1044 SHORT COMMUNICATIONS

The Condor95:1044-1048 0 The Cooper Ornithological Society 1993

EVIDENCE OF INTRASPECIFIC VOCAL IMITATION IN SINGING HONEYEATERS (MELIPHAGIDAE) AND GOLDEN WHISTLERS (PACHYCEPHALIDAE) ’

MYRON CHARLES BAKERY Department of Zoology, Universityof WesternAustralia, Nedlands,Perth, WesternAustralia 6007

Key words: Intraspecijic vocal imitation; song song patterns, is well known in (Kroodsma and learning;bindsong; Singing Honeyeater; Golden Whis- Baylis 1982, Mundinger 1982). Such songsharing may tler. be confined to small neighborhoodsof a few individ- uals (Payne 1983) or encompass larger regions con- Vocal imitation, by which two or more individuals of sistingof hundreds or thousandsof individuals (Baker the same speciespopulation come to share the same and Cunningham 1985). For functional interpretations ofimitation it is important to determine whether mem- bers of a speciesregularly exhibit vocal convergence among neighborsor larger groupsand how accurateis ’ Received 29 March 1993. Accepted 22 June 1993. the vocal copying that occurs.Imitation has a number 2 Present address: Biology Department, Colorado of interpretations and has led to several hypotheses State University, Fort Collins. CO 80523. that aid our understandingof social behavior and cul- 30a 31a

3 2

1 0.5 SEC # I

3 2 1

30a x 30b 30bx31b -I 31ax31b

FIGURE 1. Neighboring male Singing Honeyeaters that matched songs. Songs 30a and 30b are separate utterancesof one male; 3 1a and 3 1b are separateutterances of the neighboringmale. Correlations within birds (0.95 1, 0.888, respectively)and between birds (0.899) are similar in magnitude. SHORT COMMUNICATIONS 1045

51a

3 - 2 I ’

1 1 0.5 SEC , 8

50b 51b

50a x 50b 50a x 51a 51ax51b

FIGURE 2. Neighboring Singing Honeyeaters that did not match songs. Songs 50a and 50b are separate utterancesof one ; 5 la and 5 1b are separateutterances of the neighbor. Correlations within birds (0.887, 0.89 1, respectively)are higher than between birds (0.424).

tural evolution (Payne 1981, Mundinger 1982, Baker I tape. Analysis of the recordings was done with a and Cunningham 1985). digital signalprocessor (Kay ElemetricsDSP 5500) and The purpose of the present study was to document a Macintosh II computer implementing Canary soft- intraspecificvocal imitation (Type III vocal imitation: ware (version 1.O, Cornell Laboratory of Ornithology, Kroodsma 1982) in natural populations of two Aus- BioacousticsResearch Program). For spectrographic tralasianspecies of birds, the SingingHoneyeater (Me- display of vocalizationsof SingingHoneyeaters, I used liphaga virescens) and the Golden Whistler (Pachy- DSP settingsof DC-4kHz frequencyrange and 200 Hz cephala pectoralis). transform size. For Golden Whistlers, I used DSP set- tings of DC-8 kHz frequency range and 234 Hz trans- METHODS form size. Singing Honeyeaters and Golden Whistlers were tape Correlations between spectrogramswere done with recorded at locationsin Western from 5 Oc- Canary software. The processof correlatingtwo sound tober-8 November 1991 during the breeding season. spectrogramscan be visualized as follows (Clark et al. Seventy Singing Honeyeaters were recorded in Ned- 1987):(1) imagine two separatedspectrograms sharing lands, near Perth. Twelve SingingHoneyeaters and 10 the same frequency by time axes, one to the left and Golden Whistlers were recorded in and adjacent to one to the right as you observe them, (2) in stepwise John Forrest National Park 25 km east of Nedlands. increments of time, gradually move one of the spec- Twenty-five Singing Honeyeaters and 20 Golden trograms acrossthe other, (3) at each increment, cal- Whistlers were recordedon Rottnest Island 25 km west culate a correlation coefficient between the two spec- of Nedlands in the Indian Ocean. trograms. A peak correlation value results when the Recording was accomplished with a Marantz cas- two spectrogramsare overlaid in a way that they are setterecorder (PMD 201) Sennheisermicrophone (MD most similar. This peak value can be used as an index 402-K) mounted in a 40 cm parabola, and TDK Type of similarity. SHORT COMMUNICATIONS

10a 12a

6: 4- *I “I) # bF h i 0.5 SEC

6- 4-

FIGURE 3. Neighboring Golden Whistlers that matched songs.Songs 10a and lob are separateutterances of one bird; songs 12a and 12b are separateutterances of the neighbor. Correlations within birds (0.873, 0.820, respectively) are slightly greater than between birds (0.759).

Because birds were not color-banded, I selected recorded. In three casesno matching occurred,and in neighboringbirds that were interactingaggressively and four casesmatching was observedduring counter-sing- counter-singingin closeenough proximity to allow re- ing episodes. cording of both. SingingHoneyeaters were very abun- Correlational analysesof spectrograms(Clark et al. dant, and I could often hear other neighborsmatching 1987) revealed the accuracywith which some neigh- in addition to the two being recorded. My impression bors matched their songsand how different were the was that neighborhoodsof 2-5 birds with matching songsof neighborsthat did not match. Matching and songswere common in this species.Golden Whistlers non-matching of neighborswere judged in relationship occupiedlarger territories than did SingingHoneyeat- to the similarity of repetitions of songswithin individ- ers, and I rarely could hear clearly more than two birds uals. at a time. Songs of neighboring Singing Honeyeaters that matchedappeared highly similar spectrographicallyand RESULTS correlational analyses supported this subjective im- pression(Fig. 1). Non-matching neighborssang struc- Song sharing between neighboring Singing Honeyeat- turally differing songsand correlation values for neigh- ers and between neighboring Golden Whistlers was bor’s songs were about half the magnitude of the apparent to the unaided ear. Spectrographicanalyses correlationswithin birds (Fig. 2). revealed the detail and accuracy of vocal similarity. Songsof neighboringGolden Whistlers that matched Twenty-five casesof two counter-singingneighboring appeared similar by visual inspection of sonagrams, Singing Honeyeaters were recorded. In 11 cases no and the correlationvalues supported this interpretation matching occurred,and in 14 casesmatching was ob- (Fig. 3). There was, however, a consistent trend of served during countersingingepisodes. Seven casesof slightly less accuracyof neighbor matching than was two counter-singingneighboring Golden Whistlerswere the case for Singing Honeyeaters. When neighboring SHORT COMMUNICATIONS 1047

33a 36a

FIGURE 4. Neighboring Golden Whistlers that did not match songs.Songs 33a and 33b are separateutterances of one bird; songs36a and 36b are separateutterances of the neighbor. Correlations within birds (0.709, 0.879, respectively)are higher than between birds (0.156). pairs of Golden Whistlers did not match songs (Fig. known. Thus, suchpotentially important factors as the 4), the degree of mismatch was large by correlational ages of the birds involved in sharing songs,the time analysis.The example shown in Figure 4 is instructive course of vocal convergence,or the sites of origin of becausea cursoryvisual examination of the neighbor’s the individuals could not be usedto suggesthypotheses songs suggeststhey are more similar than the corre- on how or why vocal convergenceoccurred. lation value indicates.This impressionresults from the There are anecdotalreports of other speciesin these more obvious temporal pattern similarities and the less two Families mimicking vocalizationsof other species. apparentmajor differencesin frequencyof the two songs. Chisholm (1946) identified interspecific vocal imita- tion by two speciesof the Pachycephalidae,the - DISCUSSION tit (Falcunculus fronatus) and the Gray Shrike-thrush (Cokuichincla harmonica). Gilliard (1958) noted the The main conclusion drawn from this study is that common occurrenceof imitation of human voice by a SingingHoneyeaters and Golden Whistlers are capable member of the Meliphagidae, the Parson Bird (or Tui, of modifying their songs to achieve matching with Prosthemadera novaeseelandiae). These subjective re- neighbor’s songs.The general pattern of convergence ports were not documented with spectrographicevi- in songstructure among neighboringbirds is generally dence,but it seemsunlikely that the observationswere acceptedas indirect evidence of songlearning (Kroods- vastly misunderstood. In spite of these early obser- ma 1982). Experiments on learning have not been car- vations, however, there have been no follow-up stud- ried out in these species,however, nor has the process ies. of vocal modification from the pre-convergencesong A more recent report (Bruce 1988) indicates the ex- structureto the matching structurebeen documented. istenceofdialects in the songsofYellow-throated Hon- Therefore, the conclusion is tentative. The history of eyeaters(LichenostomusflavicoNis) in Tasmania. How- the individuals recorded in the present study was un- ever, the documentation of vocal variation and spatial 1048 SHORT COMMUNICATIONS scaleof songsimilarity was inadequatefor comparisons BRUCE,P. J. 1988. Spatial and individual variation to the Singing Honeyeater patterns reported here. in the songs of the Yellow-throated Honeyeater In applying the Canary software to this analysis,an Lichenostomusflavicollis. Emu 88:65-69. important point is raised.When differencesin two songs CHISHOLM,A. G. 1946. Nature’s linguists:a study of are primarily a result of an offset in the frequencyaxis the problem of vocal mimicry. Brow, Prior, An- but the shapesof the elementsare similar, the resulting derson Pty., Ltd., Melbourne, Australia. low correlation seemssomewhat misleading. This is a CLARK, C. W., P. MARLER,AND K. BEEMAN. 1987. consequenceof applying a visual analysis to the two Quantitative analysisof vocal phonology: spectrogramsrather than hearing the two songs.The an application to Swamp Sparrow song.Ethology latter perception makes the difference more apparent. 76:101-115. In any case, the question of how the birds perceive GILLIARD,E. T. 1958. Living birds ofthe world. Dou- similarity and difference is an experimental issuethat bleday, Garden City, NY. can be addressedby manipulation of featuresof songs KROODSMA,D. E. 1982. Learning and the ontogeny and presentationof the altered stimuli via the playback of soundsignals in birds, p. l-23. In D. E. Kroods- paradigm (e.g., Baker 1991). ma and E. H. Miller teds.], Acoustic communi- cation in birds, vol. 2. Academic Press,New York. I thank the Zoology Department, University of KROODSMA,D. E., ANDJ. R. BAYLIS. 1982. Appendix: Western Australia, for hospitality during my sabbatical a world survey of evidence for vocal learning in leave. Mike Johnston, Bob Black, Win Bailey, Dale birds, p. 31 l-337. In D. E. Kroodsma and E. H. Roberts, Darren Murphy, Perry deRebeira, Ron John- Miller [eds.], Acoustic communication in birds, stone and Mike Craig all contributed to my research vol. 2. Academic Press, New York. activities in many helpful ways. Financial assistance MUNDINGER,P. C. 1982. Microgeographicand mac- wasprovided by the National ScienceFoundation (U.S.) rogeographicvariation in the acquired vocaliza- through grants BNS-87-06526 and INT-9 l-13640. tions of birds, p. 147-208. In D. E. Kroodsma and E. H. Miller [eds.], Acoustic communication in birds, vol. 2. Academic Press, New York. LITERATURE CITED PAYNE,R. B. 1981. Population structureand social behavior: models for testing the ecologicalsignif- BAKER, M. C. 1991. Responsesof male Indigo and icance of songdialects in birds, p. 108-120. In R. Lazuli Buntingsand their hybrids to songplayback D. Alexander and D. W. Tinkel [eds.], Natural in allopatric and sympatric populations. Behav- selectionand social behavior: recent researchand iour 119~225-244. new theory. Chiron Press, New York. BAKER, M. C., AND M. A. CUNNINGHAM. 1985. The PAYNE,R. B. 1983. The social context of songmim- biology ofbird-song dialects.Beh. Brain Sci. 8:85- icry: song matching dialects in Indigo Buntings 133. (Passerinacyanea). Anim. Behav. 3 1:788-805.

The Condor95:1048-1050 0 The Cooper Ornithological Society 1993

NOVEL METHOD FOR ESTIMATING HOURS OF OVIPOSITION, ILLUSTRATED BY DATA ON GRAY CATBIRDS ’

DAVID M. SCOTT Department of Zoology Universityof WesternOntario, London, Ontario N6A 5B7, Canada

Key words: Hour of egg-laying;oviposition; Gray Without mechanical devices, such as used by Haftom Catbird; Dumetella carolinensis. (1966) to determine laying times of Purus spp., most investigatorshave relied either on two daily visits to The hour of egg-layinghas been studiedless than many a nest that bracketedegg-laying (Skutch 1952, Brackbill other featuresof nesting, e.g., clutch size or incubation 1958) or on observationson the arrival of a female at period. Lack of information on the hour of laying may her nest to lay (Nolan 1978, Muma 1986) and her be due to the difficulty of obtaining precise records. departureafter laying (Muma 1986). These procedures are time-consuming. Also, because some specieslay inconveniently closeto sunrise,few investigatorsof life histories routinely record the hour of laying. This is ’ ’ Received 4 April 1993. Accepted 16 June 1993. unfortunate becausethere is much interspecific vari-