Costs of reproduction in the wild: path analysis of natural selection and experimental tests of causation.
For organisms with extended parental care (e.g., birds), manipulating postlaying clutch size (Lack 1947) has become the standard protocol to study costs of reproduction (Nur 1988). However, this manipulation, which focuses on the costs of postlaying provisioning of the young, ignores any costs that might arise from energy invested before eggs are laid. We describe the fitness effects of several "mechanistic manipulations" that alter reproductive investment during early vitellogenesis (Sinervo and Licht 1991a,b) of Uta stansburiana, an iguanid lizard without extended parental care. Because energy in this lizard is invested prior to laying, current reproductive investment is conveniently indexed by attributes of the clutch at laying (e.g., egg mass, clutch size, total clutch mass, relative clutch mass = RCM = clutch mass of eggs/total mass of female and eggs [Vitt and Congdon 1978; Vitt 1981; Vitt and Price 1982] or analysis of covariance [ANCOVA] of clutch mass controlling for body size [Sinervo et al 1991]). Clutch size and total clutch mass can be increased by administering exogenous follicle stimulating hormone [FSH] [ILLUSTRATION FOR FIGURE 1 OMITTED]. Conversely, clutch size and total clutch mass can be decreased by direct ovarian manipulation involving surgical ablation of yolking follicles. A third manipulation, administration of exogenous corticosterone selectively channels more investment into female, not male offspring [ILLUSTRATION FOR FIGURE 2 OMITTED].
Although such experimental manipulations greatly facilitate the detection of costs in the wild, detection of costs arising from natural variation is, however, problematic. Phenotypic correlations between current effort and survival or future reproductive success presumably have an underlying genetic basis (Reznick 1985). However, in a natural setting many other environmental factors such as density and food availability can potentially alter current reproductive investment, and these environmental factors may also simultaneously affect survival. Such joint causal effects acting on both current effort and future reproductive success may establish spurious correlations between these two traits that are unrelated to genetically based trade-offs. Whether such environmental effects obscure detection of genetically based costs of reproduction is currently a subject of debate in the field of ecology and evolution (Partridge and Harvey 1985; Partridge 1992; Reznick 1985, 1992).
We provide a path analytical (Li 1975) approach to understanding how such environmental correlations are established and how these confounding environmental influences can be removed to reveal natural selection on reproductive traits per se. Path analysis provides a method for structuring cause-and-effect relationships among variables in a multiple regression framework (Li 1975; Mitchell 1993). Because path analysis is really only a hypothesis of cause and effect relationships, it is strongly recommended that results from path analysis be augmented with experimental verification of the cause and effect pathways. We compare results from our experimental manipulations to verify the cause-and-effect relationships that are hypothesized in our path analysis of natural variation.
We provide strong inference regarding the existence of costs of reproduction in the wild that is based on (1) the fitness consequences of natural variation in reproductive investment derived by path analysis, (2) experimental verification of these cause-and-effect relations, and we further complement these two approaches with (3) studies of year-to-year changes in reproductive investment and egg size. Based on observed patterns of natural selection from methods (1) and (2), we predict the direction of evolution. Predicted shifts in current effort correspond with observed shifts across generations. Because the underlying trait involved in all three manipulations (egg size) has been shown to have a heritable basis in natural populations (Sinervo and Doughty 1996), we infer that year-to-year changes in reproduction are due in part to rapid short-term evolution in response to acute natural selection, events that coincide with termination of a severe drought and a rise in numbers of natural predators on side-blotched lizards.
MATERIALS AND METHODS
Detailed description of the study site and husbandry of females is found in Sinervo and Licht (1991a,b), Sinervo et al. (1992), Sinervo and Doughty (1996), Doughty et al. (1994); Doughty and Sinervo (1994). During 1991 through 1994, we studied natural selection on reproductive investment of side-blotched lizards (U. stansburiana), a small iguanid lizard (3 to 10 g) that matures in one year in the inner Coast Range of California (near Los Banos Grandes, Merced Co.). Females typically lay a large first clutch (five eggs, range three to nine eggs, average total clutch mass 2.4 g), and smaller clutches later in the season (four eggs, range two to eight eggs) (Sinervo et al. 1992). Four clutches can be produced from March through August, at monthly intervals.
Because side-blotched lizards have no parental care after egg laying, the total mass of yolk that is produced in each clutch provides a convenient index of reproductive investment. The "costs of reproduction" that we present in this study are based on estimates of reproductive investment on the first clutch and how such investment impacts survival to subsequent clutches and the number of clutches produced. We have not included estimates of the survival of these offspring to maturity that would provide a comprehensive estimate of future reproductive success. We have maintained a large-scale study of natural selection on offspring survival from 1989 to the present (Sinervo et al. 1992). Short-term estimates of clutch production are consistent with comprehensive measures of future reproductive success that include survival of offspring to maturity (Sinervo et al. 1992).
Mechanistic Manipulations of Reproductive Investment
We manipulated reproductive investment by direct ovarian manipulation (follicle ablation) (Sinervo and Licht 1991 a,b); ovarian stimulation using exogenous gonadotropin (Sinervo and Licht 1991a); and an as yet undescribed manipulation of total clutch mass using exogenous corticosterone [ILLUSTRATION FOR FIGURE 1-2 OMITTED]. All surgeries were performed after injecting 0.02 ml of 0.2% lidocaine subcutaneously at the incision site and cooling the lizard in a crushed ice bed. The protocol provides appropriate anaesthesia and allows for rapid recovery (2-5 min for total recovery). Lizards were returned to their territory after recovery (within 24 h).
A reduction in total clutch mass should enhance survival or future reproductive success in female side-blotched lizards (U. stansburiana) because energetic investment in current reproduction is decreased and potential physiological and ecological costs (e.g., foraging effort, energy acquisition, and yolk synthesis) associated with large reproductive investment are ameliorated. We surgically exposed the ovaries of early to midstage vitellogenic females and removed yolk from half of their follicles (yolkectomy), which effectively terminates their growth. We performed this experimental manipulation in 1992 and 1993.
Conversely, an increase in total clutch mass using FSH should decrease survival or future reproductive success according to theories of costs of reproduction (Williams 1966, Reznick 1985) because potential costs associated with reproduction are exacerbated. We enhanced clutch mass of a second and third group of females using exogenous FSH and exogenous corticosterone. Plasma levels of exogenous FSH (ovine FSH delivered in an Elvax pellet, which allows slow delivery of the large gonadotropin protein through the polymer matrix) were designed to match levels in previous laboratory studies (Sinervo and Licht 1991a).
A third manipulation of clutch size, corticosterone supplementation, increases energy invested into the clutch; however, energy is selectively channeled into female offspring, not male offspring. Exogenous corticosterone was delivered in silastic implants that were designed to result in plasma levels near the upper physiological levels observed for field active lizards (Wilson and Wingfield 1992; DeNardo and Licht 1993; DeNardo and Sinervo 1994). We performed this experimental manipulation in the spring of 1991, 1992, and 1993. Controls for the surgical protocols consist of a large group of unmanipulated females (see below) and sham-manipulated females that undergo the surgical protocols but receive only a saline implant.
We studied the survival of individually marked females (toe clipped) from surgical manipulation to the production of the first clutch (censused at 1 mo in implant experiments) or second clutch (censused at 6-10 wk in ablation experiments). In early March, we surgically implanted females with hormone supplements (corticosterone: N = 18 in 1991, N = 60 in 1992, and N = 42 in 1993; FSH: N = 30 in 1993) or sham implants (early March, N = 19 in 1991, N = 61 in 1992, and N = 55 in 1993). Surgical reduction of clutch size on a separate group (N = 41 in 1992 and N = 45 in 1993) took place 10 d after implant experiments.
Comparative Data on Natural Selection on Reproductive Investment
We complement our experimental analysis (Table 1) of costs of reproduction associated with female survival with [TABULAR DATA FOR TABLE 1 OMITTED] an analysis of the effects of natural variation in reproductive investment on future reproductive success. Because local environmental quality can obscure the detection of costs of reproduction, we measured adult density by mapping territories of females on our main study site. From these territorial maps, we determined the number of neighbors each female overlapped with and used these values as estimates of local density experienced by each female during vitellogenesis on each clutch. Moreover, because size and age confound interpretation of estimates of clutch mass (Sinervo and Licht 1991a,b), we used path analysis to remove the effects of size and confounding effects of density on current investment and future reproductive success.
We marked a large cohort of unmanipulated females in mid-April (e.g., N = 104 in 1991, N = 107 in 1992, N = 134 in 1993, N = 79 in 1994) and followed them throughout the reproductive season to determine habitat use and survival patterns by daily censuses. In 1991, we mapped only the locations of individual females. For 1991, we estimated the number of neighbors by determining the number females located within 8 m of the focal female (typical major axis of the convex polygon circumscribing a female home range). If all females were located more than 8m away, we scored that females as being solitary.
From 1992 to 1994, we obtained accurate estimates of the number of neighbors by mapping female home ranges. To estimate home-range size and overlap of female home ranges, we walked daily transects (total of 10 passes) and made observations on active female lizards (DeNardo and Sinervo 1994a,b). Females were mapped using a compass bearing and distance to nearest reference objects (Doughty et al. 1994). Home-range overlap was estimated using a software program developed for the Macintosh (MacTurf) available from B.S. upon request (Sinervo 1993). The number of other females observed on the focal female's home range was used as an index of her "local density."
We obtained clutches of eggs from these females by bringing them into the laboratory after they had ovulated their eggs (yolk was already invested in the clutch). Ovulation is readily determined by abdominal palpation and is 100% accurate. After the brief laboratory stay required to obtain their eggs (3-10 d), females were released back on their territory and survival to the production of subsequent clutches was determined by daily censuses. Comparison of mortality patterns of these females that were briefly brought into the laboratory with a large cohort of females that remained in the wild for the duration of the study (N = 70, 1992) indicates no effects of the brief stay in the laboratory on mortality. For example in 1992, 79 of 148 females (53%) that had a brief stay in the laboratory survived during a 49-d interval from early March to late April (e.g., until production of the second clutch). Likewise, 35 out of 70 females (50%) that remained on a separate control plot for the duration of the study survived during the comparable 48-d interval. In addition, our estimates of 1 mo survival of females receiving hormonal implants (Table 1) were not confounded by the laboratory stay because censuses took place prior to obtaining eggs in the lab. Residence in the laboratory does not influence prelaying reproductive costs, and residence is unlikely to distort estimates of postlaying costs of reproduction.
Not all females marked in April survive to reproduce and we do not obtain estimates of current reproduction on every female ([less than] 5% of females lay eggs on the outcrop). Only females with complete reproductive indexes (total clutch mass of the first clutch and number of clutches) were used in path analyses (N = 60 in 1991, N = 65 in 1992, N = 54 in 1993, and N = 55 in 1994). Along with the path analysis, we present univariate plots that describe simple regression of the covariates: (1) density and (2) postlaying mass on dependent variables (3) clutch mass (current investment), and (4) clutches produced (future reproductive success).
Path Analysis of Current Reproductive Investment and Future Reproductive Success
Environmental causes obscure detection of natural selection that forms that basis of costs of reproduction (e.g., a correlation between current effort and future reproductive success). If an individual remains in the same location throughout life, there is a strong a priori expectation that local conditions experienced by an individual will also have a correlated effect on future reproductive success because of spatial and temporal autocorrelation. Path analysis decomposes simple correlations into hypothesized "causal relationships" between two or more traits. Our path diagram reflects our hypothesis concerning how environmental "causes" might lead to a correlation between current investment and future reproductive success. Our goal is to first remove the hypothesized environmental causes (e.g., density and size) and then analyze residuals to detect selection on reproductive traits that are independent of such environmental causes.
For example, if females aggregate at high density in high-quality locations and such locations remain good during the reproductive season, then current investment should be positively correlated with future reproductive success (aggregation density is positively correlated with both current investment and future success). If conditions on the first clutch are good (positive path coefficient between density and current investment) and conditions on later clutches deteriorate owing to depletion of resources by females in the high-density aggregations (negative path coefficient from density to future reproductive success), then we might expect a negative correlation to be set up between current investment and future reproductive success. Alternatively, local environmental effects experienced by a female may affect current investment but not future success (and vice versa). In this case, we would expect no correlation between these traits to be established by local environmental conditions. Thus, the "predicted" environmental correlation (double-headed arrow) between current effort and future reproductive success is due two paths (arrows in path diagram, [ILLUSTRATION FOR FIGURE 3 OMITTED]) that connect at the common cause density.
If density is correlated with clutch mass and density is correlated with future reproductive success, we can estimate the degree to which density is responsible for the observed correlation between current investment and future reproductive success by computing: (1) the predicted "path" between density and clutch mass and (2) the predicted "path" between density and future reproductive success. Statistical inference is not used on such values predicted from multiple regression equations. These values are plotted for heuristic reasons. The correlation between such predicted values reflects our hypothesis concerning density as a common cause of current and future reproductive success. Residuals for both current investment and future reproductive success that remain unexplained by the predicted effects of the common cause of density reflect the costs of reproduction that are independent of density and are drawn as a completely separate correlation to the right of the main path diagram (double-headed arrow, [ILLUSTRATION FOR FIGURE 3 OMITTED]).
Although there are strong a priori expectations of a correlation between postlaying mass and clutch mass (e.g., Sinervo 1990, larger females lay a larger total clutch mass), we have no a priori reasons for assuming a correlation between postlaying mass and future success as indexed by survival to the production of subsequent clutches. We do not necessarily expect larger females to have higher or lower survival. Survival selection on body size could be directional and favor large or small females or selection could be stabilizing in which intermediate-sized females have the highest survival. Indeed, analysis of the correlation between postlaying mass and our index of future reproductive success was not significant. Thus, we compute a path between body size and current investment, but not between body size and future reproductive success.
Effects of Manipulations on Reproductive Investment and Female Condition
Determining the time course of selection arising from the effects of corticosterone, follicle ablation, and FSH can be made from inferences concerning the effect of each manipulation on condition before and after a female lays her clutch. Such estimates of condition are useful in interpreting observed patterns of selection. Females undergoing the follicle ablation protocol are lighter per unit of body length prior to oviposition (e.g., with ovulated eggs), and these females produce a less massive clutch compared with control females (ANCOVA on prelaying body mass of 1992 females: 0.31 g lighter, F = 6.75 df = 1,82; P [less than] 0.0001, holding effects of body size as measured by snout-vent length [SVL] constant with SVL as the covariate, F = 10.53; df = 1,82; P [less than] 0.0001; ANCOVA on clutch mass of 1992 females: 0.84 g lighter holding effects of body size as measured by SVL constant, F = 21.00; df = 1,82; P [less than] 0.0001 with SVL as the covariate, F = 145.22; df = 1,82; P [less than] 0.0001). After oviposition, females undergoing the follicle abalation protocol are generally heavier and in better condition than control females (average mass of control females = 4.42 g), (ANCOVA on postlaying body mass of 1992 females: 0.18 g heavier holding effects of body size as measured by SVL constant, F = 11.56; df = 2,87; P [less than] 0.0001 with SVL as the covariate, F = 82.42; df = 1,87; P [less than] 0.0001). We might expect that enhanced female condition would also enhance survival or allocation to reproduction on subsequent clutches. Indeed, follicle ablation enhanced survival to the second clutch in all three years (Table 1), however, the manipulation does not increase clutch mass on the second or third clutches (data not shown).
Corticosterone made females more massive immediately prior to oviposition (1.2 g heavier than controls in 1991, controls are 5.28 g prior to oviposition, ANCOVA on pre-laying body mass of 1992 females: effect of corticosterone F = 12.38; df = 1,19; P = 0.01, effect of SVL F = 11.15; df = 1,19; P [less than] 0.01). Differences in "investment" present prior to oviposition is selectively channeled into female, not male offspring ([ILLUSTRATION FOR FIGURE 2 OMITTED], data from 1992 [not shown] provide similar results). The potential sources of such "investment" in female offspring is provided in the Discussion. For a given egg size, female hatchlings coming from females receiving corticosterone are 0.036 g more massive than female hatchlings, which weigh 0.41 g, coming from control females (test for effect of corticosterone on female hatchling mass, F = 8.82 g; df = 1,37; P [less than] 0.01, effect of covariate for egg mass F = 80.64; df = 1,37; P [less than] 0.0001). No such effects are observed in male offspring (test for effect of corticosterone on male hatchling mass F = 0.00; df 1,37; P [greater than] 0.95, effect of covariate egg mass F = 24.98; df = 1,37; P [less than] 0.0001). It is noteworthy that corticosterone does not change egg size per se but corticosterone makes females produce eggs that yield larger female hatchlings.
Neither corticosterone implants or sham surgery had significant affects on postlaying body condition (1992 females: effect of corticosterone F = 3.22; df = 2,99; P [greater than] 0.07, sham surgery F = 1.68; df = 2,99; P [greater than] 0.19, covariate for SVL F = 27.93; df = 1,99; P [less than] 0.0001). Although we cannot rule out other long-term effects, corticosterone implants did not affect clutch mass or post-laying condition on the second clutch (data not shown). Thus, if corticosterone has an effect on costs of reproduction, it is likely that it operates during the clutch that is manipulated. We did not have sufficient recoveries of females receiving FSH to perform similar analyses of body condition of females. However laboratory studies indicate that exogenous FSH adds two eggs to the clutch and 0.6 g to total clutch mass (Sinervo and Licht 1991a).
Natural versus Experimental Variation: Selection Favoring the Production of Large Clutches in 1991
In 1991, we detected significant effects of density on (1) clutch mass on the first clutch and (2) the number of clutches produced during the reproductive season (future reproductive success). However, because density increased clutch mass ([ILLUSTRATION FOR FIGURE 3B OMITTED], 1991) but decreased future reproductive success ([ILLUSTRATION FOR FIGURE 3B OMITTED], 1991), a negative correlation is established between clutch mass and future reproductive success that obscures the detection of selection on reproductive investment per se. When the confounding effects of density and body size are removed [ILLUSTRATION FOR FIGURE 3A-D OMITTED], we detected a significant positive correlation between clutch mass and future reproductive success that indicates that females producing the largest clutches had the highest survival to subsequent clutch production. These comparative data are corroborated by the high survival observed in females receiving corticosterone, which increases both clutch mass and survival compared with control females on the same outcrop (Table 1). High investment in 1991, whether arising from factors associated with density or from experimental manipulation of investment using corticosterone implants, enhanced future reproductive success.
The End of the Drought: Selection Favoring Production of Small Clutch Mass in 1992
In 1992, density was positively correlated with both clutch mass ([ILLUSTRATION FOR FIGURE 3B OMITTED], 1991) and future reproductive success ([ILLUSTRATION FOR FIGURE 3B OMITTED], 1992). A positive correlation is established between clutch mass and future success that obscures detection of selection on reproductive investment per se. When confounding effects of density and body size are removed ([ILLUSTRATION FOR FIGURE 3A-D OMITTED], 1992), we detected a significant negative correlation between clutch mass and future reproductive success that indicates that females producing the smallest clutch mass had the highest survival to subsequent clutch production. These comparative data are corroborated by the higher survival observed in females that had follicles ablated during early vitellogenesis compared with control females on the same outcrop. We ameliorated costs acting after females produced their first clutch by reducing clutch mass on the first clutch. However, females receiving exogenous corticosterone in 1992, during the earliest stages of vitellogenesis, had rates of survival comparable to females receiving a sham implant even though females receiving exogenous corticosterone produced a significantly larger total clutch mass. Thus, although we did not measure significant natural selection during the earliest stages of vitellogenesis on the first clutch (e.g., no selection against females receiving exogenous corticosterone), we did measure selection later in the season (e.g., selection against controls versus females receiving a follicle ablation arising from survival to production of the second clutch).
Evidence for Acute Natural Selection against Females Producing Large Clutches
In 1993, we detected a significant positive effect of density on future reproductive success, but no effect of density on clutch mass ([ILLUSTRATION FOR FIGURE 3A-D OMITTED], 1993). Thus no correlation between clutch mass and future reproductive success is established by density. Whereas, density confounds estimates of clutch production per se, density is not confounded with clutch mass. Moreover, we did not detect selection on clutch mass arising from the number of clutches produced [ILLUSTRATION FOR FIGURE 3E OMITTED].
If there was no selection observed in postlaying survival of unmanipulated females, then why is there such a large decrease in clutch size [ILLUSTRATION FOR FIGURE 4 OMITTED] and egg size [ILLUSTRATION FOR FIGURE 5 OMITTED] observed between 1992 and 1993? This drop in clutch size is readily explained if natural selection acted on unmanipulated females (e.g., natural variation) before we could index their total clutch mass (e.g., clutch mass can only be determined at laying). Experimental confirmation of this hypothesis is provided by both of our prereproductive manipulations that enhance total clutch mass (FSH and corticosterone). Females receiving either exogenous corticosterone or FSH had decreased survival to the production of the first clutch compared with controls on the same outcrop. In this case, we can reliably induce increased current investment and observe the cascading costs of such increased investment in current reproduction. Why did we induce such prelaying costs in 1993 but not in 1992? The difference in the effects of clutch mass on prelaying survival from 1992 to 1993 is consistent with a significant between-year difference in the survival rate of controls during the critical prelaying period (e.g., 82% in 1992 versus 53% in 1993, Table 2).
[TABULAR DATA FOR TABLE 2 OMITTED]
Variation in Density Effects among Years and within Seasons
The differences in the sign of the density effects among years ([ILLUSTRATION FOR FIGURE 3 OMITTED]. e.g., positive in 1992 to mainly negative in 1993-1994) are likely to be related to the degree of aggregation behavior observed in female side-blotched lizards. In 1993 and 1994 the female aggregations are significantly larger than in 1992 (0.73 [+ or -] 0.11 neighbors for 1992, 1.29 [+ or -] 0.17 and 2.61 [+ or -] 0.26 for 1992, 1993, and 1994, respectively). Although we did not measure female home ranges in 1991, animals were much more rarefied than in 1992, 1993 and 1994 and at half the overall density. The quality of the areas where females aggregate at high density may have declined, perhaps owing to the increase in the average size of female aggregations from 1991-1992 to 1993-1994. We also observed an overall decline in reproductive indices in high-density aggregations after 1992 (whereas the correlations between density and current reproduction or future reproductive success were largely positive in 1991-1992 and the correlations between density and the reproductive indices in 1993-1994 were largely negative, [ILLUSTRATION FOR FIGURE 3 OMITTED]). Although gross year-to-year differences in density might explain changes in the effects of density on reproduction among years, it is also clear that density effects can change within a single season (e.g., effects of density on total clutch mass versus number of clutches). Density is calculated once in the spring as females begin reproduction. In years in which there is a high density early (e.g., 1993 and 1994), one might expect females at the highest densities to do poorly on the first clutch. But owing to extremely high mortality after the first clutch (e.g., 1993), females that survived at "good sites" may enjoy high survival after the first wave of attrition. The locations of these high-density aggregations is consistent year after year.
Mechanistic Manipulations of Clutch Mass and Potential Costs of Reproduction
In nature, corticosterone levels in female U. stansburiana are highest early in the reproductive season when females are producing clutches with a very large total clutch mass (Sinervo and Licht 1991a; Sinervo et al. 1992), and the levels decline later in the reproductive season when females are producing a lower total clutch mass (Wilson and Wingfield 1992). Glucocorticoid secretion (e.g., corticosterone) may also be causally related to gonadotropins like FSH (Wilson and Wingfield 1992), because injection of gonadotropins in reptiles stimulates adrenal secretion of corticosterone (Callard and Callard 1978). Enhancement of clutch size (by 1.7 eggs) and total clutch mass (by 0.6 g) by gonadotropin supplementation (FSH) acts on the process of "follicular selection" in the ovary during the earliest stages of vitellogenesis (Jones 1978; Sinervo and Licht 1991a, b). One function of corticosterone during reproduction in female U. stansburiana may involve energy mobilization (Wilson and Wingfield 1992). However, corticosterone has additional regulatory roles. We observed that the increased clutch mass in females receiving exogenous corticosterone is selectively channeled into female offspring, not male offspring. In addition, corticosterone is purported to be a "stress hormone" that modulates social and environmental stress (reviewed in Johnson et al. 1992). As such, the mechanisms of enhanced clutch mass via increased female offspring size may be related to regulation of stress and cascading effects on allocation to offspring. The potential source of "investment" in female offspring remain unresolved and could take at least three forms: (1) changes in the proportional allocation of egg components (lipid, water, albumin, etc.); (2) sex-dependent differences in metabolism of egg energy reserves governed by maternally provided factors; or (3) maternally provided factors that alter postlaying rates of water uptake and hatchling size.
Hormonal manipulation of effort using exogenous corticosterone and gonadotropin appear to exact their costs of reproduction during prelaying phases of the life history. Indeed, we did not detect any long-term effects of exogenous corticosterone on body condition indices per se. Measurable phenotypic effects were restricted to total clutch mass, and selection arising from such enhanced investment (Table 1, 1993). For example, a novel result of our study is that elevated corticosterone appears to enhance survival of females during the drought year of 1991 while at the same time enhancing reproductive investment. Moreover, females that naturally produced large clutches in 1991 had high survival. Such effects presumably arise from corticosterone's role as a stress hormone and as such have a physiological basis (reviewed in Johnson et al. 1992). Our observation that exogenous corticosterone can ameliorate costs of reproduction suggests that it has a central role in the regulation of stresses associated with reproductive effort (e.g., energy invested in reproduction and energy acquisition during reproduction). Likewise, we have detected positive effects of corticosterone on male survival (DeNardo and Sinervo, unpubl.). However, corticosterone's precise role is likely to result from a complex interaction between physiologically and ecologically mediated costs, and the hormone does not always have beneficial consequences for survival. We observed negative effects of corticosterone on future reproductive success during periods of high natural mortality that were correlated with increased predatory snake activity.
The decreased clutch size (by 2.5 eggs) and total clutch mass (0.85 g) achieved by follicle ablation simulates the natural process of follicular atresia that occurs during later stages of vitellogenesis in that energy not allocated to offspring is available for future reallocation (N.B., a fraction of the yolk is removed during yolkectomy that amounts to 0.11 g of yolk that is unavailable for reallocation, Sinervo and Licht 1991a,b). Mechanistically, we would predict that this manipulation has the potential for more long-term effects. Indeed, the enhanced body condition of postlaying females that underwent the follicle ablation protocol reflects a reallocation of resources acquired during prelaying phases to postlaying body condition. This has a cascading effect on postlaying survival (Table 1) not prelaying survival (data not shown). The biological significance of the 0.18-g enhancement of postlaying body mass of females undergoing the yolkectomy procedure could be dramatic. Surgical inspection of these females after laying reveals that they have visible fat reserves compared with unmanipulated females (Sinervo, unpubl. data). If we assume that the 0.18 g is stored entirely as lipid, we can compare the energy savings from yolkectomy to the daily energy expenditure of a female U. stansburiana. Using the energy equivalent of 37.66 kJ [g.sup.-1] for lipid (Derickson 1976), the energy savings from yolkectomy is 6.8 kJ greater than controls. Using body mass-field metabolic energy regressions for uta (e.g., Nagy 1983), this could amount to the approximately 10-20 d of total metabolic energy expenditure of free-ranging uta.
The mechanistic manipulations we describe are not in the strict sense complementary because the mode of action on energy mobilization are temporally quite different, and their effect on phenotype are also distinct. On the one hand, both exogenous gonadotropin and corticosterone act during early prelaying vitellogenesis, but their effect on female phenotype is different: gonadotropin increases total clutch mass through its effects on clutch size, and exogenous corticosterone enhances total clutch mass through its effects on offspring size. Both manipulations tend to affect prelaying survival. Follicle ablation increases investment per offspring, but total investment in the clutch is reduced. The "energy" surplus from such ablation-induced reallocation is later acting on the female and seems to uniformly enhance survival after reproduction (Table 1, also see results by Landwer 1994 in which survival of females is enhanced to the next year).
Evolutionary Response to Natural Selection
From these experimental and comparative measurements of natural selection, we would predict an evolutionary response to selection from 1991 to 1992 that arises because (1) females with large clutch mass are more likely to survive, and their phenotype will be overly represented in the next year; and (2) these females are also more likely to produce more offspring that mature the following season (because the females survive to produce more offspring on clutches 2-4). These predictions are completely consistent with a large and significant increase in clutch mass observed between 1991 and 1992 that is mainly due to a large increase in egg size [ILLUSTRATION FOR FIGURES 4-5 OMITTED].
From experimental and comparative measurements of natural selection acting on midseason female survival in 1992 (e.g., after production of the first clutch, see above) and early season survival 1993 (e.g., before production of the first clutch), we would predict an evolutionary response to selection from 1992 to 1993 that arises because (1) females that produce large clutch mass in 1992 are less likely to survive to the next season and are underrepresented in the 2-yr-old population, and (2) these females also produce fewer offspring that mature the following season because they contribute fewer offspring on clutches 2-4 (e.g., clutch size is heritable). Moreover, if maturing females in 1993 were genetically similar to females from 1992 in their clutch production (e.g., egg size or total clutch mass is heritable), then offspring of 1992 females that would have naturally laid a large clutch mass in 1993 may have been selectively eliminated by natural selection acting on these offspring during the prelaying period of 1993, prior to the production of their first clutch. Indeed, when we induced females to produce a high investment by enhancing exogenous corticosterone and FSH, we also induced high mortality during prelaying phases, prior to the production of the first clutch. These predictions from natural variation and confirmation from experimental variation are consistent with a large and significant decrease in average clutch mass [ILLUSTRATION FOR FIGURE 4 OMITTED] and egg size [ILLUSTRATION FOR FIGURE 5 OMITTED] in the population that was observed between 1992 and 1993, an evolutionary response to natural selection.
It is noteworthy that we observed a fourfold increase in the activity of predatory snakes between 1992 and 1993 that is associated with the end of a severe drought [ILLUSTRATION FOR FIGURE 6 OMITTED]. Prior to 1993, very few hatchling snakes and very few adult snakes were observed on our study site. Presumably, the high rainfall during the spring of 1992 resulted in the much higher recruitment of juvenile snakes. Experimentally inducing a large reproductive investment in utas during periods with naturally high pre-laying mortality and a rise in the numbers of natural predators (1993) exacerbates costs of reproduction, but this cost of reproduction is not seen during periods with low prelaying mortality (1992) (Tables 1-2). Conversely, during the latest stages of the drought when snake abundance was low, we observe a gradual increase in both egg size and total clutch mass (1989-1992). This rise in clutch size is due to selection that favors survival of females that produce large clutches (1991, this study) and due to selection on egg size per se that favors females that produce large eggs (see Sinervo et al. 1992 for data on offspring survival selection during 1989-1991, data on 1991-1994 is unpubl.).
Consistent results from two experimental manipulations (exogenous corticosterone and FSH) in 1993, suggests that our experimental protocols are unlikely to result from non-specific hormonal effects unrelated to costs of reproduction or effects that might arise from surgical manipulation. Survival of females receiving sham implants were comparable to unmanipulated females (Table 1). In addition, the manipulation of clutch mass by corticosterone enhanced survival in 1991, a year in which females that naturally laid large clutch mass had high survival.
Natural Selection on Costs of Reproduction in Free-Ranging Lizards
Regardless of the precise ecological mechanisms and agents of natural selection (e.g., snake predation, drought, density, etc.), we have strong support for the existence of natural selection on reproductive investment in side-blotched lizards based on (1) the correlations between current reproductive investment and future reproductive success that are estimated in the absence of several potential confounding environmental effects, (2) corroborative evidence from our mechanistic assessment of the trade-off, and (3) year-to-year changes in clutch mass and egg size consistent with the predicted evolutionary response to natural selection. Although the severe drought that California experienced during the late 1980s until 1991 was a unique event, and it may not be possible to repeat our serendipitous observations even in the event of another drought, it is entirely likely that the periodic droughts on time scales measured in decades would have similar effects on microevolutionary change in side-blotched lizards.
Why is there no selection observed on natural variation in 1993 and 1994, years with the highest snake activity on site? This observation may be quite informative with regards to the likelihood of detecting costs of reproduction in other systems. The very early natural selection described during 1993 (e.g., prelaying mortality) eliminated a large fraction of the phenotypic variation that was present in 1992. Thus, phenotypic variation was high and opportunity for selection was high (variance in relative fitness, Arnold and Wade 1984) in both 1992 and early in 1993. However, those behavioral phenotypes that were at risk of mortality were eliminated later in 1993, and thus the opportunity for selection after 1993 and 1994 [ILLUSTRATION FOR FIGURES 4-5 OMITTED] is reduced.
We can detect selection acting on females in our experimental amelioration of costs in 1993 by follicle ablation (Table 1). In this case, we generate additional phenotypic variation for selection to act on by experimentally augmenting the variance in total clutch mass. Thus, in order for selection to be observed there must be (1) sufficient phenotypic variation, (2) differential survival arising from such phenotypic variation, and 3) genetically based variance underlying these phenotypic effects (Brandon 1990). Likewise, the experimental reduction in clutch mass in 1992 generates variance in total clutch mass that can be operated on by the agents of natural selection. In 1992, there was also a large natural variance in total clutch mass [ILLUSTRATION FOR FIGURE 5 OMITTED] that allowed us to detect selection against females that laid the largest clutch mass.
If we had performed experimental manipulations of reproductive traits and not measured selection on natural variation, we might mistakenly conclude that selection was operating on phenotypic variation in 1993 and 1994 (we can detect it with follicle ablation). However, because there is little opportunity for selection after the acute early season natural selection of 1993, this conclusion that is merely derived from experimental data would be flawed. Likewise, we caution other researchers from extrapolating patterns of selection and evolutionary change from purely experimental manipulations. Comparative data are essential for making such inferences.
Detecting the Ghost of Selection
Most of the dramatic selection observed in our population occurred during the early spring of 1993 and such costs largely arise from "prelaying selection" on current reproduction, not "postlaying selection" on current reproduction. Classic manipulations of postlaying clutch size (e.g., in birds) that ignore such prelaying costs of reproduction may underestimate the magnitude of costs in natural populations. If such prereproductive investment, however small, occurs during times of high predatory activity or during times of food shortage, then selection could exert a tremendous impact on the genetic structure of the population.
The experimental protocols that we describe are an advance over the single-trait clutch or litter size manipulations used in avian (Lack 1954; Hogstedt 1980; Nur 1984a,b, 1986, 1988; Lima 1987; Gustafsson and Sutherland 1988) and mammalian studies (Fleming and Rauscher 1978; Boutin et al. 1988). We manipulate clutch size and egg size, the focal life-history traits that determine total clutch mass, by hormonal and surgical manipulations. Our manipulations of the clutch and egg size trade-off encompass the suite of physiological costs not necessarily considered when one simply performs a single-trait manipulation (e.g., adding or removing eggs from a clutch in organisms with parental care). Our multitrait manipulations involve identifiable hormonal and physiological mechanisms (Sinervo and Licht 1991a,b) that appear to reflect underlying genetic trade-offs (Sinervo and Doughty 1996).
In addition, estimates of costs that are based on purely natural phenotypic correlations may also underestimate costs in nature because females that commit to a large reproductive investment during the earliest stages of reproduction may die before it is possible to measure their projected levels of current reproductive investment. For example, in 1993, we have evidence of stringent natural selection that removed three-quarters of the phenotypic variation in the population compared with the previous year 1992 [ILLUSTRATION FOR FIGURE 5 OMITTED]. However, such selection would not be apparent in natural variation because females producing a large total clutch mass were eliminated before they produced their first clutch. We have shown how such costs, which are difficult to detect by analysis of phenotypic data, can be readily detected with carefully designed experimental manipulations. Moreover, costs of reproduction estimated from phenotypic correlations of natural variation are limited because environmental effects may mask the trade-off (Hogstedt 1980). In addition, a purely experimental approach in which an attempt is made to randomize such environmental effects, while successfully controlling for what may be important confounding effects, also precludes measurement of salient environmental effects on reproduction (e.g., density) or putative agents of selection (e.g., predation). In this regard, a combination of path analysis that provides hypotheses of cause and effect, along with manipulations that experimentally verify these hypotheses, is an ideal methodology for assessing costs in nature. The next step is to merge this approach with manipulations of the environment to determine agents of selection. For example, we can only speculate as to the causes of selection in 1993. Our natural-history observations suggest snake predation may have a major role in influencing costs. Snake-exclusion experiments might verify the causal role of snakes as agents of selection, and density manipulations might verify the role of density as an environmental modulator of costs of reproduction.
On Measuring Costs of Reproduction in the Wild
Reznick (1985, 1992) has partitioned measurements of costs into four categories (1) phenotypic correlations between an index of reproductive investment and some cost such as decreased growth, survival, or fecundity; (2) experimental manipulations of current reproduction and the resultant costs; (3) genetic correlations between an index of reproductive investment and some cost; and (4) selection experiments in which selection on some index of reproductive investment results in a correlated response in other traits that is reflected as a cost of reproduction. Reznick (1985, 1992) concluded that only genetic correlations and selection experiments provide the essential ingredient necessary to address theoretical studies of life-history evolution. Only these methods establish that the costs of reproduction are causally related to genetically based trade-offs. Costs of reproduction reflected in phenotypic correlations are composed of genetic effects, environmental effects, and genotype x environment effects that may mask genetically based trade-offs (Reznick 1985, 1992). Reznick (1985) also concludes that results from experimental manipulations measure plasticity in the life-history traits and thus, may be more related to genotype x environment interactions than genetically-based trade-offs. For example, levels of parental care are readily modified by parental behaviors, and such phenotypic plasticity may overwhelm the expression of important genetically based tradeoffs involving costs of reproduction (Reznick 1985; Linden and Moller 1989) or overwhelm the detection of costs of reproduction from simple manipulations of clutch size achieved by adding or removing eggs from a nest. These views have lead to a spirited debate (Partridge 1992; Reznick 1985, 1992) concerning the relevance of experimental manipulations in addressing costs of reproduction.
Although we agree in a large part with the views presented by Reznick (1985, 1992) (e.g., showing the existence of genetically based trade-offs), we also believe there are different classes of experiments. For example, an experimental manipulation in which individuals are not allowed to reproduce (Partridge 1992; Reznick 1985, 1992) is in fact an environmental manipulation that has physiological ramifications - not a direct manipulation of the organism's physiology. Such "environmental" manipulations have great utility in detecting the phenotypic consequences of such effects on fitness. We propose that experimental manipulations of life-history traits be partitioned into the following classes.
(1) One can experimentally manipulate the environment and thus the life history to see whether a change in one trait (resulting from the environmental manipulation) results in a correlated change in another trait. For example, Hirshfield (1980) manipulated food availability and temperature (three levels of each) in growing rice-paddy fish and estimated the correlated effects on reproductive investment and cascading effects on survival. Clearly, these results can be influenced by genotype x environment effects and thus are susceptible to Reznick's (1985) basic criticism - plasticity resulting from genotype x environment effects does not necessarily demonstrate how correlated traits might respond to selection.
(2) One can hold the environment constant and instead experimentally manipulate the mean or variance of a single trait under investigation (Mitchell-Olds and Shaw 1987; Schluter 1988; Wade and Kalisz 1990). Examples of this approach involve adjusting clutch size (references above), tail length in widowbirds (Andersson 1982), or offspring size by yolk removal from eggs (Sinervo and McEdward 1988; Sinervo 1990; Sinervo and Huey 1990). Such manipulations help verify causal links between traits and fitness. Although these manipulations have great utility in establishing the causal relationships between a single trait and other components of fitness (Endler 1986; Mitchell-Olds and Shaw 1987; Schluter 1988; Wade and Kalisz 1990), they don't necessarily lead us to inferences about the causality between correlated traits involved in genetically based trade-offs. As Bernardo (1991) and Reznick (1992) pointed out, the methods of allometric engineering have great utility in establishing the causal relationship (or lack of a relationship) between maternal investment per offspring and offspring performance and thus fitness. Reznick placed certain limits on the use of such techniques to understand life-history trade-offs. Such single-trait manipulations are useful in understanding the effects of one trait on other fitness effects, particularly traits down-stream of the manipulation.
(3) One can experimentally manipulate a single physiological mechanism that links two or more traits and study simultaneous effects on both traits. If the aim of experiments is to establish multitrait causal relationships (e.g., trade-offs), manipulation must occur upstream of the cause-and-effect pathways that link physiological and hormonal traits involved in reproduction. This type of manipulation maintains the essence of any physiologically based genetic correlations. We have also begun studies that demonstrate that such multitrait "mechanistic manipulations" are also related to underlying genetic correlations between life-history traits involved in reproductive trade-offs. For example, the experimentally induced trade-off achieved by manipulation of clutch size and cascading effects on egg size are also reflected in a genetic correlation between these traits in free-ranging side-blotched lizards (Sinervo and Doughty, in press).
The experimental manipulations of clutch size, egg size, and total clutch mass address the causal basis of the physiologically based trade-off between clutch and egg size [ILLUSTRATION FOR FIGURE 2 OMITTED]. The trade-off between offspring size and offspring number is also described in terms of the evolutionary trade-off that involves the conflicting selective pressures: fecundity selection among females and survival selection among their offspring (Lack 1954). A female that lays a large clutch of small eggs might be expected to leave more surviving offspring than a female that lays a smaller clutch of larger eggs. However, fecundity selection in females that favors the production of small eggs is counterbalanced by the potentially enhanced survival of large offspring and thus the production of large eggs. The components involved in the physiological (Sinervo and Licht 1991a,b) and evolutionary trade-offs (Sinervo et al. 1992) can be tested separately or combined into a comprehensive test. Different protocols are required to test these various heirarchical levels of the trade-off. Survival selection acting on offspring size (Sinervo et al. 1992) presumably acts in part through the correlations between offspring size and performance traits more closely related to fitness (Arnold 1987). Such effects can be assessed experimentally by single-trait manipulations (Sinervo 1990; Sinervo and Huey 1990; Sinervo et al. 1992). Thus, one component of the evolutionary trade-off, survival selection, is readily amenable to single-trait manipulations that involve removing yolk from freshly laid eggs.
However, the physiological trade-off involved in the second component of the evolutionary trade-off, fecundity selection, requires more elaborate protocols that manipulate two traits simultaneously (Sinervo and Licht 1991a,b). Sinervo et al. (1992) in their manipulation of both fecundity of the female parent (e.g., direct ovarian manipulation) and offspring size in freshly laid eggs merge such mechanistic multitrait manipulations and single-trait manipulations to understand life-history trade-offs. Moreover, understanding the second major paradigm of life-history theory, costs of reproduction, entails even more elaborate protocols (this study).
Evolution of Clutch Mass and Egg Size
In the current study, we have also not ruled out the possibility that year-to-year variation in egg production are phenotypically plastic. Indeed our corticosterone manipulation provides a putative mechanism underlying phenotypic plasticity in egg size and total clutch mass. Certainly, the experimentally induced changes in reproductive investment (e.g., corticosterone and its enhancement of female offspring size) would warrant caution with regards to inferring year-to-year changes as purely the result of evolved changes. However, given the demonstrated heritable variation in egg size (Sinervo and Doughty, in press), an evolutionary response to natural selection on reproduction remains the most parsimonious explanation for year-to-year patterns of change.
This study raises a number of questions concerning the genetic basis of observed changes in life-history variation of side-blotched lizards from 1989 to 1994. How is it possible that natural selection can shift the population mean so rapidly? The heritability of egg size as measured on free-ranging side-blotched lizards is 0.61 (data from 1989-1991, Sinervo and Doughty, in press). This heritability, based on parent-offspring regressions, was measured in free-ranging lizards and we used elaborate controls for maternal effects of egg size (e.g., size manipulation of daughters) and oviposition maternal effects (e.g., egg incubation conditions were controlled and offspring were randomly released with respect to the female parent's home range). Given the high heritability for egg size in the years preceeding the acute natural selection, we would expect to see a large evolutionary response to selection in the short term. In the long term, why is additive genetic variation not depleted under such strong selection (Falconer 1981)? The differences in clutch production and egg size seem to be related to color polymorphisms of females and males that we have recently discovered. Female and male color morphs are also associated with striking differences in physiology, behavior, and life history (Sinervo unpubl. data). Noteworthy in this regard is the striking bimodal distribution of clutch mass observed in females in 1992 [ILLUSTRATION FOR FIGURE 5 OMITTED]. If the life-history traits have a mendelian pattern of inheritance, substantial additive genetic variation can be maintained by frequency-dependent selection acting on the polymorphisms (Sinervo and Lively, 1996) in which morphs have an advantage when rare.
We wish to thank P. Statdler and R. Schrimp for generously giving us permission to work on their land, P. Doughty and T. Frankino for snake data, K. Allred, K. Zamudio, J. Graff, L. Lenart, C. Giorni for assistance with gravid females in the lab and field, and to P. Niewiarowski and D. Roff for constructive comments during the review process. This research was supported by National Science Foundation grants BSR 89-19600 and DEB 93-07999 to B.S. and an National Institutes of Health postdoctoral fellowship to D.D.
ANDERSSON, M. 1982. Female choice selects for extreme tail length in a widowbird. Nature 299:818-820.
ARNOLD, S. J. 1987. Genetic correlation and the evolution of physiology. Pp 189-211 in M. E. Feder, A. F. Bennett, W. W. Burggren and R. B. Huey, eds. New directions in ecological physiology. Cambridge Univ. Press, New York.
ARNOLD, S. J., AND M. J. WADE. 1984. On the measurement of natural and sexual selection: theory. Evolution 38:709-719.
BERNARDO, J. 1991. Manipulating egg size to study maternal effects on offspring traits. Trends Ecol. Evol. 6:36-37.
BOUTIN, S., R. A. MOSES, AND M. J. CALEY. 1988. The relationship between juvenile survival and litter size in wild muskrats (Ondatra zibethicus). J. Anim. Ecol. 57:455-462.
BRANDON, R. N. 1990. Adaptation and Environment. Princeton University Press, Princeton, NJ.
CALLARD, I. P., AND G. V. CALLARD. 1978. The adrenal gland in reptilia, part 2, physiology. Pp. 370-418 in I. C. Jones and I. W. Henderson, eds. General, comparative, and clinical endocrinology of the adrenal cortex. Academic Press, New York.
DENARDO, D. F., AND P. LICHT. 1993. Effects of corticosterone on social behavior of male lizards. Horm. Behav. 27:184-199.
DENARDO, D. F., AND B. SINERVO. 1994a. Effects of corticosterone on activity and territory size of free-ranging male lizards. Horm. Behav. 28:53-65.
-----. 1994b. Effects of steroid hormone interaction on territorial behavior of male lizards, Horm. Behav. 28:273-287
DERICKSON, K. W. 1976. Lipid storage and utilization in reptiles. Am. Zool. 16:711-723.
DOUGHTY, P., AND B. SINERVO. 1994. The effects of habitat, time of hatching, and body size on dispersal in Uta stansburiana. J. Herpetol. 28:485-490.
DOUGHTY, P., G. BURGHARDT, AND B. SINERVO. 1994. Dispersal in the side-blotched lizard: A test of the mate-defense dispersal theory. Anim. Behav. 47:227-229.
ENDLER, J. A. 1986. Natural selection in the wild. Princeton University Press, Princeton, NJ.
FALCONER, D. S. 1981. Introduction to quantitative genetics, Longman, New York.
FLEMING, T. H., AND R. J. RAUSCHER. 1978. On the evolution of litter size in Peromyscus leucopus. Evolution 32:45-55.
GADGIL, M., AND W. H. BOSSERT. 1970. Life historical consequences of natural selection. Am. Nat. 104:1-24.
GUSTAFSSON, L., AND W. J. SUTHERLAND. 1988. The costs of reproduction in collared flycatcher Ficedula albicollis. Nature 335:813-815.
HIRSHFIELD, M. F. 1980. An experimental analysis of reproductive effort and cost in the Japanese Medaka Oryzias latipes. Ecology 61:282-292.
HOGSTEDT, G. 1980. Evolution of clutch size in birds: An adaptive variation in relation to territory quality. Science 210:1148-1150.
JOHNSON, E. O., T. C. KAMILARIS, G. P. CHROUSOS AND P. W. GOLD. 1992. Mechanisms of stress: A dynamic overview of hormonal and behavioral homeostasis. Neurosci. Biobehav. Rev. 16:115-130.
JONES, R. E. 1978. Control of follicular selection, Pp. 763-788 in R. E. Jones, ed. The vertebrate ovary. Plenum Press, New York.
LACK, D. 1947. The significance of clutch size. Ibis 89:302-352.
-----. 1954. The natural regulation of animal numbers. Clarendon Press, Oxford.
LANDWER, A. J. 1994. Manipulation of egg production reveals costs of reproduction in the tree lizard (Urosaurus ornatus). Oeocologia 100:243-249.
LI, C. C. 1975. Path analysis - A primer. Boxwood Press, Pacific Grove, CA
LIMA, S. L. 1987. Clutch size in birds: A predation perspective. Ecology 68:1062-1070.
LINDEN, M., AND A. P. MOLLER, 1989. Costs of Reproduction and covariation in life history traits in birds. Trends Ecol. Evol. 4:367-371.
MITCHELL, R. J. 1993. Path analysis: pollination. Pp. 211-231 in S. M. Scheiner and J. Gurevitch, eds. Design and analysis of ecological experiments. Chapman and Hall, London.
MITCHELL-OLDS, T., AND R. G. SHAW. 1987. Regression analysis of natural selection: Statistical and biological interpretation. Evolution 41:1149-1161.
NAGY, K. A. 1983. Ecological energetics. Pp. 24-54 in R. B. Huey, E. R. Pianka and T. W. Schoener, eds. Lizard ecology: Studies of a model organism. Harvard Univ. Press, Cambridge, MA.
NUR, N. 1984a. The consequences of brood size for breeding blue tits I. Adult survival, weight change and the cost of reproduction. J. Anim. Ecol. 53:479-496.
----. 1984b. The consequences of brood size for breeding blue tits II. Nestling weight, offspring survival and optimal brood size, J. Anita. Ecol. 53:497-517.
-----. 1986. Is clutch size variation in the blue tit (Parus caeruleus) adaptive? An experimental study. J. Anim. Ecol. 55:983-999.
-----. 1988. The costs of reproduction in birds: An examination of the evidence. Ardea 76:155-168.
PARTRIDGE, L. 1992. Measuring reproductive costs. Trends Ecol. Evol. 7:99-100.
PARTRIDGE, L., AND P. H. HARVEY. 1985. Costs of reproduction. Nature 316:20-21.
PEASE, C. M., AND J. J. BULL. 1988. A critique of methods for measuring life history trade-offs. J. Evol. Biol. 1:293-303.
REZNICK, D. 1985. Costs of reproduction: An evaluation of the empirical evidence. Oikos 44:257-267.
-----. 1992. Measuring costs of reproduction. Trends Ecol. Evol. 7:42-45.
SCHLUTER, D. 1988, Estimating the form of natural selection on a quantitative trait. Evolution 42:849-861.
SINERVO, B. 1990. The evolution of maternal investment in lizards: An experimental and comparative analysis of egg size and its effects on offspring performance. Evolution 44:279-294.
-----. 1993. The effect of offspring size on physiology and life history: Manipulation of size using allometric engineering. Bioscience 43:210-218.
-----. 1994. Experimental manipulations of clutch and egg size of lizards: Mechanistic, evolutionary, and conservation aspects. Pp. 183-193 in J. B. Murphy, K. Adler, and J. T. Collins, eds. Captive management and conservation of amphibians and reptiles. Contrib. Herpetol. No. 11. Society for the Study of Amphibians and Reptiles, Ithaca, NY.
SINERVO, B., AND P. DOUGHTY. 1996. Interactive effects of offspring size and timing of reproduction on offspring reproduction: Experimental, maternal, and quantitative genetic aspects. Evolution 50:1314-1327.
SINERVO, B., AND R. B. HUEY. 1990. Allometric engineering: An experimental test of the causes of interpopulational differences in locomotor performance. Science 248:1106-1109.
SINERVO, B., AND P. LICHT. 1991a. The physiological and hormonal control of clutch size, egg size, and egg shape in Uta stansburiana: Constraints on the evolution of lizard life histories. J. Exp. Ecol. 257:252-264.
-----. 1991b. Proximate constraints on the evolution of egg size, egg number, and total clutch mass in lizards. Science 252:1300-1302.
SINERVO, B., AND C. LIVELY. 1996. The rock-paper-scissors game and the evolution of alternative male strategies. Nature 380:240-243.
SINERVO, B., AND L. R. McEDWARD. 1988. Developmental consequences of an evolutionary change in egg size: An experimental test. Evolution 42:885-899.
SINERVO, B., R. HEDGES AND S. C. ADOLPH. 1991. Decreased sprint speed as a cost of reproduction in the lizard Sceloporus occidentalis: Variation among populations. J. Exp. Biol. 155:323-336.
SINERVO, B., P. DOUGHTY, R. B. HUEY, AND K. ZAMUDIO. 1992. Allometric engineering: A causal analysis of natural selection on offspring size. Science 258:1927-1930.
STEARNS, S. C. 1989. The evolutionary significance of phenotypic plasticity. Bioscience 39:436-445.
VITT, L. J. 1981. Lizard reproduction: Habitat specificity and constraints on relative clutch mass. Am. Nat. 117:506-514.
VITT, L. J., AND J. D. CONGDON. 1978. Body shape, reproductive effort, and relative clutch mass in lizards: Resolution of a paradox. Am. Nat. 112:595-608.
VITT, L. J., AND H. J. PRICE. 1982. Ecological and evolutionary determinants of relative clutch mass in lizards. Herpetologica 38:237-255.
WADE, M. J., AND S. KALISZ. 1990. The causes of natural selection. Evolution 44:1947-1955.
WILLIAMS, G. C. 1966. Natural selection, the costs of reproduction, and a refinement of Lack's principle. Am. Nat. 100:687-690.
WILSON, B. S., AND J. C. WINGFIELD. 1992. Correlation between female reproductive condition and plasma corticosterone in the lizard Uta stansburiana. Copeia 1992:691-697.
|Printer friendly Cite/link Email Feedback|
|Author:||Sinervo, Barry; DeNardo, Dale F.|
|Date:||Jun 1, 1996|
|Previous Article:||Evolutionary potential and local genetic differentiation in a phenotypically plastic trait of a cyclical parthenogen, Daphnia magna.|
|Next Article:||Interactive effects of offspring size and timing of reproduction on offspring reproduction: experimental, maternal, and quantitative genetic aspects.|