MALE COURTSHIP VOCALIZATIONS AS CUES FOR IN THE SATIN (PTILONORHYNCHUS VIOLACEUS)

CHRISTOPHER A. LOFFREDO AND GERALD BORGIA Departmentof Zoology,University of Maryland,College Park, Maryland 20742 USA

AI3STRACT.--MaleSatin (Ptilonorhynchus violaceus) court femalesat specialized structurescalled bowers. Courtship includes a complexpattern of vocalizationsin which a broad-band,mechanical-sounding song is followed by interspecificmimicry. We studiedthe effect of male courtshipdisplays on male mating successin Satin Bowerbirds.Data from 2 yearsof field researchshowed low between-maledifferences in mechanicalcomponents of courtshipsong and high variability betweenmales in mimeticsinging. Older malessang longer and higher-qualitybouts of mimicry than did youngermales. In one year, courtship song featureswere correlatedwith male mating success.The resultssuggest that female SatinBowerbirds use male courtship vocalizations in their mate-choicedecisions. We discuss hypothesesabout assessment of male age and dominancefrom courtshipvocalizations and suggestthat thesesongs have evolvedas a result of selectionfor male displaycharacteristics that provide femaleswith information about the relative quality of prospectivemates. Re- ceived27 June1985, accepted20 September1985.

MALESatin Bowerbirds(Ptilonorhynchus vio- ingnessto copulateor by flying away. Given laceus)build specializedstructures called bow- the complexity of male display and the atten- ers that are used as sites for courting females tion femalespay to displayingmales, it is likely and for mating (Vellenga 1970,Donaghey 1981, that male displayshave an important effect on Borgia 1985). Femalesraise their young unas- female mating decisions,yet there have been sistedby males,and malesdo not associatewith no detailedstudies of displayin any bowerbird females after mating. Males decorate their species. bowers with a variety of natural objectsand The relationshipbetween male display and attempt to steal decorationsand destroy the female choice in polygynous species,such as bowers of other males. bowerbirds, in which males do not provide ma- Much effort has been focused on under- terial benefitsto femalesor their offspring,has standingthe role of the bower and its decora- been the subject of considerable speculation tions in female mating decisions.Borgia (1985) (Fisher 1930, Zahavi 1975, Davis and O'Donald demonstrateda skewed distribution of matings 1976, Emlen and Oring 1977, Halliday 1978, among male Satin Bowerbirds, differences Borgia 1979, West-Eberhard 1979, Bradbury among males in bower decoration and con- 1981, Hamilton and Zuk 1982, Bradbury and struction, and a consistent pattern of female Gibson 1983, Kodric-Brown and Brown 1984, preference for males with well-constructedand Borgia et al. 1985). One view of the sexual se- well-decorated bowers. In addition, male mat- lection process is that females choose mates ing successis influenced negatively by bower based on characters that indicate the relative destruction and decoration-stealing (Borgia ability of malesto sire fit (e.g. viable) offspring 1986a,Borgia and Gore 1986). (Zahavi 1975,Halliday 1978,Borgia 1979, Ham- Male courtship vocalizations are another ilton and Zuk 1982, Thornhill 1980, Borgia et component of male sexual display that may al. 1985). An alternative model suggeststhat have important effectson female mating deci- male sexualdisplays are arbitrary resultsof the sions. Male courtship is intense, involving runaway processof (Fisher dance movements and postures that are coor- 1930).This implies that femalepreferences have dinated with the simultaneousproduction of evolved that have no effect on enhancingthe complex vocalizations. The female crouches viability of the offspring(Lande 1981,Kirkpat- within the bower, closely observing the male rick 1982, Arnold 1983). during the display, and controls the outcome Developing hypothesesthat separate these of the courtshipeither by signaling her will- modelshas proven very difficult. One method 189 The Auk 103:189-195. January 1986 190 LOFFREDOAND BORGIA [Auk, Vol. 103 is to consider the of male display Analyzer (Model 4512). A Kay ElemetricsDigital (Borgia 1986a).Models that predict gainsin off- Sono-Graphwas used to prepare spectrograms. spring viability require that male displayspro- Analysis-of-variance tests (one-way ANOVA), vide cues that allow females to assess characters Spearman rank correlations (r.), and the Student- importantto demonstratingthe quality of males Newman-Keuls (SNK) procedure for multiple comparisonsof meanswere used for statisticalcom- as sires. For example, female preference for parisons(Helwig and Council 1982). Means are ex- older males may produceoffspring with high pressedas g + 1 SD. survivorship becausethese males have dem- Becauseof the low amplitude of mimetic sounds onstratedhigh viability through their own sur- (seebelow), it wasnot possibleto accuratelymeasure vival (Trivers 1972, Howard 1974, Halliday frequencies.However, it was apparentthat quality 1978, Wilbur et al. 1978). If such a model is of mimetic songsdiffered between males,and so a operating, then we expect that cues exist in subjectivemethod was used to produce an index of malesthat indicate male age and that these are qualitativedifferences among mimetic songsof dif- correlated with male mating success.Alterna- ferent males.All mimetic songswere given a score tively, females may prefer dominant (Borgia (1, 2, or 3) for "structuralintegrity": high-scoredmi- meticsongs were of relativelypure tonal structure, 1979) or disease-resistantmales (Hamilton and free from broad-bandand other atonalsounds pro- Zuk 1982). By contrast,runaway models make duced by the singer, and were free from acoustical no predictions about the nature of male dis- distortionsresulting from productionby the singer plays and are untestableon that basis. of other soundsduring mimetic singing.Figures 2 Our objectiveshere are to describeintrapop- and 3 represent several high- and low-scored mi- ulational variation in the courtship-display metic songs. vocalizations of Satin Bowerbirds, to look for To validate subjectivescoring of mimetic songs, correlates between male vocal patterns and two volunteerswere each given the same random other male characteristicssuch as age and bow- sampleof spectrogramstaken from the analyzeddata er quality, and to usethese results to determine and were askedto sort the spectrogramson the basis if the minimum conditions for male assessment of this three-level subjectivescoring system. There was completeagreement between the resultsof the models have been met. two independenttrials and the resultsobtained by the authorfor the samesample (r. = 1.00,P < 0.001). METHODS

Field researchwas carried out at Wallaby Creek, RESULTS BeauryState Forest, New SouthWales, , from August through December in 1982 and 1983. The Descriptionof the male courtshipsong.--The study site is characterizedby open-canopy,subtrop- song used by male Satin Bowerbirdsduring ical eucalyptusscrub forest and patchesof rain forest. courtship(and in other specificbehavioral set- Recordingsof maleSatin Bowerbird courtship songs tings; see below) consistsof a long seriesof were obtained with Uher 4000L and Marantz PMD harsh-soundingbuzzes and pulses followed 220 tape recorders,using small condensermicro- immediatelyby a boutof interspecificmimicry. phonessuspended in the foliage1 m from the bower The two major portions of the courtshipsong platform.Only sexualdisplay songs at or near the henceforth will be termed the "mechanical" and bower were recorded,although male Satin Bower- use a variety of calls in other behavioralset- "mimetic" portions. tings and at placesaway from the bower. Observa- The courtshipsong lasts24.0 + 8.2 s. Within tions of focal displaying maleswere made by field the mechanicalportion there are 2-6 (œ= 4.2) observers from blinds stationed within 10 m of the wide-band "long buzzes"(Fig. 1), eachof which bower. In addition, a camera system was used to is 0.35-1.50 s long with a frequency range of monitoractivities continuously at bowersand to ob- 0.06 --- 0.01 to 7.39 + 0.02 kHz. Each long buzz tain recordsof male mating success(see Borgia 1985). is separatedfrom the next by one or more "short We recorded vocalizations in 1983 for 22 individ- buzzes" (harsh soundswith a frequency range ual males(20 adultsand 2 juveniles).Recordings from of 4.00 +__0.03 to 7.20 ___0.05 kHz, each of which 1982 for 2 additional juveniles and 11 of the same lasts for less than 0.20 s) and several short adultswere includedin the analysis,for a totalof 24 individuals.A meanof 12.0vocal samples were ana- (<0.10 s) pulselike sounds(Fig. 1). The me- lyzedfor eachmale, with a totalof 288samples. Spec- chanicalportion of the song lasts 17.0 + 3.2 s. trographic sound analysiswas performed using a For eachmechanical song analyzed, the follow- PrincetonApplied ResearchFFT Real Time Spectrum ing variables were measured:the duration of January1986] BowerbirdCourtship Vocalizations 191

16-

0 • • 4 Time (sec) Fig. 1. Componentsof the mechanicalportion of male courtshipsong. Time (sec) Fig. 3. Lewin'sHoneyeater mimetic song: (A) high the mechanical song; the numbers of long and (B) low scoresfor structuralintegrity. The sounds buzzes,short buzzes,and pulses;and the peak between 4.0 and 6.0 kHz are cicada calls. frequency(kHz), low frequency(kHz), and du- ration of each long buzz. ic calls are given, the kookaburra call always The mimetic portion of the courtship song precedesthe honeyeater. consistsof mimetic songs of either or both of The kookaburramimetic song is a structur- two avianspecies: the LaughingKookaburra (Da- ally complexseries of notes of gradually rising celonovaeguineae) and the Lewin's Honeyeat- frequency and intensity (see Fig. 2A) and is er (Meliphagalewinii). No other speciesare given for 2.9 + 1.6 s. The honeyeatercall con- regularly imitated by Satin Bowerbirdsat Wal- sists of a long train of simple, highly discrete laby Creek, although individuals in other pop- notes of constant frequency (2.80-3.36 kHz). ulations are known to mimic other bird species The notes are repeated rapidly for 4.0 ___2.7 s (Chaffer1984). For the individualsin this study, (seeFig. 3A). For each mimetic song analyzed, when both kookaburraand honeyeatermimet- the total mimetic, kookaburra,and honeyeater durations were measured. Male courtshipsongs, age, and matingsuccess.- For all 12 song variables measured, ANOVA testsindicated that a significant proportion of variability in these measurementswas a result of between-male differences. For these vari- ables,the averageF-value was 4.24 (range:2.26- 6.51; each F significant at P < 0.05), and the averageproportion of variability explained by differences among individuals (r 2) was 0.312 (range: 0.224-0.459). We tested to determine if differences in the Time (sec) time of year at which the songswere recorded or if differencesin behavioralsettings affected song characteristics.Means for one- and two- week intervals were calculatedfor song vari- ables in both years. Multiple comparisonsof means (SNK) revealed no significant differ- ences.Behavioral contexts in which recordings were obtained included male courtshipsongs given to femalesat the bower site, to juvenile and adult malesat the bower, at empty bowers (i.e. no other conspecificspresent), and from Time (sec) tree perches.Multiple comparison tests (SNK) Fig. 2. LaughingKookaburra mimetic song:(A) of mean song characteristicsby behavioral set- high and (B) low scorefor structuralintegrity. Note ting indicated that only songs given from tree band of cicada calls between 3.5 and 4.0 kHz. perchesdiffered significantly from other songs. 192 LOFFREDOAND BORGIA [Auk, Vol. 103

TABLE 1. Comparisonsof male age with mechanicaland with mimetic song variables.SI = structural integ- rity.

1982 1983

r, P r, P Mechanicalsong (n = 13) (n = 22) Mechanical duration 0.022 NS - 0.062 NS Peak frequency 0.167 NS 0.046 NS Low frequency 0.124 NS 0.065 NS Frequencyrange 0.072 NS -0.166 NS No. of long buzzes -0.073 NS -0.073 NS No. of short buzzes 0.071 NS 0.011 NS Mimetic song (n = 13) (n = 22) Mimetic duration 0.217 0.001 0.621 0.001 Kookaburra duration 0.813 0.002 0.459 0.023 Kookaburra SI - 0.457 0.001 -0.743 0.001 Honeyeaterduration 0.225 0.034 0.381 NS Honeyeater SI - 0.334 0.001 -0.621 0.002

Data on tree songsconstituted a very small part Even when juvenile maleswere excludedfrom of the sample, and these data were excluded analysis,adult males showed significant age- from further analysis. correlateddifferences in mimetic singing. Having demonstrated intermale variability In 1982,there was a significantcorrelation of in male courtshipsinging, we then divided the male mating successand the number of short individuals into seven age categoriesbased on buzzes in the mechanical song (Table 2), but known historiesof uniquely color-bandedin- this correlation was not significant in 1983. dividuals (see Borgia 1985). Male age was cor- None of the other mechanical song variables related significantlywith mimetic but not with measured was significantly correlated with mechanicalsong features(Table 1). Older males mating successin either year. Mimetic songs tended to give songswith significantlylonger showeda similar pattern (Table 2): in 1982,male mimetic portions than did younger males.The matingsuccess was significantly correlated with mimetic songs also varied qualitatively with total mimetic duration and with kookaburra age: in both 1982 and 1983, older males gave structural integrity, while in 1983 there were kookaburra and honeyeater calls of higher no significantcorrelations of any mimetic song structural integrity than did younger males. variables and male mating success.

TABLE2. Comparisonsof male mating successwith mechanicaland with mimetic song variables.$I = structural integrity.

1982 1983

r, P r, P Mechanicalsong (n = 11) (n = 22) Mechanical duration - 0.484 NS -0.357 NS Peak frequency 0.355 NS 0.067 NS Low frequency 0.429 NS 0.232 NS Frequencyrange 0.050 NS 0.126 NS No. of long buzzes 0.104 NS -0.069 NS No. of short buzzes -0.706 0.003 -0.363 NS Mimetic song (n = 11) (n = 22) Mimetic duration 0.774 0.040 0.007 NS Kookaburra duration 0.270 NS 0.073 NS Kookaburra SI - 0.635 0.024 0.080 NS Honeyeaterduration 0.522 NS 0.226 NS Honeyeater SI - 0.079 NS -0.234 NS January1986] BowerbirdCourtship Vocalizations 193

TABLE3. Comparisonsof mimetic songtypes with male age and mating success.

Mating success Percentageof Age 1982 1983 songswith: rs P rs P r. P Both mimetic types 0.220 NS -0.086 NS -0.239 NS Only kookaburra 0.282 NS -0.272 NS 0.050 NS Only honeyeater -0.237 NS 0.387 NS 0.025 NS No mimicry -0.349 NS 0.275 NS 0.033 NS (n = 24) (n = 11) (n = 22)

Male Satin Bowerbirds do not invariably long buzzesin 1982)were correlatedwith dec- mimic both kookaburra and honeyeater calls orations. each time they sing a courtshipsong. To test whether this variability is related to either age DISCUSSION or mating success,we compared male age and mating successwith the proportion of songs We have describedthe relationship between containing only kookaburra calls, those with intrapopulational patterns of male Satin Bow- only honeyeatercalls, those with both types, erbird courtshipsongs and characteristicssuch and those with no mimicry at all. No signifi- as male age and bower quality. Male age and cant correlationswere found for either year mating successwere positively correlated in (Table 3). There were no birds in the studythat both years(Borgia 1986b), so it appearsthat fe- "specialized" exclusively in either one of the male Satin Bowerbirds prefer older males as mimetictypes, and all birds sangmimetic songs. mates.The significantcorrelation of male mat- Spearmanrank correlationswere used to test ing successwith age-correlatedfeatures of songs for relationshipsbetween song features and in 1982, but not in 1983, suggeststhat females bower quality and between song features and may use songsto assessmale age. Male age is bower decorations(Table 4). Of the songvari- also positively correlated with bower quality ablesmeasured, only the mechanicalsong du- (Borgia 1986b), which could provide another ration was significantly correlatedwith bower mechanismby which females can assessmale quality (but only in 1983).Mechanical song du- age.However, the low correlationof courtship ration was alsocorrelated significantly with the song variableswith measuresof bower quality total numberof bower decorations(both years), indicatesan independent effectof eachof these but no othersong variables (except honeyeater factors on male mating success.These results structuralintegrity and the low frequency of are consistentwith the hypothesisthat male

TABLE4. Comparisonsof songvariables with decorationsand bower quality. SI = structuralintegrity.

Bower quality Total no. decorations 1982 1983 1982 1983 Song feature P P P P Mechanical duration NS 0.033 0.030 0.038 Peak frequency NS NS NS NS Low frequency NS NS 0.033 NS Frequencyrange NS NS NS NS Number long buzzes NS NS NS NS Number short buzzes NS NS NS NS Mimetic duration NS NS NS NS Kookaburra duration NS NS NS NS Kookaburra SI NS NS NS NS Honeyeater duration NS NS NS NS Honeyeater SI NS NS 0.001 NS (n = 11) (n = 22) (n = 11) (n = 22) 194 LOFFREDOAND BORGIA [Auk, Vol. 103 courtship vocalizationsfunction as age mark- dancing at bowers. Intermale differenceshave ers, independent from other aspects of male not yet been quantified for these dances.It is display,and as indicatorsof male geneticqual- likely that female choiceof male characteristics ity. dependson the assessmentof interactingchar- Severalobservations suggest that Satin Bow- acters.The failure to explain a large proportion erbirds learn and practicemimetic singing. In of variation in male mating successwith vocal other populations of Satin Bowerbirds mimetic display data may be due to the absenceof data songsexhibit much variation both within and on dancing.In addition, the lack of significant between populations, while other song fea- correlationsof mating successwith vocal dis- turesare rather invariable (Loftredoand Borgia play data in 1983 may be due to the early com- in prep.). Further, male bowerbirds, especially mencementof breeding that year, which re- young males, spend a great deal of time both suited in a highly atypical breeding season. within and outsideof the breeding seasonper- The present study suggeststhat male court- forming their displays(Vellenga 1970, Chaffer ship songscontribute to the processof female 1984, Borgia in prep.). Together with our ob- mate choice by allowing females to estimate servationsof age-relatedincreases in the qual- male age. Differences among males in court- ity of songs,these resultssuggest that learning ship vocalizationsexplain some of the varia- and practice may be important in the use of tion between males in their mating success. mimetic songs,as appearsto be the casein oth- Theseresults are consistent,at least for one year er avian speciesthat use interspecificmimetic of this study,with modelsthat suggestthat fe- singing (Nottebohm 1972,Payne 1973,Howard malesassess male age and use this as an indi- 1974,Catchpole 1980). The use of mimetic songs cator of male quality as a sire. We cannot ex- by Satin Bowerbirds may have resulted from clude the possibility that the songs have selection for reliable age cues (Zahavi 1977, evolved through runaway sexual selectionin Borgia 1979, Kodric-Brown and Brown 1984). which females have evolved preferences for These songsare reliable becausethey require certain features of male display vocalizations learning and practice, and thus young males independent of male quality. This is an inter- cannot easily "cheat" (if females prefer older esting possibility becauseit differs from the males) by sounding like older individuals. tightly genetically programmed features com- Another model concerning the role of male monly discussedin runaway sexual-selection courtshipsongs in influencing female choiceis models. It suggeststhat female preference for the male dominance model. Dominant males exaggeratedmale display has led to selection may be preferredas mates because they are able in males for the ability to learn complicated to demonstratetheir vigor relative to other songs. males through aggressive interactions (Alex- ander 1975,Cox and LeBoeuf1977, Borgia 1979, ACKNOWLEDGMENTS Borgiaet al. 1985). This prediction is supported This researchwas supported by funds from the by observationsthat female Satin Bowerbird University of Maryland and the Computer Science mating decisionsare affectedby indicatorsof Center, and by funds to G.B. from the National Sci- male dominance such as bower destruction ence Foundation (BNS 81-13477 and BNS 83-08154). (Borgia1986a) and feather stealing(Borgia and The National ZoologicalPark (Washington,D.C.) al- Gore 1986). Loud vocal displays may attract lowed use of their analysisfacilities, and thanks go more attention from competingmales than do to EugeneMorton and Miles Robertsfor their sup- low-amplitude displays;thus, a male's ability port and technicalassistance. M. Aliaback,C. Ander- to give a loud courtshipsong without interrup- son,C. Drea, K. Drea, I. Kaatz,K. Lundy, S. Mesnick, tion from neighboring males could provide P. Murphy, and C. Nations were invaluable field evidence to a female of a male's dominance sta- assistants.The Hayes,Mulkay, and Bell familiespro- tus (Borgia1979; but seeFoster 1983, Bradbury vided many forms of assistanceand, with the New SouthWales Forestry Commission, allowed accessto and Gibson 1983). We could not quantify the their property.We thank M. J. Littlejohnand J. Hook vocal data adequately to test this hypothesis. and membersof the University of Melbourne Zool- However,after listeningto many callswe heard ogy Departmentfor their support.Valuable criti- no apparent amplitude differences among cismsof the manuscriptwere madeby B. Beehler,A. males. H. Brush,S. Carter-Porges,R. Dooling, A. Houde, E. Courtshipsongs are coordinatedwith male Morton, M. Roberts, W. Schleidt, and J. Sullivan. January1986] BowerbirdCourtship Vocalizations 195

LITERATURE CITED FISHER,R.A. 1930. The geneticaltheory of natural selection. Oxford, Clarendon Press. ALEXANDER,R. D. 1975. Natural selection and spe- FOSTER,M. S. 1983. Disruption, dispersion, and cialized chorusingbehavior in acousticalinsects. dominancein lek-breedingbirds. Amer. Natur. Pp. 37-77 in Insects,science, and society(D. Pi- 122: 52-72. mentel, Ed.). New York, Academic Pressß HALLIDAY,T. 1978. Sexual selection and mate choice. ARNOLD, S.J. 1983. Sexual selection: the interface Pp. 180-213 in Behaviouralecology: an evolu- of theory and empiricismßPp. 67-107 in Mate tionary approach(J. R. Krebsand N. B. Davies, choice (P. Bateson,Ed.). CambridgeßEngland, Eds.). Sunderland, Massachusetts,Sinauer Assoc. CambridgeUniv. Press. HAMILTON, W., & M. ZUK. 1982. Heritable true fit- BORGIA,G. 1979. Sexual selection and the evolution nessand bright birds: a role for parasites?Sci- of mating systems.Pp. 19-80 in Sexualselection ence 218: 384-387. and reproductive competition (M. Blum and A. HELWIG,J. T., & K. A. COUNCIL(Eds.). 1982. SAS Blum, Eds.). New York, Academic Press. user'sguide: statistics, 1982 ed. Cary,North Car- 1985. Bower quality, number of decora- olina, SAS Inst. Inc. tions, and mating successof male Satin Bower- HOWARD,R.D. 1974. The influence of sexual selec- birds (Ptilonorhynchusviolaceus): an experimental tion and interspecificcompetition on mocking- analysisßAnim. Behav.33: 266-271. bird song(Mimus polyglottos). Evolution 28: 428- --. 1986a. Bower destruction and sexual com- 438. petition in the Satin Bowerbird (Ptilonorhynchus KIRKPATRICK,N. 1982. Sexual selection and the evo- violaceus).Behav. Ecol. Sociobiol.in press. lution of female choice. Evolution 36: 1-12. 1986b. Why do bowerbirds build bowers? KODRIC-BROWN,A., & J. H. BROWN. 1984. Truth in Sci. Amer. in press. advertising:the kinds of traitsfavored by sexual , & M. A. GORE. 1986. Feather stealing in the selection. Amer. Natur. 124: 309-323. Satin Bowerbird (Ptilonorhynchusviolaceus): male I.•NDE, R. 1981. Models of speciationby sexualse- competition and the quality of displayßAnim. lectionof polygenictraits. Proc. Natl. Acad. Sci. Behav. in press. 78: 3721-3725. , S. G. PRUETT-JONES,& M. A. PRUETT-JONES.NOTTEBOHM, F. 1972. The originsof vocallearning. 1985. The evolution of bower-building and the Amer. Natur. 106: 116-140. assessmentof male quality. Z. Tierpsychol.67: PAYNE,R. B. 1973. Behavior, mimetic songs, and 225-236. songdialects, and relationshipsof the parasitic BRADBURY,J. 1981. The evolution of leks. Pp. 138- indigobirds(Vidua) of Africa. Ornithol. Monogr. 169 in The evolution of social behavior (R. Alex- No. II. ander and D. Tinkle, Eds.). Concord, Massachu- THORNHILL,R. 1980. Competitive, charming males setts, Chiron Press. and choosyfemales: was Darwin correct?Florida ß & R. GIBSONß 1983. Leks and mate choice. Entomol. 63: 5-30. Pp. 109-138 in Mate choice (P. Bateson, Ed.). TRIVERS,R.L. 1972. Parental investment and sexual Cambridge,Englandß Cambridge Univ. Press. selection.Pp. 136-179 in Sexualselection and the CATCHPOLE, C. K. 1980. Sexual selection and the descentof man (B. G. CampbellßEd.). Chicago, evolution of complex songs among European Aldine Press. warblersof the genusAcrocephalus. Behaviour 74: VELLENCA, R. 1970. Behaviour of the male Satin 149-165. Bowerbird at the bower. Australian Bird Bander CHAFFER,N. 1984. In questof bowerbirds.New York, I: 3-11. Rigby Publ. WF,ST-EBERHARD, M.J. 1979. Sexual selection, social Cox, C., & B. LEBOEUF. 1977. Female incitation in competition,and evolution.Proc. Amer. Phil. Soc. the northern elephant seal. Amer. Natur. 111: 123: 222-234. 217-255. WILBUR,H. M., D. I. RUBENSTEIN,& L. FAIRCHILD. 1978. DAVIS,J. W., & P. O'DONALD. 1976. Sexual selection Sexual selection in toads: the roles of female for a handicap:a critical analysisof Zahavi's the- choiceand male body size. Evolution 32: 264- ory. J. Theoret. Biol. 57: 345-354. 270. DONAGHEY,R. 1981. Parentalstrategies in the Green ZAHAVI, A. 1975. Mate selection--a selection for a ( crassirostris)and the Satin handicap.J. Theoret. Biol. 53: 205-214. Bowerbird ( Ptilonorhynchusviolaceus ). Unpub- ß 1977. Reliabilityin communicationsystems lished Ph.D. dissertation, Melbourne, Australia, and the evolution of altruism. Pp. 253-260 in Monash Univ. Evolutionaryecology (G. Stonehouseand C. Per- EMLEN,S., & L. ORING. 1977. Ecology,sexual selec- rins, Eds.). London, Macmillan Press Ltd. tion, and the evolution of mating systemsßSci- ence 197: 215-223.