Kellogg Biological Station, Michigan State University, Hickory Corners, Michigan 49060, USA; 2Department of Biology, Kalamazoo College, Kalamazoo, Michigan 49006, USA Life-history theory predicts that as parents in- crease their investment in individual offspring, the survival rates of those offspring should also increase (Winkler and Wallin 1987, Stearns 1992, Bernardo 1996). For oviparous taxa, parental investment may be directed toward eggs or hatchlings. Birds are ideal organisms to test general hypotheses of parental in- vestment owing to the substantial interspecific var- iation in relative investments made by parents to- ward eggs and hatchlings and because of the ability of researchers to isolate and quantify these invest- ments. The altricial-precocial continuum, based on the development of young at hatching (Nice 1962), provides a convenient description of interspecific patterns in relative energy investment directed to- ward eggs versus hatched young. Precocial young hatch from energy-rich eggs and require little parental care after hatching, whereas altricial young hatch from relatively energy-poor eggs and require considerable parental care after hatching (Sotherland and Rahn 1987). We can use these patterns of energy investment in eggs versus young to make predictions about the relative impor- tance of each of these stages in determining offspring performance in different taxa along the life-history continuum. Parents of altricial young spend a large amount of effort during the breeding season feeding and caring for young relative to parents of precocial young. Therefore, we predict that variation in the quality of parental care after hatching should have a larger influence on offspring performance in altricial species than in precocial species. However, because altricial young hatch from energy-poor eggs, factors associated with eggs should have a smaller influence on the performance of altricial species relative to pre- cocial species. Thus, based on interspecific variation in patterns of energy investment, we predicted that the importance of the egg stage to subsequent per- formance of young birds will increase with increas- ing precocity of the young. One measure of parental investments in eggs is size, with large eggs presumably requiring a greater energy investment by females than do small eggs. 3 Present address: Department of Zoology and Ge- netics, Iowa State University, Ames, Iowa 50011, USA. E-mail: wloureed@iastate.edu 4 Present address: Department of Biology, Clarion University, Clarion, Pennsylvania 16214, USA. Although ornithologists have described the relation- ship between egg size and offspring performance for some species, interpretations of such descriptions are not straightforward because factors such as pa- rental quality and habitat quality may be correlated with egg size (see Magrath 1992, Williams 1994). One way to isolate egg-size effects is to experimentally manipulate the size of eggs or offspring (Sinervo 1990, Janzen 1993). An alternative is to isolate the ef- fects of egg size on nestling survival by adopting a fostering design in which offspring are moved among nests, thereby randomizing factors not asso- ciated with egg size. Studies that used a fostering design to elucidate the influence of egg size on offspring survival have focused largely on colonial seabirds (e.g. Nisbet 1978, Amundsen and Stokland 1990, Reid and Boer- sma 1990, Bolton 1991, Meathrel et al. 1993, Bollinger 1994, Amundsen et al. 1996) and have produced mixed conclusions. Few studies have used a fostering design to explore the influence of egg size on surviv- al of passerines (Schifferli 1973, Magrath 1992, Smith et al. 1995), a group whose young are altricial. Al- though we predict that egg size will have relatively little influence on nestling survival in these birds, at least one study of a bird with altricial young found a significant effect of egg size on juvenile survival (Schifferli 1973). Here, we present the results of an experiment in which we used a fostering design to determine the influence of egg size on nestling growth and survival in a free-living passerine, the Red-winged Blackbird (AgeIaius phoencieus). Methods.--The Red-winged Blackbird is a socially polygynous species that breeds throughout North America in both wetland and upland habitats. We studied a population of Red-winged Blackbirds nest- ing at the Experimental Ponds Facility of Michigan State University's Kellogg Biological Station in southwestern Michigan in 1993. The Experimental Ponds Facility consists of 18 replicate ponds, each 29 m in diameter, 2 m in depth, and separated by 5 m of grass. A dense ring of cattails (Typha latifoIia) grows around the margin of each pond, and a large number of Red-winged Blackbirds nest in the cat- tails. We selected 10 ponds with the highest breeding densities of blackbirds to use in our cross-fostering study. A total of 104 female Red-winged Blackbirds nested on these 10 ponds. We estimated the size of each egg laid by females nesting on these ponds by measuring linear dimen- sions (length and breadth) with calipers, or by mea- suring egg mass on an electronic balance within 24 h of laying. For a subset of these eggs we measured linear dimensions (mm) and mass (g) and calculated the following linear equation: egg mass = 0.000510 x [egg length x (egg breadth) 2] + 0.19 (1) (n = 380 eggs, r 2 = 0.97). We used this equation to estimate initial mass for eggs in which only length and breadth were measured. These masses, whether directly measured or indirectly estimated, were used for further analyses of egg-size effects. We moved eggs of known size between nests with the goal of randomizing factors other than egg size that may influence nestling growth and survival. Most of the variation in egg size of Red-winged Blackbirds occurs among rather than within nests (Muma and Ankney 1987). We focused our study on understanding the importance of egg-size variation among females and not within broods. Therefore, we exchanged whole clutches of eggs between nests and used individual nests as the experimental unit in all analyses. Our cross-fostering consisted of trading clutches between nests that were completed within 24 h of each other. We used 50 nests in the cross-fos- tering study. All experimental clutches consisted of four eggs, and all switches were done within a pond. We switched eggs within 48 h after the fourth egg was laid. Once egg switches were completed, we did not disturb nests until 10 days after the initiation of in- cubation, at which time we monitored nests closely to determine hatching. Nestlings were weighed to the nearest 0.1 g on a spring balance within 24 h of hatching. Nests were revisited and surviving nest- lings were weighed again at two days of age and just prior to fledging (i.e. leaving the nest) at 10 days of age. All nestlings used for this experiment hatched within 10 days of one another. Because hatching date had no effect on nestling performance over this time period, we did not include hatching date as a variable in our analyses. We focus our analyses of egg-size effects on nest- ling survival, nesting success, and nestling mass. Nestling survival was calculated as the proportion of nestlings surviving from hatching to fledging and does not include mortality during the egg stage. Nesting success was defined as a complete nest fail- ure (no fledglings produced) or as a success when at least one of the young survived to fledging. In addition to testing for egg-size effects, our cross-fostering design allowed us to assess whether females that lay large eggs are better able to produce fledglings than those that lay small eggs. A previous study on this population of Red-winged Blackbirds found that males made fewer than 5% of the total feeding visits to nestlings (Turner and McCarty 1998); therefore, we concentrate our discussion of pa- rental quality on females. We simultaneously evalu- ated the effects of two independent variables (1) mean egg size a female received in the fostering (egg-size effects), and (2) mean egg size a female produced (female-quality effects), on nestling sur- vival, nesting success, and nestling mass. Egg-size and female-quality effects on nestling survival and nestling masses were evaluated with a multiple lin- ear regression. We used Type III sums of squares to test whether egg size or female quality influenced nestling performance. The multiple regression tested whether each factor influenced the dependent vari- able after first accounting for the effects of the other factor(s). Nestling masses were log-transformed to normalize the residuals and to compare the slopes of the relationships for each age. Nesting success is a binomial response variable, and we evaluated egg- size effects and female-quality effects with a multi- ple logistic regression analysis. The above analyses tested for a linear effect of egg size and female quality on nestling performance. A significant linear relationship would suggest direc- tional selection. In addition to directional selection, we tested for nonlinear selection (i.e. disruptive or stabilizing selection) by squaring each independent variable and adding them to the multiple regression model (Lande and Arnold 1983, Brodie et al. 1995). Results.--Eggs varied in mass from 3.06 to 5.01 g ( = 4.04 _+ SD of 0.38 g, n = 196), and egg mass was normally distributed about the mean (Shapiro-Wilk test, W = 0.98, P = 0.31). The mean egg mass at in- dividual nests varied from 3.18 to 4.72 g ( = 4.03 +_ 0.35 g, n = 50) and also was normally distributed about the mean (W = 0.98, P  0.50). Based on a hi- erarchical analysis of variance, 24% of the overall variation in egg size resulted from differences among ponds, 56% of the variation was due to dif- ferences among nests within ponds, and the residual 20% of variation was due to variation among indi- vidual eggs within nests. Thus, the largest compo- nent of variation was due to variation among females in the average size of eggs that they produced. On average, 3.78 eggs hatched from the original 4- egg clutch, and of the eggs that hatched, an average of 1.96 survived to fledging at each nest (overall mor- tality rate of 48%). The multiple regression model in- dicated that the number of fledglings produced in each nest was not related to egg size (F = 0.32, df = 1 and 47, P  0.20) or female quality (F = 0.004, df = 1 and 47, P  0.20). Neither egg size nor female quality were under significant nonlinear selection with respect to nestling survival (P  0.20 for both terms). The nest-failure rate in our experimental nests was 16%, and all failures appeared to have resulted from starvation or abandonment rather than from preda- tion (i.e. no nests were found empty or damaged). We found little evidence that nest failure rate was related to egg size. The average egg mass was 3.86 SD of 3.5 0.5 FIG. 1. Average egg mass (g) Average nestling mass in foster Red- winged Blackbird nests at i (squares), 2 (circles), and 10 (triangles) days posthatching as a function of av- erage egg mass for that nest. A semilog plot was used to allow for comparison of slopes among nestling ages. Regression equations are: In day 1 mass = 0.2396 x egg mass + 0.3338 (r 2 = 0.41); In day 2 mass = 0.2267 x egg mass + 0.8094 (r 2 = 0.18); and In day 10 mass = 0.0782 x egg mass + 3.1341 (r 2 = 0.05). 0.29 g for failed nests and 4.06 -+ 0.35 g for successful nests (t = 1.54, P = 0.13). Statistical significance aside, the magnitude of the differences was not large: the average egg size in nests that failed to produce fledglings was 5% lower than that in nests that suc- ceeded in producing fledglings. The logistic multiple regression model simultaneously accounting for fe- male quality and egg-size effects explained only 5.4% of the variation in nesting success, and neither female quality nor egg size explained a significant amount of the variation in nesting success (female quality, X 2 = 0.0006, P > 0.20; egg size, X 2 = 2.16, P = 0.14). Neither egg size nor female quality were un- der significant nonlinear selection with respect to nesting success (P > 0.20 for both terms). The effects of egg size and female quality on nest- ling mass were evaluated simultaneously. Egg size had a positive effect on initial nestling mass, but this effect weakened as nestlings grew (Fig. 1). Nestling mass on days 1 and 2 were positively related to egg size (day 1, F = 26.03, df = 1 and 41, P < 0.0001; day 2, F = 5.90, df = 1 and 29, P = 0.02). However, 10 days after hatching, egg mass had little effect on nestling mass (F = 1.39, df = 1 and 39, P > 0.20). Effects of female quality on nestling mass were neg- ligible throughout the nestling period (day 1, F = 1.13, df = 1 and 41, P > 0.20; day 2, F = 1.80, df = 1 and 29, P = 0.19; day 10, F = 0.02, df = 1 and 39, P > 0.20). Discussion.--Although the mass of Red-winged Blackbird eggs varies considerably (40% difference between the largest and smallest eggs), we found lit- tle evidence to suggest that egg size is an important determinant of survival or growth of nestlings. We acknowledge, however, that our study lacks the pow- er to detect small effects. For example, nests com- posed of eggs larger than the median size fledged 13% more offspring than did nests composed of eggs that were smaller than the median size. Given the large amount of variability we observed in female re- productive success, effects of this size are not statis- tically detectable without much larger sample sizes, but these differences potentially represent strong se- lection differentials and may be important in the evolution of egg size. Thus, our study highlights the importance of very large sample sizes in investiga- tions aimed at testing whether egg size is under nat- ural selection. Our results show clearly that even if the small egg- size effects on offspring performance were real, the survival and growth of nestlings are still influenced primarily by factors other than egg size. Egg size and female quality together accounted for less than 1% of the variation in nestling survival. Although nestling mass at one and two days of age was related to egg size, this effect disappeared by day 10. On average, experimental nestlings increased their mass by 65% during the first day after hatching (Fig. 1). This sug- gests that any variation among nests in average growth rates likely overwhelms any differences in initial size. A previous study of this population showed that nestling growth was dependent on food availability (Turner and McCarty 1998). Because growth rates of altricial nestlings generally are a function of food provisioning (Searcy and Yasukawa 1995) and other aspects of female efficacy (i.e. para- site loads; Moller 1990, Roby et al. 1992), differences in female quality should have large effects on nest- ling growth and survival. Egg size has often been used as a measure of fe- male quality. Females that lay large eggs are as- sumed to be of better quality than females that lay small eggs (Amundsen and Stokland 1990, Reid and Boersma 1990, Bolton 1991); however, our experi- ment provides no support for this idea (see also Amundsen et al. 1996). In this population, a female's ability to rear nestlings was independent of the size of her eggs. Although it is likely that variation in fe- male quality compensated for the initial differences in nestling size, this variation was not related to the size of the eggs a female produced. Overall, our results are consistent with the idea that the sensitivity of offspring performance to egg- size variation is related to the degree of precocity of hatched young. Females that raise altricial young in- vest relatively little in the egg stage compared with females that raise precocial young. Because of this fact, factors associated with the egg stage should have less of an influence on performance of altricial young relative to that of precocial young. Clearly, testing the generality of this pattern will require ad- ditional studies on the consequences of egg-size var- iation in a wide range of avian taxa. Acknowledgments.--We thank Beth Strunk for help with the field work. Early drafts of this paper were significantly improved by E. Borer and the ELVIS dis- cussion group at Iowa State University. Partial sup- port for this study was provided by NSF Research Training Grant DIR-9113598 to the Kellogg Biologi- cal Station, the Department of Zoology and Genetics at Iowa State University, and the Pymatuning Labo- ratory of Ecology. This is contribution number 869 of the Kellogg Biological Station. LITERATURE CITED AMUNDSEN, t., S. LORENTSEN, AND t. TVERAA. 1996. Effects of egg size and parental quality on early nestling growth: An experiment with the Ant- arctic Petrel. Journal of Animal Ecology 65:545- 555. AMUNDSEN, t., AND J. N. STOKLAND. 1990. Egg size and parental quality influence nestling growth in the Shag. Auk 107:410-413. BERNARDO, J. 1996. Maternal effects in animal ecol- ogy. American Zoologist 36:83-105. BOLLINGER, P. B. 1994. Relative effects of hatching or- der, egg-size variation, and parental quality on chick survival in Common Terns. Auk 111:263- 273. BOLTON, M. 1991. Determinants of chick survival in the Lesser Black-backed Gull: Relative contri- butions of egg size and parental quality. Journal of Animal Ecology 60:949-960. BRODIE, E. D. III, A. J. MOORE, AND E J. JANZEN. 1995. Visualizing and quantifying natural selection. Trends in Ecology and Evolution 10:313-318. JANZEN, E J. 1993. An experimental analysis of nat- ural selection on body size of hatchling turtles. Ecology 74:332-341. LANDE, R., AND S. J. ARNOLD. 1983. The measure- ment of selection on correlated characters. Evo- lution 37:1210-1226. MAGRATH, g. D. 1992. The effects of egg mass on the growth and survival of Blackbirds: A field ex- periment. Journal of Zoology (London) 227:639- 653. MEATHREL, C. E., J. S. BRADLEY, g. D. WOOLLER, AND I. J. SKIRA. 1993. The effects of parental condition on egg-size and reproductive success in Short- tailed Shearwaters Puffinus tenuirostris. Oecolo- gica 93:162-164. MOLLEI, A. P. 1990. Effects of parasitism by a hae- matophagous mite on reproduction in the Barn Swallow. Ecology 71:2345-2357. MUMA, K. E., AND C. D. ANKNEY. 1987. Variation in weight and composition of Red-winged Black- birds. Canadian Journal of Zoology 65:605-607. NICE, M. M. 1962. Development of behavior in pre- cocial birds. Transactions of the Linnaean Soci- ety of New York 8:1-211. NISBET, C. t. 1978. Dependence of fledging success on egg-size, parental performance and egg-com- position among Common and Roseate terns, Sterna hirundo and S. dougallii. Ibis 120:207-214. RED, W. V., AND P. D. BOERSMA. 1990. Parental qual- ity and selection on egg size in the Magellanic Penguin. Evolution 44:1780-1786. ROBY, D. D., K. L. BRINK, AND K. WITTMAN. 1992. Ef- fects of blowfly parasitism of Eastern Bluebird and Tree Swallow nestlings. Wilson Bulletin 104:630-643. SCHIFFERLI, L. 1973. The effects of egg weight on the subsequent growth of nestling Great Tits Parus major. Ibis 115:549-558. SEARCY, W. A., AND K. YASUKAWA. 1995. Polygyny and sexual selection in Red-winged Blackbirds. Princeton University Press, Princeton, New Jer- sey. SINERVO, B. 1990. The evolution of maternal invest- ment in lizards: An experimental and compar- ative analysis of egg size and its effects on off- spring performance. Evolution 44:279-294. SMITH, -I. G., t. OHLSSON, AND K. J. WETTERMARK. 1995. Adaptive significance of egg size in the Eu- ropean Starling: Experimental tests. Ecology 76: 1-7. SOTHERLAND, P. g., AND -I. RAHN. 1987. On the com- position of bird eggs. Condor 89:48-65. STEARNS, S.C. 1992. The evolution of life histories. Oxford University Press, Oxford. TURNER, A.M., AND J.P. McCARTY. 1998. Resource availability, breeding site selection, and repro- ductive success of Red-winged Blackbirds. Oe- cologia 113:140-146. WILLIAMS, T. D. 1994. Intraspecific variation in egg size and egg composition in birds: Effects on off- spring fitness. Biological Reviews of the Cam- bridge Philosophical Society 68:35-59. WINKLER, D. W., AND K. WALLIN. 1987. Offspring size and number: A life history model linking ef- fort per offspring and total effort. American Naturalist 129:708-720. Received 26 January 1998, accepted 18 September 1998. Associate Editor: J. S. Sedinger