Thescalingand selectionofsexuallydimorphiccharacters:an exampleusingtheMarbledTeal
AndyJ.Green
Green,A.J.2000.Thescalingandandselectionofsexuallydimorphic characters: an exampleusingtheMarbled Teal. –J.Avian Biol.31:345–350.
Current theory and empirical evidence suggests that, if a character is sexually dimorphic as a result of sexual selection, it should be positively allometric (i.e. relatively larger in larger individuals), whereas ifthe dimorphism isthe result of natural selection(e.g.nichedivergence),itshould beisometric. Ishowhowthiscan be used to study the selective forces responsible for dimorphic morphological characters,using the monochromaticMarbled Teal Marmaronetta angustirostrisas anexample. Inabsolute terms, first-year maletealshaveahigher body mass,wing length, head length and billlength than females. In relative terms (controlling for bodysize),malesstillhavelongerwings,headsandbills.ThescalinginMarbled Teal suggeststhatbill andheaddimorphismsareduetosexualselection,whereaswing dimorphism isdue to natural selection. Tail length issexually monomorphic but positivelyallometric, possiblybecauseofadisplayfunction. Suchscalingstudiesare easyto carry out, and provide a useful complement to direct investigation of the influenceofvariation inthesizeofdimorphic characters onmatingsuccess,foraging efficiencyetc.
Department ofApplied Biology, Estacio´nBiolo´gicadeDon˜ana,CSIC, Avda.Mar´ıa
Luisas/n,41013Sevilla,Spain.E-mail:
Sexualdimorphism inavianmorphology canevolveby eithersexualornatural selection,anditisgenerallynot self-evident which isresponsible (Sigurjo´nsdo´ttir 1981, Livezey and Humphrey 1984, Nudds and Kaminski
1984, Price1984, Shine1989, Savalli1995, Craigand Villet1998).Good morphometricdata areavailablefor many species (e.g. for ducks, Cramp and Simmons
1977,OwenandMontgomery 1978,Livezeyand Humphrey 1984, Fox et al. 1992, Livezey 1993), yet therearerelativelyfewstudiestestingtheimportanceof dimorphic characters for e.g.mating successor inter- sexualnicheseparation (Nudds andKaminski 1984, Holmberg etal.1989,Sorenson and Derrickson1994). Thusforthegreatmajorityofbirdspecies, nodirect evidenceisavailabletoidentifytheselective forces responsiblefordimorphic characters. However,inthis paper Iillustrate howmorphometricdata canbeused tostudy differencesinthe‘‘scaling’’or‘‘allometry’’of characters (i.e.thechangeintheirrelativesizeinlarger birds) in bird populations, which provides valuable indirectevidenceoftheselective forcesunderlying dimorphism.
Thescalingofbirdmorphology hasattractedconsid- erable attentionat inter- and intra-specific levels (Prange et al. 1979, Peters 1983, Boag 1984, Calder
1984, Schmidt-Nielsen 1984, Livezey and Humphrey
1986,Rayner 1988,Harvey and Krebs 1990,Livezey
1990,Bjo¨rklund 1994),and mechanical, developmental and other explanationshave been sought for the pat- terns observed. However, littleattentionhasbeenpaid totheuse ofintraspecificscalinginmatureindividuals (‘‘staticallometry’’)inexplainingsexualdimorphism in birds.If sexualselectionactstoexaggerateacharacter, current theoryandempiricalevidencesuggestsitshould bepositivelyallometric(i.e. relativelylargerinlarger birds). Furthermore,ifsexualselection on acharacter actsmore strongly ononesex,leadingtodimorphism, the allometry should be more positive in that sex (EmersonandVoris1992, Petrie1992, Simmonsand Scheepers1996, SimmonsandTomkins1996). Incon- trast,ifnatural selection(i.e.nichedivergence)is re- sponsible fortheobserved sexualdimorphism,nosuch positiveallometryorsexualdifferencein scalingis expected.
© JOURNALOF AVIAN BIOLOGY
Inthispaper Iaddress dimorphism inrelative char- acter size, but not dimorphism in body size per se (Ranta etal.1994).Withstandardmorphometricdata, the scaling ofeach character can becompared within andbetweensexes,andthisallowsustoidentifywhich characters arelikelytobesexuallyselected,and which arelikelytobenaturally selected.Thesehypotheses can then betested byfieldor laboratoryexperiments. For example, the importanceofpositively allometric char- acters in mate choice can be tested by manipulative experiments (e.g.Møller1990,JonesandHunter 1993). I illustrate this technique using the Marbled Teal Marmaronetta angustirostris, which is highly unusual amongst northernhemisphere members ofthesubfam- ilyAnatinae (sensuLivezey1986,1997)inlackingovert sexual differences in plumage (i.e. it is almost monochromatic, Cramp and Simmons 1977, Livezey
1996). Thisworkconstitutes thefirstdetailedstudyof sexualdimorphism inthisgloballythreatened(Collaret al.1994) species.Firstly,Iaskwhatdifferencesexist betweenthesexesofMarbled Tealinstandardmorpho- metricmeasures.Secondly,Iconsiderwhichof these differencespersistwhentaking into account thediffer- enceinoverallsizebetweensexes (e.g.domaleshave longerbillssimplybecausemalesarebigger,oraretheir billsrelativelylongerforagivensize?).Afteridentifying thescalingofcharacters ineachsex,Ithencomparethe natureof thesexualdimorphismobservedforagiven characterwiththesexualdifferencein its scaling,to determine thelikelynature ofevolutionaryforceslead- ingtotheobserved sexualdimorphism.
Methods
Marbled Tealbecomesexuallymature intheirfirstyear (Cramp andSimmons1977).TheMarbled Tealstudied (N=56)wereallrecentlyfledgedbirdsmeasured on10
September 1996.Theywerebirdsrescuedin1996from a concrete-lined canal in Alicante, Spain within two weeksofhatching (Navarroetal.1995)then reared in captivityattheCentreforNature ProtectionandStudy in Valencia. They werereleased into the wildshortly after measuring.
Birdsweresexedbycloacal inspection and weighed tothe nearestgramonanelectronicbalance.Measure- ments used were folded wing length (distance from carpal joint to the tip of the longest primary), tarsal length(fromthemiddlepoint ofthejointbetweentibia and tarsometatarsusbehind tothemiddle point ofthe jointbetweentarsometarsusandmiddletoe),maximum head length, bill length (culmen) and maximum tail length. Folded wingand tail length weremeasured to thenearestmmwitharuler.Maximum headlength,bill and tarsus weremeasured to the nearest 0.1mmwith digital calipers. Cranium length (excluding billlength)
wasalsoestimatedbysubtracting billlengthfrom maximum head length.
Statisticalanalyses
Absolute sexualdimorphism (not accounting forbody size)wasstudiedwithtwosamplet-testsusingLevene’s testtoconfirmhomogeneity ofvariances.
Sexualdimorphism controlling forsizewasanalysed via generalizedlinearmodelsusingGLIMsoftware (Crawley1993),includingsex (factor)andbodymassor winglength(continuous variableasanindexofsize)as predictor variablesandwithamorphometricdependent variable. A normal error and identity link functions were usedintheGLIM models,andallmorphometric variableswereloge transformed.Thesignificanceofthe sexualdifferenceintheslopebetweenYand logmass
orlogwing wastestedbyaddinganinteraction term betweensexand thelinear predictor. Winglength was chosenasalinearmeasure ofsizetoaccompany body mass sincewinglength isthe best linear predictor of aviansize, atleastinpasserines(Gosleretal.1998). Furthermore,although winglengthcanbeconfounded byageordate(OwenandMontgomery1978,Ormerod and Tyler1990),thiswasnot thecaseinthisstudy.
Theallometric slopesofeachcharacter inrelation to body mass were calculated using reduced major axis (RMA) modelstoplacethelineofbestfitinbivariate plots ofloge transformeddata. Theresultsofordinary least squares regression (OLS) and RMA regressions are increasingly divergent as r decreases, and slopes become seriously underestimated by OLS methods. RMA istherecommended regression method forscal- ingstudies, particularlywherethere isno information
about error variances, and X and Y variables have differentunitssuchas lengthandmass(Ricker1984, Rayner 1985,McArdle 1988,Pagel and Harvey 1988, LaBarbera1989,Sokal and Rohlf 1995,but seeJoli- coeur 1990). The RMA slope is equal to the ratio betweenthestandarddeviations ofYand X(and also totheOLSslopedividedbyr).SeeMcArdle (1988)for methods usedtocalculate thestatistical significanceof thedeviationbetweenobservedRMAslopes andthose expectedunderthenullhypothesisof isometry(forthe purpose ofthispaper, Iconsider isometrytobewhere, inlog-logbivariate plots, linear measures scalewith a slopeofonewitheachother, and withaslopeofone third against body mass).
Instudies whereitispossibletotake morphometric measuresofalargenumber ofcharacters (e.g.inskele- talstudies),itcanbeappropriate tousePC-1froma principalcomponent analysisofacharacter setasan indicator of body size, then use RMA to study the scalingofother characters ofinterest against PC-1(see e.g.Livezeyand Humphrey 1986).
Table1. Differenceinmeanmeasurements inmaleandfemale Marbled Teal, tested with two sample t-tests. Mass isin g, other measures areinmm.**P0.01, ***P0.001.
Males (N=26)Females (N=30)
Table 3. Scaling ofMarbled Teal linear morphometric mea- suresagainst body mass,bysex.
Males (N=26)Females (N=30)
Measure / Mean / S.D. / Mean / S.D. / t54 / Measure / RMA slope / r / RMA slope / rWing / 0.371 / 0.337 / 0.242 / 0.260
Mass / 410.1 / 28.9 / 386.4 / 32.0 / 2.90** / Tarsus / 0.483 / 0.365 / 0.361 / 0.066
Wing / 202.9 / 5.2 / 198.0 / 4.0 / 3.99*** / Tail1 / 0.580* / 0.376 / 0.482 / 0.047
Tarsus / 38.0 / 1.3 / 37.4 / 1.1 / 1.96 / Head / 0.507* / 0.434 / 0.368 / 0.522
Tail / 61.41 / 2.2 / 62.2 / 2.5 / 1.37 / Bill / 0.489* / 0.619 / 0.352 / 0.225
Head / 96.5 / 3.5 / 90.8 / 2.8 / 6.75***
Bill / 45.4 / 1.6 / 43.6 / 1.3 / 4.71*** / 1formales,N=24.Both body massand thelinear measure
Cranium / 51.1 / 2.9 / 47.2 / 2.3 / 5.66*** / wereloge transformed.Thedeviation oftheRMA slopefrom
that expected under isometry (1/3) was tested for statistical
1N=24. / significance.*P0.05.
Results
Absolutedimorphism
MaleMarbled Tealwereheavier,andhadlongerwings, heads,billsandcraniums(Table1) thanfemales.No significant difference was observed between sexes in tarsus ortaillength.
Relativedimorphismcontrollingforsize
When controlling for body mass, males had signifi- cantly longer wings, heads, bills and craniums than females,whereas there wasno significant differencein tailortarsuslength(Table2).Whencontrolling statisti- callyforwinglength asanalternative indexofoverall size,malesstillhadsignificantlylongerheads,billsand craniums,whereasfemales hadsignificantlylongertails (Table 2). Partial correlation coefficients (controlling forsex)withbodymasswerestatisticallysignificantfor allmeasuresexceptfortarsus andtaillength.Similarly, correlations with wingweresignificant except for bill length (Table 2).
Scaling
Whenrelatedtobodymass,wingandtarsus lengthdid not deviate significantly from isometry for either sex (Table 3).For males, tail length, head length and bill length were all positively allometric (Table 3). For females,nocharacter deviatedsignificantlyfromisome- try (Table 3).For all characters,the allometric slope wasmorepositiveformalesthan forfemales(Table3, Fig. 1). However, differences in slopes between sexes wereonlysignificantinthecaseofbilllength(Table2, Fig.1).
Discussion
Empiricalevidence inintegratedstudiesofinsects,am- phibians and mammals sofar supports the hypothesis that sexuallyselectedcharacters arepositivelyallomet- ric whereasnaturally selected characterstendtobe isometric(Emerson andVoris1992,Green 1992,Petrie
1992, SimmonsandScheepers1996, Simmonsand Tomkins 1996).However, there have been almost no tests ofthis hypothesis inbirds. The frontal shieldof
Table 2. Multiple regression ofmorphometricmeasurements inMarbled Teal, using identity link function and normal error distributioninGLIM.
YMass
Wing
Sex
F2,53r2
Wing / 4.87* / – / (−)8.87** / 10.7*** / 0.288Tarsus / 2.16 / – / (−)1.69 / 3.03 / 0.164
– / 16.7*** / 0.00 / 10.8*** / 0.290
Tail / 1.37 / – / (+)1.69 / 1.59 / 0.059
– / 11.7** / (+)10.2** / 6.92** / 0.214
Head / 15.6*** / – / (−)32.3*** / 37.2*** / 0.584
– / 6.22* / (−)26.0*** / 28.5*** / 0.518
Bill / 10.0** / – / (−)12.5*** / 17.8*** / 0.402
– / 1.83 / (−)12.4*** / 12.0*** / 0.312
Cranium / 8.14** / – / (−)21.1*** / 22.4*** / 0.459
– / 4.90* / (−)17.3*** / 19.9*** / 0.428
TheYvariable and thepredictors Mass and Wingareloge transformed.F-values arepresented forthepartial effectsofsex (factor oftwolevels)and body massorwing.F2,53 and r2 refertothewholeregression. (+) represents caseswherethefitted value of the dependent variable was higher for females whilst controlling for Mass or Wing, and viceversa. **P0.01,
***P0.001.N=56exceptforlogTail(N=54).Theeffectofanadded interaction term(Sex×MassorSex×Wing)wasnot significantwiththeexception ofY=Bill,forwhichSex×Mass had asignificanteffect(F1,52=4.74,P0.05).
the MoorhenGallinulachloropusispositively allomet- ric,apparentlybecauseofitsimportanceindefenceof breedingterritories(Petrie1988). Insteamer-ducks (Tachyeres),wingsareusedinmale-male combat and, compared with other skeletal characters, wing bones arepositivelyallometricin malesbutnotin females (Livezeyand Humphrey 1984,1986).Thereisevidence that extremesexualtaildimorphism insomespeciesof widowsandbishopsEuplectesis aresultofsexual selection,andpositiveallometryof maletailshasbeen shown at the interspecific level(Savalli 1995,but see Craig and Villet1998).
Iftheincreased expression ofacharacter issexually selected,thecharacter is predictedtobepositively allometric,essentiallyowing totherelativelygreater benefitsobtainableforalargerindividual frommaking a giveninvestment ina character (Green 1992,Petrie
1992).Thereareseveralreasonswhythesebenefitsmay increase in larger individuals, although no empirical studies have so far distinguished between them. One possible explanationon the basis of age (that larger animalsareolderandtherefore investrelativelymorein reproduction since their life expectancy is lower) is excluded from thecurrent study because there wasno agevariation inthesample ofMarbled Teal.Ifsexual selectionactsmore strongly inonesexthan theother, as is usually the case, the character should be both relatively larger and more positivelyallometric inthat sex.Intheory,matechoicecouldalternatively selectfor reduced character size, although no such cases have been documented (seeAndersson 1994for review).In the sexin which smaller characters were sexually se- lected, the character would be expected to be both relativelysmallerand negativelyallometric inthat sex. On the other hand, wherenatural selection leadsto dimorphism inacharacter, thereismuchlessreason to
Fig. 1. Scaling ofbilllength against body mass inMarbled
Teal. Lines are plotted by reduced major axis (males: y=
0.873+0.489x; females: y=1.682+0.352x). ---El---, males;
–•–,females.
expect positiveallometryanda sexualdifferencein allometric slopes (although these are possible under sometheoretical conditions: seeGreen1992).Thus,the scaling relationships of the characters concerned provideapotential meanstodistinguishbetweensexual and natural selectionascausesofdimorphism.
Thisisthefirstavianstudytoexplicitlycompare the sexualdimorphism inasetofmorphometriccharacters withthescalingofthosesamecharacters atanintraspe- cific level,inorder toinvestigatethelikelycausesof dimorphism. FirstyearmaleMarbled Tealaresignifi- cantly larger than females in most absolute morpho- metricmeasurements, withtheexception oftarsus and taillength.Similarresultshavebeenobtained forsexual dimorphism inboth firstyearand older Marbled Teal at Slimbridge, UK (A. Chisholm unpubl. data), with the additionalfindingthat maleshave agreater abso- lutebill height(atthefrontedgeofthenostrils).This apparent sexual dimorphism is not merely a conse- quenceofdifferencesinoverallsizebetweenthesexes: for a given size,males have longer wings, heads and bills.Itisnotjusttheculmenwhichisrelativelylonger in males, but the cranium itself. For a given wing length, females have relatively longer tails, although thisappears toreflecttheir shorter wings.
Theevidencefrom scalingsuggeststhat sexualselec- tionisthemaincauseofsexualdimorphism inbilland head morphometry in the Marbled Teal. Heads and bills arepositivelyallometricinmalesbutisometricin females,thesexualdifferencein allometricslopebeing statistically significantforbilllength.Wingsaredimor- phic, but are isometric in both sexes,suggesting that natural selectionisresponsible forthesexualdifference inwingmorphology.Tarsuslengthis sexually monomorphic andisometric.Taillengthis sexually monomorphic,buttailsarepositivelyallometricinboth sexes (although significantly so only in males). It is possible that tail length is sexually selected in both sexes(seeJones and Hunter 1993).
Sexualdimorphism inmass,winglengthandbilland head length such asthat observed inMarbled Teal is typical of that observed in other Anatinae species (Cramp andSimmons1977,Livezey1990,1993,Foxet al.1992)andsuggeststhat, whereastheMarbled Tealis unusualamongst northern hemisphereAnatinae inits monochromatism, thereis nothingparticularlyunique about thedimorphism initsmorphometrics.Although malesmaymake alimitedcontributiontobrood rear- ing(Green 1997),Marbled Teal show the serial monogamy and much greater reproductiveinvestment offemalesthat iscommon totheAnatinae (Batt etal.
1992, Green 1998). Little previous attempt has been madetoinvestigatethecausesof sexualdimorphismin Anatinae.Nudds and Kaminski (1984)suggested that billdimorphism isselectedbynatural selection(reduced intersexualcompetitionforfood)basedon divergent distributionsof billlength in each sex.Sorenson and
Derrickson(1994) found no evidence that the larger tails of male Northern Pintail Anas acuta confer a mating advantage.
So far there is little evidence available from field studiestoassessthecontribution ofsexualornatural selectiontothesexualdimorphism observedinMarbled Teal. Males display their heads repeatedly to females during group courtship (Cramp and Simmons 1977, Green inpress),and largerheads and billsmayconfer anadvantage through femalechoiceormale-malecom- petition. Thisissupportedbythesexualdifferencesin plumage,whicharemostpronounced inthegreater development ofacrestonthenapeandforehead ofthe male, whilst billcolorationdiffers between sexes (NavarroandRobledano1995,Green inpress).Wings arenot displayed overtlywhenMarbled Tealcourt on thewater,anditseemslesslikelythat theyaresexually selected,although somecourtship occasionally appears tooccuronthewing(Navarro andRobledano 1995). Wing sizecan also be dimorphic to provide aerody- namiccompensation forexaggerated,sexuallyselected characters such as tails (Andersson and Andersson
1994), but tail length is not exaggerated in Marbled Teal compared with other Anatinae (Cramp and Sim- mons 1977). Tails are always visible during Marbled Tealcourtship (Johnsgard1965) and,onthewater, Marbled Tealalwaysholdtheirtailswell abovethe horizontalinfullviewtoconspecifics(pers.obs.).
Sexualselection thus offersthemost likelyexplana- tion for the bill and head dimorphism observed in Marbled Teal,and alternatives based onnatural selec- tionarenotsupportedbyscaling.Scalingsuggeststhat natural selectionismorelikelytoexplaintheobserved wingdimorphism (e.g.viasexualdifferencesinmigra- tion strategies, Carbone and Owen 1995). However, further research isrequired to confirm the hypotheses made onthebasisofsuchscalingstudies.
Scaling studies such as the one conducted here are easy to carry out, often with pre-existing data, and provide ausefulcomplement to direct investigation of theinfluenceofvariation inthesize ofdimorphic characters on mating success, foraging efficiencyetc. For example,theycouldbeusedapriori toidentify characters mostworthyofmanipulationinmatechoice experiments. In order to clarify the relationships be- tweenscalingandsexualselectionin birds,itwouldbe enlightening tocarry out studies similar tothisoneof sexualdifferencesinscalinginbirdspeciesinwhichthe significanceofdimorphic characters formating success hasbeenstudiedinmoredetail(e.g.Møller1990,Jones and Hunter 1993).
Acknowledgements –TheConseller´ıade MediAmbient,Gen- eralitat Valenciana allowed me to study their Marbled Teal andJ.Belliure,J.C.Dolz,N.Ramo´n,N.Selva,C.E.Smith, M.Va´zquezand C.Viedmahelpedmetomeasure them.The WildfowlandWetlands Trust provided essentialequipment. J. A.Amat,J.Figuerola, L.Hillstro¨m,B.Hughes,T.C.Michot,
R. E. Simmons and an anonymousreferee made comments that helpedmetoimprove themanuscript.
References
Andersson, M.1994.Sexualselection. –Princeton University
Press,Princeton.
Andersson, S.and Andersson, M. 1994.Tail ornamentation,
sizedimorphism and wing length in the genus Euplectes
(Ploceinae). –Auk 111:80–86.
Batt,B.D.J,Afton, A.D.,Anderson, M.G.,Ankney, C.D.,
Johnson, D.H.,Kadlec,J.A.andKrapu,G.L.(eds)1992.
Ecology and managementofbreeding waterfowl. – Uni-
versityofMinnesota Press,Minneapolis and London.
Bjo¨rklund, M. 1994.Allometric relations in three speciesof
finches(Aves:Fringillidae). – J. Zool., Lond. 233:657–
668.
Boag,P.T.1984.Growth andallometry ofexternalmorphol-
ogyinDarwin’sfinches(Geospiza)onIslaDaphne Major,
Galapagos.–J.Zool., Lond. 204:413–441.
Calder,W.A.1984.Size,function, andlifehistory. –Harvard
University Press,Cambridge, Mass.
Carbone, C.andOwen,M.1995.Differential migration ofthe
sexesofPochard Aythya ferina:results from aEuropean
survey. –Wildfowl46:99–108.
Collar,N.J.,Crosby,M.J.andStattersfield, A.J.1994.Birds
toWatch 2.TheWorld ListofThreatenedBirds.Birdlife
Conservation Series No. 4. – Birdlife International,
Cambridge.
Craig,A.J.F.K.andVillet,M.H.1998.Sexualdimorphism
and tail-length inwidowbirds and bishopbirds (Ploceidae:
Euplectesspp.):areassessment. –Ibis140:137–143.
Cramp, S.andSimmons,K.E.L.(eds).1977.Thebirdsofthe
western Palearctic. Vol. 1. – Oxford University Press,
Oxford.
Crawley,M.J.1993.GLIM forecologists. –BlackwellScien-
tific,Oxford.
Emerson, S.B.and Voris, H. 1992.Competing explanations
forsexualdimorphism inavoicelessBornean frog.–Func.
Ecol.6:654–660.
Fox, A.D.,King,R.andWatkin, J.1992.Seasonal variation
inweight,body measurements and condition offree-living
Teal. –BirdStudy 39:53–62.
Gosler, A.G., Greenwood,J.J.D., Baker, J.K.and David-
son,N.C.1998.Thefielddeterminationofbody sizeand
condition in passerines: a report to the British Ringing
Committee. –BirdStudy 45:92–103.
Green, A. J. 1992. Positive allometry is likely with mate
choice, competitive display and other functions. – Anim.
Behav.43:170–172.
Green, A. J. 1997.Brood attendance and brood care in the
Marbled Teal,Marmaronetta angustirostris.–J.Orn. 138:
443–449.
Green, A. J. 1998.Clutch size,brood sizeand brood emer-
genceintheMarbled TealMarmaronetta angustirostrisin
the Marismas del Guadalquivir,southwest Spain. – Ibis
140:670–675.
Green, A.J.Inpress.Marbled Teal. –In:Kear, J.andWWT
(eds).Birdfamiliesoftheworld: ducks, geese,swansand
screamers. Oxford University Press,Oxford.
Harvey, P. H. and Krebs, J. R. 1990.Comparingbrains. –
Science249:140–146.
Holmberg, K., Edsman, L.and Klint, T.1989.Female mate
preferences and male attributes in mallard ducks Anas
platyrhynchos. –Anim. Behav.38:1–7.
Johnsgard,P. A. 1965.Handbookofwaterfowl behavior. –
Ithaca, NewYork.
Jolicoeur, P.1990.Bivariate allometry: interval estimation of
theslopesoftheordinary and standardizednormal major
axes and structural relationship. – J. theor. Biol. 144:
275–285.
Jones, I.L.and Hunter,F. M.1993.Mutual sexualselection inamonogamousseabird. –Nature 362:238–239.
LaBarbera, M. 1989. Analyzing body size as a factor in ecologyand evolution. – Annu. Rev. Ecol. Syst.20:97–
117.
Livezey,B.C.1986.Aphylogenetic analysisofrecentAnseri-
form genera using morphologicalcharacters. – Auk 103:
737–754.
Livezey, B. C. 1990. Evolutionarymorphology of flightless-
nessintheAuckland Islands teal. –Condor 92:639–673.
Livezey, B. C. 1993. Comparative morphometrics of Anas
ducks, with particular reference to the Hawaiian Duck
Anaswyvilliana,Laysan Duck A. laysanensis,and Eaton’s
Pintail A. eatoni. –Wildfowl44:75–100.
Livezey, B. C. 1996. A phylogenetic analysis of modern
pochards (Anatidae: Aythyini). –Auk 113:74–93.
Livezey, B. C. 1997. A phylogenetic classification of water-
fowl (Aves: Anseriformes), including selected fossil spe-
cies.–Ann. Carnegie Mus. 66:457–496.
Livezey,B.C.andHumphrey,P.S.1984.Sexualdimorphism
incontinentalsteamer-ducks.–Condor 86:368–377.
Livezey, B.C. and Humphrey,P. S.1986.Flightlessness in
steamer-ducks (Anatidae: Tachyeres): its morphological
basesand probable evolution. –Evolution 40:540–558.
McArdle, B.H. 1988.The structuralrelationship: regression
inbiology. –Can. J.Zool. 66:2329–2339.
Møller, A.P.1990.Male tail length and femalemate choice
in the monogamous swallow Hirundo rustica. – Anim.
Behav.39:458–465.
Navarro,J.D., Green, A.J.and Aranda,J.C.1995.Status
of Marbled Teal in southern Alicante, Spain. – IWRB
ThreatenedWaterfowl Research Group Newsletter 8:7–
10.
Navarro,J. D. and Robledano,F. (Coordinators).1995.La
Cerceta Pardilla Marmaronetta angustirostris en Espan˜a.
– InstitutoNacional para la Conservacio´n de la Natu-
raleza, Madrid.
Nudds, T.D. and Kaminski, R. M. 1984.Sexualsizedimor-
phisminrelation toresource partitioninginNorth Ameri-
candabbling ducks. –Can. J.Zool. 62:2009–2012.
Ormerod, S. J. and Tyler, S. J. 1990. Assessment of body
condition in Dippers Cincluscinclus:potential pitfalls in
thederivation and useofcondition indicesbasedonbody
proportions.–Ring. Migr. 11:31–41.
Owen,M. and Montgomery,S.1978.Bodymeasurements of
Mallard caught inBritain. –Wildfowl29:123–134.
Pagel, M. D. and Harvey, P.H. 1988.Thetaxon-level prob-
lemin the evolution of mammalian brain size:facts and
artifacts. –Am.Nat. 132:344–359.
Peters, R. H. 1983.The ecological implications ofbody size.
–Cambridge University Press,Cambridge.
Petrie, M. 1988.Intraspecific variation instructures that dis- play competitive ability: large animals invest relatively more. –Anim. Behav.36:1174–1179.
Petrie,M.1992. Areallsecondarysexualdisplaystructures positivelyallometric and, ifso,why?– Anim. Behav.43:
173–175.
Prange, H. D., Anderson, J. F. and Rahn, H. 1979.Scaling ofskeletal mass to body mass inbirds and mammals. – Am.Nat. 113:103–122.
Price,T.D.1984.Theevolution ofsexualsizedimorphism in
Darwin’s finches. –Am.Nat. 123:500–518.
Ranta,E.,Laurila, A.and Elmberg, J.1994.Reinventing the wheel: analysis of sexual dimorphism in body size. – Oikos 70:313–321.
Rayner, J.M.V.1985. Linearrelationsinbiomechanics:the statistics ofscalingfunctions. –J.Zool., Lond. 206:415–
439.
Rayner, J.M. V.1988.Form and function inavian flight. – In: Johnston, R. F. (ed.) Current Ornithology. Vol. 5. Plenum Press,NewYork, pp. 1–66.
Ricker, W. E. 1984.Computationand usesof central trend lines. –Can. J.Zool. 62:1897–1905.
Savalli, U. M. 1995.The evolution of tail-length in widow- birds (Ploceidae): tests ofalternatives to sexual selection.