Ontogeny of Vocalization in the Greater Rhea

ONTOGENY OF VOCALIZATION IN THE GREATER RHEA

PAUL W. BEAVER

Department of Psychology,University of Washington,Seattle, Washington98195 USA

ABST•CT.--The call-notesof the Greater Rhea chick were classifiedaccording to their acoustical structureand behavioralcontext. Both rate of vocalizationand tonal quality of call-noteswere found to deteriorateas a function of maturation. Call-note productionwas completelyeliminated by 7 weeksof age. Greater Rhea adultsare silentexcept for suchrelatively infrequent vocalizations as "hissing"that sometimesaccompanies agonistic behavior and a disyllabic"grunt" that only the male producesas part of a courtship display. Tracheal growth and a marked reduction in the intrusion of the internal tympanic membranesinto the bronchial passagescould contribute to the deteriorationof vocalization.The use of vocalizationas the primary channel of communicationis replaced by visual display in the adults. These modalities are best suited to the ecologicalcontingencies facing the chick and adult, respectively. Received 17 March 1977, accepted24 Sep- tember 1977.

A greatdeal of ornithologicalliterature has focusedon the vocal displaysof birds (e.g. Thorpe 1961, Hinde 1969, Armstrong 1973). The attention accordedto bird vocalization is no doubt largely due to its central importance in avian social interac- tion. The suborderoscines of the passerinebirds have been the most popular species for research in bird vocalization. There has been little interest in the vocalizations of nonoscinesin general and ratite birds in particular. This is probably becausethe vocal repertoire of adult ratites is not very impressive. The Greater Rhea (Rhea americana), a large, flightlessratite native to the pampasof Argentina, is especially unimpressive.In the adult, vocal productionis restrictedto "hissing"during agonistic encountersand the disyllabic "grunt" of a sexually arousedmale. In contrast to adults, Greater Rhea chicks are highly vocal. This is unlike most other speciesof birds whosevocal repertoiresincrease in complexityas a function of development. The ontogenyof vocalizationin the Greater Rhea is a curiousexample of behavioral regressionand may posequestions for theoretical treatments of behavioral ontogeny that treat it as a cumulative and progressiveprocess.

METHODS

Eggs were laid by the Greater Rheas maintained at the Chicago Zoological Park, Brookfield, Illinois and incubated at the University of Chicago. A total of 33 chicks was hatched over 2 years, of which 20 survived past 2 months of age. Chicks were housedat the University of Chicago in a 0.9 x 1.2 m pen until 4 weeksof age, at which time they were transferredto a 90 m2 outdoorenclosure at the ChicagoZoological Park. Vocal behavior was recorded on a SONY tape recorder model TC-92. Over 20 h of vocalizationswere recorded and are available upon request to the author. Sonagramswere made on a Kay Elemetrics, Co. audio spectrumanalyzer model 6061 B. Vocal amplitude was determinedby a linear scalesound level meter placed at a distanceof approximately 1 m from the chick. Recordingswere made daily at random times for the first 4 weeks of life. Additional recordingswere also made on an intermittent basisbetween ages 4-8 weeks.

CALL-NOTE CATALOGUE

Several authors, from Hudson (1903) to Bruning (1973a), have used the phrase "plaintive cry" to describea frequentlyheard ca]] of the Greater Rhea chick. Bruning 382 The Auk 95: 382-388. April 1978 April 1978] RheaVocalization Ontogeny 383

kliz kliz B 8 8

2- ;

0.5 1'.0 œ5 i.o kllz c kHz O

) •5 0.5 1.0 15 kllz E 8-

0.5 1.0 ]'.5 l'.O Sec. Sec. Fig. 1. Type B/65 wide-bandsonagrams of GreaterRhea chick call-notes: (a) contact-whistle,(b) separation-whistle,(c) alarm-trill,(d) feeding-peep,(e) contentment-chirp,(f) separation-whistleof an older chick illustrating tonal deterioration.

also makesreference to an infrequentlyheard alarm cry producedwhen the chick was in immediatedanger. Beyond these observations no one has madea detailed catalogueof GreaterRhea chick call-notes. Basedon observationsof captivechicks in a variety of experimentalconditions, five call-notes(Fig. 1) weredetected and investigated. The call-noteswere supplied with terms that reflect both their function and sound: (1) contact-whistle; (2) separation-whistle;(3) alarm-trill;(4) feeding-peep;and (5) contentment-chirp. Contact-whistle.-•The contact-whistle call-note sounds much like a soft, short whistle.It is a brief, monosyllabiccall-note, with a durationof 0.20-0.50 s. It is also of relativelylow amplitude,rarely exceeding 75 db. The fundamentalfrequency is generallyaround 1.5-2.0 kHz. There are severalharmonics, at about2.0 kHz intervals, whichreflect the brillianceof the tone.Some chicks have a sharpdecrease in pitchat the endof the note,while othersexhibit a sharpdecrease at the beginningof the note;with othersthere are no suddenchanges in pitch.These variations appear to be specific to individuals. The contact-whistlecall-note was initially heard after the GreaterRhea embryo 384 PAUL W. BEAVER [Auk, Vol. 95 enteredthe air space(determined by candling),generally 2-5 days prior to hatching. The contact-whistlewas emitted spontaneouslyby the embryo. It was alsoelicited on several occasionsin responseto the tape recorded playback of this call as well as to such mild stimulation as the slow turning of the egg. Such in-egg vocal production may act to synchronizethe hatchingof GreaterRhea eggsthat are incubatedtogether in a clutch. Pilot researchby Bruning (1973b) indicated spectacularhatching synchrony, acceleratingthe time to hatchingof late set eggsby as much as 7 days. In-egg vocal behavior in other speciesmay be a potential variable in hatching synchrony (Orcutt 1974). The contact-whistleremains very prominent in the repertoireof the Greater Rhea chick after hatching. It was frequently emitted in the presenceof a human surrogate parent, especiallywhen the chick was following or approaching.It was also frequently emitted in the presenceof conspecificswith which the chick was reared. Since the call-noteis heard only in thesesituations I suggestthat it servesa social contact function. Separation-whistle.--The separation-whistlecall-note soundsmuch like a loud, long whistle. This call-note is the "plaintive cry" documentedby Hudson (1903) and Bruning (1973a). It is a long, monosyllabiccall-note, generallylasting 1.0-3.0s. The longest separation-whistlerecorded was 4.48 s. The separation-whistleis a high amplitude call-note, reaching 75-100 db. The fundamental frequency is generally around 2.0 kHz. There are several harmonics, at about 2.0 kHz intervals, which reflect the brilliant tonal quality of the call-note. There are severalindividual-specific variations in pitch changesduring separation-whistles.Some variations involve a sudden or gradual frequency decrease,which intensifiesthe sad or lonely quality of the call-note to the human ear. The fundamental remains level for some chicks while for othersit gradually decreases,by 0.1-1.0 kHz, over 1-3 s. Some chicksshow a sudden decreasein frequency at the onset of the call-note, while others show a sudden increase, but most chicks have neither. Some chicks show a sudden decrease in frequency at the tail of the call-note while others do not. The separation-whistleis first heard within a few hours after hatching. The separation-whistleappears to be functionally related to the contact-whistle.Whereas the contact-whistleis emitted when the surrogateparent or familiar conspecificsare nearby, the separation-whistle is emitted in a condition of social isolation or when with unfamiliar conspecifics.The separation-whistlealso appearsto be acoustically related to the contact-whistle;its distinctionis that it is longerand louder. There is no clear border between the two call-notes.If the surrogateparent is gradually displaced to some distance but the chick retains visual contact, the contact-whistle escalatesin both duration and amplitude, gradually assumingthe character of a separation- whistle. The inter-individual variations in pitch at the termination of the contact- whistle reflect thoseof the separation-whistle,again suggestinga closerelationship between the two call-notes. Alarm-trill.--The alarm-trill call-note has many acoustical features in common with the separation-whistle.It is a long, loud monosyllabiccall-note. The durationis generally somewhat shorter than that of the separation-whistle,lasting 0.5-2.0 s. The amplitude measures80-90 db. The fundamental frequencyis generallyaround 2.0 kHz. There are severalharmonics, at about2.0 kHz intervals.The distinguishing feature of the alarm-trill is a modulationof pitch throughoutthe call-note. The alarm-trill is elicited by stressingthe chick with suchprocedures as suddenly lifting it off the ground or by administeringintra-muscular injections of medicine or April 1978] RheaVocalization Ontogeny 385 vitamin supplement.The alarm-trill is not heard from all chicksas someconsistently remain silent throughout these procedures. The alarm-trill is acousticallysimilar and seemsto be functionally related to the separation-whistle.Both probably function to attract the parent to the chick. But the modulationof pitch in the alarm-trill may serveto give it somewhatgreater emphasis or urgency than the separation-whistle. Feeding-peep.---Thefeeding-peep is a very quiet, short, monosyllabiccall-note. From one to four feeding-peepsare generallyemitted in rapid succession.The duration of a singlefeeding-peep is usually0.15-0.20 s. The amplitude was not detectable above the ambient noiselevel of 60 db. The fundamental frequency is 1.0-1.5 kHz, slightlylower than that of the other call-notes.There are severalharmonics at about 1.5 kHz intervals. No inter-individual variations were detected. The feeding-peepis first heard shortlyafter the chick beganto feed. Thereafter it is commonlyheard during feeding. It is not heard in any other behavioral context. Contentment-chirp.-•The contentment-chirpis a short, monosyllabiccall-note. Its duration was brief but more variable than that of the feeding-peep,lasting 0.10-0.35 s. Its amplitude measures 70-75 db. The contentment-chirp is characterized by sharply rising pitch. The fundamentalfrequency is generallylow, starting at 0.5-1.0 kHz and rising 1.5-2.0 kHz. There are severalharmonics, spaced at about 1.0 kHz intervals. There seemed to be individual-specific variations in the acoustical structure, suchas the presenceor absenceof slight dips in pitch at the beginningor end of the call-note, but not enoughrecordings were made to permit detailed analysis. The contentment-chirp is an infrequently heard vocalization that was recorded from one to only a few times in most of the chicks. The call-note was not reliably elicited with any condition. It was sometimesproduced when the chick was reunited with the surrogateparent or a familiar conspecificafter a period of isolationand was also occasionallyheard when a chick found food.

VOCAL DECLINE

The most unusual aspect of vocalization in the Greater Rhea is its deterioration during ontogeny. Greater Rhea chicks have a variety of call-notesof brilliant tonal quality that are frequently emitted. But by the time they are adults they are relatively silent. Sucha regressionin vocal behaviordeviates considerably from most speciesof birds in which the complexityof vocalizationincreases as a function of maturation. The rate of vocalization was found to decreasesteadily with age (Fig. 2). The rates of vocalization of three call-notes, the contact-whistle, the separation-whistle, and the feeding-peep,were determinedduring the second,fourth, and sixth weekspost- hatch. The contentment-chirpand alarm-trill were not included in the analysissince they are both relatively infrequent call-notes, not reliably elicited from all chicks. However, it appearsthat both of thesecall-notes share the general pattern of decline with increasingage. The rate of contact-whistlesemitted in the presenceof the surrogate parent decreased significantly as a function of age (P < .001, ANOCOVA). The rate of feeding-peepsemitted during feeding also decreasedsignificantly with age (P < .001, ANOCOVA). Finally, the rate of separation-whistles elicited in a condition of isolation from the surrogateparent and conspecificsalso decreasedsignificantly with age (P < .001, ANOCOVA). The declineof vocalization rate for the separation-whistlewas less severe; the regressioncoefficient for the separation-whistlerate was statisticallydifferent from the regressioncoefficients for 386 P.•vnW. BE.•VER [Auk, Vol. 95

10-

ß \

øo\ 'o \ z o 5

7-14 21-28 3S-42

AGE IN DAYS feeding-peeprate; dotted line = contact-whistlerate; solid line = separation-whistlerate. the two other vocalization rates. This may reflect some extendedfunctional signifi- cance of the separation-whistle.Only the separation-whistlecall-note was still heard after 42 daysof age. The oldestage at which a separation-whistlewas elicitedwas 64 days. The call-note in this instancewas very hoarseand feeble. In addition to the decline in the rate of vocalization, there is also a deteriorationin tonal quality. Changesfirst becomeevident after 17 dayspost-hatch. Instead of being sharp or even shrill-sounding,the call-notessound increasinglyhoarse. Call-notes showeda decreasein the number of harmonicswith age (Fig. iF). Long call-notes such as the separation-whistleor alarm-trill often had breaks in the productionof a singlenote whereasit was smoothand continuousin youngerchicks. This resultedin a choppyand gutteral sound. The amplitude of the call-notesalso decreased steadily past 17 days of age.

ANATOMY

Two chicks,aged 11 and 15 days, and a 1 yr old adult female Greater Rhea, all of which had died of illness,were usedfor a pilot investigationinto the anatomical basis of vocalization. The resonantfrequencies of vocal productionare known to depend April 1978] RheaVocalization Ontoge•y 387

Fig.3. Internalsyrinx. Left: a crosssection from an 11-day-old chick; the diameter measures approx- imately0.5 cm. Right: across section from a 1-yr-old;thediameter measures approximately 2.0cm. Note thereduction inthe intrusion ofthe tympanic membranes intothe bronchial passages.

onthe length and diameter of thetrachea. The lengths of thechicks' tracheas were 12 and13 cm, and the diameter in bothcases was approximately 0.5cm. The length of theadult's trachea was 42 cm and its diameter varied from 2.5 cm measured just belowthe larynx to 2.0cm measured just above the syrinx. Greater Rhea chicks lose theirvocal ability when they reach approximately one-third of theiradult height. Perhapsthe growth of thetrachea is a contributingfactor. A crosssection incision was made through the trachea just above the syrinx, allowingfor an excellent view of the internal tympanum. The tympanum isthought to be theprimary source of vibrationand thus the majoranatomical structure in arianvocalization (Greenewalt 1968). The tympanum vibrates as air is forced past theinternal membranes which are stretched into the bronchial passages atthe point of entryinto the trachea. In theGreater Rhea, the internal tympanic membranes intrudedinto the bronchial passages toa muchgreater extent in thechicks than in the adult(Fig. 3). The internal tympanic membranes ofthe adult do not seem capable of closingthe air passagesto theextent necessary to producethe resonance of thechick call-notes.

ECOLOGY Levins(1968) postulated that oneadaptive strategy to dealwith predictable changesin ecologicalcontingencies, such as seasonal variations, isa modificationof thephenotype correspondent to the ecological parameter. Ontogeny often poses an analogousproblem to an animal.The ecological contingencies that confront an animal in itsinfancy may be quite dissimilar from those it facesin adulthood.A consid- erationof this process may account for the regressive nature of vocal ontogeny in the Greater Rhea. Thevocal channel of communicationseems well adapted to theecological con- tingenciesthat confront the Greater Rhea chick. Chicks suffer greatly from predation by theCrested Caracara (Polyborus plancus) as well as other avian and mammalian predators.These predators rely primarily on visual cues in obtaining prey. Aided by itscryptic coloration, the Greater Rhea chick may reduce the probability ofpredator detectionbyblending inwith the tall grass of its native Argentine pampas. However, 388 PAUI. W. BEAVER [Auk, Vol. 95 the resultinglow visual profile of the chick may alsoincrease the probability that the parent will losevisual contact with the young. Given this situation, the vocal channel seemsto be the optimal method of communicationfor the chicks. The variety of chick call-notes seem well suited to maintain contact with the parent and to signal when the chick is lost, in danger, or feeding. Between 6 to 10 weeks of age seemsthe turning point in vocal behavior and ecology. At this time chicks lose their vocal prowess, lose the cryptic coloration provided by the neossoptiles,and begin to outgrow the protective cover offered by tall grass.They are also probably too large at this age to suffer heavily from predation by the CrestedCaracara. In Greater Rhea juveniles ritualized visual displaysare elaborated for social communication. Descriptions of some of the adult visual dis- plays have been provided by Raikow (1969) and Bruning (1975). The visual channel of communicationseems well suited to the ecologicalcontingencies of adults, which do not suffer from natural predators. They are also large-bodiedanimals, attaining a height of 1.5 m, and are not cryptically colored.They thereforestand out noticeably from the surrounding flat, grasslandhabitat. Visual display would seem to be the most efficient method of communication in this situation.

ACKNOWLEDGMENTS

I am indebtedto the ChicagoZoological Park of Brookfield,Illinois for the loan of specimensfrom their Greater Rhea collection.Several membersof the ChicagoZoological Park staff also provided assistance. Specialthanks are due to GeorgeB. Rabb, director;Benjamin B. Beck, curator;Ray Pawley, curator;and Richard Soderlund,keeper. Anatomical photographswere taken by Richard Search, ChicagoZoological Societyphotographer. I am alsoindebted to Eckhard H. Hess, Dale Terbeek, and StanleyZerlin, all of the University of Chicago, for their loan of equipmentand laboratory space.

LITERATURE CITED

ARMSTRONG,E.A. 1973. A study of bird songß(Rev. ed.). New York, Dover Publications. BRUNING, D. F. 1973a. The Greater Rhea chick and egg delivery route. Nat. Histß 82(3): 68-75. ß 1973b. Breedingand rearing rheasin captivity. Intern. Zoo Yearbook 13: 163-174. 1975. Socialstructure and reproductivebehavior in the Greater Rhea. Living Bird 13:251-294. GREENEWALT,C. H. 1968. Bird song: Acousticsand physiology. Washington, D.Cß, Smithsonian Institution Press. HINDE, R. A. 1969 (ed.) Bird vocalizations:Their relationto currentproblems in biologyand psychology. Cambridge, England, Cambridge Univ. Press. HUDSON,W. H. 1903. Naturalist in La Plata. London, J.M. Dent & Sonsß LEVINS, R. 1968. Evolution in changingenvironments. Princeton, N.J., PrincetonUniv. Pressß ORCIJTT, Aß B. 1974. Soundsproduced by hatching Japanesequail (Coturnix coturnix japonica) as potential aids to synchronoushatching. Behaviour 50: 173-184. RAIKOW, R. 1969. Sexual and agonisticbehavior of the Common Rhea. Wilson Bull. 81: 192-206. THORPE,W. H. 1961ß Bird song:The biology of vocal communicationand expressionin birds. Cam- bridge, England, Cambridge Univß Press.