Arrival Time from Spring Migration in Male Pied Flycatchers ...

The Condor 100:702-708D The Cooper Ornithological Society 1998ARRIVAL TIME FROM SPRING MIGRATION IN MALE PIEDFLYCATCHERS: INDIVIDUAL CONSISTENCY AND FAMILIALRESEMBLANCE’JAIME POT-ITDepartamento de Biologia Animal, Universidad de Alcalri, E-28871 Alcalci de Henares (Madrid), Spain,e-mail: bnjps@bioani.alcala.esAbstract. The timing of arrival from spring migration was studied in 4 years for malePied Flycatchers (Ficedula hypoleuca) in central Spain. Heritability and repeatability analyseswere performed at several levels of familial resemblance and male age in order todetermine whether arrival times are consistent within individuals between years and aresimilar between relatives. Oldest males arrived earlier. Arrival time explained much of thevariation in laying time and number of young fledged. Age-independent variation in winglength also affected arrival date, the males with longest wings settling earlier. Arrival timeswere not repeatable within individuals across years and were not similar between relatives(parents-sons, full-sibs). Although genetic variation may exist in departure dates of longdistancemigrant birds, the close connection of an early arrival to high reproductive successmay have depleted genetic variation in arrival time. In addition, environmental variationprobably is too high to detect significant heritability in arrival times without very largesample sizes.Key words: age, Ficedula hypoleuca, heritability, migration, Pied Flycatcher, repeatability.INTRODUCTIONEach year migratory birds must find, occupy,and defend a breeding territory on their arrivalfrom winter quarters. Among passerines, a commonpattern is for males to arrive earlier thanfemales, most likely due to within-sex competitionfor access to resources vital for reproduction,and older birds arrive earlier than youngerones (Francis and Cooke 1986, Hill 1989, Lozanoet al. 1996). Early arrival at the breeding sitesincreases reproductive fitness through more nestsites (Potti and Montalvo 1991a) or better resources(Slagsvold 1986, Aebischer et al. 1996),earlier pairing (Slagsvold and Lifjeld 1988, Pottiand Montalvo 1991a) and, in promiscuous and/or facultatively polygynous species, greater opportunityto attract additional females (Alataloet al. 1981, Potti and Montalvo 1993). Early arrivalenhances male and female breeding successwhich is accentuated by a widespread tendencyfor early breeders to have a greater chance tofledge and/or recruit more offspring (Hochachka1990, Meller 1994, Aebischer et al. 1996).Although much research has been conductedon intrinsic and extrinsic factors affecting birdmigration, the evolutionary important question‘Received1998.18 August 1997. Accepted 15 April17021whether components of migratory behavior inanimals are heritable, and thus capable to respondto selection, has only recently come understudy (Berthold 1991, Dingle 1994). Experimentalwork with captive European Blackcaps (Sylviaatricapilla) has shown that some componentsof the migratory syndrome, such as migratoryrestlessness (Zugunruhe) or migratorydirection, are inherited and are probably polygenietraits with threshold effects (Berthold1991, Pulido et al. 1996).Data on individual (ontogenetic) and familialvariation in migratory behavior in wild bird populationsare scant (Rees 1989) and hard to obtain,due to high dispersal, low site fidelity, shortlife spans, and long-term work required. Here, Ireport data on one component of migratory behavior,the timing of arrival of males on thebreeding grounds, that were obtained across fouryears for a population of Pied Flycatchers (Ficedzdahypoleuca). Pied Flycatchers are small(- 12 g) long-distance migrants overwinteringsouth of the Sahara and breeding across thewestern Palearctic (Lundberg and Alatalo 1992).I investigate whether there exist individual consistencyor stability (repeatability) and familialresemblance (heritability) of male arrival time. Ifound no clear evidence that arrival date ofmales is a heritable trait in this species.

704 JAIME POTTI42681 45IMale age (years)FIGURE 1. Arrival times (means? SE) of male Pied Flycatchers in relation to age. Date 30 is April 30.Numbers above lines are sample sizes. ’vanced with increasing age (Fig. I), althoughsignificant differences were apparent only betweenyearlings and older birds (Tukey a posterioritests). Standardized arrival times wereearlier for longer-winged males (I = -0.30, IZ= 164, P < O.OOl), which might be confoundedby age due to the significantly longer wings ofolder birds (Alatalo et al. 1984a). However, arrivaltime was earliest for males with longerwings even after controlling for the effect of age(partial correlation, I = -0.21, P = 0.006). Thisindicates that the effects of increasing age andwing size on an earlier arrival date were at leastpartially independent.Being early was advantageous for male fitnessas the number of young fledged was lower formales that arrived later in the season (r = -0.38,n = 138, P < 0.001, Fig. 2). There was no costin being too early, as indicated by the nonsignificantquadratic regression term (quadratic coefficient= -0.002, P > 0.10; Fig. 2).In order to remove variation in arrival timesdue to age, I further standardized arrival timesfor familial comparisons by using the residualsof an ANOVA of standardized arrival time (seeMethods) in relation to male age (hereafter residualarrival times). Residual scores of arrivaltime were normally distributed (Kolmogorov-Smimov tests, P > 0.25) and their distributiondid not differ between young and older birds(Kolmogorov-Smirnov two-sample tests, P =0.22, n = 31 and 133 birds, respectively), norbetween presumptive fathers and sons within thesubsample of familial data (P = 0.21, n = 22families).I obtained repeated records of arrival time for39 males, 32 of which had 2 records in different,usually successive years, 4 males had 3 annualrecords, and 3 males had 4 records. The repeatability(R) of arrival times was very low and notsignificant (R = 0.04; mean number of data perindividual, no = 2.25, F38.87 = 1.10, P > 0.30).The same was true when standardized (R =0.08; Fx.8, = 1.19, P > 0.25) and residual arrivaltimes (R = 0.08; F38,87 = 1.21, P = 0.27;power > 0.10) were used. In order to removeany potential confounding effects of variation inage and arrival time among males, I selected, forfurther confirmatory analysis, repeated recordsof arrival only for males 2 or more years old.The repeatability of residual arrival time in thisdata set was again nonsignificant (R = 0.09; no= 2.18, F26,58 = 1.22, P > 0.25), indicating thatindividual adult males are not consistent in therelative timing of their arrivals on the breedinggrounds as they age.There were 32 male sons in 22 families forwhich one or more arrival times of the sons andtheir presumptive male parents were known. Althoughthe relationship was positive, the heritabilityof residual arrival times did not differ significantlyfrom zero (h* = 0.34 + 0.73, P >

HERITABILITY OF ARRIVAL DATE 7053-x2. . . .B p .. . . . . . y.. .. . .C= a, l-S.. . . . .p . . . . .CO5 -1 . I.: . . . . . .. .p -2 -.$j . . . .2 -3 -Zm5 -4 -..-20 -10 0 10 20 30Standardizedarrival dateFIGURE! 2. Relation between standardized date of arrival of male Pied Flycatchers and standardized numberof young fledged (see Methods for standardized computations). The line is the least-squares regression line(regression function: y = 0.11-0.05x).0.60, n = 22, Fig. 3). However, given the largevariances of arrival dates, the power of this testat the customary alpha level of 0.05 is very low(power = 0.07). Furthermore, that value of h* isentirely due to the “pull” of an outlier (Fig. 3);when this point is excluded a “no sense,” negativeheritability is obtained (h* = -0.50 + 1.00,P > 0.60, R = 21).There were 9 families for which the arrivaltimes of 2 or 3 (presumptive) full-sib males wereknown. The heritability of residual arrival datescalculated on the basis of (presumptive) full-sibresemblance was h* = 0.08 (no = 2.25, FS,,9 =1.09, P > 0.40).As there were more data available on layingdates, which are closely related to male arrivalI01ff = 0.34,SE = 0.73,P = 0.64 I2-I30z 0 01m 075 .! l-0 z2 I5 0-!n2oz0* ’-------------__i___________0 0q : 0-1 ’0I I,I0/-2 -1 0 1Residual arrival time of presumptive fatherFIGURE 3. Relationship between standardized arrival times of male Pied Flycatchers and their presumptivesons (see Methods for standardized computations). The line is the least-squares regression line (regressionfunction: y = 0.29 + 0.17x).

706 JAIME POTTIand pairing dates (Potti and Montalvo 1991a), Iinvestigated whether laying times for 142 maleswith 2-6 breeding records were repeatable. Therepeatability was -0.00 (F,,,,,,, = 0.99, P >0.50) after standardizing the data set for significantvariation with respect to year and femaleand male ages by using residuals of a factorialANOVA. I also correlated the average residuallaying dates in nests of sons with those of theirpresumptive male parents. The resemblance inlaying date in nests of male parents and maleoffspring was nil (r = -0.01, it = 74, P > 0.90;power = 0.05) and the same was true when thecomparison was made with average residual layingdates of their mothers (controlling for significantvariation of laying date with female age;I = -0.07, P > 0.70, IZ = 103 males in 71families; power = 0.06). Hence, there is no evidencethat arrival and laying times may be bothrepeatable or heritable in male Pied Flycatchers.However, the chances to correctly reject the nullhypothesis of no heritability of male arrival andlaying dates were very low (power = 0.05-0.12).DISCUSSIONI have presented what to my knowledge are thefirst data obtained on familial resemblance intiming of arrival to the breeding sites in a longdistancemigrant passerine. The estimates werenot confounded by the social environment (Rees1989) because Pied Flycatchers live solitarylives after they become independent of their parents,and they migrate alone (Lundberg and Alatalo1992). Using a large data set, Rees (1989)showed that arrival and departure times of winteringsocial Bewick’s Swans (Cygnus columbi-U~US [bewickii]) had a slight heritability (h* =0.10-o. 19), although in this case the estimatecould to some degree be confounded by the socialhabits of swans and their tendency to migratein groups, hence cultural, not genetic transmission.This problem did not exist in my study,and I still was unable to find evidence for geneticvariation in the arrival times from springmigration in male Pied Flycatchers.One may argue that lack of father-son resemblancein arrival dates may be due to geneticanalyses being plagued by a high frequency ofextra-pair fertilizations (EPF), which are knownto occur in the Pied Flycatcher at variable rates(Lifjeld et al. 1991, Gelter and Tegelstriim1992). No DNA fingerprinting estimates of ge-netic parentage are available for my populationand, unfortunately, the method to compute EPFrates on the basis of differences in heritabilityestimates of morphological traits using motherandfather-on-offspring regressions (Alatalo etal. 1984a) has proven unreliable to that end(Hasselquist et al. 1994, Potti and Merino 1994).However, misassigned parentage is not a problemfor repeatability estimates, which indicatedan overall inconstancy of male arrival and layingdates, and hence a likely absence of detectablegenetic variation (Boake 1989). Checking forthe existence of repeatability is a first step inproposed protocols to detect genetic variationbecause repeatability sets an upper limit to theheritability of a trait, which cannot be larger thanthe former (Falconer 1986, Boag and vanNoordwijk 1987). In addition, because geneticvariation within individuals is absent, EPFs arenot a confounding factor in the computations ofrepeatability, although they certainly are a problemfor the quantification of heritability. In spiteof this, the consistency of nonsignificant repeatabilityand heritability estimates in this study indicatesthat, whatever may be the rate of EPFsin my population, genetic variation for arrivaltime, if present, must be low.There are, however, further reasons to be cautiousabout the absence of genetic variation, becauseit may exist but be masked by genotypeenvironmentinteractions, which are likely to bepresent in such a complex “trait” as arrival date.Arrival times must depend to some degree ondistances migrated and departure dates fromwinter quarters, which in turn depend upon environmentalvariables, including photoperiodand weather conditions there. In addition, thepredictably high environmental variation met bythe birds while on passage in their migratoryjourney, and the likely important roles played inmigration performance by bird age and previousexperience with both the migration route and thearea (Montalvo and Potti 1992, Cantos and Tellerfa1994), may sum up to a high stochasticvariance in timing of arrival (Aebischer et al.1996). Body condition (Arvidsson and Neergard1991), including lack of infection by hematozoanparasites (Ratti et al. 1993), and body size(wing and first primary lengths; Alatalo et al1984b, Potti and Montalvo 1991a) also are relatedto early arrival in at least some bird species.For instance, male flycatchers with the longestwings arrived earlier independently of their

HERITABILITY OF ARRIVAL DATE 707age, which may be interpreted as evidence thata size constraint may operate on migration speedand counters the argument that late arrivals byyoung, small-winged birds are due to their relativeinexperience with the migratory journey.An analytical problem in this study was thelow power of statistical tests to detect significantheritability (Falconer 1986, Arnold 1994). Powercalculations indicate that very large samplesizes are necessary to detect a low heritabilitywith the sample variance observed. For example,to detect a low heritability of, say, 0.10 (regressionslope = 0.05) with a reasonably high(Thomas and Juanes 1996) power of 0.80 (thatis, an 80% chance of getting a significant resultgiven that the null hypothesis of no heritabilityis false), it would be necessary to have data onarrival dates for 641 families, implying about 32years of data collection at the current recruitmentrate in my population (Potti and Montalvo1991b). If heritability of arrival date was larger,say 0.50 (a common heritability for morphologicaltraits, Boag and van Noordwijk 1987), then335 father-son pairs (about 17 years of fieldwork) would have to be sampled to detect it atalpha = 0.05 and power = 0.80. Similar computationsfor detecting a low (h2 = 0.10) heritabilityof “male laying dates” result in unwieldlyfigures (> 13,000 and > 295,000 datapairs, using laying dates of mother and fatherPied Flycatchers, respectively).The fact that male arrival time in spring isclosely connected to male fitness in the Pied Flycatcher(Fig. 2) might be taken as a priori evidencein support of a predictable absence of geneticvariation in this characteristic. Geneticvariance might have been depleted by the closeassociation of early breeding to fitness due tostrong and constant selection, with all remainingvariation being enviromentally caused (“Fisher’sfundamental theorem;” Mousseau and Roff1987). The evidence presented here and backgroundinformation on this population are consistentwith this interpretation. An early date ofbreeding is the most important determinant ofrecruitment in this population (Potti and Montalvo1991b; unpubl. data), imposing a strongselection pressure to settle and breed early. Alternatively,as discussed above, low heritabilityof arrival date may be due to the predominantimportance of environmental variables affectingit (Price and Schluter 1991), i.e., arrival timesmay be subject to all the variation in the traitsthat affect it plus additional environmental variation,such that a low heritability is to be expected(Price and Schluter 1991). For instance,wing length seems to retain genetic variance inthis population (around 0.50; unpubl. data),whereas corresponding variation in arrival time,one of its correlates, is almost undetectable. Thisillustrates that the link from genes to winglength to a life-history trait such as arrival dateis long: much environmental variance is addedat each step in the path of morphological to lifehistorytraits, so that even though there may bea genetic component to arrival date via its connectionwith heritable body traits, this must bevery small (Price and Schluter 1991, Roff 1997).It is not possible to distinguish between the “exhaustionof genetic variance hypothesis” and the“environmental noise hypothesis” with the dataat hand and, most likely, both may be true.Genetic variation may underlie the proximatecontrol of departure dates in long-distance migrants,as well as in birds migrating shorter distances(Berthold 1991), but still be undetectablewith arrival times to the breeding grounds. Individualsmay not be consistent in their time ofarrival from one year to another because complexinteractions between environmental factorsaffecting arrival date and migration performance.Demonstrating among-population geneticvariation in major traits, such as c&annualrhythms (Berthold 1991, Pulido et al. 1996), canhelp to explain within- and between-speciesvariation in the urge to migrate and hence ondeparture dates to breeding or wintering areas(Ellegren 1990). However, environmental variationwill likely make it difficult to demonstratewithin-population genetic variation in arrivaltimes of migrant birds in natural settings.ACKNOWLEDGMENTSI thank the reviewers for detailed comments improvinga first draft. The study was funded by grants of theSpanish Direccidn General de Investigaci6n Cientificay Ttcnica (projects PB86-0006-C02-02 and PB91-0084-CO3-03). During preparation of the manuscript Iwas supported by project PB94-0070-C02-02.LITERATURECITEDAEBISCHER, A., N. PERRIN, M. Kamo, _I. STULIER, AND D.R. MEYER. 1996. The role of territory choice, matechoice and arrival dam on breeding success in theSavi’s Warbler Locustella luscinioides. J. Avian Biol.27:143-152.ALATALO, R.V., A. CARLSON, A. LUNDBERG, AND S.ULFsm.4m.1981. The conflict between male polygamy

708 JAIME PO-IT-1and female monogamy: the case of the Pied FlycatcherFicedzda hypoleuca. Am. Nat. 117:738-753.catchers revealed by DNA fingerprinting. Behav.Ecol. Sociobiol. 29:95-101.ALATALO, R. V., L. ~JSTAFSSON, AND A. LUNDBERG L~Z.ANO, G. A., S. PERREAULT, AND R. E. &ON. 1996.1984a. High freouencv of cuckoldrv in Pied andCollared I&atchers. dikes 42:41-4?.Age, arrival date and reproductive success of maleAmerican Redstarts Setouhapa ruticilla. J. AvianALATALO, R. V., L. GUSTAFSSON, AND A. LUNDBERG. Biol. 27:16&170. - -198413. Why do young passerines have shorterwings than older birds? Ibis 126:410-415.ARNOLD, S. J. 1994. Multivariate inheritance and evolution:a review of concepts, p. 17-48. In C. R. B.Boake [ed.], Quantitative genetic studies of behavioralevolution. Univ. Chicago Press, Chicago, IL.LUNDBERG, A., AND R. V. AL_ATALO. 1992. The Pied Flycatcher.T & A. D. Poyser, London.MILLER, A. P 1994. Phenotype-dependent arrival timeand its consequences in a migratory bird. Behav.Ecol. Sociobiol. 35:115-122.MONTALVO, S., AND J. Porrt. 1992. Breeding dispersalhvIDSSON, B. L., AND R. NFERGARLX 1991. Mate choice in Spanish Pied Flycatchers Ficedula hypoleuca. Orinthe Willow Warbler-a field experiment. Behav.Ecol. Sociobiol. 29:225-229.BEXTHOLD, l? 1991. Genetic control of migratory behaviourin birds. Trends Ecol. Evol. 6:254-257.BOAG, l? T, AND A. J. VAN N~~RDWUK. 1987. Quamitativegenetics, p. 45-78. In E Cooke and F! A.Buckley [eds.], Avian genetics. Academic Press,London.BOAKE, C. R. B. 1989. Repeatability: its role in evolutionarystudies of mating behavior. Evol. Ecol. 3:173-182.CANTOS, E J., AND J. L. TELLERS. 1994. Stopover sitefidelity of four migrant warblers in the lberian peninsula.J. Avian Biol. 25: 131-134.DINGLE, H. 1994. Genetic analyses of animal migration.p. 145-164. In C. R. B. Boake, [ed.], Quantitativegenetic studies of behavioral evolution. Univ. ChicagoPress, Chicago.DUPONT, W. D., AND W. D. PLUMMER JR. 1990. Powernis Stand. 23:491-498.MOUSSEAU, T A., AND D. A. Row. 1987. Natural selectionand the heritability of fitness components. Heredity59:181-197.PATTI, J., AND S. MERINO. 1994. Heritability estimatesand maternal effects on tarsus length in Pied Flycatchers,Ficedula hypoleuca. Gecologia 100:33 l-338.Porn, J., AND S. MEIUNO. 1995. Some male Pied FlycatchersFicedula hypoleucu in Iberia become cdllaredwith age. Ibis 137:405-409.Porn, J., AND S~MONTALVO. 1991a. Male arrival andfemale mate choice in Pied Flycatchers Ficedula hypoleucain central Spain. Grnis Stand. 22:45-54.Porn, J., AND S. MONTALVO. 1991b. Return rate, age atfirst breeding and natal dispersal of Pied FlycatchersFicedula hypoleuca in central Spain. Ardea 79:419-428.Porn J., AM) S. MONTALVO. 1993. Polygyny in Spanishand sample size calculations: a review and computer Pied Flvcatchers Ficedula hvooleuca. ,* Bird Studv . 40:program. Controlled Clinical Trials 11: 116-128.31-37:ELLEGREN, H. 1990. Timing of autumn migration inBluethroats Zawinia s. svecica depends on timing ofbreeding. Chnis Fennica 67:13-17.FALCONER D S. 1986. Introduction a la genetica cuantitativa.Cecsa Mexico D.E. Mexico.FRANCIS, C. M., AND E COOKE. 1986. Differential timingof spring migration in wood warblers. Auk 103:548-556.GELTER, H. l?, AM) H. TEGELSTR~M. 1992. High hequencyof extra-pair paternity in Swedish pied flycatchersrevealed by allozyme electrophoresis andDNA fingerprinting. Behav. Ecol. Sociobiol. 31: 1-7.GWINNER, E. 1977. Circannual rhythms in bird migration.Annu. Rev. Ecol. Syst. 8:381-405.HASSELQUIST, D., S. BENSCH, AND T VON SCHAN~Z. 1994.Estimating cuckoldry in birds: the heritability methodand DNA fingerprinting give different results. Oikos72:173-178.HILL, G. E. 1989. Late spring arrival and dull nuptialplumage: aggression avoidance by yearling males?Anim. Behav. 37665-673.HOCHACHKA, W. 1990. Seasonal decline in reproductiveperformance of Song Sparrows. Ecology 71: 1279-1288.KARLSSON L., K. F%RSSON, AND G. WALIN~ER. 1986.Ageing and sexing in Pied Flycatchers Ficedula hypoleuca.Var Fagelvaild 45:131-146.LIFJELo, J. T, T SLAGSVOLD, AND H. M. LAMPE. 1991.Low frequency of extra-pair paternity in pied fly-PRICE, T, AND D. SCHLUTER. 1991. On the low heritabilityof life-history traits. Evolution 45:853-861.Pu~mo, E, I? BERTHOLD, AND A. J. VAN NOORDWUK. 1996.Frequency of migrants and migratory activity are geneticallycorrelated in a bird population: evolutionaryimplications. Proc. Natl. Acad. Sci. 93: 14642-14647.R0r1, O., R. DUFVA, AND R. V. ALATALO. 1993. Bloodparasites and male fitness in the Pied Flycatcher.Oecologia 96:410-414.REES, E. C. 1989. Consistency in the timing of migrationfor individual Bewick’s Swans. Anim. Behav. 38:384-393.ROFF, D. A. 1997. Evolutionary quantitative genetics.Chauman and Hall, New York.SLAGSVOL~, T 1986. Nest settlement by the Pied Flycatcher.Does the female choose her mate for thequality of his house or himself! Gmis Stand. 17:210-220.SLAGSVOID, T., AND J. T L~;JELD. 1988. Plumage colourand sexual selection in the Pied Flycatcher Ficeduluhypoleuca. Anim. Behav. 36:395-407.S~ENSSON, L. 1992. Identification guide to European passerines.Svensson, Stockholm.THOMAS, L., AND E JUANES. 1996. The importance ofstatistical power analysis: an example from AnimalBehaviour. Anim. Behav. 52:856859.ZAR, J. H. 1996. Biostatistical analysis. 3rd ed. PrenticeHall, Englewood Cliffs, NJ.