University of Nebraska - Lincoln DigitalCommons@University of Nebraska - Lincoln

Papers in Natural Resources Natural Resources, School of

1992 ECTOPARASITISM AS A CAUSE OF NATAL DISPERSAL IN CLIFF Charles R. Brown Yale University, [email protected]

Mary Bomberger Brown Yale University, [email protected]

Follow this and additional works at: http://digitalcommons.unl.edu/natrespapers Part of the Natural Resources and Conservation Commons, Natural Resources Management and Policy Commons, and the Other Environmental Sciences Commons

Brown, Charles R. and Brown, Mary Bomberger, "ECTOPARASITISM AS A CAUSE OF NATAL DISPERSAL IN CLIFF SWALLOWS" (1992). Papers in Natural Resources. 471. http://digitalcommons.unl.edu/natrespapers/471

This Article is brought to you for free and open access by the Natural Resources, School of at DigitalCommons@University of Nebraska - Lincoln. It has been accepted for inclusion in Papers in Natural Resources by an authorized administrator of DigitalCommons@University of Nebraska - Lincoln. ATco/otv73(5). 1992. pp. 1718-1723 cI 992 by the Ecological Society of America

ECTOPARASITISM AS A CAUSE OF NATAL DISPERSAL IN CLIFF SWALLOWS'

CHARLES R. BROWN AND MARY BOMBERGER BROWN Departmentof Biology, Yale Universitl,P.O. Box 6666, New Haven, Connecticut06511 USA

Abstract. NestlingCliff Swallows (IHirundopyrrhonota) in southwestern Nebraska that were relativelyheavily parasitized by hematophagousfleas (Ceratoph yl/us celsus) and swal- low bugs () dispersed to nonnatal colonies to breed the subsequent year, whereas nestlingsthat were relativelylightly parasitized returnedto theirnatal colony to breed. There were no significantdifferences between dispersersand nondispersersin natal clutch size, natal brood size, relative hatchingdate, natal body mass, natal 's distance fromthe colony's center,and natal nest's age. There were no sex differencesin dispersal tendencies. Dispersing tended to move to smaller colonies to breed, and in some cases settledin breedingcolonies later than nondispersers.Ectoparasitism may be a major cause ofnatal dispersalin CliffSwallows and perhapsshould be consideredanother potential cause of dispersal in general,especially in group-livingspecies that may be associated with large numbers of highlyco-evolved ectoparasites. Keypwords: Ceratophylluscelsus; Cliff Swvallow; colonialist; ectoparasitism; pyrrhonota; natal dispersal,Nebraska; Oeciacusvicarius; social behavior.

INTRODUCTION status, natal territoryquality, natal clutch size, and In most populations, some individuals upon natal brood size have been examined in some species reaching maturitysettle and breed at or near their (reviewed by Pirt 1990). birthplace,whereas others disperse varyingdistances The potential influenceof ectoparasitismon natal to potentiallyunfamiliar breeding sites (reviewed in dispersal has not been examined in any species to our Greenwood and Harvey 1982, Chepko-Sade and Hal- knowledge.Increasing evidence indicatesthat parasites pin 1987, Johnsonand Gaines 1990). Natal dispersal have at times major effectson many aspects of host can both maintaingene flowbetween populations and behavior and ecology (e.g., Loye and Zuk 1991). This influencesocial and genetic structurewithin popula- may be especiallytrue for group-living species thatare tions, especially when dispersal occurs preferentially associated withlarge numbers of ectoparasites,such as among one sex or age class. Studies on natal dispersal thecolonial CliffSwallow (Hirundopvrrhonota).At our and philopatryhave most oftenmeasured whetherone study site in Nebraska, infestationsof ectoparasites sex is more likely to disperse (e.g., Greenwood et al. increase with CliffSwallow colony size and severely 1979, Greenwood 1980, Dobson 1982, Chepko-Sade depress nestling body condition and survivorship and Halpin 1987, Johnsonand Gaines 1990) and spec- (Brown and Brown 1986). In this paper we reportthat ulated as to whetherthe avoidance of eitherinbreeding natal dispersalin CliffSwallows can be predictedlarge- or increasedcompetition for resources in the natal area ly by the extentof ectoparasitismindividuals experi- has primarilycaused dispersal (e.g., Hoogland 1982, ence as nestlings;those that are relativelyheavily par- Moore and Ali 1984, Dobson and Jones 1985, Shields asitized disperse to another colony to breed the 1987, Payne 1991). subsequentyear, whereas those thatare relativelylight- Surprisinglyfew studies have investigatedpotential ly parasitizedreturn to theirnatal colony. We evaluate genotypicor phenotypicdifferences, other than sex and otherpotential phenotypicdifferences among dispers- age, that characterize dispersers and nondispersers ers and nondispersersand suggestthat ectoparasitism withina population. Studies of small rodentssuggested may be one cause of natal dispersal in CliffSwallows. thatdispersers differed genetically (with respectto two loci) fromnondispersers (Myers and Krebs 1971) and METHODS thatdispersers had reduced body mass relativeto non- Study and studysite dispersers(Fairbairn 1978). Among birds, no consis- CliffSwallows are small, migratorypasserines that tent intrasexual differencesbetween dispersers and nest in colonies throughoutmuch of western North nondispersershave been identified,although variables America. They arrivein the southernand coastal parts such as population density,time of birth,dominance of their breeding range in March and arrive in most other areas (including our study area) by early May. I Manuscript received 2 December 1991; revised 18 Feb- Most CliffSwallows leave in August ruarv 199T2;accepted 28 February 1992. and September for their winteringrange in South October 1992 ECTOPARASITISM AND DISPERSAL IN SWALLOWS 1719

America. The birds build gourd-shaped out of Marking and recapturingbirds mud pellets, and their nests are attached underneath In 1982 we began banding nestlingCliff Swallows overhangingrock ledges on the sides of steep cliffsor with Fish and Wildlife Service bands; on artificialstructures such as bridges. Some colony banding has continued to date. Each was banded sites and individual nests are used repeatedlyin suc- 10 d afterhatching and its body mass measured at that cessive years, whereas others may be abandoned for time with a Pesola scale. For each individual we also intervals of 1-6 or more years before they are used recorded natal clutch size, defined as the maximum again. CliffSwallows are highlysocial in all of their numberof eggs laid in its nest; natal brood size, defined activities; they feed, preen, gather mud, and migrate as the total number of nestlingsalive in the nest 10 d in large groups. after hatching; relative hatching date, a measure of The CliffSwallow bug (Hemiptera: ; Oeci- when within a colony's period of nesting the bird acls vicarius) is the most abundant ectoparasite of Cliff hatched, calculated as the number of standard devia- Swallows in our studyarea. Some nests contain up to tions forthat colony eitherbefore or afterthe colony's 2500 bugs each. The hematophagousbugs feed modal hatchingdate (see Brownand Brown 1987); the on both adult birds and nestlings.The bugs are pri- natal nest's lineardistance along the substratefrom the marilynest ectoparasites,apparently traveling on the colony's centermostnest, measured in centimetres;and adult birds only forbrief periods duringthe summer, the natal nest's age, old nests being those in which half and consequentlyare adapted to long periods during or more of the nest already existedwhen firstoccupied each year or for several consecutive years when their by birds,and new nests being those in which less than migratoryhosts are absent fromthe colonies (Loye and half of the nest existed when firstoccupied by birds Hopla 1983, Loye 1985). Swallow bugs are long-lived (after Brown and Brown 1986). Sex of nestlingCliff and begin breedingas soon as a colony is occupied by Swallows could not be determined. birds in the spring.The lifecycle and basic biology of Adult CliffSwallows were captured at colonies by 0. licariuisis relativelywell known(Usinger 1966, Loye mist-nettingat each site repeatedlythroughout each 1985). nesting season. A breeding colony is defined as any The othercommon CliffSwallow ectoparasitein our singlebridge or highwayculvert containing active nests; study area is the bird Ceratophvlluscensus (Si- study colonies were generallybetween 1 and 65 km phonaptera: Ceratophyllidae). are also hema- apart. The analyses reportedhere use only birds orig- tophagous but apparentlytravel and feed on the adult inally banded as nestlings(or in some cases recently birds more than do swallow bugs. Fleas reproduce in fledgedjuveniles) and recapturedat a breedingcolony swallow nests duringthe summer,spend the winterin the followingyear. All adults capturedsince 1986 have the nests,and clusterat the entrancesof nests the next been sexed by cloacal protuberance or presence/ab- spring,jumping on the CliffSwallows when birds are sence of a brood patch. investigatingnests early in the season (Hopla and Loye Nestlingsfrom three focal colony sites thatwere ac- 1983; C. and M. Brown,personal observations).Fleas tive each year of the study constitutedthe sample of appear to have small effectson nestlinggrowth rates birds used in this paper (see Results). These threecol- or nestlingsurvivorship (Brown and Brown 1986), but ony sitesvaried consistentlyin size each year:the small their short- and long-termeffects on adult birds are colony averaged 53 nests over the years of the study unknown.Fleas apparentlyare less able than swallow (range 6-140 nests) and was the smallest of these sites bugs to withstandmultiple-year periods of colony dis- each year;the medium-sizedcolony averaged 227 nests use by CliffSwallows, althoughthe lifecycle and basic (range90-375 nests)and was intermediatein size each biologyof C. censusis not well known (Hopla and Loye year; and the large colony averaged 1260 nests (range 1983). 125-2350 nests) and was the largestof these sites each Our studywas conducted in southwesternNebraska year. In addition to the threefocal colony sites,we also near the Universityof Nebraska's Cedar Point Biolog- searched for birds that had dispersed by mist-netting ical Station,in 1982-1991. CliffSwallows are abundant at up to 36 additional colony locations withinthe ap- in this area, and have probably increased in recent proximately200 x 60 km study area each year (not years with the constructionof highway culverts and all of these 36 sites necessarilywere active or could be bridges upon which they can nest. These birds also sampled in a given year). Colony size was determined occurredin southwesternNebraska beforethe appear- by counting or estimatingthe maximum number of ance of artificialstructures, nesting on bluffsand out- nests containingeggs. crops along the North Platte River. Colonies in our study area were situated on bridges, in highwaycul- verts,on irrigationstructures, and on naturalcliff sites Measuring ectoparasitism along the south shore of Lake McConaughy. Mean (+ 1 Extentof ectoparasitism was measuredfor each nest- SE) colony size in Keith, Garden, and Lincoln counties lingCliff Swallow 10 d afterhatching at the time it was was 392 ? 27 nests (N 564 colonies; range 2-3700 banded. All swallow bugs and fleas presentanywhere nests). Some CliffSwallows were also solitarynesters. on the bird's body were counted (Brown and Brown 1720 CHARLES R. BROWN AND MARY B. BROWN Ecology, Vol. 73, No. 5

1986). Nestlingswere sparsely feathered at thatage and 0 100 ectoparasites could be thoroughlysearched for and counted in 1 min/bird. Parasite counts fromnestlings reflected only the rel- 801 ative degree of among nests and did not 60 representthe nests' actual parasite loads. Actual par- LU asite loads could be determinedonly by collectinglarge numbers of entirenests, which was prohibitivelyde- 400 LL 00 80 0 60 ~~ 0-~~

40 NO. SWALLOW BUGS AS NESTLING

FIG.2. Percentof Cliff Swallows originally banded as nest- 20 lingsthat were recaptured the following year breeding in their natalcolony (open bars)or havingdispersed to a nonnatal 0- ***~~~~~~~~~~~~~r...*." colony(shaded bars) vs. theextent of swallow bug parasitism 0 1 2 3 4 theyexperienced as a nestlingin a small colony.N = 26 nondispersers,52 dispersers.The distributionsof dispersers lLC~ and nondisperserswith respect to parasitismdiffered signif- icantly(P < .0001,chi-square test), as did themean level of D loo- natalbug parasitismfor dispersers and nondispersers(P < .0001,Wilcoxon test). 0 80 structive.At least 0 60 for swallow bugs, however, counts on nestlingscorrelated significantly with a nest's total 40 " parasite load, based on a small sample of nests we collected(r, = 0.623, P < .00 1, N= 65). Nests in which 0 20 no swallow bugs were found on any of the nestlings averaged 199 ? 27.6 total bugs countedwithin the nest o 0- (mean ? 1 SE, N = 39 nests), whereas nests with at 2r 0 1 2 least one bug on at least one nestlingaveraged 565 ? 56.2 total bugs counted withinthe nest (N = 26 nests; C Wilcoxon rank sum test, P < .001). Fleas were too 100 mobile to be counted accuratelyin collected nests.

80 RESULTS

60 From 1983-1991 we recaptured 1580 birds at a breedingcolony that had been banded as nestlingsor juveniles the precedingyear; 1535 of these (97. 10%)had been banded at one of the threefocal colony sites. This 40 ------paper uses data from only the 1535 birds from the 00...... focal sites,because each was occupied by CliffSwallows in each year of the study,and thereforeall birds born there each year had the option of returningto their NO. FLEAS AS NESTLING natal colony the followingyear. FIG. 1. Percentof Cliff Swallows originally banded as nest- NestlingCliff Swallows on which one or more fleas lingsthat were recaptured the following year breeding in their (Fig. 1) or swallow bugs (Fig. 2) were counted tended natalcolony (open bars) or havingdispersed to a nonnatal to disperse to another to breed the next colony(shaded bars) vs. the extentof fleaparasitism they colony year, experiencedas a nestlingin a small (A), medium(B), and whereas those birds on which no fleas or bugs were large(C) colony.In (A), N= 26 nondispersers,52 dispersers; found tended to returnto breed at theirnatal colony in (B), N= 139 nondispersers,48 dispersers;in (C), N= 770 (Figs. 1 and 2). This patternheld for CliffSwallows nondispersers,106 dispersers.The distributionsof dispersers born at all three colonies of differentsizes when par- and nondisperserswith respect to parasitismdiffered signif- icantlyfor all colonies(P < .0001 foreach, chi-square tests), asitized by fleas (Fig. 1). Only the small colony could as did themean level of natalflea parasitism for dispersers be used in analyses for swallow bugs because the me- and nondispersers(P < .0001 foreach, Wilcoxon tests). dium and largecolonies were in part or fullyfumigated October 1992 ECTOPARASITISM AND DISPERSAL IN SWALLOWS 1721 in some years to remove bugs, but the fumigantwas onies moved on average to smaller colonies to breed ineffectiveagainst fleas (Brown and Brown 1986). (Table 1), whereas the birds dispersingfrom the small There were few other differencesthat distinguished colonytended to move to slightlylarger colonies (Table birds that dispersed vs. those that returnedto their 1). Birdsthat returned to theirnatal colony siteto breed natal colony. Natal clutch size, natal brood size, rela- occupied colonies that tended to be largerthan their tive hatchingdate, and body mass at 10 d of age did own natal colony had been at that site (Table 1). not differsignificantly for dispersers vs. nondispersers Settlementdates at breedingcolonies, definedas the (Table 1). There was a suggestionthat dispersers tended firstdate on which an adult was capturedor otherwise to come fromnests located fartherfrom the colony's known to be present(e.g., by our findingit dead on a center,although the differenceswere not statistically road), tended to be slightlylater for dispersers than for significant(Table 1). nondispersers,but the differencewas statisticallysig- There were no sex differencesin dispersal tendency; nificantonly forthe medium-sized colony (Table 1). of 865 birds recaptured and subsequently sexed as Birds from the same brood tended to exhibit the males, 158 (18.3%) dispersed to a nonnatal colony, same pattern of dispersal or philopatry.There were whereas 114 of 577 birds sexed as females (19.8%) 128 instancesin which two birds fromthe same brood dispersed to a nonnatal colony (X2 = 0.50, P = .48, df (putative siblings) were recapturedthe followingyear = 1). Natal nestage similarlyhad no effect:of 633 birds at a breedingcolony; in 126 of thesecases (98.4%) both fromold natal nests, 118 (18.6%) dispersed to a non- birdseither dispersed (3 cases) or both returnedto their natal colony, whereas 65 of 401 birds (16.2%) coming natal colony (123 cases). There were 16 instances in from new natal nests dispersed to a nonnatal colony which threeputative siblingswere recapturedthe fol- (X= l.00,P =.32,df= 1). lowing year at a breeding colony, and in all of these We calculated the differencein colony size between cases all three birds either dispersed (2 cases) or all a bird's firstbreeding colony and its natal colony for returnedto their natal colony (14 cases). Of the 304 dispersersand nondispersers(Table 1; negativevalues birds representedin these sib-sib pairs and triplets, indicate that the breedingcolony was smallerthan the only 14 (4.6%) dispersed to nonnatal colonies, a sig- natal colony, positive values that it was larger).Birds nificantlylower percentage of dispersers than in the that dispersed from the medium-sized and large col- overall sample (19.3%, N = 1535; X2 = 39.01, P <<

TABLE 1. Comparison of CliffSwallows that dispersed to nonnatal colony to breed vs. returnedto natal colony, for birds fromthree natal colony sites of differentsizes that were active each year, 1983-1991.

Small colony Medium-sized colony Large colony X SE N X SE N X SE N Natal clutch size Disperser 3.96 0.10 52 4.12 0.16 51 3.97 0.08 94 Nondisperser 3.92 0.15 24 3.93 0.06 157 3.84 0.04 721 Natal brood size Disperser 3.65 0.10 60 3.67 0.14 51 3.50 0.06 115 Nondisperser 3.77 0.17 30 3.59 0.06 160 3.34 0.03 807 Relative hatchingdate Disperser 0.065 0.20 46 0.33 0.18 48 0.67 0.20 87 Nondisperser 0.48 0.26 21 0.27 0.13 130 0.82 0.06 706 Natal body mass (g) Disperser 23.07 0.50 50 22.40 0.35 48 22.39 0.24 123 Nondisperser 23.21 0.59 26 22.40 0.24 127 22.75 0.09 829 Natal nest's distance fromcolony's center(cm) Disperser 331.4 106.3 43 490.7 39.3 49 305.2 19.3 90 Nondisperser 152.0 26.0 25 459.4 31.6 157 272.2 6.8 689 Size differencebetween breedingand natal colony (no. nests) Disperser 97.3 70.0 62 -133.6 35.6 51 -291.5 78.6 183 Nondisperser 32.0 8.0 31 77.8 9.2 164 287.7 18.8 1044 Date of settlementat breedingcolony (firstcapture date; 01 1= May) Disperser 45.4 2.0 62 45.4 3.3 49 39.8 1.0 180

Nondisperser 39.9 2.2 31 33.0 1.2 161 38.9 0.6 1018 * P < .05, *** P < .001; Wilcoxon tests with Bonferronicorrection for multiple comparisons. No other variables (except fleasand swallow bugs; Figs. 1 and 2) differedsignificantly between dispersers and nondispersers.Similar resultswere obtained with a multiple logistic regressionusing all variables in Table 1 plus fleas,bugs, sex, natal nest age, and natal colony site. Only fleas,bugs, and size differencebetween breedingand natal colony contributedsignificantly to the regression,and there were no significantinteraction terms among any of the othervariables. 1722 CHARLES R. BROWN AND MARY B. BROWN Ecology, Vol. 73, No. 5

.00 1, df = 1). This suggeststhat CliffSwallows are Furthermore,siblings tended to exhibitthe same pat- eithermore likelyto returnto theirnatal colony when ternof dispersal or philopatry,a resultconsistent with theyhave a survivingsibling, or thatdispersing siblings the notion thatindividuals respond to the level of par- were simplyharder forus to locate. asitism encounteredin theirnatal nest. Siblingsexhib- itingsimilar dispersal tendenciesalso were reportedin Great Tits (Dhondt 1979) and Sparrowhawks(,Accip- DISCUSSION iter uisus; Newton and Marquiss 1983), although the Dispersal to nonnatal colonies in CliffSwallows was possible effectsof ectoparasitesin the natal nest were stronglyassociated withnatal levels of ectoparasitism. not mentioned forthese species. Since dispersal could be predicted by extent of ecto- We can dismiss, fortwo reasons, the possibilitythat parasitism and since slightdifferences in the number birds that had been parasitized the precedingsummer of fleasand bugs counted on nestlings(that is, 0 vs. 1) wereaggressively excluded (forwhatever reasons) from seemed to have had a dramaticeffect on whetherthose theirnatal colonies by previouslyunparasitized birds birds dispersed the followingyear (Figs. 1 and 2), a and were thus "forced" to disperse. First,dates of first causal relationshipbetween natal ectoparasitismand capture at a breedingcolony differedsignificantly for dispersal is suggested.The swallow bug counts from dispersersvs. nondispersersat only one of the colony collected nests indicate that the nestlingsthat have at sites(Table 1); dispersersshould have consistentlylater least 1 bug on them and that subsequentlydisperse are settlementdates if they had firsttried and failed to coming fromnests that have, on average, >350 more settleat theirnatal colony. Second, we have never ob- bugs thando thenests from which nondispersers come. served birds tryingto exclude others from a colony We are unaware of any previous studies on other spe- site; upon arrival,Cliff Swallows interactonly withthe cies demonstratingsuch an apparenteffect of parasites relativelyfew individuals who happen to have occu- on natal dispersal. pied nests nearby(Brown and Brown 1989), and birds Ectoparasitismas a cause of natal dispersal in Cliff withoutnests are ignored. Swallows is also suggestedby the absence of differences These results (Figs. 1 and 2) suggestthat dispersal between dispersers and nondispersersin other vari- decisions in CliffSwallows may be determinedquite ables with which ectoparasitismmight have been as- early in an individual's life,that is, sometime before sociated(Table 1). Date ofbreeding, in particular,might it is 24 d old (the average fledgingage) while the bird be a potentially confounding variable, since later is still exposed to the ectoparasites in its natal nest. hatchedchicks of at least one sex tend to be more likely The level of parasitism in the natal nest may be used to disperse in Great Tits (Parus major; Dhondt and as a simple rule-of-thumbthat on average helps a bird Huble 1968) and Marsh Tits (P. pallstris;Nilsson 1989). avoid colony sites that are likely to be infestedfrom However, therewas no significantdifference in relative the previous summer. The fact that not all nestlings hatching date for dispersing vs. nondispersing Cliff dispersed the next year in our focal colonies could in- Swallows (Table 1). Because ectoparasitismwithin a dicate thatnot all nestswere infestedequally, and there colony is determinedrelative to when a colony starts is otherevidence of inter-nestvariation in parasiteload (which can be any time between late April and early withina colony (C. and M. Brown, unpublisheddata; Julyin our studyarea) and not by date per se, relative also see Figs. 1 and 2). An effectof ectoparasites on hatchingdate is a directlycomparable measure of sea- dispersal is perhaps not surprising,given the extremely sonal effectsfor Cliff Swallows of differentclasses both deleteriouseffects of parasites,especially swallow bugs, within and among colonies (see Brown and Brown on the birds' reproductivesuccess (Brown and Brown 1987). Thus, thereis littleevidence thatbirds hatched 1986) and behavior (Brown and Brown 1991). later in the year are inherentlymore likelyto disperse Although CliffSwallows assess the extent of ecto- independentof parasitism(Table 1). parasitismamong unoccupied nests upon theirreturn Nest age is anothervariable thatmay affectthe extent in earlyspring (Brown and Brown 1986), dispersalfrom of ectoparasitismand thereforepotentially confounds (and thus avoidance of) sites known to be infestedthe the presumed effectof parasites on dispersal. Earlier previous summermay save the birdscritical time dur- analyses indicated that old nests tended to have more ingthe colony selectionphase and enable themto begin ectoparasites(Brown and Brown 1986), althoughmore breeding earlier in the year. However, the delay in recentanalyses using largerdata sets indicate no con- presumably findinga nonnatal colony as reflectedin sistentdifference in parasite load betweenold and new slightlylater settlementdates of dispersers(Table 1) nests (C. and M. Brown, unpublisheddata). Natal nest may negate some of the ectoparasite-relatedbenefits age per se had no effecton dispersal tendency. of dispersingand could itselfrepresent a cost of dis- That dispersersare reactingto, and perhaps attempt- persal. Slightlylater settlement dates fordispersers may ing to avoid, parasites is suggestedby the birds' move- also reflectlonger term physiologicaleffects of being ment.in manycases, to smallercolonies to breed(Table parasitized as a nestling(see Chapman 1973) if these 1). Ectoparasitismvaries directlywith colony size, and effectssubsequently cause laterarrival on the breeding birds can presumablymore readilyescape parasites by grounds. occupying small colonies (Brown and Brown 1986). The resultsreported here suggestthat avoidance of October 1992 ECTOPARASITISM AND DISPERSAL IN SWALLOWS 1723 parasiteswithin the natal environmentmay be a cause Duffy,D. C. 1983. The ecologyof tickparasitism on densely of dispersal in CliffSwallows and perhaps should be nestingPeruvian seabirds. Ecology 64:110-119. Fairbairn, D. J. 1978. Dispersal of deer mice, Perompscus considered another potential cause of natal dispersal imaniculatus.Oecologia (Berlin) 32:171-193. in general.This may be especiallytrue for group-living Greenwood, P. J. 1980. Mating systems,philopatry and (e.g., colonial) species that are associated with large dispersal in birds and mammals. Animal Behaviour 28: numbersof ectoparasites(e.g., Hoogland and Sherman 1140-1162. 1976, Duffy1983, Shields and Crook 1987, Rubenstein Greenwood, P. J., and P. H. Harvey. 1982. The natal and breedingdispersal of birds. Annual Review of Ecology and and Hohmann 1989, Poulin 1991), some possiblyhigh- Systematics13:1-21. ly co-evolved with their hosts. The swallow bug and Greenwood, P. J., P. H. Harvey, and C. M. Perrins. 1979. flea found in our study area are specialized parasites The role of dispersal in the great tit (Parus major): the of swallows, rarelyoccurring on other birds, and ex- causes, consequences and heritabilityof natal dispersal. hibitvarious behavioral and physiologicaladaptations Journalof Animal Ecology 48:123-142. Hoogland, J. L. 1982. Prairie dogs avoid extremeinbreed- to their hosts' short and erratic annual availability ing. Science 215:1639-1641. (Usinger 1966, Hopla and Loye 1983, Loye 1985). Dis- Hoogland, J. L., and P. W. Sherman. 1976. Advantagesand persal away from them is apparentlyone major way disadvantages of Bank Swallow ( riparia) coloni- that CliffSwallows respond to these parasites. ality. Ecological Monographs 46:33-58. Hopla, C. E., and J. E. Loye. 1983. The ectoparasitesand ACKNOWLEDGMENTS microorganismsassociated withcliff swallows in west-cen- We thankCathy Boersma, Carol Brashears, Karen Brown, tral Oklahoma. I. Ticks and fleas. Bulletin of the Society Rachel Budelsky,Barbara Calnan, Sasha Carlisle,Miyoko forVector Ecology 8:111-121. Chu, Laurie Doss, KristenEdelmann, Ellen Flescher,Jerri Johnson,M. L., and M. S. Gaines. 1990. Evolution of dis- Hoskyn,Laura Jackson,Deborah Johnson, Bara MacNeill, persal: theoreticalmodels and empirical tests using birds Kathi Miller,Christine Mirzayan, Laura Monti, Barbara and mammals. Annual Review of Ecology and Systematics Raulston.Craig Richman, Martin Shaffer, Lora Sherman, Todd 21:449-480. Scarlett,Jessica Thomson, and KarenVan Blarcumfor field Loye, J. E. 1985. The life historyand ecology of the cliff assistance;the University of Nebraska-Lincoln for use ofthe swallowbug, Oeciacus vicarius(Hemiptera: Cimicidae). Ca- CedarPoint Biological Station; Steve Beissinger, Peter Clyne, hiersOffice de la RechercheScientifique et Technique Outre- StephenDobson, James Gibbs, John Hoogland,Bridget Mer, Serie Entomologie Medicale et Parasitologie 23:133- Stutchbury,Peter Walsh, Peter Waser, and Steve Zack for 139. helpfuldiscussion and commentson themanuscript; and the Loye, J. E., and C. E. Hopla. 1983. Ectoparasitesand mi- NationalScience Foundation (BSR-8600608, BSR-90 15734), croorganismsassociated withthe cliffswallow in west-cen- theNational Geographic Society (3244-85, 3996-89, 4545- tral Oklahoma. II. Life history patterns. Bulletin of the 91), theAmerican Philosophical Society, Yale and Princeton Society forVector Ecology 8:79-84. universities,the American Museum of NaturalHistory, the Loye, J. E. and M. Zuk, editors. 1991. Bird-parasiteinter- NationalAcademy of Sciences,and SigmaXi forfinancial actions: ecology, evolution and behaviour. Oxford Uni- support. versityPress, Oxford,England. Moore, J., and R. Ali. 1984. Are dispersal and inbreeding LITERATURE CITED avoidance related?Animal Behaviour 32:94-112. Brown. C. R., and M. B. Brown. 1986. Ectoparasitismas a Myers, J. H., and C. J. Krebs. 1971. Genetic, behavioral, cost of colonialityin CliffSwallows (Ilirundo pyrrhonota). and reproductiveattributes of dispersingfield voles MVicro- Ecology 67:1206-1218. tus pennsYlvanicusand Microtus ochrogaster.Ecological Brown. C. R., and M. B. Brown. 1987. Group-livingin cliff Monographs 41:53-78. swallows as an advantage in avoiding predators.Behavioral Newton, I., and M. Marquiss. 1983. Dispersal of sparrow- Ecology and Sociobiology 21:97-107. hawks between birthplaceand breeding place. Journal of Brown,C. R., and M. B. Brown. 1989. Behavioural dynam- Animal Ecology 52:463-477. ics of intraspecificbrood parasitism in colonial cliffswal- Nilsson, J.-A. 1989. Causes and consequences of natal dis- lows. Animal Behaviour 37:777-796. persal in the marsh tit,Parus palustris.Journal of Animal Brown C. R., and M. B. Brown. 1991. Selection of high- Ecology 58:619-636. quality host nests by parasitic cliffswallows. Animal Be- Part, T. 1990. Natal dispersal in the collared flycatcher: haviour 41:457-465. possible causes and reproductive consequences. Ornis Chapman, B. R. 1973. The effectsof nest ectoparasiteson Scandinavica 21:83-88. cliffswallow populations. Dissertation. Texas Tech Uni- Payne, R. B. 1991. Natal dispersal and population structure versity,Lubbock, Texas, USA. in a migratorysongbird, the Indigo Bunting.Evolution 45: Chepko-Sade, B. D., and Z. T. Halpin, editors. 1987. Mam- 49-62. malian dispersalpatterns. University of Chicago Press,Chi- Poulin, R. 1991. Group-livingand infestationby ectopar- cago, Illinois, USA. asites in . Condor 93:418-423. Dhondt. A. A. 1979. Summer dispersal and survival of Rubenstein,D. I., and M. E. Hohmann. 1989. Parasites and juvenile great tits in southernSweden. Oecologia (Berlin) social behavior of island feralhorses. Oikos 55: 312-320. 42:139-157. Shields, W. M. 1987. Dispersal and matingsystems: inves- Dhondt. A. A., and J. Huble. 1968. Fledging-dateand sex tigatingtheir causal connections.Pages 3-24 in B. D. Chep- in relation to dispersal in young great tits. Bird Study 15: ko-Sade and Z. T. Halpin, editors. Mammalian dispersal 127-134. patterns.University of Chicago Press,Chicgo, Illinois,USA. Dobson. F. S. 1982. Competition for mates and predomi- Shields, W. M., and J. R. Crook. 1987. co- nant juvenile male dispersal in mammals. Animal Behav- loniality:a net cost forgroup breedingin the Adirondacks? iour 30:1183-1192. Ecology 68:1373-1386. Dobson. F. S., and W. T. Jones. 1985. Multiple causes of Usinger,R. L. 1966. Monograph of Cimicidae (Hemiptera- dispersal. American Naturalist 126:855-858. Heteroptera).Thomas Say Foundation 6.