., B. N. White, and P. T. Boag. 1991. Preservation of avian

blood and tissue samples for DNA analyses. Can. J. Zool. 69:82–90.

Shaw, K. L. 1999. A nested analysis of song groups and speciesboundaries in the Hawaiian cricket genus Laupala. Mol. Phy-logenet. Evol. 11:332–341.

Shedlock, A. M., and N. Okada. 2000. SINE insertions: powerfultools for molecular systematics. BioEssays 22:148–160.

Smith, G. 1992. Introgression in fishes: significance for paleontol-ogy, cladistics, and evolutionary rates. Syst. Biol. 41:41–57.

Strecker, U., C. G. Meyer, C. Sturmbauer, and H. Wilkens. 1996.Genetic divergence and speciation in an extremely young speciesflock in Mexico formed by the genus Cyprinodon (Cyprinodon-tidae, Teleostei). Mol. Phylogenet. Evol. 6:143–149.

Sullivan, J. P., S. Lavoue, and C. D. Hopkins. 2000. Molecularsystematics of the African electric fishes (Mormyroidea: Te-leostei) and a model for the evolution of their electric organs.J. Exp. Biol. 203:665–683.

Taliev, D. N. 1955. Sculpins of Baikal (Cottidae). Academy ofScience USSR, Moscow.

Taverne, L. 1971. Note sur la systematique des poissons Mormy-riformes. Le probleme des genres Gnathonemus, Marcusenius,Hippopotamyrus, Cyphomyrus, et les nouveaux genres Pollimy-rus et Brienomyrus. Rev. Zool. Bot. Afr. 84:99–110.

———. 2000. Nouvelles donnees osteologiques et phylogenetiquessur Paleonotopterus greenwoodi, notopteride (Teleostei, Osteo-glossomorpha) du Cenomanien inferieur continental (Cretace)du Maroc. Stuttg. Beitr. Naturkde. Ser. B 293:24.

Taverne, L., and J. G. Maisey. 1999. A notopterid skull (Teleostei,Osteoglossomorpha) from the continental Early Cretaceous ofSouthern Morocco. Am. Mus. Novit. 3260:1–12.

Taverne, L. P., and D. F. E. Thys van den Audenaerde. 1985. Brien-omyrus hopkinsi new species from north Gabon (Pisces: Te-leostei: Mormoridae). Rev. Zool. Afr. 99:49–52.

Taverne, L., D. F. E. Thys van den Audenaerde, and A. Heymer.1977a. Paramormyrops gabonensis nov. gen., nov. sp. du norddu Gabon. Rev. Zool. Afr. 91:634–640.

Taverne, L., D. F. E. Thys van den Audenaerde, A. Heymer, andJ. Gery. 1977b. Brienomyrus longicaudatus new species andBrienomyrus curvifrons new species from northern Gabon (Pi-sces: Mormyridae). Rev. Zool. Afr. 91:200–208.

Teugels, G. G., and C. D. Hopkins. 1998. Morphological and os-teological evidence for the generic position of Mormyrus king-sleyae in the genus Brienomyrus (Teleostei: Mormyridae). Cop-eia 1998:199–204.

Thys van den Audenaerde, D. F. E. 1966. Les Tilapia (Pisces, Cich-lidae) de Sud-Cameroun et du Gabon etude systematique. Mus.R. Afr. Cent. Tervuren Belg. Ann. Ser. Octavo Sci. Zool. 153:123.

van Oppen, M. J. H., G. F. Turner, C. Rico, R. L. Robinson, J. C.Deutsch, and G. M. Hewitt. 1998. Assortative mating amongrock-dwelling cichlid fishes supports high estimates of speciesrichness from Lake Malawi. Mol. Ecol. 7:991–1001.

Wildekamp, R. H. 1993. A world of killies. American KillifishAssociation, Mishawaka, IN.

Willet, C. E., J. J. Cherry, and L. A. Steiner. 1997. Characterizationand expression of the recombination activating genes (rag1 andrag2) of zebrafish. Immunogenetics 45:394–404.

Winemiller, K. O., and A. Adite. 1997. Convergent evolution ofweakly electric fishes from floodplain habitats in Africa andSouth America. Environ. Biol. Fishes 49:175–186.

Yang, Z. 1994. Maximum likelihood phylogenetic estimation fromDNA sequences with variable rates over sites: approximatemethods. J. Mol. Evol. 39:306–314.

Corresponding Editor: L. Bernatchez

discovery and phylogenetic analysis of a riverine …

Documents