The Auk 109(3):609-619, 1992 FEMALE-BIASED DELAYED DISPERSAL AND HELPING IN AMERICAN CROWS CAROLEE CAFFREY Departmentof Biology,University of Californiaat LosAngeles, LosAngeles, California 90024, USA AnsTRAcr.--Observationof a resident population of WesternAmerican Crows (Corvusbra- chyrhynchoshesperis) over five yearsrevealed a socialstructure that deviatesfrom that of a majority of cooperativebreeders, including other corvids.Breeding density was unusually high (0.8 pairs/ha). Core areasoccupied by pairs and familieswere small,overlapped exten- sively with thoseof neighbors,and were not defendedagainst conspecifics. A nonbreeding flock was resident on the study site. Juvenile dispersalpatterns were highly variable; indi- viduals dispersed at various ages after two months postfledging. They either joined the nonbreedingflock or left the studyarea. Some individuals delayed dispersal for one or more years,while somereturned home after extendedabsences. As yearlings,more femalesthan males were resident in their natal core area during the breeding season.Most, but not all of thoseat home servedas helpers.More femalesthan malesassisted their parentsin breeding. The female bias in dispersaland helping is unusual.It suggeststhat the costsand benefits associatedwith these behaviors differ between the sexes,and may be different from those postulatedfor many other cooperativebreeders. Received 28 January1991, accepted 13 January 1992. ALTHOUGHMUCH diversity exists in life-his- common to many cooperativelybreeding spe- tory traits among avian cooperativebreeders, a cies, including other corvids.I use patternsof majority of speciesshare a number of ecological, spaceuse, dispersal,settlement, and breeding demographic and social characteristics.Most to characterize the unusual social organization speciesare primarily monogamouswith helpers in this population. at the nest (Brown 1987, Smith 1990). Breeding groupstend to be permanentresidents on all- STUDY AREA AND METHODS purpose territories (e.g. Lewis 1981, Brown et al. 1983,Rabenold 1985, Curry and Grant 1990); I studied a resident population on the BalboaGolf colonially breeding species are uncommon Course in Encino, California, from March 1985 to (Brown 1974, 1978,Heinsohn 1987,Emlen 1990). March 1990. The habitat consists of tracts of grass Delayed dispersalby offspringrarely occursif separatedby rows of trees,with additionalclumps of dispersaland floating are an option and, there- trees scatteredthroughout the site. The tree flora is fore, within-population nonbreeding flocksare dominated by conifers and eucalypts, but also in- cludessycamores, oaks, and severalexotics. The course uncommon (Brown 1978, Koenig and Pitelka is watered and mowed regularly, and is surrounded 1981, Koenig et al. 1992). Dispersalin most co- by parkland, cornfields, and residential areas. The operativebreeders is female-biased(e.g. Gaston climate is southern Californian Mediterranean, with 1978a, Mumme and de Queiroz 1985, Marzluff hot, dry summers and occasionalwinter rains be- and Balda 1988, Reyer 1990), as in most birds tween Decemberand February. (Greenwood 1980). The majority of helpers I captured173 free-flying individuals (63 males,54 acrossspecies, therefore, are sonsfrom previous females,56 unsexed)using large walk-in traps (n = broods(e.g. Rowley 1981,Austad and Rabenold 52) and a cannon net (n = 121). Trapped crows were 1986, Lennartz et al. 1987, Curry 1988, Rowley weighed,measured and marked.I took five morpho- and Russell 1990). logicalmeasurements: weight; total head length (from back of head to tip of bill = THL); culmen length; Here I describethe socialorganization of a bill depth at anterior end of nares;and tarsuslength cooperativelybreeding population of Western (TL), measured as distancebetween intertarsal joints American Crows (Corvusbrachyrhynchos hesper- bent at right angles. Each individual received two is). I found that breeding pairs regularly were identical, 3.5 x 6.5 cm patagial tags bearing two let- assistedby nonbreeding auxiliaries,as in other ters, a U.S. Fish and Wildlife leg band, and a unique cooperative breeders. However, their social combinationof colored leg bands.Nestlings (n = 97) structure deviates in several ways from that also were weighed, measuredand marked between 609 610 C^ROLEEC^FFRE¾ [Auk,Vol. 109 core areas based on their location: interior core areas • 96 o o E borderedon all sidesby golf course;peripheral areas E in which nestswere placedin treesalong the edge, 92 Oo o %0oOoø øø%%o Oo so that one side of the nest tree faced off the course 0 0000 0000 00 0 0 (e.g. a streetor the driving range). Individuals were classified as breeders, auxiliaries, o o or membersof the nonbreedingflock. Breedersen- T 84 : ß gaged in courtshipactivities, remained in closecon- •- •0 tact during the egg-layingstage, and were never as- sistedin nest building. For 31 of 59 breeding pairs 44 ' 4•6' 4•8' 5'0 observedover five years, at least one member was Tersus Length (mm) marked.Auxiliaries were nonbreedersregularly as- Fig. 1. Relationshipbetween tarsus and total head sociatedwith a breedingpair and their core area.Of length for individuals older than two months post- 45 individual auxiliaries, 28 were marked. Members fledging. Femalesand malesrepresented by solid and of the nonbreedingflock often were with eachother, open circles,respectively. and were not tied to a particular core area. Fourteen nonbreedingflock memberswere marked.Of marked nestlings,12 malesand 9 femaleseither were caught 32 and 38 days posthatching(mean nestling period after two monthspostfledging or eventuallywere ob- = 41.0 + SE of 0.9 days, n = 17; unpubl. data). served breeding. These individuals are included in I distinguishedyearling crowsfrom adultsby the the above sample sizes. following characteristics:yearlings have a dull brown I defined helpers as auxiliariesthat fed nestlings. castto their plumage in contrastto the glossyblack Determination that nonbreeders associated with of adults;yearling retricesand remigesare pointed breeding pairs were not acting as helpers was based or rounded in shape,in contrastto the more truncated on at least three 4-h periods of observationat nests feathersof adults;yearling tail shapeis squaredcom- after hatching.Individuals were classifiedas having pared to the rounded appearanceof adult tails (Emlen dispersedonce they no longer concentratedtheir ac- 1936). Marked individuals were sexedbehaviorally tivity in their natal core area. or by using a discriminant function based on mor- Survivorshipof nestlingsand juveniles was mea- phological measurements.Thirty individuals were suredat four stages:fledging, two weeks postfledg- sexed from observation of breeding behavior--fe- ing, two months postfledging,and one year. Two males performed all incubation and brooding and weeks is the approximatetime before fledglings first were fed on and off the nest by males,as in other come down to the ground successfully;before this crows (Good 1952, Goodwin 1976). A discriminant time they cannot generateenough lift to return to function was generated basedon the measurements trees.Two months was the earliest observedage of of these individuals plus those of 42 found or ob- independence.For the groupclassified as having sur- tained dead from local rehabilitation facilities and vived to one year, I included only those individuals sexedby dissection.The best discriminantfunction, alive at two months whose subsequentfates were obtainedby a stepwiseprocedure, correctly classified known. the sex of 96% of these 72 individuals based on total Survivorship,dispersal and settlementpatterns, and head and tarsuslength (Fig. 1; F = 102.74,df = 2 and sex-relatedbehavior were analyzedonly for marked 76, P < 0.0001). This discriminant function (Z = individuals.I usedANOVAs and t-teststo analyze 1.782THL - 0.503TL) was then used to sex 17 indi- relationships between dispersalbehavior and indi- viduals caught at least two months postfledging. vidual characteristics.Sex biasesin dispersaland Overlap in body sizewas too greatto distinguishthe helping were analyzedusing only yearlingsso that sexof nestlingsand young juveniles. eachindividual was included only once.All statistical I conductedapproximately 5,300 h of observation, testsused are two-tailed. In all casesthe most pow- concentrated during the breeding season (March erful test, appropriatefor samplesize, was used. For through June).I made observationsusing 10x bin- variableswhere no significantbetween-year differ- ocularsand a spottingscope, and my car as a blind. enceswere found, datawere pooledacross years. De- The roadsthroughout the courseprovided accessto scriptive statisticsare presented as means and stan- all areas.I censusedthe study site and noted the lo- dard errors. cationof individuals four to sixtimes per week during the breedingseason and one to three timesper week RESULTS at other timesof the year.Nest locationswere plotted on a map of the study site each year. I calculated SPATIAL DISTRIBUTION internestdistances using a GTCO Digitizer. Residents ranged widely both on and off the course;however, During the breeding season, crows were during the breeding seasonpairs concentratedtheir memberseither of breeding groupsor a resi- activity in a specificcore area. I defined two types of dent nonbreeding flock. July1992] CrowDispersal and Helping Behavior 611 Breedinggroups.--Breeding groups occupied core areas distributed throughout the course (Fig. 2). Breeding density on the course as a whole remainedrelatively constantover