CAVITY ROOSTING, PHILOPATRY, AND COOPERATIVE BREEDING IN THE GREEN WOODHOOPOE MAY REFLECT A PHYSIOLOGICAL TRAIT

J. DAVID LIGON, • CYNTHIA CAREY,2 AND SANDRA H. LIGON• •Departmentof Biology,University of New Mexico,Albuquerque, New Mexico87131 USA, and 2Departmentof EPO Biology,University of Colorado,Boulder, Colorado 80309 USA

ABSTRACT.--Cooperativebreeding in birds generally is related to a critical resource.A critical limiting factor for the Green Woodhoopoe(Phoeniculus purpureus) is the availability of roost cavities.Although predation at roost holes is the major sourceof mortality, wood- hoopoesinvariably roost in cavitiesor under loosebark. This dependenceon cavitiesappears to be relatedto the birds' inability to copephysiologically with low nighttime temperatures. This is the first evidenceto suggestthat a physiologicallimitation has played a major role in the evolutionof philopattyand possiblycooperative breeding in an avian species.Received 2 April 1987, accepted8 October1987.

THEprevalent view of the basisof cooperative Ligon 1982), the returning woodhoopoesusu- breeding in birds is that it is related to some ally flew to a secondarycavity. In the absence kind of resource(e.g. Brown 1974, Koenig and of a backuproost hole, the birds crawled under Pitelka 1981, Emlen 1982). In some cases the loosebark to spend the night. There is no evi- putative critical resource is territorial space dencethat they ever roostedin the open. More- (Woolfenden and Fitzpatrick 1984), while in over, becausewoodhoopoes do not excavate others it may be costly self-constructedstruc- their own cavities and must compete for holes, tures suchas food storagesites (Stacey and Koe- often unsuccessfully,with numerousother cav- nig 1984) or roostcavities excavatedinto living ity-dwelling animals (J. D. Ligon and S. H. Li- trees(Ligon 1970).Although the selectivebasis gon 1982, 1988), high-quality (i.e. relatively of the nondispersalelement of avian coopera- predator-proof)roost sites are scarce.The severe tive breeding is unresolved (Staceyand Ligon interspecific competition for cavities (honey- 1987), the specificfactors prohibiting dispersal bees, rodents, other birds) means that safe cav- or, alternatively, favoring group living vary ities vary in number both spatially and tem- from speciesto species,and nondispersal of porally. Becauseof the frequent takeover of young birds is a necessaryprecondition for the cavities by other species,a key element of ter- evolution of cooperativebreeding. ritory quality is the number of potential backup We studied the behavior and ecology of the or reserve roost sites. cooperatively breeding Green Woodhoopoe Although the woodhoopoesthus require cav- (Phoeniculuspurpureus) near Lake Naivasha in ities in which to roost,this behavior appearsto the Rift Valley of Kenya (0ø40'S,36ø23'E) for be costlybecause the holes used often are in several years, 1975-1984, and found that for weak wood not chosenby other species,and thesebirds tree cavitiessuitable for roosting are the birds are vulnerable to arboreal nocturnal a critical limiting factor (J. D. Ligon and S. H. mammalian predators, such as large-spotted Ligon 1978, 1982, 1988). We followed several genets(Genetta tigrina) and possiblycats (Felis hundred birds to their roosts,usually to capture libycaand F. domesticus;Ligon and Ligon 1978). them for individual marking (386 woodhoo- Driver ants (tribe Dorylini) also capture adult poes banded). Without exception, the wood- woodhoopoesin their roost holes. Primarily as hoopoesroosted under cover. They almost al- a result of nocturnal predation at roost sites, ways entered a tree cavity and usually roosted Green Woodhoopoessuffer annual mortality in groups; as many as eight birds have been ratesabout twice as high as some other well- recorded in a single cavity (J. D. Ligon and S. studied cooperatively breeding species,about H. Ligon 1982). If, during the day, the roost 40% per year for males and 30% per year for cavity was occupied by another species,e.g. females (Ligon 1981, 1983; Ligon and Ligon honeybees(Apis mellifera; S. H. Ligon and J. D. 1988).

123 The Auk 105: 123-127. January1988 124 LIGON,CAREY, AND LIGON [Auk,Vol. 105

T^BLE1. Metabolic rates and body temperaturesof h. During this perioddried air flowed continuously Green Woodhoopoesat different ambient temper- throughthe chamberat approximately725 ml/min. atures (T•). Measurementof •rO2 was conductedin a manner

Male Male Fe- similarto that describedby Dawsonand Carey(1976). I 2 male Dry air was channeledthrough flow meters,into the metabolicchambers, through tubes containing As- Body mass(g) 71.0 66.1 52.3 carite and Drierite for removal of CO2 and water va- Daytime body temperature por, respectively,and then into a Beckmanmodel 755 (øC) at Ta 27øC 42ø 42ø 40ø paramagnetic oxygen analyzer for analysis. The air Metabolic rate (ml O2.g-•.h -•) from each chamberwas sampledsequentially for 5 At T• 30.5øC 1.33 1.67 1.14 min, and room air was sampled once an hour. After At T• 27øC 1.35 1.40 1.09 steady values were obtained at the chamber air tem- At T• 19øO 1.92 2.54 1.31 perature of 30.5øC,the temperature of the constant- Body temperature (øC) at Ta temperaturebox was lowered to 27øC,and finally to 19øC, at 2400 36 ø 35 ø 33 ø 19øC.The birds were exposedto 27øCand 19øCfor at • Metabolic rates at i9øC were taken as the birds moved about. We least I h after the temperature in the chamber had were unable to obtain "resting" rates. stabilizedbefore measurementsof •O2 were made. Two or three measurements were taken at 30.5øC and 27øCand averaged to represent the value for that In view of the high risk of mortality associ- individual at those temperatures. atedwith roostingin cavities,the obviousques- Body temperatureswere taken cloacallywith a YSI tion is why woodhoopoesdo not roost among telethermometerused in conjunctionwith a YS! small- animal probe. the thorny branches and twigs of the acacia trees(Acacia xanthophloea) on the study site, as do mostother small birds,including someother RESULTS group-living speciesof similar masssuch as two Average oxygen consumption of the three speciesof Turdoidesbabblers (pets. obs.) and the birds at 30.5øCfell within 5% of the value pre- Grey-backed Fiscal Shrike (Laniusexcubitorius; dictedby the equationof Aschoffand Pohl (1970) Zack 1986). The low, 7-10øC, nighttime tem- for nonpasserinebirds measuredin the inactive peraturescharacteristic of the study site at 1,860 phase of their daily cycle (Table 1). Although m above sea level, together with the extreme we did not measure•O2 at a wideenough va- agitationshown by woodhoopoeswhen denied riety of temperaturesto describea thermoneu- accessto their roostsat dusk, suggestedthey tral zone, the similarity between the predicted might be physiologically incapable of coping and measured values and information about successfullywith low nocturnal ambient tem- typical lower critical temperatures in birds of peratures (Ligon and Ligon 1978). To address similar body mass(Calder and King 1974) sug- this possibility, we investigated thermoregu- gest that these birds were in what would be latory responsesof three captiveGreen Wood- termeda thermallyneutral condition. The hoopoesto low nocturnal ambient tempera- tures. of thesebirds remained relatively stableas am- bient temperaturewas lowered to 27øC,but in- creasedsharply when air temperaturedropped METHODS to 19øC.Although the birds were totally inac- We borrowed 2 adult male Green Woodhoopoes tive at 30.5 ø and 27øC, the elevated rates at 19øC from the Denver Zoo and 1 young female from the were coincident with frantic hopping and flut- Rio GrandeZoo in Albuquerque,New Mexico. One tering movements of the bird against the sides of the males and the female had been hatched in and roof of the metabolic chambers. These captivity;the other male was capturedfrom the wild. movements became progressively more fre- Measurementof oxygenconsumption (•O2) was quent the longer the birds remained at 19øC.To conductedusing 3.8-1metabolic chambers with black- avoid damage to the woodhoopoes,we termi- ened inner surfaces.Air temperatures within the nated the experiment at 2400 and immediately chamber were measuredwith a copper-constantan removed them from the environmental cham- thermocouple connectedto a Soltex millivolt record- er. The birdswere placedindividually in the cham- ber. The body temperature of each woodhoo- bers at 1600. The chamberswere then placed in a poe, measuredas it was taken from the chamber, constant-temperaturebox, and the birds were allowed indicatedthat despitethe increasedactivity, the to adjust to conditions in the chamber at 30.5øCfor 4 birdswere becominghypothermic (Table 1). The January1988] GreenWoodhoopoe Physiology 125

hypothermia was not an adaptive response to turnal metaboliccosts. Early selectionfor cavity low ambient temperatures; rather, it occurred roostingcould have led to an eventualloss of despite the metabolic heat production associ- thermoregulatoryabilities at low ambient tem- ated with intense locomotor activity. peratures. If cavityroosting is adaptive,low metabolic DISCUSSION rates make for more efficient use of cavities, in that 02 uptake and CO2 production in the cav- Compared with the Temperate Zone avian ities is lessthan if the birds had higher rates. speciesthat have been studied (mostly passer- This could be an important considerationin an ines), these tropical coraciiform birds demon- environment where cavities are in limited sup- strated a very different and apparently inap- ply and where several birds sometimesroost propriate pattern of response to cold air together.Nondispersal of young as a result of temperatures.In general, birds maintain nor- their dependenceon scarcecavities could have mothermia during exposureto cold by elevat- set the stagefor the eventualdevelopment of ing heat production through shivering ther- cooperativebreeding. mogenesisand by minimizing activity, which Phylogenetichistory interpretation.--The meta- would otherwise disrupt the insulative quali- bolic rates of Green Woodhoopoesprobably are ties of the feathers (Dawson and Carey 1976). characteristic of the woodhoopoe family In contrast, although the woodhoopoes re- Phoeniculidae.All eight membersof this dis- mained quiet at 30.5øCand 27øC,they became tinctivegroup are confinedto sub-SaharanAf- extremely active when air temperatureswere rica, and most occur in warm, equatorial low- lowered to 19øC.We cannot prove that these lands (Davidson and Ligon in press).Thus, an birds did not employ shivering thermogenesis, ability to maintainnormothermia through shiv- but if they did so, it was ineffective at pre- ering thermogenesismay never havebeen pres- venting hypothermia at what appears to be a ent in this lineage. This could have promoted moderate ambient temperature. The intense ac- cavity-roostingbehavior early in the group's tivity we observedat 19øCmay be the solemeans history,because heat produced by evena single these birds have of generating extra metabolic bird in a protectedcavity can raise air temper- heat. In short, in the absence of insulated roost ature sufficiently so that an increase in meta- cavities, or other birds with which to cluster, bolic heat productionis unnecessary(Kendeigh the woodhoopoesshowed little ability, phys- 1961). Limited data on a few other coraciiforms iological or behavioral, to cope appropriately also either suggestlow metabolicrates (Ken- with low nighttime ambient temperatures. deighet al. 1977,Oniki 1975)or indicateuse of The metabolic data can account for the ap- cavities to ameliorate ambient temperatures parent paradox that woodhoopoesroost in cav- (White et al. 1978, Hamas 1981). ities despite the fact that the associatedrisk of Only two membersof the woodhoopoefam- mortality is high. Interpreting the basisfor this ily, the Green Woodhoopoe(plus fellow su- situation is more difficult, however (Gould and perspeciesmembers P. somaliensisand P. damar- Lewontin 1979; see also Murray 1986). In the ensis)and the White-headed Woodhoopoe (P. caseconsidered here severalplausible interpre- bollei),occur commonly at high elevations(Da- tations exist. These fall into one of two broad vidsonand Ligon in press).These also are the categories, "adaptationist" or "phylogenetic largestspecies, 2.5-3.0 timesthe sizeof the oth- history." We briefly present these alternative ers, and are the only ones that exhibit cooper- perspectives and interpretations without at- ative breeding. Large body size may increase tempting to marshal argumentsfor either. resistanceto low ambient temperatures, but it Adaptationistinterpretations.--The key distinc- also reduces the number of cavities the birds tion between an adaptive interpretation and an potentially coulduse for roosting,thus favoring interpretation based on phylogenetic history is long-termretention of young birds in their na- that in the former, low standard metabolic rates tal territories, where known roost sites are avail- (SMR) are presumed to be adaptive. For exam- able. Retention of young birds is a necessary ple, low metabolic demands reduce foraging precursorof cooperativebreeding. requirements,or lower growth rates of young The high mortality rate of Green Woodhoo- birds reduce their risk of starvation. A second poes as compared with other well-studied co- possibilityis that cavity roostingdecreases noc- operative breeders makes their metabolic pat- 126 LIGON,CAREY, ^iXlD LIGON [Auk,Vol. 105 tern appear to be a constraint: The birds BROWN,J. L. 1974. Alternate routes to sociality in apparently must roost in cavities, and our evi- jays--with a theory for the evolution of altruism dence suggeststhat most mortality resultsfrom and communal breeding. Am. Zool. 14: 63-80. nocturnalpredation on roostingbirds. It is im- CALDERßW. A., & J. R. KING. 1974. Thermal and possible to determine, however, whether the caloric relations of birds. Pp. 260-415 in Avian biologyßvol. 4 (D. S. Farrierand J. R. King, Eds.). dependenceon cavitiesis primarily a constraint New Yorkß Academic Press. or an adaptation. Knowledge of roost habits of DAVIDSONßN. K., & J. D. LIGON. In press. Order other species of phoeniculids and of Green Coraciiformes, Phoeniculidae. In The birds of af- Woodhoopoesin areas where nighttime tem- rica, vol. 3 (E. K. UrbanßC. H. Fry, and S. Keithß peratures are warm, together with knowledge Eds.). New Yorkß Academic Press. of standardmetabolic ratesof the other species, DAWSONßW. R., & C. CAREY. 1976. Seasonal accli- would provide valuable insights;even such in- matization to temperature in cardueline finches. formation, however, would not lead to un- I. Insulative and metabolic adjustments.J. Comp. equivocal falsification of either alternative. Physiol. 112: 317-333. Moreover, elements of each interpretation EMLEN,S. T. 1982. The evolution of helping. I. An ecologicalconstraints model. Am. Nat. 119: 29- probably are correct, and they most likely in- 39. fluence each other. Thus, at present, about all GOULDßS. J., & R. C. LEWONTIN. 1979. The spandrels that can be concluded is that contemporary of San Marco and the Panglossianparadigm: a Green Woodhoopoesrequire accessto roostcav- critique of the adaptationistprogramme. Proc. R. ities, and that this requirement may have had Soc. London B 205: 581-598. a profound effecton the socialbehavior of these HAMAS,M.J. 1981. Thermoregulatory development birds. First, most woodhoopoesmove only 1-2 in the Belted Kingfisher. Comp. Biochem. Phys- territories when they disperse, occupying ter- iol. 69A: 149-152. ritories where they probably know the location KENDEIGH,S.C. 1961. Energy of birds conservedby and quality of cavities(Ligon and Ligon 1988). roostingin cavities.Wilson Bull. 73: 140-147. ßV. R. DOL'NIK, & V. M. GAVRILOV. 1977. Avi- Second, the scarcityof safe cavities could have an energetics.Pp. 127-204 in Granivorous birds promoted the development of extended or per- in ecosystems(J. Pinowski and S.C. Kendeigh, manent tolerance of young birds by their par- Eds.). CambridgeßEngland, Cambridge Univ. ents (Ligon 1981), which in turn may have set Press. the stagefor the eventual evolution of the suite KOENIG,W. D., & F. A. PITELK&. 1981. Ecological of behaviors that characterize cooperative factors and kin selection in the evolution of co- breeding. operativebreeding in birds. Pp. 261-280 in Nat- ural selection and social behavior: recent re-

ACKNOWLEDGMENTS searchand new theory (R. D. Alexanderand D. W. Tinkle, Eds.). New York, Chiron. We thank J. Moore, J. Mallen, and W. Aragon of LIGONßJ.D. 1970. Behaviorand breeding biology of the Rio Grande Zoo in Albuquerque, New Mexico, the Red-cockadedWoodpecker. Auk 87: 255-287. and J. Wortman, S. Haeffner, and B. Caverider of the 1981. Demographicpatterns and communal Denver Zoo for generously loaning us the Green breedingin the GreenWoodhoopoe (Phoeniculus Woodhoopoesused in this study.W. Koenig, S. Zack, purpureus).Pp. 231-243 in Natural selectionand G. Farley, P. StaceyßE. Toolsonßand B. McNab pro- social behavior: recent research and new theory vided discussionand help with the manuscript. R. E. (R. D. Alexander and D. W. Tinkle, Eds.). New Leakeyand G. R. Cunningham-vanSomeren of the Yorkß Chiron. National Museum of Kenya madepossible the Ligons' 1983. Cooperation and reciprocity in avian researchin Kenyaßand K. G. Ruchiamiof the Office socialsystems. Am. Nat. 121: 366-384. of the Presidentof Kenya granted researchpermits. --, & S. H. LIGON. 1978. The communal social The field studies of Green Woodhoopoeshave been systemof the Green Woodhoopoein Kenya. Liv- supportedby the National ScienceFoundationß The ing Bird 17: 159-197. National GeographicSocietyß The American Museum --, & --. 1982. The cooperative breeding of Natural History, the U.S. National Fish and Wild- behaviorof the GreenWoodhoopoe. Sci. Am. 247: life Laboratory,and the University of New Mexico. 126-134. --, & --.. 1988. Territory quality: key de- LITERATURE CITED terminant of fitness in the group-living Green Woodhoopoe.Pp. 229-254 in The ecologyof so- ASCHOFF,J., & H. POHL. 1970. Rhythmic variations cial behavior (C. Slobodchikoff,Ed.). New York, in energy metabolism.Fed. Proc. 29: 1541-1552. Academic Press. January1988] GreenWoodhoopoe Physiology 127

LIGON, S. H., & J. D. LIGON. 1982. Abstract, 100th a challenge to the habitat saturationmodel of meetingof The AmericanOrnithologists' Union. cooperativebreeding. Am. Nat. 130: 654-676. Chicago,Illinois. WHITE, F. N., G. A. BARTHOLOMEW,& J. L. KINNEY. MUI•P,A¾,B. G., JR. 1986. The influenceof philosophy 1978. Physiologicaland ecologicalcorrelates of on the interpretationof interspecificaggression. tunnel nestingin the EuropeanBee-eater, Merops Condor 88: 543. apiaster.Physiol. Zool. 51: 140-154. ONIKI, Y. 1975. Temperaturesof some Puerto Rican WOOLFENDEN,G. E., & J. W. FITZPATRICK. 1984. The birds, with note of low temperature in todies. Florida ScrubJay: demography of a cooperative- Condor 77: 344. breeding bird. Princeton, New Jersey,Princeton STACEY,P. g., & W. D. KOENIG. 1984. Cooperative Univ. Press. breedingin the Acorn Woodpecker.Sci. Am. 251: Z^c•<,S. 1986. Behaviourand breeding biology of 114-121. the cooperatively breeding Grey-backed Fiscal , & J. D. LIGON. 1987. Territory quality and Shrike Laniusexcubitorius in Kenya. Ibis 128: 214- dispersaloptions in the Acorn Woodpecker,and 233.