OLD AGE AND BREEDING BEHAVIOR IN A TROPICAL PLOCEUS CUCULLATUS UNDER CONTROLLED CONDITIONS

N. E. COLLIAS,1 E. C. COLLIAS,1'2 C. H. JACOBS,1 C. R. Cox, 1,3AND F. A. MCALARY 1 •Departmentof Biology,University of California,Los Angeles, California 90024 USA, 2NaturalHistory Museum of LosAngeles County, 900 ExpositionBoulevard, Los Angeles, California 90007 USA, and 3TheLos Angeles Zoo, 5333 ZooDrive, LosAngeles, California 90027 USA

ABSTR^CT.--Manytropical speciesof passefinebirds have potentially long reproductive lives in both nature and aviaries,but rate of breeding by old individuals may eventually decline. ^ group of individually color-bandedAfrican Village Weavers(Ploceus cucullatus), a polygynousspecies, was kept in aviariesat LosAngeles for 22 years.At 14-18 yearsof age, old malesbuilt significantlyfewer nestsper year, displayedto femalesless often, had fewer mates,and generallyfathered fewer clutchesthan did the samemales when 9-13 yearsold. One male reached24 yearsof age,when he rarely sangand no longer wove when provided with nest materials.The females,after reaching"middle age" (9-13 yr), laid significantly fewer eggsper year,and still fewer during "old age" (14-18 yr). When 14 or moreyears old, weavers of both sexesrested much more than did young adults 4-6 years old in the same aviary. Individualsdiffered greatly in the declineof breedingbehavior with age.Some individual Village Weaversbred at a surprisinglygreat age for passefinebirds, e.g. fertile copulations by one 19-year-oldmale and by one 18-year-oldfemale. Theseseem to be the oldestbreeding agesyet recordedfor any small passerinebird. Long reproductivelife of a small number of individualsthat contributedisproportionately to the next generationwould give a selection pressurefor longevity, and also could help explain the evolution of small clutch size (2-3 eggs)in the Village Weaverand in many other tropicalbirds. Received 7 January 1985, accepted 5 December 1985.

ETHOLOGICALgetontology deals with the be- annual survival rates were 70-90% in over 30 havioral aspectsof the ecology and evolution species.By contrast,among small land of of life spans. Our object is to give a the north Temperate Zone, adult survival per quantitative descriptionof how old age influ- year seemsto be about 40-60% (Ricklefs 1969, encesthe breeding behavior of a tropical pas- Cody 1971:490). sefine bird under controlled conditions. Like The relatively greatadult survival known in people, rarely, if ever, die of "old age," naturefor at leastsome small tropicalland birds but rather of increasedsusceptibility to some raisesthe questionof how long and how suc- stressful condition. The reactions and behavior cessfully such old birds are able to breed. If of old animals must be compared simulta- somelong-lived individualscontinue to repro- neouslywith thoseof younger control animals duce over a number of years,and excludeoth- to evaluateadequately the effectsof advanced ers from breeding, their contribution to suc- age. Our data deal with maximum breeding ceedinggenerations may be much greaterthan ages,but we try to relate our results to actual that of the average adult of the same species. breeding agesin nature. We can find no pre- Furthermore,the questionof senescenceamong vious study that presents experimental and small tropical land birds has not been studied quantitative data on the effectsof old age on systematically. breeding behavior of wild birds under con- For over 20 yearswe studied the behavior of trolled conditions. a breeding colony of a tropical passefinebird, Recent studies suggest high adult survival the African Village Weaver (Ploceuscucullatus), among small land birds in the tropics(Fogden in outdoor aviaries at the University of Cali- 1972, Snow and Lill 1974, Fry 1980, Lewis 1980, fornia, Los Angeles (UCLA). The birds are po- Collias and Collias 1981, Willis 1983), where lygynousand breedin colonieswith manynests

408 The Auk 103: 408-419. April 1986 April 1986] BehaviorofOld Weaverbirds 409 in one tree. The male weaves the nest from TABLE1. Sex and age compositionof birds in two long stripsthat he tearsfrom the leavesof palms experimentalaviaries. (M = number of males, F = number of females.) or tall grasses.Each male occupiesand defends from one to a few small branches on which he Aviary Aviary builds his nests. He endeavors to attract un- 1 2 Total mated femalesto his nest with specialdisplays M:F M:F M:F and vocalizations.If a female acceptsthe nest, 1977 a she lines it with grass tops and often with Older birds (12-15 yr) 3:3 4:3 7:6 feathersas well. She incubatesthe eggsand, in Young adults (2-3 yr) 4:4 5:5 9:9 this subspecies,does virtually all the feeding Total 7:7 9:8 16:15 of the young. The male is normally sexually 1979 mature by 2 years of age, the female when 1 Old birds (14-17 yr) -- 6:6 6:6 year old. Young adults(4-5 yr) 5:5 3:4 8:9 The birds in our aviaries freely competed Total 5:5 9:10 14:15 with each other for food, nest materials, and 1980 mates. It gradually became evident that many Old birds (15-18 yr) 2:2 2:2 4:4 of these Village Weaverswere able to breed at Young adults (5-6 yr) 4:5 4:4 8:9 a surprisingly advanced age. A decline in Total 6:7 6:6 12:13 breeding behavior of some very old adults • Startingfigures: during 1977 all the malesof the seemed evident, but controlled observations two aviarieswere interchangedat either 4- or 6-week were needed. In 1977, 1979, and 1980 we spe- periods. cifically compared groups of old and young adults, but in 1978 the birds were kept indoors becauseof constructionactivities by the uni- the overall sex ratio was kept about equal each year. It was more difficult to keep age ratiosconstant, be- versity near the outdoor aviaries. We con- causeby the nature of the experimentthe oldestbirds trolled numbersof birdsper aviary, sexand age tend to die. We included in the experimentalanalysis ratios, food supply (ample ad libitum),and of only thosebirds that lived through the breeding sea- course predation. In nature, it would be ex- son of any given year. Each of the two outdoor avi- tremely difficult to compare the behavior of aries contained both older and younger (control) younger and very old birds exposedto exactly birds, exceptfor one aviary in 1979.The resultsfrom the same conditions. But such comparisonscan both aviarieswere merged for statisticalanalysis. All readily be made on captive birds in aviaries the birds in each aviary, regardlessof their ages,were under controlled conditions. We have pre- exposedto the samenumber of adults of the opposite sex. sented two preliminary reports (Collias et al. Each bird carried two colored leg bands on each 1980, 1981). leg, with the same individually distinctive color combinationon each leg. Eachbird was named from MATERIALS AND METHODS its leg bands;thus, Male OG had an orange over a green band on each leg. The birds were fed on par- The Village Weaversbelonged to the West African akeet seed mixture, fresh lettuce, cuttlebone,grit, a race (P. c. cucullatus);the original stock came from vitamin supplement (Vionate), meal worms regular- ,where we studied the behavior of this sub- ly, and cricketswhen the birds were raising broods. speciesin the field (Collias and Collias 1970). Dr. Freshwater was given daily. The 5-10 malesusually Jean Delacour sent us 18 males and 8 females in the kept in eachaviary commonlybuilt up to 100 or more fall of 1958, from which the birds of this experiment nestsduring the breeding seasonof April to Septem- were descendedover three generations. The breed- ber. During the cool winter months we brought the ing behavior of the birds in our five outdooraviaries birds indoors, where they were maintained in cages (from 9.2 m long x 5.2 m wide x 5.2 m high, to 7.9 x or small aviaries. 6.1 x 4.2 m) at UCLA was essentiallysimilar to that With the exceptionof one male (RA), all the birds seen in nature. The birds were allowed to choose we report on here were hatched in our aviaries on their own mates. known dates. Recordswere kept routinely for each Usually, a similar number of males and females male of the number of nests he built, number of mates were kept in each aviary (Table 1). Males typically he secured,and number of clutcheshe presumably weighed 41-51 g (•?= 46, n = 9), females31-39 g (•?= fathered. Of several hundred copulations observed, 36, n = 9). Becausethe sex ratio has a profound effect only twice was a female seen to copulate with a male on amount of breeding activity (Collias et al. 1971), other than her own mate. We made sample counts 1 410 COLL•^SET ^L. [Auk,Vol. 103

T^I•I•E2. Individual recordsof numbersof nestsbuilt (N), matesobtained (M), and clutchesfathered (C) by older and youngeradult male Village Weaversduring the last four yearsof the study.a In 1978birds were indoors and not allowed to breed.

1977 1979 1980 Age in Life span 1977 (yr) N M C N M C N M C Older males Wbr 12 17.9 31 8 14 17 4 12 23 7 27 yrB 12 16.7 33 4 15 7 1 1 1 1 1 yrA 12 15.4 25 5 7 2 0 0 0 0 0 OG 12 15.2 25 2 8 0 0 0 0 0 0 AR 12 14.6 25 9 9 3 0 0 * Brb 12 14.1 13 3 13 4 2 5 * Awr 15 16.2 10 0 0 * Younger adult males OY 2 24 1 1 19 3 9 14 4 9 GW 2 14 0 0 10 2 6 7 3 9 YW 2 10 0 0 3 0 0 6 0 0 WO 2 15 1 1 20 3 11 14 1 6 AO 2 6 0 0 15 2 8 0 0 0 RG 3 20 3 4 10 1 2 26 4 19 WB 3 18 2 6 14 6 19 8 5 12 YA 3 16 3 8 4 1 1 0 ! 2 WR 3 25 3 9 *

a An asterisk indicates that the bird died. h each day of the frequency with which each male In 1977, the older males (all but one then 12 gavenest-invitation displays to females,counting each yr old) exceededthe younger adult males (2-3 time the male hung beneath the nest entrance flap- yr old) in number of nests built, number of ping his wings. For each female, we recordedher mates, and number of clutches fathered (Tables mates and the number of clutches and eggs laid. In 2 and 3). In 1979, these same younger males occasionaltests for fertility of old birds, we cracked (then 4-6 yr old) significantlyexceeded the old open the eggsand checkedfor presenceor absence of an embryo. "Old age" was arbitrarily defined as males(then 14 yr old) in number of nestsbuilt. about 14 years. Time samplesof specificbehaviors The change from 1977 to 1979, comparing sur- engagedin by every bird alsowere taken during the vivors only, was statisticallysignificant for all last two years. To stimulate more nest building and aspectsof behavior mentioned. Frequency of more nest acceptancesby the females,we generally nest-invitation displays to the female by the removed clutches.This was done shortly after the malesvaried greatly,but here too the old males eggswere laid to prevent wide variationsin the in- declined significantly from 1977 to 1979 when terval before a female laid her next clutch. Prompt comparedwith the young males(rank-sum test, removal of clutches(experimental predation) made P = 0.004). The old males dropped from an av- the data more comparablewithin and between the erage of 168 displays.month-l-male -1 to 47, different years, and allowed us to test the reproduc- tive capacityof the females in relation to their age. whereasthe young malesstayed essentially the In September1980 we terminated the experiments. same in both years (64:61). Although the males normally mature at 2 yearsof age,these differencescould be a func- RESULTS tion of increasing experience of the young adultsas well asaging of old adults.The change Effectsof ageon amountof breeding.--Oldand with age from 1979 to 1980 was not significant, young adults were comparedin two ways for but only four old malessurvived; three of them breeding performance. First, they were com- had poor recordsin both years (Table 2), while pared during the sameyear to control for dif- in 1980 one vigorous 15-year-old male (Wbr) ference in circumstancesin different years. Sec- had the most mates and fathered 59% of the ond, the breeding behavior of the same clutchesin his aviary (with 2 old and 5 young individual was comparedduring different times adult males). in its life. When each of eight old maleswas compared April1986] BehaviorofOld Weaverbirds 411

TABLE3. Breedingbehavior of older and youngeradult male Village Weavers;same birds (survivors)each year. Averagesper male are given in parentheses.

Total

1977 a 1979 1980 Age of older: younger males (yr) 12-13:2-3 14:4-5 15:5-6 No. of older: youngermales 7:9 6:8 4:8 No. of nests built Older males 162 (23.1) 33 (5.5) 24 (6.0) Youngermales 148(16.4) 95 (11.9) 75 (9.4) Probabilityb NS (0.06) 0.001 0.041 NS NS No. of mates Older males 31 (4.4) 7 (1.2) 8 (2.0) Youngermales 13 (1.4) 18 (2.2) 18 (2.2) Probabilityb 0.024 0.002 NS NS NS No. of clutches fathered Older males 66 (9.4) 18 (3.0) 28 (7.0) Youngermales 29 (3.2) 56 (7.0) 57 (7.1) Probability• 0.02 0.001 NS NS NS ' In 1977only one male was older than 12 years. bProbabilities are basedon rank-sumtest (Dixon and Massey1969: 344), comparingtotals for older and younger males within and between years. againsthis own record, the males built signif- clutches fathered. The term "bird-years ob- icantly more nestson the averageeach breed- served" (Table 4) refers to the fact that the dif- ing seasonafter reaching "middle age" (9-13 ferent birds were observed for 1-4 seasons each yr) than as young adults (3-8 yr) (Table 4). Av- during each 5- or 6-yr period of life, varying erage annual number of nests built per male with each bird. Not all birds were allowed to declined greatly (P = 0.01) when the males breed every year, but the data were grouped to reached "old age" (14-18 yr). The average number of mateseach male obtained per year TABLE5. Numbers of clutchesand of eggs (in pa- alsodecreased significantly after age 14, but not rentheses)laid during the last four years of the the averagenumber of clutchesfathered (Table study by individual older and younger adult fe- 4), becauseof one exceptionalold male (Wbr), male Village Weavers.a In 1978birds were indoors who fathered two-thirds of all the clutches and not allowed to breed. fathered by the eight old males. For the other Age Life seven males together, there was a significant in span decline (P = 0.016) from middle to old age in 1977 (yr) 1977 1979 1980 Older females TABLE4. Changesin breeding behavior by eight WA 15 18.4 9 (24) 2 (5) 2 (2) adult male Village Weaversas they aged. Oar 13 16.2 7 (14) 0 (0) I (1) brG 12 17.5 2 (3) 0 (0) 4 (4) Yearly averages Bya 12 15.1 I (1) 2 (4) 0 (0) Aba 12 14.4 13 (27) 1 (2) * Bird- permale a Aya 12 13.9 5 (9) 0 (0) * years Nests No. of Clutches Age observed built mates fathered Youngeradult females RY 3 7 (15) II (31) 8 (21) Young adult AB 3 8 (16) 7 (13) 8 (17) (3-8 yr) 20 15 2.7 5.5 YB 3 7 (13) 6 (12) 9 (20) Middle age LW 3 3 (6) 5 (10) 8 (16) (9-13 yr) 29 23* 3.7 7.8 RO 2 7 (13) 9 (25) 9 (20) Old age GY 2 10 (21) 5 (15) 9 (25) (14-18 yr) 13 6* 1.5' 3.7b OL 2 4 (8) 9 (20) 7 (13) YL 2 9 (19) 10 (24) 10 (23) a ß = significantchange (P < 0.05) from preceding LG 2 3 (6) 5 (8) 4 (8) age by Wilcoxonmatched-pairs signed-rank test. • Not significant(P = 0.078); see text. ' An asterisk indicates that the bird died. 412 COLLIASETAL. [Auk,Vol. 103

TABI•E6. Number of clutchesand eggslaid by older and younger adult female Village Weavers;same birds (survivors)each year. Averagesper female are given in parentheses.

Total

977' 1979 1980

Age of older :younger females(yr) 12-15:2-3 14-17:4-5 15-18:5-6 No. of older: younger females 6:9 6:9 4:9 No. of clutches Older females 37 (6.2) 5 (0.8) 7 (1.8) Younger females 58 (6.4) 67 (7.4) 72 (8.0) Probability• NS 0.004 <0.001 NS 0.001 No. of eggs Older females 78 (13.0) 11 (1.8) 7 (1.8) Younger females 117 (13.0) 158 (17.6) 163 (18.1) Probability• NS 0.005 < 0.001 NS < 0.001 a In 1977 only one female (WA) was older than 13 years. bProbabilities are based on rank-sum test (Dixon and Massey 1969: 344), comparing totals for older and younger femaleswithin and between years. show general trends by yearly averages per males becausethe females no doubt laid many male for young, middle-aged, and old adults. more clutches than birds in nature would, un- The clutches fathered by old males dropped der this strong experimental "predation." significantlyfrom 1977to 1979,including Male Interest in males or their nests may persist Wbr, when the birds went from 12 to 14 years longer than egg laying in females that were of age (Table 3). previously good layers. During the breeding In 1977 the average number of clutchesand seasonof 1980, a 16-year-old female accepted eggslaid by the older females(all but one were nestsfour times without laying eggs,although then 12-13 yr old) was about the same as the she copulated with the male owner each time. number laid by young adult females (2-3 yr She laid only one egg all season.This egg was old). By 1979 the same older females laid sig- addled and could not be checked for fertility. nificantly fewer clutchesand eggsthan did the Another female still visited nests but was not same young females (Tables 5 and 6). The av- seento copulateand laid no eggsat 15 yearsof erage number of clutches and eggs laid each age in the last breeding seasonshe was ob- seasonby eight females over their lifetime de- served. She carried a little lining material into creasedsignificantly when they reached mid- the nests.Very old femalesincubated their eggs, dle age and again in old age (Table 7). and in the caseof three females (15-18 yr old) The modal clutch was two eggs.We removed allowed to incubate, kept the eggs covered for the eggsof eachfemale assoon as she had com- about the same percentageof time as did four pleted her clutch, and this proceduremay have younger females that were incubating eggs at contributed to an earlier decline in reproduc- the sametime. The oldest female given the op- tive behavior of the femalescompared with the portunity to raise young was 12 years old, and she fed and cared for her nestlings. Individual Village Weavers varied greatly in T^m•E7. Average number of clutchesand eggslaid per year by eight adult female Village Weaversas the apparent effect of old age on their breed- they aged. ing. The oldest female averagedtwice as many clutchesand almost four times as many eggs Bird- Yearlyaverages per seasonas did the best of seven other very per femalea years old females, all of whom were at least 2 years Age observed Clutches Eggs younger. During the 1980breeding season,two Young adult (3-8 yr) 8 9.7 20.7 15-year-oldmales failed to establishterritories Middle age (9-13 yr) 24 5.3* 11.2' and did not breed. That same season another Old age (14-18 yr) 15 2.4* 5.8* 15-year-oldmale built only one nest, obtained a * = significantdecline (P < 0.05) from preceding only one mate, and fertilized only one clutch. ageby Wilcoxonmatched-pairs signed-rank test. But a fourth 15-year-old male held a territory, April1986] BehaviorofOld Weaverbirds 413

TABI•E8. Percentageof instantaneoustime samplesin which old and young adult male Village Weavers engaged in various activities.

1979 1980

Old Young P Old Young P No. of males (territorial: nonterritorial) 4.'3 3:0 2:3 6:3 Resting (perch, preen, sleep) Territorial 62 55 NS 46 26 0.01 • Nonterritorial ------81 62 0.05 a Forage,eat, drink 11 8 NS 11 7 NS Building by territorial males 20 21 NS 21 24 NS Singing 1 0 NS 3 8 0.07 Displayof nest 3 7 0.058 10 15 NS Precopulatoryand copulatorybehavior 1 4 0.055 2 2 NS Total samples 200 200 110 110 120 120 aDifferences statistically significant. Probabilities are basedon relativefrequencies and binomial test(Siegel 1956). dominated the six other males in his aviary, and copulated,although he died near the end and built more nests (23 during that season) of that season(1976). One 20-year-oldmale did than any other male in his aviary. He also ob- not breed, but despite an old wing injury, dom- tained 7 mates, and fathered 27 clutches, far inated the other nine males in his aviary. His more than any of the other 13 malesthat season wing had beenbroken in a fight and had healed in either of the aviaries (Table 2). By 1983, in imperfectly. Our oldest female laid 24 eggs at his 18th year and living in the Los Angeles 15 yearsof age. When she was 18 yearsold, she Zoo, he had outlived all the other old males still lined nests,laid fertile eggs,and incubat- but had not yet built a nest that season,al- ed. though four young adults in his aviary at the Frequencyof variousbehaviors in old and young zoo had done so. He died later that season of adults.--Instantaneoussamples of all activities unknown causes. by eachbird were taken by scanningthe aviary Some Village Weavers continue to breed, or every 6 min for 1-h periods(ARmann 1974). In retain high dominance, at a surprisingly ad- 1979 sampleswere taken for 20 h and in 1980 vanced age. One wild-trapped adult was still for 11 or 12 h in the two aviaries (Tables 8 and fertile at a minimum age of 19 years when last 9). checked for fertility. At 20 years old he still Resting includes perching, self-preening, or held a territory, wove nests, courted females, sleeping. In 1980 the old birds of both sexes

TABLE9. Percentageof instantaneoustime samplesin which old and young adult female Village Weavers engaged in various activities.

1979 1980

Old Young P Old Young P No. of females 5 4 4 9 Rest(perch, preen, sleep) 72 60 0.078 57 27 0.0003a Forage, eat, drink 11 13 NS 20 14 NS Visit male's territory 4 8 0.076 4 7 NS Gather nest materials 1 2 NS 2 6 NS Inside a nest 11 17 0.071 12 38 0.00007' Total reproductiveactivities 16 27 0.012a 17 52 <0.00003a Total samples 200 200 110 110 120 120

a Differencesstatistically significant. Probabilities are basedon relative frequenciesand binomial test (Siegel 1956). 414 COLL•ASETAL. [Auk,Vol. 103 rested significantly more than did the young testis(Collias et al. 1961). Both old and young adults(Tables 8 and 9). Sleepis objectivelyde- malesshowed the usualseasonal cyclic changes fined here as tucking the beak back amongthe in bill color and plumage color before and after scapular feathers. Although not within the the breeding season,when allowed to breed. A conventionalbounds of statisticalsignificance, 24-year-old male still developed black bill col- the old males tended to sing less, to display oration, indicating that he was still secreting their nestsless often to visiting females, and to testosterone. engage in less precopulatoryand copulatory Causesof deathin the aviarybirds.--The aver- activity than did younger adult males (Table age age of 35 birds that had attained at least 8). Old males more often failed to establish a one year of age in our aviarieswas 11.2 years, territory. Males without territories never ob- but somelived much longer. Nine males(26%) tain mates. There was little or no difference and 8 females(23%) reached 14 yearsof age or between the old and young territorial males in more.Of 17 birds that died in the last five years building activity. of the study, 3 died during a severeheat wave In 1980 the old females, compared with the and 4 from the cold one winter night when younger adult females, were inside nests less their heat lamp failed. These 7 birds included often (Table 9). In both years the young fe- 3 old (15-19 yr) and 3 young adult (6-7 yr) malesengaged in significantlymore reproduc- femalesand one old (16 yr) male. One young tive activity than did the old females. adult male subjectto convulsionsdrowned in In May 1982, over a two-week period, the the water pan, and another young adult male frequenciesof various activitiesby our oldest was killed in a territorial fight. An 8-year-old bird, a 23-year-oldmale, were comparedwith male died from cancer of the liver, while our thoseof a 7-year-oldadult male in his prime. oldest bird, a male, died of an impacted intes- These males were hand raised in isolation to tine. The remaining 6 birds, which were 14-19 adulthood,both were very tame,and both could yearsold, died of disease,especially from vital, be observedclosely without disturbing them. bacterial, and protozoan infections of the in- They were maintained indoors in separatein- testine. dividual cages(92 cm long, 42 cm wide, 43 cm None of the birds died of "old age," al- high) in different buildings. Both males were thoughadvanced age may haveincreased their in full breeding plumage at the time of the test. susceptibility,because all of the birds that died All activitieseach bird performed were record- of infections were old. Combinations of factors ed during 1-min time samplesevery 5 min for probablyare often the causeof death in caged 1-h periods.All hours of the day from 0700 to birds.Autopsy of one 15-year-oldmale showed 1900 were sampled for a total of 12 h per bird. "microscopicextended areasof atrophic muscle Each activity was sampled for occurrenceor fibers in the heart, fibrosis of the endocardium nonoccurrence for a total of 144 1-min time and of liver vessels,and severely congested samplesper bird. lungs" (autopsyby H. Niewisch, D.V.M.). This The very old male restedmore often (54%vs. male becameso ill that he stopped eating and 1% of time samples) and moved about much had to be force fed. Another died of an im- less than the younger male. He showed very pactedintestine and enteritis(R. Kray, D.V.M.) little interest in nest materials (4:38) and did when 24 yearsold. He was the oldestmember not weave at all, whereas the younger male of his specieswe know of for whom the dates wove in 38%of the time samples.The old male of hatching and death are both known. His sang only about half as often as the younger heart(examined by R. J. Gayek,D.V.M.) showed adult male, and ate only about a third as often. no histopathology, but there was arterioscle- The old male showedno differencein drinking rosisof the coronaryartery with 90%closure. or bathing and preened himself about as often as did the young adult male. Bill and plumagecycles in advancedage.--Dur- DISCUSSION ing the nonbreeding season,the lower man- dible of the male Village Weaver turns a pale We draw four main conclusions from our re- ivory. Blackening of the bill in the breeding suits.(1) Old age in both male and female Vil- season,singing, territoriality, and nest build- lage Weavers is eventually accompaniedby a ing are stimulated by male hormone from the decline in generalactivity and in breeding be- April!986] BehaviorofOld Weaverbirds 415 havior. (2) Very old malesare generally chosen four yearsas a breeder,then levelsoff, and may lessoften as matesthan are young adult males. decline slightly after eight yearsas a breeder. (3) There is great variability between individ- Most researchon the physiologicalbasis of uals in length of life under supposedlyfavor- aginghas been done with mammals,especially able conditions. (4) Some individuals of this humans (Finch and Hayflick 1977, Behnke et tropicalspecies continue to live and breed in al. 1978, Walford 1983). Decline in general ac- captivity to a surprisinglygreat age for a pas- tivitywith old agecould be caused by a general serine bird. reduction in enzyme activity (Gafne 1981). Re- (1) Declinein activityand breedingbehavior in cent evidence indicates that in old female rats oldage.--Our oldestweaverbirds rested a great the decline in the ability to releasegonadotto- deal (Tables 8 and 9), which suggestsa low en- pinsfrom the anteriorpituitary results mainly ergy level. Botkinand Miller (1974)and Calder from decrementsin hypothalamicfunction, but (1985) reviewed the questionof age-indepen- a decreasein pituitary responsivenessto hy- dent adult mortality in birds and concludedon pothalamicgonadotropin-releasing hormone or both empiricaland theoreticalgrounds that it of the ovaries to gonadotropin,or both, may is more reasonable to assume, at least in older contribute to loss of regular estrous cycles birds, that mortality is age-dependent. Among (Meites 1982). Rather similar factorsmay hold nonpasserinebirds, Coulsen and Wooler (1976) in the aging of reproductivecapacity and be- found a significantlylower annual survival in havior in birds. both sexesof breedingBlack-legged Kittiwakes The decline in activity and reproductive ca- (Rissatridactyla) among older birds (6-17 yr pacityof very old birdssuggests that theywill breedingexperience) than amongyounger birds be at a disadvantagein competitionunder nat- (1-5 yr breeding experience).Among tropical ural conditions,but they may have compensa- passerinespecies, Willis (1983)reported signif- tory assetsfor competingsuch as greater ex- icantlylower survivorshipin color-bandedmale perience and wariness, relatively high SpottedAntbirds (Hylophylaxnaevioides), 7-10 dominance,and ownership of a familiar terri- or more yearsold, in Panamaforests. Two very tory, all of which would aid reproduction.Ac- old males sang little and had lost their mates cording to McClure (1974: 294), summarizing to younger males. returns from more than 100,000 Ploceidae Very old male Village Weavers generally ringed in Asia, only young individuals are showeda declinein breedingbehavior (Tables readilycaught. He states,"Old ringedbirds will 2-4) related to a decline in nest building. Old remain in an area for many yearsand never be femaleVillage Weaverslaid fewer eggs(Tables captured again, even though they are seen 6 and 7) and showed less interest in the nests readily." In a 15-year study of Black-capped woven by the males(Table 9) than did younger Chickadees(Parus atricapillus) in Missouri, El- adult females. der and Zimmerman(1983) found significantly Few data are available on the decline of higherannual survival based on resightingsof breeding with advancedage among passerine color-banded individuals compared with re- birds in nature, such as we observed with our capturedata in the samepopulation. The birds weaversin aviaries.Great Tits (Parusmajor) have becamewary of trapsas they grew older. There been studied a very long time near Oxford, En- is need for caution in the interpretation of gland.The averageage of adultsis 1.9 yearsfor banding returns. females, 2.5 for males. Most die before the age Medawar (1957: 22) defined senescence"as of 5 or 6. The number of young that a female that which predisposesthe individual to death raisesto independenceincreases steadily until from accidental causes of random incidence." the female is about 3 or 4 years old, but there- He suggestedthat senescenceconsists of an in- after breeding successsteadily declines (Per- nate portion, indicatedby the geneticdecrease rins 1979). Dhondt (1985) found evidence that in vitality and in the growth capacityof tissues GreatTits sometimesforego breeding once they with agingbeginning virtually at birth, and an reach5 yearsof age.In a 15-yearstudy of color- environmental portion, comprised of the ac- bandedFlorida Scrub Jays (Aphelocoma coerules- cumulated sum of the effects of recurrent stress cens)Fitzpatrick and Woolfenden(1984) found or injury or infection.Medawar explainedthe that successin producing offspring that reach evolution of senescenceby suggesting that breeding age improvesannually over the first geneswill be selectedfor if they confer high 416 COLL•ASETAL. [Auk,Vol. 103 fitness at a time of life when the individual speciesof frequently banded speciesof North concernedhas high reproductive value, where- Americanpassefine birds lived 10 yearsor more as genes that act later in life will be relatively in the wild (Clapp et al. 1983, Klimkiewicz et unimportant (see also Wilson 1975). However, al. 1983). it seemspossible that a postreproductivephase The life span of adult birds in captivity also of life could evolve by means of family selec- varies greatly within a species.Flower (1938) tion if old individuals in someway continue to mentioned severalspecies of passefinebirds in aid the survival or reproduction of their de- which an exceptional individual lived for 20 scendantsafter the latter reach maturity. There years or more in captivity. Of 167 individuals is no extensive evidence as yet of a postrepro- of 27 speciesof Ploceidae (weaverbirds and al- ductive phase of life for most birds in nature. lies) living in zoos in 1981 [listed by the Inter- We have given experimentalevidence for a def- national SpeciesInventory System (ISIS); Seal inite decline in breeding capacitywith old age and Olsen 1978], the median age was between in captive Village Weavers, and have cited 4 and 5 years. Twelve birds were over 10 years evidence for senescencein nature among two old, however, and 3 were over 15 years old. other passefinespecies, Great Tits and Spotted The oldest,a Ploceuscucullatus, was over 20 years Antbirds. old. The oldest Ploceus cucullatus in the wild (2) Old malesless successful in competitionfor that we know of was from the South African mates.--In the Village Weaver the female gen- race P. c. spilonotusand lived over 14 years erally selectsher mate. After reaching 14 years (Adrian Craig pets. comm.).Of 437 individuals of age, the malesbuilt fewer nestsand obtained of 47 estrildid finches() living in var- fewer mates in this polygynous speciesthan ious zoos in 1981, the median age was 3 years when they were younger (Table 4). In the mo- (ISIS). Only 2 birds were over 10 years old; the nogamous Rock Dove (Columbalivia), individ- oldest was between 15 and 20 years of age. In uals of both sexesover 7 years of age were def- nature, the tiny estrildid finchesalso have rel- initely and significantly discriminated against atively short lives (Morel 1973, Woodall 1975). (Burley and Moran 1979). In general, longevity in different species of (3) Greatindividual variability in maximumlife birds correlateswell with body size (Linstedt span.--We have shown that individual Village and Calder 1976, Brown and Pomeroy 1984). Weavers under presumably favorable condi- We conclude that there is tremendous vari- tions vary greatly in length of life. Great vari- ability among individuals of bird speciesin po- ability amongindividuals in maximumlife span tential, as well as in actual, life span. The cause may turn out to be a general principle. Coulsen and significanceof this variation are problems and Horobin (1976) found that Arctic Terns for investigation. (Sternaparadisaea) ringed on the Fame Islands, (4) Breedingto a relativelyadvanced age among Northumberland, usually began breeding at tropicalpasserine birds.--Part of the explanation about 4 years of age, with a further life expec- for the length of the reproductivelife of some tancy of 7 years. But one 29-year-old tern suc- Village Weaversmay be in the tropical habitat. cessfully reared young. One of the oldest re- Animal species in which population size is corded living birds was a Royal Albatross variable and frequently below the carrying ca- (Diomedeaepomophora sanfordi) that was still pacity of the environment usually have short breedingin New Zealandwhen at least55 years lives, and individuals of specieswhere popu- old (Robertson 1980, John Warham pets. lation size is fairly constantand at or near car- comm.). Fisher (1975), however, found that of tying capacityusually have relatively long lives 742 Laysan Albatrosses (Diomedeaimmutabilis) (Pianka 1970). Pianka believes the latter pat- banded on Midway Island in the Pacific,only tern is probably more characteristicof tropical 3% lived to be more than 40 years old. species;however, investigation of many more In 70 yearsof banding returns over the world speciesis needed to validate the hypothesis. for severalmillion birds, mostly from the north Among 95 American Goldfinches (Carduelis Temperate Zone, Rydzewski (1978) listed 86 tristis)5 years or older, banded in the U.S. and speciesof passefine birds in which some indi- Canada, there was only one bird as old as 9 viduals lived for at least 10 years. But the vast years (Middleton and Webb 1984). The five majority of individuals lived a much shorter oldest goldfinches recovered in Canada were time. Occasional individuals in 22% of 141 only 5 or 6 years old. Only one 9-year-old bird April1986] BehaviorofOld Weaverbirds 417 appearedamong 253 Black-cappedChickadees 1985), we here emphasizethe importance of color-bandedin Missouri over a 15-yearperiod one factor--life span--that has been relatively (Eider and Zimmerman 1983). In 12 other stud- neglected. ies a maximumage of 8-10 yearswas reported for this species(Elder and Zimmerman 1983). ACKNOWLEDGMENTS In contrast, in Trinidad Snow and Lill (1974) found that ages of 10 years or more were not We are very grateful for supportof this research by the National ScienceFoundation (BNS 75-02329 uncommon in two species of color-banded AO1 to UCLA and N.E.C. and BNS 77-11585 to the manakins (Pipridae). One male White-bearded LosAngeles County Museum of Natural Historyand Manakin (Manacus manacus)was still resident E.C.C.)and by the University of California at Los on the courtship display grounds at a mini- Angeles(UCLA 1623to N.E.C.).Jean Delacour kindly mum age of 14 years. A male of this species sent us the original stock used in establishingour copulatedwhen at least 11 years old, and a fe- captiveweaverbird colony. We thank the following male when 10 years old. investigatorswho helpedgather data on the breed- Many tropicaland subtropicalspecies of birds ing of the weaverbirdsin our experimentalcolony: have helpers at the nest, and such cooperative JaniceK. Victoria, Lloyd F. Kiff, RobertJ. Shallen- or communal breedersare long-lived (Rowley berger,Carl E. Rischer,Michael Brandman,Jeffrey 1983, Brown and Pomeroy 1984). For example, Fujimoto,Trudy CarsonBell, ThomasHaglund, and Barry Higbee. Diane Riska,Bruce Tedford, and Al- in Arizona a Pinyon Jay (Gymnorhinuscyano- exandraZaugg-Haglund helped care for the birds cephala)was still breeding at 15 years (R. P. during the experimenton old age.We thank Russell Balda pets. comm.) and a Gray-breastedJay P. Balda,Paul Boyer,Jerram L. Brown,W. A. Calder (Aphelocomaultramarina) at 16 yearsof age (J. L. III, A. J. F. K. Craig, Marvin Jones,John K. H. Lu, Brown pets. comm.). In 1983 the oldest passer- Bill Maynard,Bruce Miller, Arthur C. Risser,Charles ine bird in the San Diego Zoo was a communal H. Sawyer,Roy Walford,and JohnWarham for use- breeder, a 26-year-old Apostlebird (Struthidea ful information; Wilfred J. Dixon and Donald Guth- cinerea) from Australia (Marvin Jones pets. rie for statistical advice; and the UCLA and the Los comm.). In the cooperativelybreeding Florida AngelesZoo veterinarystaff for autopsyreports. Jon- athan Atwood, Elizabeth Flint, Diane Riska, William Scrub Jay there is evidence, based on lifetime M. Shields, and Curtis S. Adkisson made helpful breedingsuccess, that mostyoung birds reach- commentson the manuscript. ing breedingage are producedby relativelyfew long-lived breeders (Fitzpatrick and Woolfen- den 1984, Woolfenden and Fitzpatrick 1984). LITERATURE CITED This finding indicatesa selection pressurefor longevity in nature. ALTMANN,J. 1974. Observationalstudy of behavior: Among birds, there is generally strong in- samplingmethods. Behaviour 49: 227-267. dividual competition for nest sites and mates. BEHNKE,A., C. E. FINCH,& G. B. MOMENT(Eds.). 1978. Many tropical bird species lay very small Biology of aging. New York and London, Ple- num Press. clutchesand sufferhigh nestpredation (Skutch BOTKIN,D. B., & R. $. MILLER.1974. Mortality rates 1985). High adult survival means that fewer and survival of birds. Amer. Natur. 108: 181-192. placesare available for young birds and that BROWN,L. H., & D. E. POMEROY.1984. The age struc- fewer young birds are needed to replacelosses ture of populationsof wild birds in tropicalAf- in the adult population. Thus, long reproduc- rica, asdemonstrated by plumagecharacters and tive life might help explain the small clutch marking techniques.Proc. 5th Pan-African Or- size (2 or 3 eggs) in the Village Weaver, as in nithol. Congr: 97-119. many other tropical passetines(Collias et al. BURLEY,N., & N. MORAN. 1979. The significanceof 1981). Long reproductive life with repeated age and reproductive experience in the mate small clutchesand nesting attempts making up preferencesof feral pigeons,Columba livia. Anim. Behav. 27: 686-698. lossesto predation may well be a major char- CALDER,W. A., III. 1985. The comparativebiology acteristic of many tropical passefine species. of longevity and lifetime energetics.Exper. Ger- This would distinguish them from most pas- ontol. 20: 161-170. sefinebirds of the north TemperateZone. While CLAPP,R. B., M. K. KLIMKIEWICZ,& A. G. FUTCHER. it is evident that the evolution of clutch size in 1983. Longevity records of North American birds depends on the interaction of many fac- birds: Columbidae through Paridae. J. Field Or- tors(Cody 1966,Lack 1966,Murray 1985,Skutch nithol. 54: 123-137. 418 COLLIASET AL. [Auk,Vol. 103

CODY,M. 1966. A general theory of clutch size. Evo- FRY, C. H. 1980. Survival and longevity among lution 20: 174-184. tropical land birds. Proc. 4th Pan-African Orni- 1971. Ecologicalaspects of reproduction.Pp. thol. Congr.: 333-343. 462-512 in Avian biology, vol. 1 (D. S. Farner GAFNE,A. 1981. Location of age-related modifica- and J. R. King, Eds.). New York, Academic Press. tions in rat muscle glyceraldehyde-3-phosphate COLLIAS,N. E., M. BRANDMAN,J. K. VICTORIAßL. F. dehydrogenase.J. Biol. Chem. 256(17):8875-8877. KIFF, & C. E. RISCHER. 1971. Social facilitation KLIMKIEWICZ, M. K., R. B. CLAPP, & A. G. FUTCHER. in weaverbirds:effects of varying the sex ratio. 1983. Longevity records of North American Ecology52: 829-836. birds: Remizidaethrough Parulinae. J. Field Or- --, & E. C. COLLIAS. 1970. The behaviour of the nithol. 54: 287-294. West African Village Weaverbird. Ibis 112: 457- LACK,D. 1966. Population studiesof birds. Oxford, 480. Clarendon Press. --, & ---. 1981. Survival and intercolony LEwis,D. M. 1980. Population biology of a com- movements of White-browed Sparrow Weavers munal bird speciesPlocepasser mahali. Unpub- Plocepassermahali. Scopus 5: 61-65. lished Ph.D. disse•'tation,Austin, Univ. Texas. ß --, C. H. JACOBS,& C. R. Cox. 1980. LINSTEDT,$. L., & W. A. CALDER.1976. B•)•y'size Breeding behavior by very old (15 to 20 years) and longevity in birds. Condor 78: 91-94. weaverbirds (Ploceidae, Passeriformes). Amer. MCCLURE,H.E. 1974. Migration and survival of the Zool. 20: Abstr. 339. birds of Asia. Bangkok, U.S. Army Component , , --, & . 1981. Old age and SEATO Medical Res. Lab. behavior in a tropical passerinespecies (Ploceus MEDAWAR,P. B. 1957. The uniquenessof the indi- cucullatus).99th Annual Meeting, Amer. Orni- vidual. Londonß Methuen and Co. thol. Unionß Edmonton, Alberta, Canada, p. 13 MEITES,J. 1982. Changesin neuroendocrinecontrol (Abstr.). of anterior pituitary function during aging. Sixth --, P. J. FRUMKES,D. S. BROOKS,&: R. J. BARFIELD. Geoffrey Harris Memorial Lecture.Neuroendo- 1961. Nest-building and breeding behavior by crinology 34: 151-156. castratedVillage Weaverbirds (Textor cucullatus). MIDDLETONßA. L. A., & P. WEBB.1984. Longevity of Amer. Zool. 1: Abstr. 101. the American Goldfinch. J. Field Ornithol. 55: COULSEN,J. C., & J. HOROBIN. 1976. The influence 383-386. of age on the breeding biology and survival of MOREL,M.-Y. 1973. Contribution& l'6tude dynamic the Arctic Tern Sternaparadisaea. J. Zool. 178:247- de la population de Lagonostictasenegala L. (Es- 260. trildidae) & Richard-Toll (S•n•gal). Mere. Mus. --, R. D. WOOLER. 1976. Differential survival Natl. Hist. Nat., Ser. A., Zool. 93: 1-155. rates among kittiwake gulls Rissatridactyla (L.). MURRAY, B.G. 1985. Evolution of clutch size in J. Anim. Ecol. 45: 205-213. tropical speciesof birds. Pp. 505-519 in Neotrop- DHONDT, A. A. 1985. Do old Great Tits forego ical ornithology (P. A. Buckley, M. S. Foster,E. breeding?Auk 102: 870-872. S. MortonßR. S. Ridgely, and F. G. Buckley,Eds.). DIXON, W. J., & F. J. MASSEYßJR. 1969. Introduction Ornithol. Monogr. No. 36. to statisticalanalysis, 3rd ed. New York, Mc- PERRINS,C. M. 1979. British tits. Londonß Collins. Graw-Hill. PIANK•, E.R. 1970. On "r" and "k" selection. Amer. ELDER,W. H., & D. ZIMMERMAN.1983. A comparison Natur. 104: 592-597. of recaptureversus resighting data in a 15-year RICKLEFS,R.E. 1969. An analysisof nesting mortal- studyof survivorshipof the Black-cappedChick- ity in birds. SmithsonianContrib. Zool. 9: 1-48. adee. J. Field Ornithol. 54: 138-145. ROBERTSON,C. J. R. 1980. Albatrossesof Australasia. FINCH, C. E., & L. HAYFLICK(Eds.). 1977. Handbook Australasia Seabird Group Newsletter No. 14, of the biology of aging. New Yorkß Van Nos- July. trand Reinhold Co. ROWLEY,I. 1983. Commentary.Pp. 127-133 in Per- FISHERßH. J. 1975. Longevity of the LaysanAlba- spectivesin ornithology (A. H. Brush and G. A. tross,Diomedea immutabilis. Bird-Banding 46: 1-6. Clarkß Jr., Eds.). Cambridge, Cambridge Univ. FITZPATRICKßJ. W., •r G. E. WOOLFENDEN.1984. Some Press. aspectsof lifetime reproductive successamong RYDZEWSKI,W. 1978. The longevity of ringed birds. FloridaScrub Jays. 102nd Annual MeetingßAmer. Ring 8 (96-97): 218-262. Ornithol. Union, Lawrence, Kansas. Abstr. 41. SEAL,U.S., & J. M. OLSEN. 1978. The International FLOWERß$. $. 1938. Further notes on the duration SpeciesInventory System.Avicult. Mag. 84: 171- of life in animals. IV. Birds. Proc. Zool. Soc. Lon- 173. don, Series A, 108: 195-235. SIEGELß$. 1956. Nonparametric statisticsfor the be- FOGDEN,M. P.L. 1972. The seasonalityand popu- havioral sciences. New York, McGraw-Hill Book lation dynamicsof equatorial forest birds in Sa- Co. rawak. Ibis 114: 307-342. SKUTCH,A. F. 1985. Clutch sizeß nesting successß April 1986] Behaviorof Old Weaverbirds 419

and predation on nests of neotropical birds re- SpottedAntbirds (Hyplophylax naevioides). J. Field viewed. Pp. 575-594 in Neotropical ornithology Ornithol. 54: 413-414. (P. A. Buckley, M. S. Foster,E. S. Morton, R. S. WILSON,E. O. 1975. Sociobiology.The new synthe- Ridgely, and F. G. Buckley, Eds.). Ornithol. sis. Cambridge, Massachusetts,Harvard Univ. Monogr. No. 36. Press. SNOW,D. W., & A. LILL. 1974. Longevityrecords for WOODALL,P. F. 1975. On the life history of the someneotropical land birds. Condor 76: 262-267. Bronze Mannikin. Ostrich 46: 55-86. WALFORD,R.L. 1983. Maximumlife span.New York, WOOLFENDEN,G. E., •r J. W. FITZPATRICK. 1984. The Avon Books. Florida ScrubJay. Demography of a cooperative- WILLIS,E.O. 1983. Longevitiesof somePanamanian breeding bird. Princeton,New Jersey,Princeton forestbirds with note of low survivorshipin old Univ. Press.