Top Banner
Phylogeography and phenotypic diversification in the Patagonian fish Percichthys trucha: the roles of Quaternary glacial cycles and natural selection DANIEL E. RUZZANTE 1 *, SANDRA J. WALDE 1 , PATRICIO J. MACCHI 2 , MARCELO ALONSO 2 and JUAN P. BARRIGA 3 1 Department of Biology, Dalhousie University, Halifax, Nova Scotia, B3H 4J1, Canada 2 Centro Regional Universitario Bariloche, Universidad Nacional del Comahue, Bariloche, Rio Negro, Argentina 3 Instituto de Investigaciones en Biodiversidad y Medioambiente (CONICET – Universidad Nacional del Comahue), Bariloche, Río Negro, Argentina Received 4 March 2011; accepted for publication 4 March 2011Current patterns of genetic and morphological diversity are the product of historical climatic and geomorphological events, and of contemporary selection processes acting upon this diversity. Here we examine the phylogeographic and phenotypic patterns of diversity within Percichthys trucha, a widely distributed Patagonian fish species complex that inhabits Andean and steppe freshwater environments. Molecular analysis (mtDNA control region) of 21 populations distributed throughout its latitudinal range revealed little evidence of phylogeographic structure and no evidence of species-level genetic divergence east of the Andes. The complex, however, exhibits high levels of intra- and interpopulation phenotypic variation. Patterns of among-population divergence in morphology were most easily explained by differences in predation pressure among populations; dorsal fin spines (commonly a defensive characteristic) were longer in environments with greater densities of potentially piscivorous fish. Trophic characters were highly variable within populations, suggesting an important role for resources in generating within-population morphological variation. The very shallow levels of divergence shown by the molecular data most likely reflect the historical mixing of populations as a result of the climatic and landscape changes that affected Patagonia throughout the Quaternary. The phenotypic divergences, in contrast, are probably the result of differing contemporary selection regimes acting on currently disjoint populations. © 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 103, 514–529. ADDITIONAL KEYWORDS: adaptive radiation Patagonia predation resource competition Salmonidae. Los patrones de diversidad genética y morfológica que pueden observarse en poblaciones existentes son el producto de la influencia conjunta de procesos históricos (climáticos, geomorfológicos) y de la selección natural. En este trabajo examinamos los patrones de diversidad filogeográfica y fenotípica en Percichthys trucha, una especie o complejo de especies de amplia distribución en Patagonia andina y esteparia. Análisis molecular (Region de Control ADN mitocondrial) de 21 poblaciones a lo largo y ancho del rango distribucional del grupo reveló poca evidencia de estructura filogeográfica (estructura poco profunda) y ninguna evidencia de divergencia genética a nivel de especie al este de los Andes. El complejo exhibe sin embargo, altos niveles de variación fenotípica tanto intra-, como interpoblacional. Los patrones de divergencia morfológica entre poblaciones se correlacionan con diferencias interpoblacionales en la intensidad de predación; las espinas dorsales (comúnmente una característica defensiva) son más largas en ambientes con mayor densidad de peces potencialmente piscívoros. Los caracteres tróficos exhiben alta variación intrapoblacional sugiriendo que los recursos tróficos cumplen un rol importante en *Corresponding author. E-mail: [email protected] Biological Journal of the Linnean Society, 2011, 103, 514–529. With 3 figures © 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 103, 514–529 514
16

Phylogeography and phenotypic diversification in the Patagonian fish Percichthys trucha: the roles of Quaternary glacial cycles and natural selection

Dec 15, 2022

Download

Documents

Gladys Galende
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Phylogeography and phenotypic diversification in the Patagonian fish Percichthys trucha: the roles of Quaternary glacial cycles and natural selection

Phylogeography and phenotypic diversification in thePatagonian fish Percichthys trucha: the roles ofQuaternary glacial cycles and natural selection

DANIEL E. RUZZANTE1*, SANDRA J. WALDE1, PATRICIO J. MACCHI2,MARCELO ALONSO2 and JUAN P. BARRIGA3

1Department of Biology, Dalhousie University, Halifax, Nova Scotia, B3H 4J1, Canada2Centro Regional Universitario Bariloche, Universidad Nacional del Comahue, Bariloche,Rio Negro, Argentina3Instituto de Investigaciones en Biodiversidad y Medioambiente (CONICET – Universidad Nacionaldel Comahue), Bariloche, Río Negro, Argentina

Received 4 March 2011; accepted for publication 4 March 2011bij_1682 514..529

Current patterns of genetic and morphological diversity are the product of historical climatic and geomorphologicalevents, and of contemporary selection processes acting upon this diversity. Here we examine the phylogeographicand phenotypic patterns of diversity within Percichthys trucha, a widely distributed Patagonian fish speciescomplex that inhabits Andean and steppe freshwater environments. Molecular analysis (mtDNA control region) of21 populations distributed throughout its latitudinal range revealed little evidence of phylogeographic structureand no evidence of species-level genetic divergence east of the Andes. The complex, however, exhibits high levelsof intra- and interpopulation phenotypic variation. Patterns of among-population divergence in morphology weremost easily explained by differences in predation pressure among populations; dorsal fin spines (commonly adefensive characteristic) were longer in environments with greater densities of potentially piscivorous fish. Trophiccharacters were highly variable within populations, suggesting an important role for resources in generatingwithin-population morphological variation. The very shallow levels of divergence shown by the molecular data mostlikely reflect the historical mixing of populations as a result of the climatic and landscape changes that affectedPatagonia throughout the Quaternary. The phenotypic divergences, in contrast, are probably the result of differingcontemporary selection regimes acting on currently disjoint populations. © 2011 The Linnean Society of London,Biological Journal of the Linnean Society, 2011, 103, 514–529.

ADDITIONAL KEYWORDS: adaptive radiation – Patagonia – predation – resource competition –Salmonidae.

Los patrones de diversidad genética y morfológica que pueden observarse en poblaciones existentes son el productode la influencia conjunta de procesos históricos (climáticos, geomorfológicos) y de la selección natural. En estetrabajo examinamos los patrones de diversidad filogeográfica y fenotípica en Percichthys trucha, una especie ocomplejo de especies de amplia distribución en Patagonia andina y esteparia. Análisis molecular (Region de ControlADN mitocondrial) de 21 poblaciones a lo largo y ancho del rango distribucional del grupo reveló poca evidenciade estructura filogeográfica (estructura poco profunda) y ninguna evidencia de divergencia genética a nivel deespecie al este de los Andes. El complejo exhibe sin embargo, altos niveles de variación fenotípica tanto intra-, comointerpoblacional. Los patrones de divergencia morfológica entre poblaciones se correlacionan con diferenciasinterpoblacionales en la intensidad de predación; las espinas dorsales (comúnmente una característica defensiva)son más largas en ambientes con mayor densidad de peces potencialmente piscívoros. Los caracteres tróficosexhiben alta variación intrapoblacional sugiriendo que los recursos tróficos cumplen un rol importante en

*Corresponding author. E-mail: [email protected]

Biological Journal of the Linnean Society, 2011, 103, 514–529. With 3 figures

© 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 103, 514–529514

Page 2: Phylogeography and phenotypic diversification in the Patagonian fish Percichthys trucha: the roles of Quaternary glacial cycles and natural selection

la generación de variación morfológica dentro de poblaciones. Los bajos niveles de divergencia moleculary de estructura filogeográfica son probablemente el resultado de la mezcla histórica de individuos ypoblaciones como consecuencia de los cambios climáticos y geográficos (paisaje) que afectaron la regiónPatagónica durante el Cuaternario. Las divergencias fenotípicas por el contrario, son el resultado dediferencias en los regímenes de selección natural operantes en poblaciones de Percichthys truchaactualmente disjuntas.

PALABRAS CLAVE: competencia por recursos – Cuaternario – filogeografia – predacion – radiación adapta-tiva – Patagonia – Salmónidos – trucha criolla (Percichthys) – variación fenotípica.

INTRODUCTION

Current patterns of genetic and morphological diver-sity are the product of historical processes that haveconstrained, moulded and dispersed that diversity,and of the selection processes (biotic and abiotic) thatare currently acting upon it. The Quaternary glacialcycles are among the most important of the majornatural historical processes known to have influenceddiversity patterns worldwide, and their effects onphylogeographic patterns have been well documentedfor the biota of the northern hemisphere (Avise, 2000;Hewitt, 2000). Comparable information for the south-ern hemisphere is sparser (Beheregaray, 2008), buthas recently been accumulating, particularly for thefauna and flora of Patagonia (Ruzzante et al., 2008;Zemlak et al., 2010; Pardiñas et al., 2011, Sérsic et al.,2011).

It is critical to recognize that, concomitant withshifts in climate during the Quaternary, the biota ofPatagonia experienced large-scale and importantchanges in geography and landscape. First, and mostobviously, as the glaciers advanced, some existinghabitat was eliminated or rendered uninhabitable,leading to the elimination of local populations andshifts in the distribution of most species. Icefieldswere mostly restricted to the Andes through muchof their latitudinal range, but glaciers calved intothe Pacific Ocean south of 39°S, and reached theAtlantic Ocean in southernmost Patagonia and Tierradel Fuego during the most extensive glaciations(Clapperton, 1993; Sugden et al., 2005; Rabassa,Coronato & Martínez, 2011). However, the total areaof exposed and potentially habitable land actuallyincreased dramatically during glacial periods, asthe sea level fell. In fact, Patagonia doubled itscurrent surface area during the most recent glacialcycle (Cavallotto, Violante & Hernández-Molina,2011; Ponce et al., 2011). Thus, there was consider-able potential for terrestrial refugia, and perhapseven the potential for the expansion of some popula-tions during glacial periods. Glaciers largely elimi-nated the current headwater lakes and upper reaches

of the major river systems of Patagonia (Rabassaet al., 2011). However, the exposure of much of thecontinental shelf to the east would have also allowedfor the extension and perhaps expansion of aquatichabitats. The changing size of these aquatic envi-ronments and the routes of the now-submergedriver systems would have had major effects on thedistribution of diversity within the aquatic biota ofPatagonia.

Current ecological forces, i.e. the physical envi-ronment and biotic factors, such as competitionand predation, act upon and may modify the patternsof morphological and/or genetic diversity producedby historic processes. For example, previously diver-gent populations may converge phenotypically whenexposed to a similar environment, or marked pheno-typic differences might emerge among currently iso-lated populations for which there is no evidence ofphylogeographic structure. Differences among habi-tats in resources, competitive regimes, and numberand type of predators can rapidly lead to divergentmorphology through phenotypic plasticity and/orselection, and differences in competitive and pre-dation regimes among lakes have repeatedly beenassociated with morphologically and behaviourallydistinct fish populations (Milano et al., 2002, 2006;Vamosi, 2002; Langerhans et al., 2004; Andersson,Johansson & Söderlund, 2006; Eklöv & Svanbäck,2006; Svanbäck & Persson, 2009).

In the present study we examine phylogeographicpatterns within Percichthys. Although the genus ispresent on both sides of the Andes in Patagonia, herewe focus primarily on populations in Argentina,where they are found in both Andean and steppefreshwater environments. Our goal is to interpret themolecular genetic and phenotypic diversity withinthe group in the light of known historic and currentforces. We thus look at the effects of historic changesin aquatic landscapes as well as the potential conse-quences of current environments on the observedpatterns of genetic and morphological diversity. Per-cichthys exhibits such high levels of intra- and inter-population phenotypic variation in characters such as

PHYLOGEOGRAPHY OF PERCICHTHYS TRUCHA 515

© 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 103, 514–529

Page 3: Phylogeography and phenotypic diversification in the Patagonian fish Percichthys trucha: the roles of Quaternary glacial cycles and natural selection

mouth structure, body depth, and dorsal fin spine,that it has historically been divided into variousspecies (Ringuelet, Aramburu & Alonso de Aramburu,1967; López-Albarello, 2004). The number and iden-tities of recognized morphological species has variedthrough time. López-Albarello (2004) collapsed threeof the species originally identified by Ringuelet et al.(1967) (Percichthys trucha, Percichthys vinciguerrai,and Percichthys altispinnis) into a single species(P. trucha), and created a new one (Percichthyslaevis), to conclude that three species of Percichthysare present in Patagonia east of the Andes: P. trucha,Percichthys colhuapensis, and P. laevis. Here wedescribe the available molecular genetic informationfor the group along with information on morphologi-cal diversity and its ecological correlates. Our objec-tives were to: (1) interpret current phylogeographicpatterns in Percichthys in the light of the historicchanges in the aquatic environment in Patagonia;and (2) determine if there is molecular supportfor multiple species as described by Ringuelet et al.(1967) or López-Albarello (2004). A finding of sig-nificant genetic divergence and concordance betweenpatterns of genetic and morphological diversitywould lend support to the proposal that severalspecies of Percichthys are present east of the Andes.Alternatively, the absence of geographic patterningin genetic diversity coupled with shallow phy-logeographic structure would indicate that only asingle species of Percichthys is found in theregion.

MATERIAL AND METHODSCOLLECTION OF FISH

A total of 25 locations spanning the latitudinaland altitudinal range of Percichthys east of theAndes, and including four locations west of theAndes in Chile, were sampled between 1996 and2007 (Fig. 1). Molecular genetics and morphologicalanalyses were conducted on samples from 21 and 12of these locations, respectively, with eight of themanalysed for both molecular and morphological diver-sity. The four locations for which only morphologicaldata are available all belong to the Limay riverbasin (Fig. 1). The eight locations for which bothgenetic and morphological data were availableinclude populations from the six major Patagonianriver drainages with headwaters east of the Andes:(1) Limay or Negro drainage (Lakes Quillén andEspejo); (2) Puelo (Lake Puelo); (3) Futaleufú (LakesRivadavia and Futalufquen-Kruger); (4) Chubut-Chico-Senguerr (Lake Musters); (5) Baker (LakePueyrredón); and (6) Santa Cruz (Lake Argentino).The Puelo, Futaleufú, and Baker rivers drain into

the Pacific Ocean, whereas the rest drain into theAtlantic. Details on sampling procedures were pre-sented in Ruzzante et al. (1998, 2003, 2006, 2008).Briefly, fish were collected with sets of gillnetsplaced at between three and five sites per lake (2–5nights per lake, depending on the catch). Nets wereset before dusk and hauled in after dawn the fol-lowing morning. Fish were weighed, sexed, andassessed for reproductive status immediately uponretrieval. Blood or muscle tissue samples were takenand stored in ethanol for DNA analysis. Stomachswere removed, and stomach contents and fish werepreserved in 4% formaldehyde. Estimates of catchper unit effort (CPUE) were based on gillnet surfacearea and the number of hours of deployment.

MOLECULAR ANALYSIS

DNA extraction and mitochondrial DNA sequencingDetailed procedural information on DNA extrac-tion and amplification is available in Ruzzanteet al. (2006, 2008). Briefly, a fragment of the mtDNAcontrol region (~380 bp) was amplified from allsamples using the polymerase chain reaction (PCR)and standard methods on an MJ PTC-225 Thermocy-cler in 25-mL volumes using 2 mL of DNA extract astemplate. The primers L19 and MT16498H (Jerry &Baverstock, 1998) were used under the followingconditions: an initial denaturing cycle of 94 °C for5 min, 35 cycles with denaturation at 94 °C for1 min, annealing at 51 °C for 1 min, and extension at72 °C for 1 min, followed by a final extension step of5 min at 72 °C. PCR products were purified usingQIAGEN MinElute 96 PCR purification plates.For all samples, sequencing was conducted in bothdirections.

VISUALIZATION OF MOLECULAR DATA: GENGIS

The pattern of phylogeographic structure was visual-ized with GenGIS v1.08, a free downloadable bioin-formatics application that provides a 3D graphicalinterface for the merging of information on moleculardiversity (DNA sequences) with the geographic loca-tion from which the sequences were collected (http://kiwi.cs.dal.ca/GenGIS/Main_Page; Parks et al., 2009).A permutation test implemented in the application isdesigned for the testing of phylogeographic structure.In the present study we tested whether the lati-tudinal position of different haplotypes is correlatedwith their phylogenetic relatedness according to themaximum likelihood phylogram constructed withMEGA v5 (Tamura et al., 2011), and assuming defaultsettings of uniform rates among sites and a nearest-neighbour interchange (NNI) as the tree inferenceoption. Disagreement between the geographic gra-

516 D. E. RUZZANTE ET AL.

© 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 103, 514–529

Page 4: Phylogeography and phenotypic diversification in the Patagonian fish Percichthys trucha: the roles of Quaternary glacial cycles and natural selection

dient and the ordering of haplotypes in the treeis assessed by counting the number of crossings(a reflection of rank-order differences) that occurbetween the two orderings, and comparing this countwith the counts obtained from 1000 random per-mutations of the leaf labels of the tree. The result-ing P value reflects the number of replicates thathave equal or fewer crossings than the true tree, andP � 0.05 was taken as an indication of phylogeo-graphic structure.

MORPHOLOGICAL AND DIET ANALYSIS

Morphological measurements were carried out on atotal of N = 1067 individuals, from 12 populations, allinhabiting lakes. All measurements were performedon formalin-preserved material by the same person.We measured standard length (SL), head length (HL),length of the upper jaw (UJ), depth of the caudalpeduncle (CP), and length of the longest (usuallythe second) spine of the first dorsal fin (DF) (Fig. 2).Characters were selected based on their likely

Figure 1. Sampling locations (25) for Percichthys trucha. The species is found throughout Patagonia in Andean as wellas steppe freshwater environments. Collections span the altitudinal and latitudinal distributional range for this specieseast of the Andes, and took place between 1996 and 2007. Molecular genetic variation (Control Region (CR) haplotypes)was examined in individuals collected from 21 sites, phenotypic variation was examined in individuals collected from 12sites, with overlapping molecular and phenotypic information available for eight sites. Key: �, sites with molecularinformation (13); +, sites with molecular and morphological information (8); �, sites with morphological but no molecularinformation (4). The figure associates the phylogenetic relationship among Percichtys mtDNA CR haplotypes with thegeographic distribution of those haplotypes. Colours reflect latitude, going from red in the north to dark blue in the south.There is no relationship between the latitude (or population) at which haplotypes were found and the phylogeneticrelationships among haplotypes (permutation test between latitude and genetics: P > 0.333). The results suggest thoroughmixing among ancestral Percichthys populations throughout Patagonia east of the Andes. Mixing among populations fromdifferent drainages is likely to have taken place at times of high meltwater discharge during glacial stabilization and/ortermination periods (Martínez & Kutschker, 2011) on the current Patagonian steppe, and/or the exposed continental shelf(Ponce et al., 2011). Locations: 1, River Tunuyán (El Carrizal Reservoir); 2, River Atuel (El Nihuil Reservoir); 3, RiverColorado; 4, River Negro; 5, River Itata (Chile); 6, River Andalién (Chile); 7, River Bio-Bio (Chile); 8, Lake Blanca; 9, LakeRuca Choroi; 10, Lake Quillén; 11, Lake Falkner–Villarino; 12, Lake Espejo; 13, Lake Correntoso; 14, Lake Morenito; 15,Lake Puelo; 16, Lake Rivadavia; 17, Lake Futalaufquen-Kruger; 18, Lake Musters; 19, Lake La Plata; 20, Lake Silvia(Chile); 21, Lake Pueyrredón (Argentine)/Cochrane (Chile); 22, Rio Chico (near highway 3); 23, River Chalia (in TresLagos); 24, Lake Viedma (Lake de los Toros); 25, Lake Argentino.

PHYLOGEOGRAPHY OF PERCICHTHYS TRUCHA 517

© 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 103, 514–529

Page 5: Phylogeography and phenotypic diversification in the Patagonian fish Percichthys trucha: the roles of Quaternary glacial cycles and natural selection

relationship with feeding and/or swimming ability(HL, UJ, CP), as well as predator avoidance (DF), andrepresent the most repeatable of the phenotypic traitsused in previous studies (Ruzzante et al., 1998, 2003).These traits have also been used as diagnostic vari-ables for species identification within the Percichty-idae (Ringuelet et al., 1967). In addition, we measuredthe length of the four longest gill rakers on the firstleft branchial arch. In early collections (1996, 1998),gill rakers were drawn with a camera lucida attachedto a stereomicroscope and then measured. Later col-lections (2000, 2001, and 2005) were digitally photo-graphed under a stereomicroscope and measuredusing image analysis software. Prior to analysis, allvariables were standardized to a common fish sizeusing the relationship:

y x b SL SLi i i= ( ) − ( ) − ( )[ ]{ }log log log ,mean (1)

where yi and xi are the adjusted and original valuesfor the character in individuals i (i = 1, . . . , N), SLi isthe individual standard length, and b is the regres-sion coefficient of the logarithm of x on the logarithmof SL. Allometric relationships did not differ betweensexes for any lake, but they differed slightly amonglakes for some of the variables. Variables were there-fore standardized with sexes pooled, using individuallake allometric relationships. For the comparisonsamong populations we conducted a principal compo-nent analysis on standardized variables and ANOVA.

We obtained information on diet for the 12 popula-tions where morphological variation was measured.For each prey category present in the fish stomachs i(i.e. family, genus, or species), we calculated an indexof relative importance, RI, based on prey numberand size (Pinkas, Oliphant & Iverson, 1971): RIi =(Ni + Vi)Fi, where Ni is the percentage of total preyitems that were prey of type i, Vi is the percentage oftotal prey volume occupied by prey of type i, and Fi

is the percentage of non-empty guts that containedat least one prey of type i. For each population,RI indices were then expressed as a proportion of

the sum of all RI indices (proportional RI) for thatpopulation (Appendix). Diet analyses for the lakesthat were visited twice were conducted separately foreach year of collection, and then the proportionalRI values were averaged across the years. Diet datawere analysed using principal component analysis.

RESULTSPHYLOGEOGRAPHIC PATTERN OF PERCICHTHYS

TRUCHA IN PATAGONIA

Analysis of N = 107 haplotypes from the mtDNAcontrol region for the individuals collected from 21lakes and rivers shows no evidence of geographicpatterning in the distribution of genetic diversitywithin Percichthys east of the Andes (Fig. 1). InFigure 1, the haplotypes are linked to the population(lake or river) the individuals were collected from.Haplotypes found in more than one location are indi-cated by the presence of multifurcating nodes asterminals. Haplotype colour in this figure reflectslatitude, with the northernmost locations shown inred and the southernmost locations shown in darkblue. Related haplotypes are widely distributed acrosslatitude, as can be seen in the crossing of the linesconnecting sampling location, and the terminal nodesin the maximum-likelihood haplotype tree. No latitu-dinal or geographical pattern emerges among hap-lotypes collected from locations east of the Andes(Fig. 1). Three closely related haplotypes collectedfrom a single Chilean location (location 5, River Itata;Fig. 1) are slightly more divergent, as reflected bytheir relatively long branches in the maximum like-lihood haplotype tree (see also Ruzzante et al., 2006),although the permutation test designed to examinethe relationship between latitudinal position andgenetic relationship overall indicated no significantevidence of structure (P > 0.333). The relationshipbetween mitochondrial control region haplotypes andthe location from which they were sampled thereforeprovides no evidence of phylogeographic structureor of species-level differentiation among populationsof Percichthys east of the Andes in Patagonia.

MORPHOLOGICAL DIFFERENCES AMONG POPULATIONS

AMONG AND WITHIN DRAINAGES

As stated above, previous authors identified severalmorphological distinct groups within Percichthys eastof the Andes, with characters different enough to bedesignated as distinct species (Ringuelet et al., 1967;López-Albarello, 2004). Here we address the questionof morphological variability in the light of the littlegenetic variation described above.

A principal components analysis, using the fivestandardized traits, suggests that morphologically,

Figure 2. External morphological measurements ofPercichthys trucha; CP, caudal peduncle depth; DF, lengthof dorsal spine; HL, head length; SL, standard length;UJ, length of upper jaw.

518 D. E. RUZZANTE ET AL.

© 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 103, 514–529

Page 6: Phylogeography and phenotypic diversification in the Patagonian fish Percichthys trucha: the roles of Quaternary glacial cycles and natural selection

the 12 populations fall into three groups (Fig. 3). Theseparation is primarily along the first component,which accounted for 61% of the total variance, andwas most strongly influenced by differences in dorsalfin spine length (DFloading = 0.907), and to a lesserextent by gill raker and upper jaw lengths (Table 1).The first group (shortest dorsal spines, and relativelylong gill rakers and upper jaw) consisted of the north-ern populations (Limay and Futalaufquen basins),the second group (intermediate dorsal fin spine, gillrakers, and upper jaw) contained the two southernand one central lake populations (Argentino, Puey-rredón, and Puelo), and the third group (longestdorsal fin spines) comprised the only lake located onthe Patagonian steppe (Musters).

The second principal component (23.4% of the vari-ance) was influenced most strongly by gill rakerlength (MGRloading = -0.875) and, to a lesser extent, bydorsal spine length and upper jaw length (Table 1).Most of the variation along this axis was amongthe northern lakes (Fig. 3). The most northerly popu-lation, from Lake Ruca Choroi, had the shortestgill rakers, but otherwise there seemed to be nogeographic pattern. If the dorsal fin spine is excludedfrom the principal component analysis, there are nodistinct population groupings along either PC1 orPC2 (result not shown).

FISH COMMUNITY COMPOSITION, DENSITY, AND

MORPHOLOGICAL DIFFERENCES AMONG PERCICHTHYS

SP. POPULATIONS FROM 12 LAKES

The composition of the fish community differed sig-nificantly among the 12 lakes. Galaxias platei (puyen)were present in eight lakes, silverside (pejerrey)Odontesthes hatcheri were present in five lakes, andcatfish (bagre) Diplomystes viedmensis were presentin one lake (Table 2). All twelve lakes contained intro-duced salmonids, but differed in abundance andspecies composition (Table 2). Rainbow trout (Onco-rhynchus mykiss) were present in 11 of the 12 lakes,and were usually the most abundant salmonid(Table 2). Brook trout (Salvelinus fontinalis) were

present in eight of the 12 lakes, and were oftenabundant, and brown trout (Salmo trutta), althoughcollected from nine of the 12 lakes, were usuallyfound in small numbers. Lake trout, Salvelinusnamaycush, were found only in Lake Argentino.

Our CPUE estimates are only rough approxi-mations of population density; much more inten-sive sampling, including temporal replication, wouldbe required to get accurate abundance estimates.However, they do provide an index of the variationamong lakes in densities of Percichthys and otherspecies. There was almost a 20-fold difference amonglakes in CPUE for Percichthys, from a low of 1.0 fishper unit effort in Lake Espejo to a high of 19 fish perunit effort in Lake Musters (Table 3). Salmonid den-sities also varied among lakes, with CPUEs rangingfrom 0.6 to 4.3. Single estimates were averaged forlakes with temporal replication (Espejo, Quillén, andRivadavia). Lake Morenito is a shallow lake, used asthe breeding area for fish from Lake Moreno, to whichit is connected through a short passage (Buria et al.,2007). Individuals spend most of their time in thelarger Lake Moreno, and thus we used a CPUE esti-mate for Lake Moreno (CPUE = 1.92), obtained forcollections not used in the present study.

Variation among populations in morphology (PC1scores) was highly correlated with the variationin density of Percichthys (CPUEPercichthys) (r = -0.90,d.f. = 9, P < 0.0001), and of salmonids (CPUEsalmonide)(r = -0.90, d.f. = 9, P � 0.0002; Fig. 3C). Thus, indi-viduals from populations of Percichthys in lakes withhigh densities of conspecifics or of introduced salmo-nids tended to have longer dorsal spines than indi-viduals from lakes with low densities of conspecificsand/or salmonids.

DIET AND MORPHOLOGICAL DIFFERENCES

AMONG POPULATIONS

A principal component analysis of diet (based onthe proportional RI of prey types) indicated thatvariation in the importance of three prey types(Odonata, Amphipoda, and Chironomidae) was

Table 1. Principal component loading for five phenotypic traits measured across all 12 Percichthys trucha populationsand percentage of the total variance explained by each of the principal components

Phenotypic trait PC 1 PC 2 PC 3 PC 4 PC 5

Head length (HL) 0.113 -0.110 -0.422 0.051 0.891Upper jaw length (UJ) 0.231 -0.231 -0.797 0.240 -0.449Caudal peduncle length (CP) 0.004 -0.116 -0.222 -0.967 -0.064Dorsal fin Spine length (DF) -0.907 -0.395 -0.128 0.071 0.002Mean gill raker length (MGR) 0.334 -0.875 0.349 0.020 0.014% of total variance explained 61.4 23.4 8.3 5.2 1.8

PHYLOGEOGRAPHY OF PERCICHTHYS TRUCHA 519

© 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 103, 514–529

Page 7: Phylogeography and phenotypic diversification in the Patagonian fish Percichthys trucha: the roles of Quaternary glacial cycles and natural selection

Figure 3. A, among-lake variation in morphology. Principal component analysis based on five phenotypic traits: lengthof dorsal fin spine (DF); length of upper jaw (UJ); length of the head (HL); depth of caudal peduncle (CP); and length ofgill rakers (average of four longest rakers; MGR). Scatterplot of PC1 versus PC2 (together these two PCs explain ~85%of the total phenotypic variation). Length of the dorsal fin spine (DF) and mean gill raker length (MGR) are the traitswith the highest loading on PC1 (-0.907) and PC2 (-0.875), respectively. Lake abbreviations are as listed in Table 2. B,length of the dorsal fin spine (DF) plotted against standard length of the fish. C, scatterplot of catch per unit effort (CPUE)in numbers of individuals [�, P. trucha; +, Salmonids) versus morphology (principal component 1)]. Fish density correlateswith PC1 scores (morphology), for which the length of the dorsal fin spine (DF) has the highest loading. D, among-lakevariation in the diet of P. trucha. The principal component analysis is based on the relative index (%) of prey items listedin Table 4. Scatterplot of PC1 versus PC2. Together these two principal components explain 74% of the total variance inthe diet of P. trucha among lakes. The two diet items with the highest loading on PC1 are Odonata (-0.888) andChironomid larvae and pupae (0.376), and on PC2 are Amphipoda (0.746) and Chironomid larvae and pupae (-0.555). SeeTable 4. Key for panels A, B and C: black, Percichthys trucha from Limay and Futalaufquen systems pooled (Limay – lakesRuca Choroi, Quillén, Falkner–Villarino, Espejo, Correntoso, and Morenito; Futalaufquen – lakes Rivadavia andFutalaufquen); brown, P. trucha from Lakes Puelo, Pueyrredón, and Argentino; blue, P. trucha from Lake Musters. Inpanel C, no CPUE was available for Lake Argentino where P. trucha were caught by seine netting.

520 D. E. RUZZANTE ET AL.

© 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 103, 514–529

Page 8: Phylogeography and phenotypic diversification in the Patagonian fish Percichthys trucha: the roles of Quaternary glacial cycles and natural selection

Tab

le2.

Fis

has

sem

blag

eco

mpo

siti

on:n

um

ber

offi

shca

ptu

red

and

stan

dard

len

gth

[mea

n(S

D)

and

ran

gein

mm

]of

nat

ive

and

non

-nat

ive

fish

cau

ght

ingi

ll-n

etsa

mpl

ing

in12

Pat

agon

ian

lake

ssp

ann

ing

the

dist

ribu

tion

ofP

erci

chth

ystr

uch

a

Lak

eP

erci

chth

ystr

uch

aO

nco

rhyn

chu

sm

ykis

sS

alve

lin

us

fon

tin

alis

Sal

mo

tru

tta

Sal

veli

nu

sn

amay

cush

Od

onte

sth

esh

atch

eri

Gal

axia

spl

atei

Dip

lom

yste

svi

edm

ensi

s

Ru

caC

hor

oi(r

c)N

um

ber

offi

sh15

1611

––

–2

–M

ean

len

gth

(SD

)25

1(7

4)30

2(6

6)33

6(5

1)–

––

298

–R

ange

125–

395

185–

435

255–

400

––

–27

0–32

5–

Qu

illé

n96

&98

(ql)

Nu

mbe

rof

fish

171

4411

12–

1–

–M

ean

len

gth

(SD

)24

2(7

1)34

6(8

5)28

4(6

1)49

7(6

8)–

345

––

Ran

ge84

–375

120–

490

180–

355

350–

593

––

––

Fal

kner

–Vil

lari

no

(fv)

Nu

mbe

rof

fish

2213

171

––

1–

Mea

nle

ngt

h(S

D)

321

(43)

260

(138

)25

4(8

2)42

0–

–20

0–

Ran

ge19

5–39

010

0–47

016

0–42

0–

––

––

Cor

ren

toso

(cr)

Nu

mbe

rof

fish

2112

42

––

–M

ean

len

gth

(SD

)32

6(3

2)37

1(1

35)

281

(43)

433

(110

)–

––

–R

ange

230–

370

105–

485

255–

345

355–

510

––

––

Esp

ejo

96&

98(e

j)N

um

ber

offi

sh82

4441

6–

–41

1M

ean

len

gth

(SD

)32

7(6

4)37

9(1

08)

308

(81)

431

(131

)–

–18

5(3

6)16

1R

ange

149–

420

115–

525

157–

473

265–

565

––

115–

275

Mor

enit

o(m

r)N

um

ber

offi

sh50

–2

–15

––

Mea

nle

ngt

h(S

D)

311

(34)

–26

3(1

1)–

267

(38)

––

Ran

ge19

6–39

0–

255–

270

–21

0–33

0–

PHYLOGEOGRAPHY OF PERCICHTHYS TRUCHA 521

© 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 103, 514–529

Page 9: Phylogeography and phenotypic diversification in the Patagonian fish Percichthys trucha: the roles of Quaternary glacial cycles and natural selection

Tab

le2.

Con

tin

ued

Lak

eP

erci

chth

ystr

uch

aO

nco

rhyn

chu

sm

ykis

sS

alve

lin

us

fon

tin

alis

Sal

mo

tru

tta

Sal

veli

nu

sn

amay

cush

Od

onte

sth

esh

atch

eri

Gal

axia

spl

atei

Dip

lom

yste

svi

edm

ensi

s

Pu

elo

(pu

)N

um

ber

offi

sh17

59

–18

––

34–

Mea

nle

ngt

h(S

D)

271

(72)

NA

–N

A–

–N

A–

Ran

ge95

–460

NA

–N

A–

–N

A–

Riv

adav

ia98

&00

(rv)

Nu

mbe

rof

fish

174

145

256

–39

37M

ean

len

gth

(SD

)20

9(8

3)30

0(8

1)25

8(5

7)55

0(9

5)–

241

(62)

200

(47)

Ran

ge89

–450

105–

447

172–

380

395–

650

–13

6–37

112

7–29

0

Fu

tala

ufq

uen

–Kru

ger

(fu

)N

um

ber

offi

sh71

331

29–

–5

–M

ean

len

gth

(SD

)21

8(8

4)33

337

040

9(6

8)–

–14

4–

Ran

ge90

–386

122–

460

–19

5–55

0–

–11

7–19

5–

Mu

ster

s(m

u)

Nu

mbe

rof

fish

166

38[2

0]*

––

–28

5[6

7]*

––

Mea

nle

ngt

h(S

D)

230

(64)

321

(85)

––

–20

7(3

8)–

–R

ange

85–3

8422

5–55

0–

––

138–

332

––

Pu

eyrr

edón

(pe)

Nu

mbe

rof

fish

5420

––

–19

19–

Mea

nle

ngt

h(S

D)

298

(45)

429

(73)

––

–29

1(4

8)29

4(6

4)–

Ran

ge17

0–38

032

0–63

0–

––

203–

365

175–

334

Arg

enti

no

(ar)

Nu

mbe

rof

fish

204

37–

–17

–1

–M

ean

len

gth

(SD

)24

5(6

9)23

0(1

01)

––

382

(90)

–14

5–

Ran

ge80

–384

101–

403

––

260–

516

––

*Act

ual

nu

mbe

rof

fish

mea

sure

dou

tof

thos

eca

ptu

red.

522 D. E. RUZZANTE ET AL.

© 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 103, 514–529

Page 10: Phylogeography and phenotypic diversification in the Patagonian fish Percichthys trucha: the roles of Quaternary glacial cycles and natural selection

primarily responsible for distinguishing diets of thedifferent populations (Fig. 3D). Populations in theLimay river basin were distinguishable from southernor steppe populations along PC1 (51% of total vari-ance). Odonata were the prey item with the highestloading on PC1 (PC1 Odonataloading = -0.888; Table 4):Limay river basin populations tended to rely moreheavily on Odonata. The second principal component(23% of total variation) mostly reflected differences inthe importance of Chironomidae versus Amphipoda inthe diet (Table 4). For example, Percichthys fromLake Musters fed mostly on Amphipoda and to alesser extent also on Cladocera, with little contribu-tion from Chironomidae, whereas individuals fromLake Puelo relied mostly on Chironomidae (Fig. 3D;Table 4).

We estimated correlations between diet and mor-phology after the exclusion of the length of the dorsalfin spine from the suite of morphological traits, asthis trait is unlikely to be related to resource use.In this analysis PC1 (45% of total variation) formorphology chiefly reflected variation in gill rakerlength and, to a lesser extent, upper jaw length(loadingMGR = 0.902; loadingUJ = 0.382). There was a

negative correlation between morphology (PC1morph)and diet (PC1diet) (r = -0.61, d.f. = 10, P < 0.036), indi-cating that populations that relied more heavily onOdonata tended to have shorter gill rakers andshorter jaws. Diet varies seasonally, but differencesin sampling time among lakes were minor (all insummer – January and February).

MORPHOLOGICAL DIVERSITY WITHIN VERSUS

DIVERGENCE AMONG POPULATIONS

We used ANOVA to partition the variance for eachmorphological trait between the among- and within-population levels. Only the dorsal fin spine (DF) wasmore variable among populations than within popu-lations, with 72% of the total variance for this traitexplained by differences among populations (Table 5).Two traits are much more variable within thanamong populations: gill raker length and caudalpeduncle depth (Table 5).

DISCUSSION

In this study, we show what might be considered aparadox: a high level of morphological variabilityamong and within Percichthys populations, sufficientlyhigh that variants from different drainages have been

Table 3. Catch per unit of effort (CPUE) in numbers ofindividuals

CPUE

Ruca Choroi 1.44Quillén 2.07Falkner–Villarino 1.00Espejo 1.35Correntoso 1.25Morenito (Moreno) 1.92Puelo 8.5Rivadavia 3.38Futalaufquen–Kruger 2.68Musters 19.00Pueyrredón 3.00Argentino Not available

Table 4. Principal component loading for the six most important prey items in the diet of Percichthys trucha in 12 lakesdistributed throughout the range for the species in Patagonia

Prey type Comp1 Comp2 Comp3 Comp4 Comp5

Odonata -0.888 -0.145 0.132 -0.052 0.061Amphipoda 0.117 0.746 0.168 -0.214 0.437Chironomid Larvae & pupae 0.376 -0.555 0.343 -0.472 0.116Chilina 0.116 -0.137 -0.780 0.124 0.144Tricoptera 0.184 -0.070 0.389 0.824 0.004Cladocera 0.078 0.289 -0.009 -0.139 -0.842Percentage of Total Variance explained 50.7 23.2 13.6 9.6 2.2

Table 5. Percentage of total variance explained by differ-ences among and variation within populations in anANOVA framework for each of the five measured morpho-logical characters

Morphological character

Amongpopulations(% variance)

Withinpopulations(% variance)

Dorsal fin (DF) 72 28Head length (HL) 41 59Upper jaw (UJ) 35 65Mean gill raker length

(MGR)19 81

Caudal peduncle (CP) 12 88

PHYLOGEOGRAPHY OF PERCICHTHYS TRUCHA 523

© 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 103, 514–529

Page 11: Phylogeography and phenotypic diversification in the Patagonian fish Percichthys trucha: the roles of Quaternary glacial cycles and natural selection

deemed to be different species, yet we find no evidenceof deep divergence in the entire haplotype tree forPercichthys east of the Andes, and only a very shallowphylogeographic structure for the region. We arguehere that this pattern has been produced by two verydifferent processes, operating on different time scales.

The morphological trait identified as best distin-guishing groups of populations, and the only traitthat differed more among than within populations,was the length of the dorsal fin spine. Long dorsalspines have been associated with predator avoidancein fish (Januszkiewicz & Robinson, 2007), and wefound that variation among populations in spinelength correlates with the density of conspecifics andintroduced salmonids. Small Percichthys are vulner-able to predation by larger conspecifics, and perhapsalso by some salmonids. Longer spines could be adefensive trait that has evolved in populations sub-ject to high predation pressure, or could be a develop-mental response to predators, as has been shownfor sunfish (Januszkiewicz & Robinson, 2007).Superimposed on the phenotypic differences amongpopulations, there was strong evidence of markedintrapopulation phenotypic diversity involving mainlygill raker length, a trait often linked to resourceacquisition in fish. Thus, it appears that the relativeimportance of the various processes responsible forthe generation of within-population diversity maydiffer from those involved in allopatric differentiationamong populations (see also Calsbeek & Cox, 2010).Regardless of which processes have produced pheno-typic diversification within and among Percichthyspopulations east of the Andes, our molecular analysisindicates that the phenotypic differences do notcorrelate with genetic differentiation (as assessedfrom variation at the mitochondrial control region)(Fig. 1). Below we discuss the potential reasons for,and implications of, our results.

ABSENCE OF PHYLOGEOGRAPHIC STRUCTURE WITHIN

THE PERCICHTHYS TRUCHA COMPLEX

As we found previously with a smaller data set(Ruzzante et al., 2006), analysis of 21 populations ofPercichthys, encompassing their full distributionalrange east of the Andes as well as some westernpopulations, demonstrates little to no phylogeo-graphic structure. The very shallow structure east ofthe Andes suggests that there must have been rela-tively recent mixing of populations throughout theregion. We suspect that this mixing took place via twonon-exclusive mechanisms. Firstly, there may havebeen exchange among now disjunct lakes that werepart of larger proglacial lakes formed in front of theice during the retreat of the glaciers. Secondly, andprobably more importantly, individuals were probably

able to move between current drainages via the manybraided and deltaic connections that formed onthe exposed continental shelf during glacial periods(Martínez & Kutschker, 2011; Ponce et al., 2011).

Species respond in different ways to repeatedglacial advances and retreats, depending, in part, oncharacteristics such as cold tolerance and dispersalability. Recent studies suggest that the variousterrestrial Patagonian taxa did show distinctiveresponses, with some species surviving glacial periodsin one to a few southern refugia, whereas otherssurvived in and subsequently recolonized from north-ern areas (Sersic et al., 2011, Pardiñas et al., 2011).Likewise, there is evidence that some aquatic speciesappear to have survived in southern refugia (Zemlaket al., 2008, 2010), but it is likely that others weredriven extinct locally. Percichthys is a relatively warmwater-adapted Patagonian fish that reaches very highdensities in warmer steppe lakes and reservoirs. If itsurvived in southern drainages through the glacialcycles, the refugia must have been on the expandedPatagonian surface that included exposed areas of thecurrently submerged continental shelf (Cavallottoet al., 2011; Ponce et al., 2011).

Several lines of evidence suggest that the exposedcontinental shelf may have provided refugial habitatas well as opportunities for the movement of Percich-thys and other freshwater fauna among river drain-ages, from the Colorado River in the north to theSanta Cruz and Gallegos Rivers in the south. Bathy-metric images provide evidence for the presence ofendorheic basins with a circular morphology and adeeper centre than periphery in the present-day SanJorge, San Matías, and San José gulfs (Ponce et al.,2011). These areas would have been exposed duringglacial periods, and were probably filled with shallow,relatively warm water. Shallow, productive basins arean ideal habitat for Percichthys. The exposed shelf isrelatively homogeneous and flat, with a west–eastgradient of generally < 1%, and is thus conduciveto channel shifts and overflow during wetter periods.In addition, there is sedimentary evidence for theformation of braided river systems (which typicallyhave unstable, shifting channels) and deltaic fronts(e.g. the Colorado and Negro systems; Martínez &Kutschker, 2011). Thus, there was considerable poten-tial for the large, currently disjoint, Patagonian riversystems on the continental shelf to have merged con-tinually or intermittently during the extended fullglacial periods that lasted tens of thousands of years,and probably also during the relatively short glacialtermination periods, when river flows were high, upto ten times greater than today (Cavallotto et al.,2011; Martínez & Kutschker, 2011). The shallow phy-logeographic structure that we see in Percichthys isthus likely to have been maintained through periodic

524 D. E. RUZZANTE ET AL.

© 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 103, 514–529

Page 12: Phylogeography and phenotypic diversification in the Patagonian fish Percichthys trucha: the roles of Quaternary glacial cycles and natural selection

mixing, produced by large-scale landscape changesthat occurred as a function of changing climatethrough the Quaternary (Rabassa, Coronato &Salemme, 2005; Rabassa, 2008). The extent of mixingprobably varied among freshwater species, and wespeculate that species that inhabited the Patagoniansteppe and/or the shallow lake and river environ-ments on the now submerged continental shelf weremost susceptible.

PHENOTYPIC VARIATION

The remaining question, then, is what processes werelikely to have produced such divergent morphologyamong and within populations. The lakes we sur-veyed span the latitudinal and elevational range ofPercichthys in Patagonia east of the Andes, and differgreatly in the abiotic conditions that are associatedwith productivity. Deep Andean lakes of glacial ori-gin are ultraoligotrophic–oligotrophic, with nitrogenlevels sufficiently low so as to limit productivity (Sotoet al., 1994; Diaz et al., 2007). Steppe lakes, on theother hand, tend to be warmer, to have higher levelsof dissolved nutrients, and to be much more pro-ductive (Quirós & Drago, 1999; Diaz et al., 2007).Differences in productivity are likely to lead to differ-ences in both resource/competitive and predationregimes.

By far the most important morphological characterdifferentiating populations was the length of thedorsal fin spine, suggesting that predation (predatordefence or predator avoidance) may underpin muchof the among-population morphological diversity(see also Reimchen 1983 and Reimchen & Nosil 2002,and references therein). Percids typically erect theirdorsal spine in response to piscivorous fish (Ylönenet al., 2007): longer spines presumably reduce preda-tion risk, perhaps in part through increases in appar-ent size to gape-limited predators. Predators havebeen shown to induce morphological changes in fishthrough water-born chemicals, including the induc-tion of longer dorsal spines in sunfish (Januszkiewicz& Robinson, 2007), and greater body depth in sunfishand other species (Brönmark & Miner, 1992; Langer-hans et al., 2004; Januszkiewicz & Robinson, 2007).Body depth can also be related to predation riskthrough its effects on swimming performance andescape success (Domenici et al., 2008). The principalpredation threats to young Percichthys in the studylakes are conspecifics (large Percichthys are partiallypiscivorous) and, perhaps, introduced salmonids(Macchi et al., 1999). The much longer spines in high-density populations of Percichthys, such as the steppeLake Musters, probably result from an induced orevolutionary response to high predation intensity. Thelack of genetic differentiation of this population,

together with a plausible ecological explanation for itsmorphological divergence argues that the Percichthysspecies described for this lake, P. colhuapensis, is nomore than a morphotype of P. trucha that developsunder particular environmental conditions, i.e. highdensities of conspecific and salmonid predators and/orcompetitors.

The other characters that differed somewhatamong populations (head and upper jaw lengthsand, to a lesser extent, mean gill raker length) areusually related to feeding, and differences probablyresulted from variation in resources among thelakes. Resource availability and type can inducevariation in trophic morphology in fish: individualsthat feed primarily on zooplankton tend to developmore streamlined bodies, and a head morphologythat can efficiently consume small pelagic prey(many long, closely-spaced gill rakers), whereasthose that feed on benthic prey tend to developdeeper bodies, and sometimes longer, more robustjaws (Adams, Woltering & Alexander, 2003; Anders-son et al., 2005; Yonekura, Kohmatsu & Yuma, 2007;Berner et al., 2008). In oligotrophic Andean lakes,Percichthys feed primarily on benthic macroinverte-brates, although small crustaceans (e.g. cyclopoidcopepods and cladocerans) are consumed by juveniles(Ruzzante et al., 2003). Lake productivity affects theage at which young percids begin feeding on largerbenthic invertebrates (Persson, 1987; Huss, Byström& Persson, 2008), and variation in the timing of dietshifts can lead to differences in trophic morphology.The most productive lake in our study (LakeMusters) was the only lake in which small Crusta-ceans (Cladocera) were a significant part of the dietof adult Percichthys, perhaps because of a greateravailability of plankton or perhaps because of moreintense competition for benthic resources, and thevery distinctive morphology of individuals in thislake might therefore be resource related as well aspredation related.

Other marked differences among lakes in the typeof benthic prey consumed by Percichthys probablyalso reflect variation among lakes in resourceavailability/abundance. (As all fish were collected inthe summer, in January and February, differences indiet across lakes are not likely to be greatly con-founded by seasonal differences in composition of preycommunity). Some of the variation in diet could beassociated with variation in morphology: for instance,populations that relied heavily on Odonata tended tohave relatively short gill rakers and jaws comparedwith those that did not feed on Odonata. We do notknow the nature of any links between diet and trophicmorphology for Percichthys: adult morphology isalmost certainly influenced by the diet of early devel-opmental stages, and diet can also be affected by

PHYLOGEOGRAPHY OF PERCICHTHYS TRUCHA 525

© 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 103, 514–529

Page 13: Phylogeography and phenotypic diversification in the Patagonian fish Percichthys trucha: the roles of Quaternary glacial cycles and natural selection

predation regime. Thus, some combination of differ-ences in predation and resource regimes, both relatedto lake productivity, are likely to be responsible forthe morphological diversity within the species. Whatis very clear is that very different morphologies haveemerged in different aquatic environments withoutconcomitant genetic divergence.

Percichthys is also known to have variable morphol-ogy within lakes (Ruzzante et al., 1998, 2003; López-Albarello, 2004). Our analysis showed that twocharacters (gill raker length and caudal peduncledepth) were much more variable within than amongpopulations. Both traits are commonly linked toresource use. Competition for resources may promotediversification of resource use and associated diver-gence in phenotype (Lack, 1947; Schluter & McPhail,1992; Schluter, 1994), although other processessuch as predation can also play a significant role(Jablonski & Sepkoski, 1996; Rundle, Vamosi &Schluter, 2003; Langerhans et al., 2004; Nosil &Crespi, 2006; Meyer & Kassen, 2007). The populationwith the greatest interindividual variation in trophicmorphology, the steppe Lake Musters (D. E. Ruz-zante, unpubl. data), was also the most productivelake with the highest density of potential predators(Percichthys and salmonids). Thus, ecological factors(resource competition and/or predation) are likelyto be responsible for within-population morpho-logical variation as well as the differences amongpopulations.

SPECIATION WITHIN PATAGONIAN PERCICHTHYIDAE?

Several of the morphological characters examined inthis study (upper jaw length, body depth, and dorsalspine length) have been used to define differentspecies within the genus Percichthys (Ringuelet et al.,1967; López-Albarello, 2004). We found similarly highlevels of within- and among-population variability inthese traits, and identified three broad morphologicalgroups: one encompassing all Percichthys populationsfrom the Limay and Futalaufquen river basins; oneencompassing Perichthys from lakes Argentino, Puey-rredón, and Puelo; and one type from Lake Musters(Fig. 3A). In a recent and thorough attempt to sort outspecies designations within Argentine Percichthys,López-Albarello (2004) collapsed most of the previousspecies into a single species, P. trucha, but designatedthe Percichthys from Lake Argentino as P. laevis,whereas those from Lake Musters were presumed tobe P. colhuapensis. We also found these populations tobe morphologically distinct. However, the moleculardifferences (mtDNA and nuclear sequences) amongindividuals and populations that we describe here andin previous studies (Ruzzante et al., 2006, 2008) aremuch smaller than would be expected for species, or

even subspecies, designation. All Percichthys popula-tions in Patagonia east of the Andes thus appear tobelong to the same species: P. trucha. We argue thatdifferent processes produced the spatial patternsof genetic versus morphological variation. Shiftingaquatic landscapes during the Quaternary mixed thepopulations, producing a very shallow phylogeo-graphic structure east of the Andes, whereas ecologi-cal factors (perhaps differences in predation andcompetition regimes) most likely account for thecurrent morphological differences. We expect thatpopulations of P. trucha do diverge over time inresponse to different ecological pressures in differentenvironments, and that current P. trucha populationsin Patagonia may be in some intermediate stage ofan adaptive radiation. Divergence at the molecularlevel appears to be considerably short of speciation,however, perhaps because of the relatively youngage of individual populations, or perhaps becauseof inconsistency through time in the direction orstrength of selection pressures (e.g. Svanbäck &Persson, 2009). We suspect that these processesmay have been repeated multiple times in the past:populations in different environments underwentpartial but incomplete divergence caused, at least inpart, by natural selection, but the divergence was thenlost as climate change altered the landscape, allowinghaplotypes from different populations to mix.

ACKNOWLEDGEMENTS

We thank Guillermo Ortí (George Washington Uni-versity) and Evelyn Habit (Universidad de Con-cepción) for insightful comments on the article, andRobert Beiko and Donovan Parks (Faculty of Com-puter Science, Dalhousie University) for assistancewith GenGIS and Figure 1. We also thank the Com-mittee for Research and Exploration of the NationalGeographic Society, Washington, for generous supportfor fieldwork in 2000–01 (NGS 6799–00) and 2007–08(8168–07), NSERC Discovery grants, and a SpecialResearch Opportunity award (SROPJ/326493–06) toD.E.R. and S.J.W. as well as Universidad Nacionaldel Comahue, FONCYT, and CONICET (Argentina)grants for the support of fieldwork. We also acknow-ledge an NSF-PIRE award (OISE 0530267) for thesupport of collaborative research on PatagonianBiodiversity granted to the following institutions(listed alphabetically): Brigham Young University,Centro Nacional Patagónico, Dalhousie University,Instituto Botánico Darwinion, Universidad Austral deChile, Universidad Nacional del Comahue, Univer-sidad de Concepción, and University of Nebraska. Wethank Evelyn Habit for collections from Chile, andJuana Aigo, Miguel Battini, Víctor Cussac, Amalia

526 D. E. RUZZANTE ET AL.

© 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 103, 514–529

Page 14: Phylogeography and phenotypic diversification in the Patagonian fish Percichthys trucha: the roles of Quaternary glacial cycles and natural selection

Denegri, Maria Lattuca, Daniela Milano, and PabloVigliano for assistance in the field in Argentina,and Tyler Zemlak for assistance in the laboratory. Wethank Mauricio Uguccioni and José ‘Pepe’ De Giustofor their logistical support for more than a decade ofthe fieldwork in Patagonia.

REFERENCES

Adams CE, Woltering C, Alexander G. 2003. Epigeneticregulation of trophic morphology through feeding behaviourin Arctic charr, Salvelinnus alpinus. Biological Journal ofthe Linnean Society 78: 43–49.

Andersson J, Byström P, Persson L, De Roos AM. 2005.Plastic resource polymorphism: effects of resource availabil-ity on Arctic charr (Salvelinus alpinus) morphology. Biologi-cal Journal of the Linnean Society 85: 341–351.

Andersson J, Johansson F, Söderlund T. 2006. Inte-ractions between predator- and diet-induced phenotypicchanges in body shape of crucian carp. Proceedings of theRoyal Society of London B 273: 431–437.

Avise JC. 2000. Phylogeography: the history and formation ofspecies. Cambridge, MA: Harvard University Press.

Beheregaray LB. 2008. Twenty years of phylogeography: thestate of the field and the challenges for the Southern Hemi-sphere. Molecular Ecology 17: 3754–3664.

Berner D, Adams DC, Grandchamp A-C, Hendry AP.2008. Natural selection drives patterns of lake–streamdivergence in stickleback foraging morphology. Journal ofEvolutionary Biology 21: 1653–1665.

Brönmark C, Miner JG. 1992. Predator-induced phenotypi-cal change in body morphology in Crucian Carp. Science258: 1348–1350.

Buria L, Walde SJ, Battini MA, Macchi PJ, Alonso M,Ruzzante DE, Cussac VE. 2007. Movement of a nativeperch (Percichthys trucha) in a montane Patagonian lakeduring spawning and pre-spawning periods. Journal FishBiology 70: 215–230.

Calsbeek R, Cox RM. 2010. Experimentally assessing therelative importance of predation and competition as agentsof selection. Nature 465: 613–616.

Cavallotto JL, Violante RA, Hernández-Molina FJ. 2011.Geological aspects and evolution of the Patagonian Conti-nental Margin. Biological Journal of the Linnean Society103: 346–362.

Clapperton CM. 1993. Quaternary Geology and Geomorph-ology of South America. Amsterdam: Elsevier.

Diaz M, Pedrozo F, Reynolds C, Temporetti P. 2007.Chemical composition and the nitrogen-regulated trophicstate of Patagonian lakes. Limnologica 37: 17–27.

Domenici P, Turesson H, Brodersen J, Brönmark C.2008. Predator-induced morphology enhances escape loco-motion in crucian carp. Proceedings of the Royal Society ofLondon B 275: 195–200.

Eklöv P, Svanbäck R. 2006. Predation risk influences adap-tive morphological variation in fish populations. AmericanNaturalist 167: 440–452.

Hewitt G. 2000. The genetic legacy of the Quaternary iceages. Nature 405: 907–913.

Huss M, Byström P, Persson L. 2008. Resrouce hetergene-ity, diet shifts andintra-cohort competition: effects on sizedivergence in YOY fish. Oecologia 158: 249–257.

Jablonski D, Sepkoski JJ Jr. 1996. Paleobiology, commu-nity ecology, and scales of ecological pattern. Ecology 77:1367–1378.

Januszkiewicz AJ, Robinson BW. 2007. Divergent walleye(Sander vitreus)-mediated inducible defenses in the cen-trarchid pumpkinseed sunfish (Lepomis gibbosus). Biologi-cal Journal of the Linnean Society 90: 25–36.

Jerry DR, Baverstock PR. 1998. Consequences of a cat-adromous life-strategy for levels of mitochondrial DNA dif-ferentiation among populations of the Australian bass,Macquaria novemaculeata. Molecular Ecology 7: 1003–1013.

Lack D. 1947. Darwin’s finches. Cambridge: Cambridge Uni-versity Press.

Langerhans RB, Layman CA, Langerhans AK, DeWittTJ. 2004. Predator-driven phenotypic diversification inGambusia affinis. Evolution 58: 2305–2318.

López-Albarello A. 2004. Taxonomy of the genus Percichthys(Perciformes: Percichthyidae). Ichthyological Exploration ofFreshwaters 15: 331–350.

Macchi PJ, Cussac VE, Alonso MF, Denegri MA. 1999.Predation relationships between introduced salmonids andthe native fish fauna in lakes and reservoirs in northernPatagonia. Ecology of Freshwater Fish 8: 227–236.

Martínez O, Kutschker A. 2011. The ‘Rodados Patagónicos’(Patagonian Shingle Formation) of Eastern Patagonia: envi-ronmental Conditions of gravel sedimentation. BiologicalJournal of the Linnean Society 103: 336–345.

Meyer JR, Kassen R. 2007. The effects of competition andpredation on diversification in a model adaptive radiation.Nature 446: 432–435.

Milano D, Cussac VE, Macchi PJ, Ruzzante DE, AlonsoMF, Vigliano PH, Denegri MA. 2002. Predator associatedmorphology in Galaxias platei in Patagonia lakes. Journalof Fish Biology 61: 138–156.

Milano D, Ruzzante DE, Cussac VE, Macchi PJ, FerrizRA, Barriga JP, Aigo JA, Lattuca ME, Walde SJ. 2006.Latitudinal and ecological correlates of morphological varia-tion in Galaxias platei (Pisces: Galaxiidae) in Patagonia.Biological Journal of the Linnean Society 87: 69–82.

Nosil P, Crespi BJ. 2006. Experimental evidence thatpredation promotes divergence in adaptive radiation. Pro-ceedings of the National Academy of Sciences, Washington103: 9090–9095.

Pardiñas UFJ, Teta P, D’Elía G, Lessa EP. 2011. Theevolutionary history of sigmodontine rodents in Patagoniaand tierra del Fuego. Biological Journal of the LinneanSociety 103: 495–513.

Parks DH, Porter M, Churcher S, Wang S, Blouin C,Whalley J, Brooks S, Beiko RG. 2009. GenGIS: a geospa-tial information system for genomic data. Genome Research19: 1896–1904.

Persson L. 1987. Competition-induced switch in young ofthe year perch Perca fluviatilis: an experimental test of

PHYLOGEOGRAPHY OF PERCICHTHYS TRUCHA 527

© 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 103, 514–529

Page 15: Phylogeography and phenotypic diversification in the Patagonian fish Percichthys trucha: the roles of Quaternary glacial cycles and natural selection

resource limitation. Environmental Biology of Fishes 19:235–239.

Pinkas L, Oliphant MS, Iverson ZL. 1971. Food habitsof albacore bluefin tuna and bonito in California water.Department of Fish and Game Fisheries Bulletin 152:1–105.

Ponce JF, Rabassa J, Coronato A, Borromei AM. 2011.Paleographic evolution of the Atlantic coast of Pampa andPatagonia from threLast Glacial Maximum to the MiddleHolocene. Biological Journal of the Linnean Society 103:363–379.

Quirós R, Drago E. 1999. The environmental state of Argen-tinean lakes: An overview. Lakes & Reservoirs: Researchand Management 4: 55–64.

Rabassa J. 2008. Late Cenozoic Glaciations in Patagonia andTierra del Fuego. In: Rabassa J, ed. The Late Cenozoic ofPatagonia and Tierra del Fuego. Developments in quater-nary sciences. Vol. 11. Amsterdam: Elsevier, 151–204.

Rabassa J, Coronato A, Salemme M. 2005. Chronology ofthe Late Cenozoic Patagonian Glaciations and their corre-lation with biostratigraphic units in the Pampean Region(Argentina). Journal of South American Earth Sciences 20:81–103.

Rabassa J, Coronato A, Martínez O. 2011. Late CenozoicGlaciations in Patagonia and Tierra del Fuego: an updatedreview. Biological Journal of the Linnean Society 103: 316–335.

Reimchen TE. 1983. Structural relationships between spinesand lateral plates in threespine stickleback (Gasterosteusaculeatus). Evolution 37: 931–946.

Reimchen TE, Nosil P. 2002. Temporal variation in diver-gent selection on spine number in threespine stickleback.Evolution 56: 2472–2483.

Ringuelet RA, Aramburu RH, Alonso de Aramburu A.1967. Los Peces Argentinos de Agua Dulce. Gobernación:Provincia de Buenos Aires. Comisión de InvestigaciónCientífica.

Rundle HD, Vamosi SM, Schluter D. 2003. Experimentaltest of predation’s effect on divergent selection during char-acter displacement in sticklebacks. Proceedings of theNational Academy of Sciences USA 100: 14943–14948.

Ruzzante DE, Walde SJ, Cussac VE, Macchi PJ, AlonsoMF. 1998. Trophic polymorphism, habitat and diet segre-gation in Percichthys trucha (Pisces: Percichthyidae) in theAndes. Biological Journal of the Linnean Society 65: 191–214.

Ruzzante DE, Walde SJ, Cussac VE, Macchi PJ, AlonsoMF, Battini M. 2003. Resource polymorphism in a Patago-nia fish Percichthys trucha (Percichthyidae): phenotypic evi-dence for interlake pattern variation. Biological Journal ofthe Linnean Society 78: 497–515.

Ruzzante DE, Walde SJ, Cussac VE, Dalebout ML,Seibert J, Ortubay S, Habit E. 2006. Phylogeography ofthe Percichthyidae in Patagonia: roles of orogeny, glaciation,and volcanism. Molecular Ecology 15: 2949–2968.

Ruzzante DE, Walde SJ, Gosse JC, Cussac VE, Habit E,Zemlak TS, Adams EDM. 2008. Climate control onancestral population dynamics: insight from Patagonian fishphylogeography. Molecular Ecology 17: 2234–2244.

Schluter D. 1994. Experimental evidence that competitionpromotes divergence in adaptive radiation. Science 266:798–801.

Schluter D, McPhail JD. 1992. Ecological characterdisplacement and speciation in sticklebacks. AmericanNaturalist 140: 85–108.

Sérsic AN, Cosacov A, Cocucci AA, Johnson LA, PoznerR, Avila LJ, Sites JW Jr, Morando M. 2011. Emergingphylogeographic patterns of plants and terrestrial verte-brates from Patagonia. Biological Journal of the LinneanSociety 103: 475–494.

Soto D, Campos H, Steffen W, Parra O, Zuniga L. 1994.The Torres del Paine lake district (Chilean Patagonia): acase of potentially N-limited lakes andponds. Archiv furHydrobiologie Supplement 99: 181–197.

Sugden DE, Bentley MJ, Fogwill CJ, Hulton NRJ,McCulloch RD, Purves RS. 2005. Late-glacial glacierevents in southernmost South America: a blend of ‘northern’and ‘southern’ hemispheric climatic signals. GeografiskaAnnaler 87A: 273–288.

Svanbäck R, Persson L. 2009. Population density fluctua-tions change the selection gradient in Eurasian perch.American Naturalist 173: 507–516.

Vamosi SM. 2002. Predation sharpens the adaptive peaks:survival trade-offs in sympatric sticklebacks. AnnalesZoologici Fennici 39: 237–248.

Ylönen H, Kortet R, Myntti J, Vainikka A. 2007. Predatorodor recognition and antipredatory response in fish: does theprey know the predator diel rhythm? Acta Oecologia 31: 1–7.

Yonekura R, Kohmatsu Y, Yuma M. 2007. Difference in thepredation impact enhanced by morphological divergencebetween introduced fish populations. Biological Journal ofthe Linnean Society 91: 601–610.

Zemlak TS, Habit EM, Walde SJ, Battini MA, AdamsEDM, Ruzzante DE. 2008. Across the southern Andes onfin: glacial refugia, drainage reversals & a secondarycontact zone revealed by the phylogeographic signal of Gal-axias platei in Patagonia. Molecular Ecology 17: 5049–5061.

Zemlak TS, Habit E, Walde SJ, Carrea C, Ruzzante DE.2010. Surviving historical Patagonian landscapes andclimate: molecular insights from Galaxias maculatus.BMC, Evolutionary Biology 10: 67. Available at http://www.biomedcentral.com/1471-2148/10/67.

528 D. E. RUZZANTE ET AL.

© 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 103, 514–529

Page 16: Phylogeography and phenotypic diversification in the Patagonian fish Percichthys trucha: the roles of Quaternary glacial cycles and natural selection

AP

PE

ND

IX

Tab

leA

1.P

ropo

rtio

nal

rela

tive

indi

ces

(pro

port

ion

alR

Is)

for

each

prey

cate

gory

byla

kepo

pula

tion

Lak

eO

don

ata

(nym

phs)

Am

phip

oda

Ch

iron

omid

larv

ae&

pupa

eTr

ich

opte

raA

egla

spE

phem

erop

tera

Sam

asta

cus

Ch

ilin

asp

.D

iplo

don

sp.

Cla

doce

raP

lan

cton

Oli

goch

aete

sO

rgan

icm

atte

rF

ish

Ru

caC

hor

oi0.

542

0.02

40.

222

00.

020

0.19

20

00

00

00

0Q

uil

lén

0.47

60.

356

0.08

50.

077

0.00

30.

002

00

00

00

00.

001

Vil

lari

no–

Fal

kner

0.86

20

00.

008

00

0.01

90

00

00

00.

017

Esp

ejo

0.84

20.

102

0.00

10.

025

0.00

30

0.02

40

00

00

00

Cor

ren

toso

0.42

90.

462

00.

045

00

0.06

30.

001

00

00

00

Mor

enit

o0.

656

0.02

20.

147

00

0.01

50.

001

0.14

50

00

00

0.01

5P

uel

o0.

005

0.00

10.

840

0.01

40

0.02

80

0.03

00.

074

00

00

0.00

3R

ivad

avia

0.11

70.

347

0.24

60.

143

00

00.

001

00.

005

00.

015

0.11

40.

001

Fu

tala

ufq

uen

–Kru

ger

0.84

50.

001

0.01

60.

028

00.

001

00

00

0.09

70.

012

00

Mu

ster

s0.

001

0.58

20.

016

0.00

40

00

0.00

20

0.38

90.

001

00

0P

uey

rred

ón0

0.04

20.

307

0.64

30

00

0.00

10

00

00

0.00

7A

rgen

tin

o0

0.00

30.

127

00

00

0.60

20

00

0.21

40

0.04

8

PHYLOGEOGRAPHY OF PERCICHTHYS TRUCHA 529

© 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 103, 514–529