SOURCES AND CONSEQUENCES OF VARIATION IN QUANTITATIVE TRAITS IN TREE SWALLOWS TACHYCINETA BICOLOR David Anthony Wiggins B.S., University of Oklahoma, 1983 M.Sc., Brock University, 1985 THESIS SUBMITTED IN CONFORMITY WITH THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in thegepartment of Biological Sciences @ David Anthony Wiggins SIMON FRASER UNIVERSITY April 1990 All rights reserved. This work may not be reproduced in whole or in part, by photocopy or other means, without permission of the author.
APPROVAL Name : Degree : DAVID ANTI-IONY WIGGINS, Doctor of Philosophy Title of Thesis: SOURCES AND CONSEQUENCES OF VARIATION IN QUANTITATIm TRAITS IN THE TREE SWALLOW TACHYCINETA BIC0IX)R Examining amnittee: Chairman: Dr. M. Mackauer, Prof-essor Dr. N.A.M. Verbeek, Professor, Senior Supervisor, Dept. &i olosical Sciences, SFU Dr. ~.c.\yde Biological S - D~.N.M. Smith, Associate Professor, Dept. of Zoology, UBC Dr. B. ~gitber~, Associate Professor, Dept. of ~iol&al. Sciences, SFU, Public Examiner Dr. D. Schluter, Assistant Professor, Dept. of Zoology, UBC, Public Examiner External Examiner Dept. of Biological iversity, Florida,
PARTIAL COPYRIGHT LICENSE I hereby grant to Slmon Fraser Unlverslty the right to lend my thesis, project or extended essay (the ;It10 sf whlch is shown below) to users of the Simon Fraser Unlversl ty LI briry, and to make partlal or single copies only for such users or in response to a request from the library of any other unlverslty, or other educational institution, on its own behalf or for one of Its users, I further agree that permission for multiple copying of this work for scholarly purposes may be granted by me or the Dean of Graduate Studies. It is understood that copying or publication of this work for flnanclal gafn shall not be allowed without my wrltten permisslon, Author: (s tgnaelroe)
Abstract Quantitative genetic theory provides a framework from which to model the evolution of quantitative traits. Estimating the heritability of quantitative traits is the initial step in understanding how, and if, such traits respond to episodes of natural selection. Unfortunately, reliable heritability estimates are difficult to obtain in natural populations, due to a lack of control over environmental variables. In this thesis, I examine the heritability and environmental sources of variation in quantitative traits in a population of Tree Swallows (Tachycineta bicolor) in southeastern British Columbia. I also apply current models of the measurement of natural selection to assess the reproductive and survival consequences of such variation. During three consecutive field seasons I performed cross-fostering, brood size, and food manipulation experiments designed to control for or induce environmental sources of variation in body size. In each of the three field seasons, I found significant heritability estimates for tarsus length among control broods. The tarsus length of cross-fostered offspring resembled that of their true parents, but not that of their foster parents. Brood size experiments were complicated by high levels of predation during the 1987 field season. Fledglings in enlarged broods were lighter in mass but had tarsus lengths similar to young in control and reduced broods. Food abundance during the inflection period of nestling growth did not appear to have \ significant effects on heritibility estimates. These experiments support the idea that heritability estimates derived from natural populations of birds are relatively free from maternal effects and genotype-environment correlations and interactions. The repeatability of egg mas; was 0.77, suggesting an additive genetic component to egg mass variation. Laying date was also highly repeatable and a preliminary heritability estimate was suggestive of a genetic basis to laying date. Selection analyses revealed significant disruptive selection on female wing length and tarsus length, associated with overwinter survival and reproductive success, respectively. Selection also favored early laying date via
enhanced female reproductive success and overwinter survival. Analyses on the association between fitness (reproductive success and overwinter survival) and male body size revealed no significant selection. The results of this thesis suggest that despite inherent methodological and statistical problems, the application of quantitative genetic models may provide valuable insights into the evolution of quantitative traits.
Acknowledgements I thank the staff and directors of the Creston Valley Wildlife Management Area for allowing me to conduct this research on Management Area lands and for countless bits of help over the years. Santo Wood, the human alarm-clock, provided a consistent source of conversation at 6 a.m. on those mornings that I had planned to sleep in. Lawrence (LW) Schalla was a 24 hour source of help in overcoming such emergencies as nest predation and loss of partners at the pool table. My fellow toilers in the Creston Valley were certainly an interesting source of distraction over the years - the tales I could tell. I would especially like to thank whichever one of you it was that I saw several times late at night during the summer of 1989, always jumping out of my truck headlights and disappearing into the woods just in time to avoid identification. Barb Beasley, Danny Fan, Sue Holroyd, Jim Karagatzides, John Keen, Liz McGowan, Joanne Siderius, and Steve Wilson all helped ebb the tide of fieldwork. I especially thank L. W. Schalla for his keen logistic sense and for all those bad videos. Of course, I can't complete my.-a- Creston Valley acknowledgements without thanking Aunt Dorothy for all those amazing pieces of apple pie B la mode. My supervisory committee, Jamie Smith, Nico Verbeek, and Ron Ydenberg provided much direction and suggestion at times of need. For help with the analyses I thank Todd Arnold, Rob Butler, Peter Grant, David Hussell, Fran James, Antero Jhinen, Dolph Schluter, Jamie Smith, Nico Verbeek and Ron Ydenberg. Financial support was provided by a Grant-in-Aid of Research from Sigma.Xi, by Simon Fraser University Graduate Fellowships, and by the Natural Sciences and Engineering Research Council of Canada (grant to Nico Verbeek). Finally, I am thankful for having Liz McGowan at my side throughout the course of this thesis work - all of your support, whether intellectual, financial, or recreational, made the completion of this degree a joy.
Table of Contents Approval Page Abstract Acknowledgements Table of Contents List of Tables List of Figures Chaptcr 1. General Introduction Literature Cited 11... lll v vi... Vlll X 1 5 Chapter 2. Heritabililty of body size in cross-fostered Tree Swallow broods Introduction Methods Results Discussion Literature Cited Chapter 3. Food availability, growth, and heritability of body size in nestling Tree Swallows Introduction Methods Results Brood sizes and grou th Heritability of body size Discussion Food abundance and growth Environmental conditions and heritability
vii Literature Cited Chapter 4. Natural selection on body size and laying date in Tree Swallows Introduction Methods Morphology Manipulations during the study Results Discussion Morphology Laying date Literature Cited Chnptcr 5. Sources of variation in the egg mass of Tree Swallows Introduction Methods Results Discussion Literature Cited Chapter 6. Clutch size, offspring quality, and parental survival in Tree Swallows Introduction Methods Results Discussion Literature Cited Chapter 7. General Conclusions Literature Cited
viii List of Tables Page Table 1. Differences in body size of female and male Tree Swallows in 1986. 9 Table 2. Parametric correlation matrix for body size traits in female and male Tree Swallows. Table 3. Repeatability of body size measurements of adult Tree Swallows. Table 4. Mid-parentfoffspring heritability estimates for body mass and tarsus length in control and cross-fostered Tree Swallow broods. Table 5. Single parent/offspring heritability estimates for body mass and tarsus length in control and cross-fostered Tree Swallow broods. Table 6. Comparison of body mass (g) and tarsus length (rnrn) of deprived, fed and control Tree Swallow broods at days 9 and 16 post-hatch. Table 7. Comparison of wing length (mm) and bill length (mm) of deprived, fed and control Tree Swallow nestlings at day 16 post-hatch. 26 Table 8. Parma! feeding mfcs (visitsh) at experimem! (N = 8) and control (N = 8) nests on days 6 and 7 post-hatch. 27 Table 9. Heritability estimates for tarsus length at days 9 and 16 and body mass at day 16 post-hatch. 29 Table 10. Heritability estimates for wing length and bill length at day 16 post-hatch. 30 Table 11. Standardized directional (S) and stabilizing (C) selection differentials associated with overwinter mortality in adult Tree Swallows. 40 Table 12. Selection on body size associated with reproductive success. 41 Table 13. Selection differentials, associated with overwinter survival, on laying date in female Tree Swallows. 44 Table 14. Selection differentials, associated with reproductive success, on laying date in female Tree Swallows. 45 - Table 15. Mean Tree Swallow egg mass (g) relative to laying sequence and clutch size. 55
Table 16. Clutch size, day 16 brood size, nestling mass, and nestling tarsus length of expanded, control, and reduced Tree Swallow broods.
List of Figures Page Figure 1. Regressions of experimental offspring (brood means) tarsus length on true and foster mid-parent tarsus length. 15 Figure 2. Growth in mass of control, deprived, and fed young from day 5 to day 9 post-hatch. 24 Figure 3. Scatterplot of relative fitness on tarsus length among female Tree Swallows in 1986. 42 Figure 4. Egg mass relative to laying sequence (A) and clutch size (B) in 170 Tree Swallow clutches. 56
CHAPTER 1 : GENERAL INTRODUCTION The question of how and when selection acts on quantitative traits has received considerable attention recently (see Dingle and Hegmann 1982; Endler 1986). Inherent in this question is the assumption that variation in quantitative traits has a considerable genetic basis (Stearns 1976). When the observed phenotypic variation has a significant additive genetic component, natural selection can cause evolutionary change in ecologicdly important waits. Many recent studies have presented evidence of genetic variation underlying intraspecific variation in quantitative traits of birds (see review in Boag and van Noordwijk 1987). However, at the same time, considerable environmental influences on such traits have also been confirmed (see references in Chapter 3). Therefore, the nature of variation in quantitative traits remains unclear and attempts to model trait evolution are consequently hindered. Originally developed for use with artificial breeding programs, quantitative genetic theory has been applied to natural populations of plants and animals over the past 20 years. Quantitative (or polygenic) traits, such as body mass or height, show continuous phenotypic variation, the result of a large number of genes acting additively. Total phenotypic variation is expressed as: Vp=Va+Vd+Vi+Ve where Vp is the total phenotypic variance, V, is the additive genetic variance, Vd is the nonadditive genetic variation due to dominance effects, Vi is the non-additive genetic variation due to epistatic interactions, and V, is the environmental variance. It is important to note that the terms on the right side of the equation are theoretical constructs and are estimated by indirect, statistical methods.
The repeatability of a trait provides an initial indication of whether further quantitative genetic analysis may be fruitful. Repeated measurements of the same individual provide an indication of the importance of measurement error or changes in the trait between measurement periods. ANOVA partitions the observed phenotypic variance into between and within-individual components, which provide an estimate of the intraclass correlation coefficient (= repeatability; Sokal and Rohlf 1981; Lessells and Boag 1987). If repeatability values are low with environmental sources of variation accounting for most of the observed variation, then further genetic analyses are not likely to be productive. The heritability (h2) of a trait refers to the proportion of total phenotypic variance that is due to the effects of additive genes (i.e., VJV,). Heritability estimates are based on phenotypic resemblances between relatives and range theoretically from 0 to 1, indicating large environmental and additive genetic variance components, respectively. Several methods can be used to estimate heritability (Falconer 1981). The most common analyses utilize regressions of offspring on mid-parents or on single parents. Mid-parent analysis provides the best estimate since it is free of the potential effects of assortative mating and the increased error associated with single-parent methods. One reason for estimating heritabilities is to assess the potential for natural selection to alter phenotypic variation. When a large additive genetic component underlies phenotypic variation, natural selection can rapidly alter trait distributions both within and between generations. One of the central questions being addressed in the study of quantitative traits is how trait variation affects fitness (Endler 1986). Answers to this question can provide valuable information about adaptation, speciation, sexual dimorphism, and geographic variation in quantitative traits. Recent theoretical advances have provided a methodology for measuring the magnitude and form of natural selection on quantitative traits (Lande and Arnold 1983; Arnold and Wade 1984a,b; Schluter 1988). When combined with data on inheritance, such
measures may provide insights into the adaptive significance and evolution of quantitative traits in natural populations (Price and Boag 1987). In addition to examining the link between natural selection and genetic variation, quantitative genetic studies also address a number of other evolutionary questions. Estimation of heritabilities provides information on the potential rate of change of the traits under study. In addition, although practically difficult, experiments on the effects of variable environments on trait development and variation can be performed in natural populations. Finally, long-term studies provide insights to the process of evolution. Only rccently, for example, have we begun to assess the frequency and long-term effects of natural selection on quantitative traits in natural populations. One of the difficulties in applying quantitative genetic techniques to natural populations is in designing an appropriate experimental protocol. Environmental sources of variation are easily controlled for in laboratories but such studies are not likely reflective of the patterns of inheritance under natural conditions. Several sources of environmental variance, including maternal effects, genotype-environment interactions and correlations, and variance in parental and habitat quality can inflate or decrease estimates of additive genetic variance. In this thesis, I explore the nature, and to a lesser extent, the consequences of variation underlying several quantitative traits in Tree Swallows (Tachvcineta bicolor). Due to their relatively rapid and determinate growth, birds are particularly good organisms for such studies. I chose to study Tree Swallows largely because they occupy artificial nest boxes, which allowed me to collect data on a large number of breeding pairs and also simplified the experimental transfer of eggs and nestlings between nests. I studied body size, clutch initiation ("laying") date, and egg mass, and while none of these traits is likely a major
contributor to lifetime reproductive success, they have been shown to be important in determining seasonal reproductive success and survival in many species of birds. Although avian body size has been measured for a number of reasons (James 1982), most recent studies, including this thesis, have concentrated on linking within-population variation in body size with natural selection (Boag and van Noordwijk 1987). Such studies have the potential to reveal many of the forces that shape the evolution of body size in natural populations (e.g. Price 1984a, b). In addition, the measurement of selection on particular aspects of body size may lead to the identification of the ecological importance of such traits. Unfortunately, body size is a composite "character" and is sometimes difficult to characterize. Among birds, however, several external morphological characters correlate well with multivariate (e.g., principal components) measures (Rising and Somers 1989). Consequently, I used four external measurements (body mass, tarsus, wing and bill length) as indicators of body size in Tree Swallows. The major goal in Chapters 2 and 3 is to identify the sources of variation in body size. Cross-fostering (Chapter 2) and food deprivation (Chapter 3) experiments were designed to reduce and enhance environmental sources of variation in nestling growth and final size. Chapter 4 considers the consequences of variation in body size and laying date by measuring selection during periods of reproduction and overwinter mortality. The sources of variation in egg mass are investigated in Chapter 5. Unfortunately, time constraints did not allow me to assess the consequences of intraclutch variation in egg mass. In Chapter 6, I consider how nestling body size at fledging (i.e., "final size") may be affected by parental foraging and reproductive (clutch size) decisions. Finally, I conclude in Chapter 7 by consolidating the results of Chapters 2-6 and relate these findings to the emerging idea that the evolution of quantitative traits can be traced accurately in natural populations.
Literature Cited Arnold, S. J. and M. J. Wade. 1984a. On the measurement of natural and sexual selection: Theory. Evolution 38:709-719. Arnold, S. J. and M. J. Wade. 1984b. On the measurement of natural and sexual selection: Applications. Evolution 38:720-734. Boag, P. T. and A. J. van Noordwijk. 1987. Quantitative genetics. pp. 45-78. In F. Cooke and P. A. Buckley (eds.), Avian genetics. Academic Press, London. Dingle, H. and J. P. Hegmann. (eds.). 1982. Evolution and the genetics of life histories. Springer-Verlag, Berlin. Endler, J. A. 1986. Natural selection in the wild. Princeton University Press, Princeton. Falconer, D. S. 1981. Introduction to quantitative genetics. 2nd. ed. Longman, London. James, F. C. 1982. The ecological morphology of birds: a review. Ann. Zoo. Fennici 19:265-275. Lande, R. and S. J. Arnold. 1983. The measurement of selection on correlated characters. Evolution 37:1210-1226. Lessells, C. M. and P. T. Boag. 1987. Unrepeatable repeatabilities: a common mistake. Auk 104:116-121. Price, T. D. 1984a. Sexual selection on body size, territory and plumage variables in a population of Darwin's Finches. Evolution 38:327-341. Price, T. D. 1984b. Life history traits and natural selection for small body size in a population of Darwin's Finches. Evolution 38:483-494. Price, T. D. and P. T. Boag. 1987. Selection in natural populations of birds. pp. 257-287. h F. Cooke and P. A. Buckley (eds.), Avian genetics. Academic Press, London. Rising, J. D. and K. M. Somers. 1989. The measurement of overall body size in birds. Auk 106:666-674.
Schluter, D. 1988. Estimating the form of natural selection on a quanititative trait. Evolution 42:849-861. Sokal, R. R. and F. J. Rohlf. 1981. Biometry. 2nd. ed. W. H. Freeman and Co., San Francisco. Steams, S.C. 1976. Life history tactics: A review of the ideas. Q. Rev. Biol. 51:3-47.
CHAPTER 2: HERITABILITY OF BODY SIZE IN CROSS-FOSTERED TREE SWALLOW BROODS Introduction Studies of the heritability of body size have facilitated investigations of natural selection in natural populations of birds (review in Price and Boag 1987). Many studies have documented a genetic basis to the resemblance in morphology between offspring and parents (summary in Boag and van Noordwijk 1987). In addition, several recent experiments have shown effects of the rearing environment on the size and shape attained by Red-winged Blackbirds (Agelaiua ghoeniciug; James 1983; James and NeSmith 1988), Tree Swallows (Tachvcineta bicolor; Quinney et al. 1986), and Zebra Finches (Poephila -; Boag 1987). Thus, it is clear that both genetic and environmental effects play significant roles in the determination of body size in nestling birds. However, environmental effects, such as variation in parental quality and maternal effects, may also influence the estimation of parent-offspring resemblance. One way to control for such effects is to cross-foster offspring among broods and to compare the size of offspring to both their true and foster parents (Smith and Dhondt 1980; Dhondt 1982; Alatalo and Lundberg 1986). I report here the results of a cross-fostering experiment in a population of Tree Swallows. Methods Data were collected from late April until July in 1986 and in 1987 at the Creston Valley Wildlife Management Area in southeastern British Columbia (49 05'N, 11635'W). Tree Swallows begin arriving in the Creston area in late March and begin laying eggs in early May. Nest boxes were erected in early March 1986 at approximately 20m intervals around
a large marsh. Nest contents were checked each morning (0800-1200) to determine laying dates and to number individual eggs. To standardize the measurements of adult body mass, I captured females on the nest while they incubated eggs, 6-13 days following clutch completion. Males were captured while feeding young, 3-10 days after hatching began. Measurements taken on all birds included: body mass (with electronic balance to the nearest 0.02g); tarsus length (with dividers to the nearest 0.lmm) from the tibiotarsal joint to the furthest distal, undivided scute; wing chord (adults) or length of ninth primary (nestlings; with ruler to the nearest mm); and bill length (with dividers to the nearest 0.lmm) from the anterior margin of the nares to the tip of the upper mandible. Nestlings were measured at 16 days of age, about 6 days before fledging. At this age, body mass, tarsus, bill and wing lengths have reached approximately loo%, 97%, 92%, and 65% of adult size, respectively. Twenty-eight experimental clutches were chosen opportunistically and were matched for size and laying date. Eggs from a "donor" nest, located off the study area, were placed in experimental nests during the clutch transfers. Control clutches (N = 18) were removed from nests for 5 minutes and then replaced. Predation of females and young lowered the sample size of experimental clutches from 28 to 15 nests. Sexual dimorphism can complicate the analysis of heritability since it may directly affect the estimate and the variation. Dimorphism was found in wing length and mass (Table 1). However, differences in mass may simply be a seasonal effect as females were weighed during incubation and males during chick feeding. Sex differences in wing length were not corrected for as offspring could not be sexed at day 16. Assortative mating can influence the heritability based upon single-parent/offspring estimates (Reeve 1961). Consequently, I correlated the four body size traits within 38 pairs of swallows for evidence of assortative mating (Table 2). Parametric correlations were weak, ranging from 0.1-0.24 and none.
Table 1. Differences in body size of male and female Tree Swallows in 1986. Units of measurement are in grams (mass), centimeters (wing), and millimeters (tarsus and bill). Females Males N Man S.D. C.V. N Mean S.D. C.V. t Mass 52 22.33 1.67 7% 32 20.12 1.38 7% 6.3" Tarsus 58 11.1 0.5 5% 38 11.2 0.6 6% 0.3 Wing 58 11.7 0.3 3% 38 12.10.3 2% 7.1* Bill 58 5.5 0.3 5% 38 5.6 0.4 7% 1.5 * P < 0.001, t-test.
Table 2. Parametric correlation matrix for body size traits in male and female Tree Swallows. Bold numbers along the diagonal refer to correlations between pairs members (N = 38). Numbers above and below the diagonal refer to correlations within males and females, respectively. Sample sizes are for correlations within sexes in 1986. Mass Tarsus Wing Bill N Mass.1Y.16.63**.03 32 Tarsus.32*.24.35*.03 38 Wing.06.12.13.18 38 Bill.ll.04.26*.10 N 52 58 58
were significant at the 0.05 level. Within-sex correlations revealed two significant correlations among males (wing length and body mass, wing length and tarsus length) and females (body mass and tarsus length, wing length and bill length; Table 2). Repeatabilities of adult measurements were estimated using analysis of variance (Lessells and Boag 1987). Measurements of adults captured more than once in 1986 and of adults captured in both 1986 and 1987 were used in the analyses. Repeatabilites of all traits were signficant and ranged from 0.28 to 0.93 (Table 3). I used ANOVA to assess phenotypic variation in body mass, tarsus length, and bill length among foster, true, and control broods. ANOVA was not carried out on nestling wing lengths as the measurement of wing length was different for adults (wing chord) and nestlings (#9 primary), thus precluding the estimation of heritability. Variation among broods encourages calculation of the heritability of traits. Significant among-brood variation was found for mass and tarsus length but not for bill length. Heritability estimates were obtained by regressing offspring values (brood 2 mean) on female and male parents separately (h = twice the slope) and on mid-parent 2 values (h = slope). Falconer's (1981 pp. 167-168) technique was used to assign weights to families of different size. Results There were no significant resemblances between parents and true or foster offspring in body mass (Tables 4 and 5). Scatterplots of true and foster offspring tarsus lengths on mid- parent values are presented in Figure 1. Significant resemblances were found between true mid-parents and their offspring in both control and experimental treatments (Table 4). However, foster offspring did not resemble their fostpr mid-parents. Single parentloffspring regressions yielded significant estimates for tarsus length between true parents and offspring, but not for foster parents (Table 5). Although there was a weak resemblance in tarsus length between offspring and their foster mother (Table 9, maternal
Table 3. Repeatability of body size measurements of adult Tree Swallows. Sexes were pooled (N = 46 individuals). Data are from birds captured in 1986 and 1987. Character Repeatability F-ratio Mass Tarsus Wing Bill
Table 4. Mid-parent/offspring heritability estimates for body mass and tarsus length in control and cross-fostered Tree Swallow broods. Sample sizes are for mass and tarsus length, respectively. -- Experimental Families Foster midparent/offspring True midparent/offspring Control Families True mid-parent/offspring S.E. S.E. S.E. Mass 0.01 0.01 Tarsus 0.16 0.16
KJ 3 2 a I- Cr,.- C k. V) w- 5 True Mid-parent Tarsus Length (mm) "'t.. 0.. 0 10.6 11.4 12.2 Foster Mid-parent Tarsus Length (mm) Figure 1. Regressions of experimental offspring (brood means) tarsus length on true and foster mid-parent tmus length. Sample sizes are 13 and 15 broods, respectively.
effects appeared to be minimal as the mother/offspring heritability estimates were similar to those for fathedoffspring. In addition, the correlation between female size (tarsus length) and egg size was low and not significant (r = 0.17, N = 57, P > 0.05). Discussion None of the heritability estimates for body mass were significantly different from zero. Most estimates of the heritability of body mass in other species have been comparatively high (Prince et al. 1970; Grant 1981,1983; Lessells 1982; Moss and Watson 1982; Boag 1983; van Noordwijk 1984; but see Smith and Zach 1979). Several factors may have contributed to these low estimates. First, the repeatability of Tree Swallow body mass was relatively low. Second, variation within broods was high as Tree Swallows often hatch asynchronously (Zach 1982; unpubl. data). Finally, I measured the mass of young at day 16. Tree Swallow young often go through a period of mass recession late in the nestling period and measurement at day 16 may not have provided a reliable estimate of adult body mass. The results show that offspring tarsus length resembles that of true parents and not foster parents. Thus, much of the observed resemblance between parents and offspring in tarsus length is due to the effects of shared genes. Similar results have been found in several other experimental studies on wild populations of birds (Smith and Dhondt 1980; Dhondt 1982; Alatalo and Lundberg 1986; Alatalo and Gustafsson 1988). Although all of the body size correlations between mates were positive, none were significant and thus assortative mating likely had a negligible effect on the heritability estimate. In addition, maternal effects did not play a role as mother-offspring estimates were either similar to or less than fatheroffspring estimates. Although nearly half of the variation in tarsus length was attributable to unknown environmental factors, my results indicate that this population would be responsive to natural selection for tarsus length.
Literature Cited Alatalo, R. V.and L. Gustafsson. 1988. Genetic component of morphological differentiation in Coal Tits under competitive release. Evolution 42:200-203. Alatalo, R. V. and A. Lundberg. 1986. Heritability and selection on tarsus length in the Pied Flycatcher (Fidecula hypoleucd. Evolution 40574-583. Boag, P. T. 1983. The heritability of external morphology in Darwin's Ground Finches (Geospizq) on Isla Daphne Major, Galapagos. Evolution 37:877-894. Boag, P. T. 1987. Effects of nestling diet protein on growth and adult size of Zebra Finches (Poephila m). Auk 104: 155-166. Boag, P. T., and A. J. van Noordwijk. 1987. Quantitative genetics. pp. 45-78. In P. A. Buckley, and F. Cooke (eds.), Avian Genetics. Academic Press, London. Dhondt, A. A. 1982. Heritability of Blue Tit tarsus length from normal and cross- fostered broods. Evolution 36:4 18-4 19. Falconer, D. S. 198 1. Introduction to Quantitative Genetics. 2nd. ed. Longman, London. Grant, P. R. 1981. Patterns of growth in Darwin's Finches. Proc. R. Soc. Lond. B 212:403-432. Grant, P. R. 1983. Inheritance of size and shape in a population of Darwin's Finches, GeosDiza gonirostris. Proc. R. Soc. Lond. B 220:219-236. James, F. C. 1983. Environmental component of morphological differentiation in birds. Science 22 1 : 184-186. James, F. C., and C. NeSmith. 1988. Nongenetic effects in geographical differences among nestling populations of Red-winged Blackbirds. pp. 1424-1433. In H. Ouellet (ed.), Acta, XIX Congressus Internationalis Ornithologici, Vol. 11. XIX International
Ornithological Congress, Ottawa, Canada. Lessells, C. M. 1982. Some causes and consequences of family size in the Canada Goose Branta ganadensi~. D. Phil. Thesis, Oxford. Lessells, C. M. and P. T. Boag. 1987. Unrepeatable repeatabilities: a common mistake. Auk 104:116-121. Moss, R. and A. Watson. 1982. Heritability of egg size, hatch weight, body weight, and viability in Red Grouse (Lagopus 1-s scoticus). Auk 99:683-686. Price, T. D., and P. T. Boag. 1987. Selection in natural populations of birds. pp. 257-287. Jg P. A. Buckley and F. Cooke (eds.). Avian Genetics. Academic Press, London. Prince, H. H., P. B. Siegel, and G. W. Cornwell. 1970. Inheritance of egg production and juvenile growth in Mallards. Auk 87:342-352. Quinney, T. E., D. J. T. Hussell, and C. D. Ankney. 1986. Sources of variation in growth of Tree Swallows. Auk 103:389-400. Reeve, E. C. R. 1961. A note on random mating in progeny tests. Genet. Res. 2:195-203. Smith, J. N. M. and A. A. Dhondt. 1980. Experimental confirmation of heritable morphological variation in a natural population of Song Sparrows. Evolution 34: 1155-1 158. Smith, J. N. M. and R. Zach. 1979. Heritability of some morphological characters in a Song Sparrow population. Evolution 33:460-467. van Noordwijk, A. J. 1984. Quantitative genetics in natural populations of birds illustrated with examples from the Great Tit, Parus major. pp. 67-79. In K. Wohrmann and V. Loeschcke (eds.). Population Biology and Evolution. Springer-Verlag, Berlin. Zach, R. 1982. Hatching asynchrony, egg size, growth and fledging in Tree
Swallows. Auk 99:695-700.
CHAPTER 3: FOOD AVAILABILITY, GROWTH, AND HERITABILITY OF BODY SIZE IN NESTLING TREE SWALLOWS Introduction Nestling birds are often subject to widely fluctuating environmental conditions during a short period of rapid growth. Field studies of growth in natural populations of small passerine birds have typically shown that feeding conditions during the growth stage have a significant impact on the final size (body mass and/or tarsus length) attained by nestlings (Ricklefs and Peters 1981; Quinney et al. 1986; van Noordwijk et al. 1988). In addition, a laboratory study by Boag (1987) found that the quality of nestling diet had a significant effect on the final size attained by Zebra Finches (Pmphila guttat&. However, studies of growth in domestic chickens have shown that under laboratory conditions, birds may compensate for periods of undernutrition by accelerating growth at a later stage ("compensatory" or "targeted" growth; Wilson and Osbourn 1960, Plavnik and Hurwitz 1985). Berthold (1976, 1977) showed that nestling Blackcaps (S~lviavica~illa) and Blackbirds (Turdus merula) raised under poor nutrient conditions had wing growth similar to that of nestlings fed on a high protein diet. Further, Smith and Arcese (1988) found that relatively poor early growth in Song Sparrow (Melospiza melodia) nestlings had no effect on their size following independence. Thus, the effects of fluctuations in the amount and quality of food on the final size attained by nestling birds are not well understood. A strong influence of environmental conditions on final size should greatly complicate the detection of heritable variation underlying body size. Nonetheless, many studies have now documented heritability underlying morphological variation in adult and nestling birds (review in Boag and van Noordwijk 1987). In this study, I investigated the effects of food
2 1 deprivation and supplementation on the growth and final size attained by nestling Tree Swallows (Tachvcineu bicolor). The following two questions were addressed: 1) does food deprivation or supplementation during the inflection period of the growth trajectory have significant effects on the size attained by nestling birds at fledging?; and 2) is heritable variation detectable within broods of deprived or supplemented nestlings? Methods The experiment was conducted in 1988 at the Creston Valley Wildlife Management Area in southeastern British Columbia where Tree Swallows nest in boxes located around a large marsh. Tree Swallows began laying eggs in early May and continued to initiate clutches until early July. However, only those pairs that initiated clutches between early May and early June were used in this study. Nest contents were checked daily and each egg was individually numbered. Females were caught on the nest while incubating, 6-12 days following clutch completion. Females were individually marked with colored, plastic leg bands and with U.S.F.&W.S. aluminum bands. Measurements taken on all females and nestlings included: 1) tarsus length (with calipers to the nearest O.lmm) from the tibiotarsal joint to the most distal, undivided scute; 2) body mass (with electronic balance to the nearest 0.02g); 3) wing chord (with ruler to the nearest mm) from the bend of the wing to the tip of the ninth primary; and 4) bill length (with dividers to the nearest 0.lmm) from the anterior margin of the nares to the tip of the bill. Both experimental and control nests were chosen opportunistically according to clutch initiation date. Nests were checked 2-3 times/day during hatching. Upon hatching, nestlings were individually marked with felt-tip pens (until day 3) and subsequently with colored plastic leg bands (from day 3 to day 16). Brood size ranged from 4 to 7 in experimental and control nests. One (brood size 4) or two (brood size 5-7) nestlings were designated "deprived" in each of 20 experimental nests. Their unmanipulated nestmates are
hereafter referred to as "fed" nestlings. Deprived nestlings were chosen from those chicks near the middle of the weight hierarchy of each brood at day 5 post-hatch. From days 5-8, deprived nestlings were removed from their nests between 1530 and 1600h and kept indoors under a heatlamp until 1930-2000h when they were returned to their nests. The day 5-8 period of the growth stage corresponds to the inflection point of both mass and tarsus growth in Tree Swallows (Zach and Mayoh 1982; unpubl. data). Nestlings in the 16 control nests were not manipulated. The body mass of both control and experimental nestlings was recorded between 1400 and 1600h from days 5-9 and on day 16 post-hatch. Tarsus length was measured on days 9 and 16, and wing and bill length were measured on day 16 post-hatch. Nestlings were not handled after day 16 as young often fledge prematurely if disturbed beyond this age. Differences between deprived, fed, and control groups were tested by comparing brood means. Paired t-tests were used to compare the means of fed and deprived nestmates. Unequal sample sizes for heritability of body mass and tarsus length at day 16 were a result of including body mass only for those females captured 6-12 days after clutch completion. While it is preferable to regress offspring values on mid-parent values for heritability estimates, I captured few males and, consequently, regressed offspring means on maternal values. As a measure of potential parental "compensation" at experimental nests, the number of parental visits was monitored at eight experimental and eight control nests on days 6 and 7 post-hatch. Observations were made in the early morning (0600-1000) when chick begging rates were expected to be highest. Observations were made from a stationary position, typically near cover, 20-30m from the nest box. Results Brood sizes and crrowth
The experiment did not have serious effects on the survival of young to day 16. The brood size of experimental and control nests was not significantly different at day 5 (experimental mean = 5.2, S.D. = 0.9; control mean = 5.2, S.D. = 0.8; t = 0.14, d.f. = 34, P > 0.05) or at day 16 (experimental mean = 4.9, S.D. = 0.9; control mean= 5.0, S.D. = 0.7; t = 0.38, d.f. = 33, P > 0.05). Reductions in brood size between days 5 and 16 were the result of the disappearance of ten nestlings and the apparent starvation of three nestlings (including one deprived nestling). The experiment had significant effects on the growth of mass and tarsus length. Body mass growth trajectories for experimental and control nestlings are in Figure 2. Prior to experimentation, at 5 days post-hatch, deprived chicks were significantly heavier than their fed nestmates (Fig. 2; paired-t = 4.83, P c 0.001). During days 6-8 post-hatch, there were no significant differences in mass between deprived and fed nestlings (all P's > 0.05). However, after the 4-day experimental period, there was a reversal in the mass hierarchy and deprived chicks were significantly lighter than both their fed nestrnates and control broods (Table 6). At day 16 there was no significant difference in the mass of deprived, fed and control chicks. Differences among experimental and control nestlings in tarsus length were similar to those for mass. At day 9, deprived nestlings had significantly shorter tarsi than both their fed nestmates and control broods (Table 6). However, by day 16 there was no significant difference in the tarsus length of the three groups. In addition, there were no significant differences between the means of fed, deprived, or control nestlings for wing length or for bill length at day 16 (Table 7). Parental feeding rates at experimental and control nests during days 6 and 7 are presented in Table 8. There was no significant difference in the rate at which control and experimental parents fed their chicks during the 2 days.
5 6 7 8 9 Day Post Hatch Figurc 2. Growth in mass of convol (m), deprived @), and fcd (0) young from day 5 to day 9 post-hatch. Points rcprcscnt mans + 1 S.E.
Table 6. Comparison of body mass (g) and tarsus length (rnm) of deprived, fed and control broods at days 9 and 16 post-hatch. Body mass Tarsus length Day 9 Da.y 16 Day 9 Day 16 N Mean S.E. N Mean S.E. N Mean S.E. N Mean S.E. Fed 20 16.96 0.33 19 21.42 0.29 20 10.4 0.1 19 11.0 0.1 Control 16 17.32 0.30 16 21.68 0.32 16 10.6 0.1 16 11.1 0.1 Except where noted, there were no significant differences between brood means. Deprived 20 16.25~ 0.28 b 19 21,09 0.43 20 10.3 0.1 19 10.9 0.1 a Deprived < Fed, paired t = 2.95, P < 0.01; Deprived <. Control, F = 5.96, P < 0.02. b Deprived < Fed, paired t = 2.37, P < 0.05; Deprived < Control, F = 5.38, P < 0.05.
Table 7. Comparison of wing length (rnrn) and bill length (mm) of deprived, fed, and control Tree Swallow nestlings at day 16 post-hatch. There were no significant differences between groups for either trait (ANOVA'S, P's > 0.05). Wing length Bill length N Mean S.E. N Mean S.E. Deprived 19 68.9 3.9 19 5.2 0.2 Fed 19 69.6 3.6 19 5.2 0.1 Control 16 69.1 4.0 16 5.3 0.2
Table 8. Parental feeding rates (visits/h) at experimental (N = 8) and control (N =8 ) nests on days 6 and 7 post-hatch. Feeding visit were monitored between 0600h and 1000h. Differences between control and experimental nests were not significant (Mann-Whitney U-tests; P's > 0.05). Day 6 Day 7 Mean S.E. Mean S.E. Experinem! 12.38 1.88 14.38 1.17 Control 14.38 1.17 14.50 1.42
Heri tabilitv of bodv size Heritability estimates for tarsus length and body mass are in Table 9. Estimates for tarsus length of control broods were significant at days 9 and 16. Although the estimate for tarsus length was similar between deprived and fed groups at day 16, the standard error associated with the estimate was higher for deprived nestlings. Consequently, only the fed nestlings had a significant heritability estimate for tarsus length. Estimates for tarsus length at day 9 were not significant for either fed or deprived groups. None of the heritability estimates of body mass were significantly different from zero at day 16. Heritability was estimated for wing length and bill length at day 16 (Table 10). Heritability estimates for wing length were not significant within any of the three groups. Bill length was heritable only within control broods, although the estimate for fed nestlings (0.35) approached significance (P = 0.09). Discussion Food abundance and growth Most studies of nestling growth in passerine birds have shown that both the size (body mass and/or tarsus length) and shape of nestlings is partly dependent on environmental conditions (Bryant 1975; James 1983; James and NeSmith 1988). Parental quality, prey abundance, and time of season are all correlated with the final size of nestlings in natural populations of birds (Ricklefs and Peters 1981; Alatalo and Lundberg 1986; Quinney et al. 1986; van Noordwijk 1987; van Noordwijk et al. 1988). Adult Tree Swallows in the Creston area feed their young for approximately 15 hours per day during early-mid June. Chick feeding rate peaks from 1000-2000h, following a slow period from 0600-1000h aeffelaar and Robertson 1986). Thus, deprived young in this experiment likely received 25-30% less food than control nestlings. Nonetheless, although significantly smaller in
Table 9. Heritability estimates for tarsus length at days 9 and 16 and body mass at day 16 post hatch. Estimates (2X the slope) are derived from regressions of brood means on maternal values. Tarsus length Body mass Day 9 Day 16 Day 16 N h2 S.E. N h2 S.E. N h2 S.E. - - Deprived 19 0.44 0.37 18 0.52 0.40 11-0.26 0.72 Fed 19 0.48 0.43 18 0.63" 0.30 11 0.06 0.50 Control 16 1.35" 0.61 16 1.11" 0.42 12-0.93 0.64
Table 10. Heritability estimates for wing length and bill length at day 16 post-hatch. Estimates (2X the slope) are based on regressions of brood means on maternal values. Wing length Bill length Deprived 17 0.44 0.37 17 0.25 0.31 Fed 17-0.59 0.53 17 0.35 0.19 Control 16-0.60 1.14 16 1.29* 0.47
both body mass and tarsus length at day 9, deprived nestlings reached a size that was similar to that of their fed nestmates by day 16. These results may be explained in two ways. First, nestling Tree Swallows may be able to compensate for poor parental feeding performance during the peak of the growth trajectory through some form of physiological response (e.g., digestive efficiency; Plavnik and Hurwitz 1985). However, the mechanism for such a physiological change is unknown and would likely come at a cost to some other growth component. A second possibility is that adults at experimental nests adjusted their feeding behaviour to compensate for the increased food demands of deprived nestlings. Although adults at experimental nests did not appear to compensate by increasing their chick feeding rates, I cannot rule out the possibility of some form of compensation by adults. For example, adults may have adjusted the size of prey loads brought to nestlings, or they may have fed deprived nestlings preferentially over their fed nestmates. There is increasing evidence that nestlings with relatively high begging rates receive significantly more meals from their parents, relative to their more silent nestmates (see references in Gottlander 1987); In addition to depriving some nestlings of food, the experimental design used in this study produced a set of offspring (fed group) that likely received "supplemental" feeding, relative to control nestlings. If parents did not adjust their absolute feeding rate during the 4-hour deprivation period, then fed young may have received a large amount of food relative to both deprived and control nestlings. Several studies have shown that the body mass of nestlings on territories provided with supplemental food is greater than that of nestlings on control territories (e.g., Hogstedt 1981; Arcese and Smith 1988). However, despite the potential for additional feedings received by fed nestlings over the 4-day experimental period, fed and control nestlings were sithilar in size for all characters at days 9 and 16. The lack of disparity in the size of fed and control nestlings following the.
experiment may have been due to the initial mass disparity at day 5, when control nestlings were significantly heavier than fed nestlings. Thus, the supplemental feedings received by the fed nestlings may have allowed them to reach a sik similar to that of control nestlings. In addition, because aerial insectivores such as Tree Swallows are subject to poor feeding conditions during inclement weather, they may be more likely to show compensatory growth than would other species. Environmental conditions and heritability A large environmental variance component during the growth stage may mask any additive genetic variation underlying body size. For example, three separate studies have found significant heritability estimates for tarsus length in Song Sparrows (Smith and Zach 1979; Smith and Dhondt 1980; Schluter and Smith 1986). However, a later study by Smith and Arcese (1988) found that the heritability estimate for tarsus length was not significantly different from zero during a year when juvenile survival was low. Similar results have been found during long-term studies of Great Tit (Parus&) populations (van Noordwijk et al. 1988); during "normal" or "good" food years, body size is heritable but during "poor" food years, heritability estimates are near zero. Thus, heritable variation underlying body size may only be detectable when environmental conditions during growth are good. In two previous years, I found significant heritability estimates for tarsus length in this population of Tree Swallows (Chapter 2; unpubl. data). This experiment attempted to simulate "poor" and "good" years of food abundance for deprived and fed nestlings, respectively, and as noted above, heritability estimates should be lower during poor food. years. However, the heritability estimates for tarsus length of deprived nestlings were similar to those for fed nestlings. The lower heritability estimates (and higher SE's) for tarsus length and bill length of deprived nestlings may be due to smaller sample sizes.
within brood means. Deprived brood means were based on either one or two nestlings whereas fed brood means were based on 3-4 nestlings. Sample size problems are particularly important in studies, such as this one, using single-parent/offspring regressions. In such cases, standard errors are typically higher than in studies using rnidparent/offspring comparisons (e.g., Smith and Zach 1979; Boag 1983) and heritability estimates, therefore, are more difficult to interpret. The heritability estimates for body mass and wing length were both associated with high standard errors. In contrast to tarsus length and bill length, both body mass and wing length have not reached final size by day 16. For a short period of time late in the growth stage, the body mass of nestling Tree Swallows exceeds that of adults, then recedes until fledging. Thus, body mass at day 16 may be a poor indicator of body mass the following breeding season. In addition, the repeatability of body mass is relatively low in Tree Swallows (Chapter 2). Wing length has reached only 5565% of adult size by day 16 and environmental variation is, therefore, much more likely to obscure any additive genetic effects at this age (Grant 1981). Contrary to most studies of growth during poor feeding conditions (see references in Rofstad 1988), my study showed that by 16 days post-hatch, nestling Tree Swallows were able to overcome poor early growth. Heritability estimates of tarsus length were not significantly affected by the experiment. However, the small sample sizes in this study may have obscured any differences between the experimental treatments. Whether nestlings physiologically compensated for poor early growth, or whether the experimental protocol was not sufficient in inducing a significant growth lag is not known. Although compensatory growth in wild birds is still poorly understood, experimental studies are likely to provide more answers than simple correlations between weather conditions and nestling size.