Latitudinal variation in clutch size lay date regressions in Tachycineta swallows: effects of food supply or demography?

Transcription

1 Ecography 37: , 2014 doi: /j x 2014 The Author. Ecography 2014 Nordic Society Oiko Subject Editor: Carten Rahbek. Accepted 1 November 2013 Latitudinal variation in clutch ize lay date regreion in Tachycineta wallow: effect of food upply or demography? David W. Winkler, Kevin M. Ringelman, Peter O. Dunn, Linda Whittingham, David J. T. Huell, Robert G. Clark, Ruell D. Dawon, L. Scott Johnon, Alexandra Roe, Suzanne H. Autin, W. Dougla Robinon, Michael P. Lombardo, Patrick A. Thorpe, Dave Shutler, Raleigh J. Roberton, Maria Stager, Marty Leonard, Andrew G. Horn, Jani Dickinon, Valentina Ferretti, Viviana Maoni, Florencia Bulit, Juan C. Reboreda, Marcela Liljethröm, Martín Quiroga, Eldar Rakhimberdiev and Daniel R. Ardia D. W. Winkler (dww4@cornell.edu), Dept of Ecology and Evolutionary Biology and Laboratory of Ornithology, Cornell Univ., Ithaca, NY 14853, USA. K. M. Ringelman, Dept of Ecology and Evolutionary Biology, Cornell Univ., Ithaca, NY 14853, USA, preent addre: Dept of Entomology and Wildlife Ecology, Univ. of Delaware, Newark, DE 19716, USA. P. O. Dunn, Dept of Biological Science, Univ. of Wiconin-Milwaukee, Milwaukee, WI 53201, USA. L. Whittingham, Dept of Biological Science, Univ. of Wiconin-Milwaukee, Milwaukee, WI, USA. D. J. T. Huell, Ontario Minitry of Natural Reource, 2140 Eat Bank Drive, Peterborough, ON K9J 7B8, Canada. R. G. Clark, Environment Canada, 115 Perimeter Road, Sakatoon, SK S7N OX4, Canada. R. D. Dawon, Ecoytem Science and Management Program, Univ. of Northern Britih Columbia, Univerity Way, Prince George, BC V2N 4Z9, Canada. L. S. Johnon, Dept of Biological Science, Towon Univ., Towon, MD 21252, USA. A. Roe, Univ. of Colorado Mueum of Natural Hitory, Boulder, CO 80309, USA. S. H. Autin, Dept of Fiherie and Wildlife, Oregon State Univ., 104 Nah Hall, Corvalli, OR 97331, USA. W. D. Robinon, Dept of Fiherie and Wildlife, Oregon State Univ., 104 Nah Hall, Corvalli, OR 97331, USA. M. P. Lombardo, Dept of Biology, Grand Valley State Univ., Allendale, MI 49401, USA. P. A. Thorpe, Dept of Biology, Grand Valley State Univ., Allendale, MI 49401, USA. D. Shutler, Dept of Biology, Acadia Univ., Wolfville, NS B4P 2R6, Canada. R. J. Roberton, Dept of Biology, Queen Univ., Kington, ON K7L 3N6, Canada. M. Stager, Dept of Ecology and Evolutionary Biology, Cornell Univ., Ithaca, NY 14853, USA, preent addre: Dept of Animal Biology, Univ. of illinoi, Urbana-Champaign, IL 61820, USA. M. Leonard, Dept of Biology, Dalhouie Univ., Halifax, NS B3H 4R2, Canada. A. G. Horn, Dept of Biology, Dalhouie Univ., Halifax, NS B3H 4R2, Canada. J. Dickinon, Cornell Laboratory of Ornithology, 159 Sapucker Wood Road, Ithaca, NY 14850, USA. V. Ferretti, Dept of Ecology and Evolutionary Biology, Cornell Univ., Ithaca, NY 14853, USA, and Fuller Evolutionary Biology Program, Cornell Lab of Ornithology, Ithaca, NY 14850, USA, preent addre: Diviión Ornitología, Mueo Argentino de Ciencia Naturale Bernardino Rivadavia, Av. Ángel Gallardo 470, Bueno Aire, C 1405 DJR, Argentina. V. Maoni, Depto de Ecología, Genética y Evolución, Facultad de Ciencia Exacta y Naturale, Univ. de Bueno Aire, Intendente Güiralde 2160, Bueno Aire, C 1428 EHA, Argentina. F. Bulit, Depto de Ecología, Genética y Evolución, Facultad de Ciencia Exacta y Naturale, Univ. de Bueno Aire, Intendente Güiralde 2160, Bueno Aire, C 1428 EHA, Argentina. J. C. Reboreda, Depto de Ecología, Genética y Evolución, Facultad de Ciencia Exacta y Naturale, Univ. de Bueno Aire, Intendente Güiralde 2160, Bueno Aire, C 1428 EHA, Argentina. M. Liljethröm, Depto de Ecología y Evolución, Facultad de Ciencia Exacta y Naturale, Univ. de Bueno Aire, Pabellón II, Ciudad Univeritaria, Bueno Aire, C 1428 EHA, Argentina, preent addre: Centro Autral de Invetigacione Científica, CADIC-CONICET, Bernardo Houay 200, Uhuaia, Tierra del Fuego, V 9410 BFD, Argentina. M. Quiroga, Aquatic Bird Lab, INALI-CONICET-UNL/UADER, Santa Fé, SF 3000, Argentina. E. Rakhimberdiev, Dept of Ecology and Evolutionary Biology, Cornell Univ., Ithaca, NY 14853, USA, and Dept of Marine Ecology, Royal Netherland Int. for Sea Reearch (NIOZ), PO Box 59, NL-1790 AB, Den Burg, Texel, the Netherland, and Dept of Vertebrate Zoology, Biological Faculty, Lomonoov Mocow State Univ., RU , Mocow, Ruia. D. R. Ardia, Dept of Biology, Franklin and Marhall College, Lancater, PA 17604, USA. In a tudy of almot net record from even wallow pecie acro the entire Wetern Hemiphere, clutch ize decline with relative laying date in each population, but the lope of thi decline grow teeper with increaing ditance from the equator. Late-laying bird at all latitude lay clutche of imilar ize, uggeting that latitudinal difference may be driven primarily by earlier-laying bird. Focued comparion of ite-year in North America with qualitatively different food availability indicate that food upply ignificantly affect mean clutch ize but not the clutch ize lay date regreion. Other tudie on the eaonality of wallow food alo indicate that teeper clutch ize lay date decline in the North are not caued by teeper earlier food peak there. The ditribution of lay date grow increaingly right-kewed with increaing latitude. Thi variation in lay-date ditribution could be due to the predominance of higher quality, early-laying (and large-clutched) individual among population at higher latitude, reulting from latitudinal variation in mortality rate and the intenity of exual election. Our reult undercore the importance of tudying clutch ize and lay date in tandem and ugget new reearch into the caue of their joint geographic variation. 670

2 Natural election act rigorouly to hape phenotypic trait cloely aociated with urvival and reproduction. Thee life hitory trait may be epecially enitive to election under changing environmental condition that directly affect population demographic. Depite the central importance of life hitory trade-off in evolutionary ecology (Stearn 1989), there ha been little reearch on the potential temporal and patial variability in thee trade-off (Ardia 2005). For many bird pecie, the only life hitory trait that i frequently quantified with certainty i clutch ize the number of egg laid in a ingle neting attempt. Clutch ize i thu the mot tudied avian life hitory trait (Ricklef 1980, Murray and Nolan 1989, Cooper et al. 2005), and latitudinal variation in avian clutch ize i one of the mot thoroughly documented example of geographic life hitory variation in any organim (Lack and Moreau 1965, Dunn et al. 2000, Martin et al. 2000). Bird breeding nearer the Equator tend to lay maller clutche than their conpecific or congener breeding at higher latitude, yet convincing general explanation for thi broad geographic pattern remain eluive. By contrat, latitudinal variation in ditribution of lay date (the date on which the firt egg of a clutch i laid) i not well documented, but it appear that timing of breeding i more tied to local environmental condition (Dunn and Winkler 1999, Huell 2003). Effort to undertand clutch ize or lay date variation individually are complicated by the fact that clutch ize i negatively correlated with lay date in many ingle-brooded bird pecie (Daan et al. 1990, Winkler and Allen 1996, Chritian et al. 2001). Thi correlation create an intereting and generative tenion: how might the inherent relationhip between clutch ize and lay date itelf affect geographic variation in average clutch ize? One of the mot broad-cale tudie to date on the relation between clutch ize and lay date wa a tudy of houe wren (Troglodyte pp.) acro the America (Young 1994). In thi tudy, Young howed that, a in many other pecie (Hochachka 1990, Chritian et al. 2001), clutch ize in houe wren decline linearly with later lay date, with the teepet decline in high latitude population and much hallower decline nearer the equator. Thi wa one of the firt demontration of continental-cale variation in the pattern of covariation of thee two key life hitory trait. Young data alo ugget a novel explanation for the geographic variation in clutch ize, namely that latitudinal difference in average clutch ize may be driven by difference in the clutch ize lay date regreion (Fig. 1). That i, if the late-eaon clutche laid by bird at different latitude are of imilar ize, then the larger average clutche of high latitude population may reult from difference only between earlier-laying bird, not in bird breeding acro the entire lay date pectrum. Data from tree wallow Tachycineta bicolor have been an important ource for patially (Dunn et al. 2000) and temporally (Dunn and Winkler 1999) aeing life hitory variation. Studie uing large ample ize have hown that tree wallow exhibit a linear decline in clutch ize with later lay date (Winkler and Allen 1995, 1996), and previou work indicate that thi clutch ize lay date regreion i remarkably contant acro northern North America (Winkler et al. 2002). A linear eaonal decline in clutch ize ha alo been noted in T. leucorrhoa (Maoni et al. 2007) in Figure 1. Latitudinal difference in the mean clutch ize might arie becaue tropical bird do not lay larger clutche early in the eaon, not becaue of any latitudinal difference in clutch ize later in the eaon (uggeted by the work of Young 1994). central Argentina. Here, we teted for latitudinal variation in the clutch ize lay date regreion throughout the pan- American ditribution of Tachycineta wallow. Tachycineta i a monophyletic genu of nine pecie endemic to the New World; we included even pecie in our analye (bicolor, thalaina, cyaneoviridi, albilinea, tolzmanni, meyeni, and leucorrhoa) repreenting each of the two or three lightly differentiated clade within the genu (Whittingham et al. 2002, Ceraale et al. 2012, Dor et al. 2012). Our tudy ite encompaed a latitudinal gradient of 120, which allowed u to tet the aforementioned hypothei that latitudinal difference in clutch ize are driven largely by difference in the clutch ize lay date regreion. Thi appear to be the larget collection of data ever aembled on the breeding biology of a ingle genu of bird acro an entire hemiphere, and the data provide an excellent opportunity to explore how geography and phylogeny affect variation in clutch ize, lay date, and the connection between them. Method We gathered data on clutche over 262 ite-year (Table 1), contributed by member of a reearch network called Golondrina de la America ( cornell.edu ). Lay date (clutch initiation date) were tandardized a the number of day from the earliet lay date in each ite-year, thu removing the effect of different breeding eaon acro the America. Swallow will lay a econd clutch of egg if the firt i detroyed, and tropical Tachycineta pecie may be double- or even triple-brooded (Stager et al. 2012); however, our analyi conidered only firt clutche of individual female for all ite-year. Length of eaon i defined a the length in day from the firt to lat clutch initiation for a given ite-year. At all our ite, the bird neting in boxe contitute well over 90% of the breeding Tachycineta in the area, and all boxe were 671

3 Table 1. Study ite and their repective propertie, arranged alphabetically by ite code. Site elevation are given in meter. Site Code Specie Lat. Long. Elev. Seaon n Fairbank, AK, USA AKS T. bicolor Beaverhill, AB, CA ALB T. bicolor Beaverhill Lake, AB, CA BBO T. bicolor Grand Bahama, BS BHM T. cyaneoviridi Hill Bank, Orange Walk, BZ BLZ T. albilinea , 2003, Chacomu, B A, AR CHA T. leucorrhoa , Carmel, CA, USA CRM T. thalaina , , Elkhorn Slough, CA, USA ELK T. bicolor George Re, MI, USA ESG T. bicolor Allendale, MI, USA GVS T. bicolor Ame, IA, USA IOW T. bicolor Ithaca, NY, USA ITH T. bicolor General Lavalle, B A, AR LAV T. leucorrhoa Long Point, ON, CA LPT T. bicolor Lee Vining, CA, USA LVC T. bicolor T. thalaina Windor, NS, CA NSC T. bicolor Yellowknife, NT, CA NWT T. bicolor Port Rowan, ON, CA ONT T. bicolor , , Corvalli, OR, USA OSU T. bicolor T. thalaina Lambayeque, Peru PER T. tolzmanni Point Reye, CA, USA PRB T. bicolor Prince George, BC, CA PRG T. bicolor , Lake Opinicon, ON, CA QUB T. bicolor , St Deni, SK, CA RGC T. bicolor , , Sacramento, CA, USA SAC T. bicolor Saticoy, CA, USA SAT T. bicolor , 1999, Sante Fe, SF, AR SFP T. leucorrhoa , Port Rowan, ON, CA (ewage lagoon) SLN T. bicolor , , Louden Co., TN, USA TEN T. bicolor Jone Beach, NY, USA TOB T. bicolor Harberton, TF, AR USH T. meyeni Redland, OR, USA WIL T. bicolor Saukville, WI, USA WIS T. bicolor Sheridan, WY, USA WYO T. bicolor

4 checked regularly, and female in them banded, o our box-baed etimate of the tart and end of the eaon, and firt or ubequent breeding attempt per female, hould repreent a accurately a poible the breeding parameter for the bird in the area. To undertand the effect of multiple cauative factor on the clutch ize lay date regreion, we conducted a mixed model analyi uing package lme4 (Bate and Sarkar 2007) in R 2.14 (R Development Core Team). In thi analyi, clutch ize (CS) wa ued a the repone variable; thu, the clutch ize lay date regreion i teted a the effect of lay date, and the effect of other potentially cauative factor on the clutch ize lay date regreion appear a a factor-by-lay date interaction. Site and Year were included a random effect. For exploration of effect of predictor on clutch ize, we ued generalized linear mixed model with Poion ditribution for count data (i.e. clutch ize) after having teted for and failing to find any evidence of overdiperion. Becaue clutch ize are count data, Poion model are more appropriate for tatitical inference than normal linear regreion ued in the pat to meaure the relationhip between lay date and clutch ize (cf. Winkler and Allen 1996, Winkler et al. 2002), though linear model can till be valuable for the viualization of overall geographic variation. In addition to lay date (lay), we included abolute latitude (degree from the equator, either north or outh lat), pecie (pp), and eaon length (Leaon) a explanatory variable, a well a the twoway interaction of thee factor with lay date. Data for elevation and ditance to nearet ocean were not ufficiently well-ditributed to reliably tet their effect. Thu, we began with the mot highly parametrized model: CS e b 1 1 b 2 lay 1 b 3 lat 1 b 4 pp 1 b 5 Leaon 1 b 6 lay 3 lat 1 b 7 lay 3 Leaon 1 b 8 lay 3 pp 1 b 9 lat 3 Leaon 1 e i and then examined the model fit to determine which term could be excluded. We identified our bet model by BIC core, and then calculated the tatitical ignificance of each parameter uing a proportional log-likelihood ANOVA. To better undertand the latitude-by-lay date interaction by empirically exploring the variation in the clutch ize lay date lope with latitude, we divided the range of latitude into four band of increaing ditance from the equator: 0 24, 24 40, 40 53, and The firt of thee wa choen to include the entire tropic, the econd (16 wide) wa choen to include the relatively large number of ite in the outhern tier of US tate (CA and TN), and the third and fourth (13 and 12 wide) were choen to plit the remaining latitude about equally, with the extreme climate of Uhuaia, Argentina, included with the northernmot ite on the other end of the globe. To evaluate difference in the relative frequency of early-laying, large-clutched bird, we calculated the kewne of relative laying date for each iteyear and then ued a generalized linear model to tet for it dependence on latitude. To tet the effect of food upply on clutch ize, lay date, and the CS lay date regreion we took advantage of pot hoc comparion of breeding parameter at one ite in Alberta where food upplie varied dramatically acro decade, and another pair of contemporaneou ite in Ontario where food availability varie dramatically acro a ditance of only a few kilometer. Reult Similar to T. bicolor and leucorrhoa, mot other Tachycineta pecie exhibited a linear eaonal decline in clutch ize, and larger clutche occurred more frequently at higher latitude (Fig. 2a). The final clutch ize lay date regreion model for fixed effect contained only tatitically ignificant term and had the following deign: CS e b 1 1 b 2 lay 1 b 3 lat 3 b 4 pp 1 b 5 lay 1 lat 1 e i Year wa not ignificant a a random effect (L 1, , p 0.63) and wa removed from the final model, but Site had a highly ignificant effect and wa retained (L 1, , p 0.001). The final model wa 6 BIC unit better than any other model, and o we report parameter etimate only for thi top model. Our final model revealed a highly ignificant effect of both lay date and latitude, and their interaction, on clutch ize (Table 2). The lope of the regreion how a pattern of decline with latitude (Fig. 2b) imilar to that reported by Young for houe wren. There are till few lope etimate available for the Southern Hemiphere, but thoe available ugget a lack of any teep lope in the outh (Fig. 2b). Neverthele, while pecie had a ignificant effect on clutch ize, the lay-by-pecie interaction term did not appear in our top model, indicating that neither pecie nor, implicitly, neither hemiphere nor phylogeny, had a ignificant effect on the lope of the clutch ize lay date regreion. Both linear regreion line (Fig. 3) and Poion prediction (Fig. 4) how a general marked increae in the teepne of the clutch ize lay date regreion lope further from the equator. Thu, latitudinal variation in the clutch ize lay date regreion may explain ome of the latitudinal variation in average clutch ize een in Tachycineta (Fig. 5). To tet whether latitudinal difference in the regreion lope might be aociated with difference in the frequencie of earlylaying, large-clutched individual, we modeled the kewne of the lay date ditribution a a function of latitude. Site at higher latitude averaged more right-kewed laydate (beta SE, p 0.01): high-latitude ite contain a ignificantly higher proportion of early-breeding individual than ite nearer the equator. Two pair of ite in our data et howed very intereting difference that may hed light on the importance of food upply in affecting the CS lay regreion. The two ite near Beaverhill Lake, Alberta (ALB and BBO) overlap patially, but they are from different period of time (Table 1), only the earlier of which wa characterized by abnormally and ditinctively high concentration of flying inect (Dunn and Hannon 1992). The bird there during the 1980 laid on average 6.35 egg per clutch and thoe preent 20 yr later averaged 5.84, a ignificant difference (t 5.13, p ) that i mirrored by a difference in mean lay date of 2.5 d (t 24.93, p ). Though the earlier period had a teeper clutch ize lay date regreion (Fig. 6), thee regreion are not ignificantly different (ANOVA 673

5 deviance 22.39, p 0.12). The other intereting contrat i between contemporary ite near Long Point, Ontario, the general ite of which (ONT) had maller clutche (5.10) than thoe at a ewage lagoon (SLN, mean CS 5.75; t 15.51, p ), where population of flying inect have been very high (Huell and Quinney 1987). Thee ite differed in mean lay date by 1.24 d (t 26.11, p ). Depite their proximity and different clutch ize and lay date, thee ite had no difference in their CS lay lope (ANOVA deviance 20.59, p 0.44). Dicuion Figure 2. (a) Mean clutch ize relative to the abolute value of latitude for Tachycineta pecie at every ite in the tudy. (b) The lope of the Poion clutch ize lay date regreion a a function of the abolute value of latitude. For both, Northern Hemiphere in blue, Southern in red. Table 2. Final generalized linear mixed model with Site a a random effect. All pecie (pp) effect are on the log cale of Poion variate, and are fit relative to a zero reference level for Bahama wallow T. cyaneoviridi. Lay date (lay) i day ince the firt clutch in each ite-year, and latitude (lat) i degree either north of outh of the equator. (CHSW: T. meyeni, MANS: T. albilinea, TRES: T. bicolor, TUSW: T. tolzmanni, VGSW: T. thalaina, WRSW: T. leucorrhoa). CS e^(b 1 b 2 lay b 3 lat b 4 pp b 5 lay lat e). Etimate SE z-value p-value (Intercept) lay date latitude ppchsw ppmans pptres pptusw ppvgsw ppwrsw lay date latitude Our analye reveal the interaction between clutch ize and lay date a an important life hitory linkage in Tachycineta wallow. Indeed, the interaction of thee two life hitory trait i trong enough that tudie of clutch ize eparated from lay date hould be purued with caution. While there are difference in mean clutch ize among pecie, thee difference are le triking than the latitudinal difference in clutch ize (Fig. 2a) and the clutch ize lay date regreion (Fig. 2b, 5). Mot importantly, Fig. 5 and Fig. 1 look remarkably imilar depite early-eaon difference, clutch ize are quite imilar at all latitude later in the eaon. Thu, throughout the entire Wetern Hemiphere, the latet-laying Tachycineta wallow lay clutche of very imilar ize regardle of latitude, uggeting the poibility that the phyiological contraint and environmental determinant facing late-laying bird may be quite imilar. However, cloer inpection of Fig. 5 in conjunction with our analyi of lay date kewne reveal coniderable complexity in the clutch ize lay date regreion. Firt, acro all latitude, the lay date ditribution are right-kewed: the mean lay date are very early relative to the length of the eaon. Thi i true even cloe to the Equator, depite the fact that the breeding eaon i much longer in the tropic. (Recall that we are comparing only firt clutche in the eaon for female acro the network.) Second, the kewed ditribution reveal that there are proportionately few late-laying bird in all population. Becaue thee later bird tend to lay imilar-ized clutche acro latitude, the latitudinal difference in clutch ize are being driven primarily by the proportion and clutch ize of earlierlaying bird. There i a great deal of theory and evidence for factor (e.g. difference in food upply, mortality rik, pace of life, etc.) that may lead temperate-neting bird to lay larger clutche than their tropical counterpart (Skutch 1949, 1985, Ahmole 1963, Ricklef 1969, Martin et al. 2000). It eem likely that thee ame factor may provide the driving force for latitudinal trend in the clutch ize lay date regreion by affecting the clutch ize of the early breeder at each ite. For the Tachycineta ytem, comparative data are jut being gathered on food upply, netling growth, incubation and proviioning behavior, chick and parental urvival, metabolic rate, etc. Although the reult of thee analye are likely to produce new inight on the CS lay regreion, the few reult already available ugget the mot promiing area for future invetigation. 674

6 AKS ALB/BBO NWT PRG OSU WIL SAC PRB CRM ELK LVC SAT WYO RGC WIS QUB LPT/ONT NSC TOB ESG GVS BHM TEN ITH Tropic of Cancer The 9 Tachycineta pecie T. bicolor T. thalaina T. euchryea T. cyaneoviridi T. tolzmanni T. albilinea T. albiventer T. leucorrhoa T. meyeni < 100 clutche clutche > 500 clutche BLZ PER LAV CHA Equator Tropic of Capricorn SFP < > USH Figure 3. Latitudinal band ued in analye, pecie range, and the linear CS lay date regreion for each ite. For each regreion the color of the axi for bicolor i black but all other are keyed to the pecie depicted; the color of the regreion line indicate the lope of the regreion; an aterik in the lower left corner indicate ignificance of the regreion; and the color of the dotted line connecting the ite label to it location indicate the ample ize. Site label and coordinate in Table 1. The mot obviou environmental factor with the potential to explain both variation in mean clutch ize and in the CS lay regreion i food upply. Depite many effort, the Golondrina network ha not yet devied reliable and effective mean to gather the large amount of comparative data on food availability needed to quantitatively tet food upply effect. Still, the comparion between ite in Alberta and Ontario ugget that, though clutch ize were higher in time or localized ite where food wa qualitatively more abundant, the CS lay lope appear to be little affected by thee large variation in available food (contra Siikamäki 1998, Both and Vier 2005 in other ytem). In wallow in Alberta and Ontario, richer food regime were aociated with larger average clutche not becaue early breeder raied their clutch ize ceiling, but becaue there were proportionally greater number of early breeder laying large clutche (Fig. 6). However, thee change in lay date kewne were not enough to generate ignificant difference in the CS lay relationhip, uggeting that local food abundance i not ufficient to explain the overarching latitudinal pattern in CS lay regreion. Of coure, early, teep temporal peak of food availability could till drive election for more bird to lay (large clutche) earlier in the north; however, limited data from everal ite acro North America ugget that food upply in northern ite i not heavily concentrated early in the eaon (Dunn et al. 2011). Indeed, it mot often increae acro the laying and brood-rearing period. Thu, it eem unlikely that the highly kewed ditribution of lay date in northern ite are due to election to capture a teep and early peak in the food upply. The proportion of early-breeding bird (and thu, the trength of latitudinal difference in CS) may be related more to the demography of the population than to any 675

7 (a) (b) (c) (d) Figure 4. Original data and prediction from the Poion CS lay date regreion for each pecie in each of the latitudinal band in Fig. 3. Note that, becaue the pecie difference in lope are not ignificant, only a ingle lope i fit in each of the band. Figure 5. Summary of the prediction from Poion regreion of clutch ize on lay date for all data in each of the latitudinal band of the ame color in Fig. 3. For each band, the mean lay date and clutch ize i alo indicated by a colored dot. pecific environmental elective force (Young 1994: 551). There i growing evidence from T. bicolor that earlierlaying bird are of higher quality than thoe laying later. Not only do early breeder lay more egg, but they have uperior flight performance (Bowlin and Winkler 2004) and immunocompetence (Haelquit et al. 2001, Ardia 2005). Furthermore, temperate Tachycineta in both hemiphere lay larger egg (Maoni et al. 2007) or egg with larger yolk (Liljethröm et al. 2012) earlier in the eaon, uggeting a eaonal decline in invetment per egg. Individual quality difference have long been known to overwhelm expected trade-off between life hitory trait (Partridge and Harvey 1988), producing ituation where, for example, bird both produce more young and have higher urvival rate (Smith 1981). Perhap temperate population generally contain a greater proportion of high-quality individual and hence more indivi dual that can breed early and lay large clutche. Adult in tropical population of Tachycineta wallow appear to have higher annual urvival rate than thoe in temperate population (Winkler unpubl.). Thu, temperate bird may be ubjected to a tronger filter for individual quality. Thi mortality-driven election for high-quality individual may be further trengthened by exual election acting on one or both exe, a temperate population generally have 676

8 (a) (b) d d (c) (d) d d Figure 6. Comparion of the CS lay date regreion for two pair of ite-record with differing food upplie. Beaverhill Lake upported enormou population of flying inect in the 1980 and normal population in the early The ite at the ewage lagoon at the bae of Long Point, Ontario, upported very high denitie of aerial inect, and other ite nearby provided normal food upplie. See text for further detail. higher level of extra-pair paternity (Ferretti 2010), and competition for early breeding may be epecially trong. (We hould alo note that, a a reult of greater adult mortality rate, temperate population alo contain more young bird. However, the teeper temperate decline of clutch ize i not caued by a higher proportion of young bird laying maller, later clutche, becaue clutch ize depend more trongly on lay date than on female age [Winkler and Allen 1996].) Undertanding geographic variation in the frequency of high-quality individual may provide inight into the overarching geographic variation in clutch ize and the clutch ize lay date regreion. How do we compare individual quality acro geography and phylogeny? Do more-variable environment in temperate region provide more opportunitie for individual of a greater diverity of qualitie? Are late-laying bird in both region of commenurate quality, a their imilar clutch ize ugget? Many bird pecie have advanced their lay date in repone to warming pring condition throughout the Northern Hemiphere (Brown et al. 1999, Dunn and Winkler 1999, 2010) and everal pecie that have not been able to do o appear to be paying fitne cot for thi lack of flexibility (Both et al. 2006, Both 2010). Our reult ugget that early-laying bird may make increaingly large contribution to population at higher latitude. Therefore, temperate bird that fail to adapt to a changing climate may pay greater cot (face tronger election) than their tropical counterpart. Clearly, effort to undertand biological linkage between clutch ize and lay date have the propect of illuminating an ecological effect on life hitory variation that may alo be of great importance in time of rapid climate change. Thu, undertanding variation in both lay-date ditribution and the linkage between lay date and clutch ize emerge a an important reearch challenge, and analye of geographic clutch ize variation baed on population mean alone will mi much of the important variation within and between wallow population. Acknowledgement We thank dozen of tudent intern with the Golondrina project for their dedicated work tudying little-known wallow, and C. Burney, C. Cooper, N. DeWitt, E. Eldermire, N. Hamm, H. Haylock, E. I. Elia, I. Lovette, J. Moh, T. Robinon, E. Romero,S. Sharbaugh, L. Sonken, D. Tzul, and P. Wrege for upport in the field and laboratory. The reearch wa upported by the following NSF grant: PIRE OISE to DWW and DRA; LTREB DEB to DWW; IOS to CV and DWW; IBN to WDR. Site-pecific funding: ite PRG NSERC of Canada, BC Knowledge Development Fund, Canada Foundation for Innovation and UNBC grant to RDD; ite CHA Ubacyt and CONICET grant to VM and JCR; ite TOB American Mueum of Natural Hitory, Rutger Univ., and Sigma Xi grant to MPL; ite ESG U. Michigan grant to MPL; ite GVS grant from Grand Valley State Univ. and the McNair Scholar Program to MPL and PAT; ite LAV Lewi and Clark Fund for Exploration and Field Reearch, Sigma Xi, AOU, Cornell Univ., Cornell Lab of Ornithology, Organization of American State (Argentina); ite LPT, ONT, and SLN Long Point Bird Obervatory (Bird Studie Canada). Reference Ardia, D. R Tree wallow trade off immune function and reproductive effort differently acro their range. Ecology 86:

9 Ahmole, N The regulation of number of tropical ocean bird. Ibi 103: Bate, D. M. and Sarkar, D Ime4: linear mixed-effect model uing S4 clae. R package ver Both, C Food availability, mitiming, and climatic change. In: Möller, A. P. et al. (ed), Effect of climate change on bird. Oxford Univ. Pre, pp Both, C. and Vier, M. E The effect of climate change on the correlation between avian life hitory trait. Global Change Biol. 11: Both, C. et al Climate change and population decline in a long-ditance migratory bird. Nature 441: Bowlin, M. S. and Winkler, D. W Natural variation in flight performance i related to timing of breeding in tree wallow (Tachycineta bicolor) in New York. Auk 121: Brown, J. L. et al Long-term trend toward earlier breeding in an American bird: a repone to global warming? Proc. Natl Acad. Sci. USA 96: Ceraale, D. C. et al Phylogeny of the Tachycineta genu of New World wallow: inight from complete mitochondrial genome. Mol. Phylogenet. Evol. 63: Chritian, J. K. et al Seaonal decline in clutch ize in European tarling: a novel randomization tet to ditinguih between the timing and quality hypothee. J. Anim. Ecol. 70: Cooper, C. B. et al Seaonal and latitudinal trend in clutch ize: thermal contraint during laying and incubation. Ecology 86: Daan, S. et al Family-planning in the ketrel (Falco tinnunculu): the ultimate control of covariation of laying date and clutch ize. Behaviour 114: 1 4. Dor, R. et al Specie tree for the tree wallow (Genu Tachycineta): an alternative phylogenetic hypothei to the mitochondrial gene tree. Mol. Phylogent. Evol. 65: Dunn, P. O. and Hannon, S. J Effect of food abundance and male parental care on reproductive ucce and monogamy in tree wallow. Auk 109: Dunn, P. O. and Winkler, D. W Climate change ha affected the breeding date of tree wallow throughout North America. Proc. R. Soc. B 266: Dunn, P. O. and Winkler, D. W Effect of climate change on timing of breeding and reproductive ucce in bird. In: Möller, A. P. et al. (ed), Effect of climate change on bird. Oxford Univ. Pre, pp Dunn, P. O. et al Geographical and ecological variation in clutch ize of tree wallow. Auk 117: Dunn, P. O. et al A tet of the mimatch hypothei: how i timing of reproduction related to food abundance in an aerial inectivore? Ecology 92: Ferretti, V Variation in extra-pair mating ytem in Tachycineta wallow: a life-hitory approach. Cornell Univ., Ithaca. Haelquit, D. et al Humoral immunocompetence correlate with date of egg-laying and reflect work load in female tree wallow. Behav. Ecol. 12: Hochachka, W. M Seaonal decline in reproductive performance of ong parrow. Ecology 71: Huell, D. J. T Climate change, pring temperature, and timing of breeding of tree wallow (Tachycineta bicolor) in outhern Ontario. Auk 120: Huell, D. J. T. and Quinney, T. E Food abundance and clutch ize of tree wallow Tachycineta bicolor. Ibi 129: Lack, D. and Moreau, R Clutch ize in tropical bird of foret and avanna. Oieau et la Revue Francaie d Ornithologie Suppl. 35: Liljethröm, M. et al Clutch invetment in the Chilean wallow (Tachycineta meyeni) hift with time of breeding and poition in the equence of laying. Condor 114: Martin, T. E. et al Parental care and clutch ize in North and South American bird. Science 287: Maoni, V. et al Breeding biology of the white-rumped wallow Tachycineta leucorrhoa in Bueno Aire Province, Argentina. Ibi 149: Murray, B. G. and Nolan, V The evolution of clutch ize. I. An equation for predicting clutch ize. Evolution 43: Partridge, L. and Harvey, P. H The ecological context of life hitory evolution. Science 241: Ricklef, R. E The neting cycle of ongbird in tropical and temperate region. Living Bird 8: Ricklef, R. E Geographical variation in clutch ize among paerine bird Ahmole hypothei. Auk 97: Siikamäki, P Limitation of reproductive ucce by food availability and breeding time in pied flycatcher. Ecology 79: Skutch, A. F Do tropical bird rear a many young a they can nourih? Ibi 91: Skutch, A. F Clutch ize, neting ucce, and predation on net of neotropical bird, reviewed. Ornithol. Monogr. 36: Smith, J. N. M Doe high fecundity reduce urvival in ong parrow? Evolution 35: Stager, M. et al Reproductive biology of a narrowly endemic Tachycineta wallow in dry, eaonal, foret in coatal Peru. Ornitol. Neotrop. 23: Stearn, S. C Trade-off in life-hitory evolution. Funct. Ecol. 3: Whittingham, L. A. et al Phylogeny of the tree wallow genu, Tachycineta (Ave: Hirundinidae), by Bayeian analyi of mitochondrial DNA equence. Mol. Phylogenet. Evol. 22: Winkler, D. W. and Allen, P. E Effect of handicapping on female condition and reproduction in tree wallow (Tachycineta bicolor). Auk 112: Winkler, D. W. and Allen, P. E The eaonal decline in tree wallow clutch ize: phyiological contraint or trategic adjutment? Ecology 77: Winkler, D. W. et al Predicting the effect of climate change on avian life-hitory trait. Proc. Natl Acad. Sci. USA 99: Young, B. E Geographic and eaonal pattern of clutchize variation in houe wren. Auk 111: