Copeia 2009, No. 2, 328–338
Density Dependence in the Terrestrial Stage of Wood Frogs: Evidence from a 21-Year Population Study Keith A. Berven1 Population regulation in pond-breeding amphibians has generally been assumed to occur through density dependence in the aquatic larval stage. However, studies examining terrestrial stage population dynamics are comparatively rare. As a consequence, the relative importance of density dependence in the terrestrial stage remains largely unknown. Here I determine the density dependence of terrestrial stage vital rates, including juvenile and adult survival, age and size at first reproduction, and reproductive traits, for a population of Wood Frogs, Rana sylvatica, in Michigan. I also examine the carry-over effects of density-dependent variation in metamorphic traits on adult demographic traits. During the 21 years of this study, the number of breeding adult males and females varied by an order of magnitude, whereas the number of metamorphic juveniles leaving the pond varied by two orders of magnitude. Adult male and female annual survival was negatively correlated with the number of males and females, respectively. Partial correlation coefficients (holding population size of the opposite sex, number of juveniles, and precipitation levels constant) indicated that variation in adult male and female survival was largely explained by variation in the number of males and females, respectively. Juvenile survival (both males and females) was strongly negatively correlated with juvenile population size but not juvenile body size, adult population size, or total precipitation. Partial correlations revealed that variation in juvenile population size was the single most important factor accounting for variation in male and female juvenile survival, and female age and size at first reproduction. Clutch size and egg size also varied with juvenile population size independent of female body size. At lower juvenile population sizes, female juveniles matured earlier, at a larger body size, and produced larger numbers of smaller eggs than at high juvenile population sizes. Although juvenile body size was positively correlated with male and female reproductive body size and total clutch weight, juvenile fitness traits, in general, were more closely related to variation in juvenile population sizes. Larval traits did not affect adult fitness traits in the terrestrial stage. A nonlinear logistic model best described the functional relationship between juvenile population size and the number of surviving males, females, and total adult biomass. As number of juveniles produced increased, the number of surviving adult males, females, and total biomass plateaued, suggesting that the terrestrial environment limited adult population size. My results demonstrate that density-dependent effects operating in the terrestrial stage may be important in regulating Wood Frog population size.
I
MPLICIT in the concept of population regulation is density dependence. Population regulation is thought to occur when the density of conspecifics limits population growth by decreasing the growth, survival, and/or fecundity of individuals. In addition, density dependence must occur within populations rather than merely among populations (Turchin, 1999). For species with complex life cycles, density-dependent processes can occur in one or multiple stages of their life cycle (Wilbur, 1980). In addition, density dependence operating in one stage may carry over and affect fitness traits in another stage (Prout and McChesney, 1985; Altwegg, 2003). Thus, determining the mechanism of population regulation in species with complex life cycles requires an understanding of how density affects different stages in the life cycle and how density-dependent effects within each stage of the life cycle interact among stages. Most pond-breeding amphibians have a complex life cycle consisting of an aquatic larval stage followed by a terrestrial juvenile and adult stage. Amphibians have long been recognized as an ideal group for investigating population regulation in general (Wilbur, 1980), as well as for investigating single stage vs. multistage population regulation (Wilbur, 1996; Hellriegel, 2000). The majority of amphibian population studies, however, have focused on the larval stage (Van Buskirk and Smith, 1991; Beebee et al., 1996; Loman, 2002). A large body of research performed in a variety of venues, including lab, field, and mesocosm studies (summarized in Skelly and Kiesecker, 2001) has demonstrated strong effects of larval density on larval growth rates,
development times, and survival. The near unanimity of these studies has led to the general consensus that pondbreeding amphibians are regulated by density-dependent processes operating in the aquatic larval stage. In addition, density-dependent variation in larval metamorphic traits may persist after metamorphosis and continue to exert strong effects on population dynamics and regulation through its effects on adult traits (Scott, 1994; Morey and Reznick, 2001). Studies of terrestrial stage population dynamics are comparatively rare. However, a few experimental studies (Pechmann, 1994; Altwegg, 2003; Harper and Semlitsch, 2007) and population studies (Berven, 1990, 1995) have demonstrated negative effects of terrestrial density on juvenile survival, growth, and reproductive development. These studies caution against the prevailing view that population regulation in pond-breeding amphibians is largely single-staged and suggest that density effects in the terrestrial stage may also play an important role in amphibian population regulation. Several aspects of amphibian ecology suggest that terrestrial stage density dependence could be important in limiting population size. For most pond-breeding amphibians, breeding sites are spatially discrete and juvenile dispersal rates among breeding sites is low (and may vary with population size), while adult fidelity to breeding sites is high (Smith and Green, 2005). Physiological restrictions associated with a permeable skin are known to limit the suitability of much of the terrestrial habitat surrounding breeding sites (Wyman, 1988). Additionally, larval meta-
1 Department of Biological Sciences, Oakland University, Rochester, Michigan 48309; E-mail:
[email protected]. Submitted: 24 March 2008. Accepted: 11 December 2008. Associate Editor: M. J. Lannoo. DOI: 10.1643/CH-08-052 F 2009 by the American Society of Ichthyologists and Herpetologists
Berven—Terrestrial stage density dependence
morphic rates often produce juvenile population sizes that vastly exceed replacement rates (Berven, 1995). As a consequence, juvenile and adult densities could exceed the habitat and resource limitations of the terrestrial environment. Under these situations density dependence may limit population size. Theoretical studies have also demonstrated the potential importance of the terrestrial stage in amphibian population dynamics. For species such as pond-breeding amphibians that use different habitats at different stages of their life cycle, adult habitat size can strongly limit population size (Halpern et al., 2005). Mathematical models incorporating amphibian demographic data have shown that population growth is more affected by reductions in juvenile and adult survival than similar reductions in larval survival (Biek et al., 2002; Vonesh and De la Cruz, 2002). Incorporating terrestrial stage density dependence into population growth models based on single-stage regulation in the aquatic stage has been shown to produce dramatic and unpredictable population dynamics (Hellriegel, 2000). These theoretical studies stress the need for a better understanding of the functional relationships between terrestrial density and vital rates in natural populations of amphibians. Few studies have documented terrestrial stage vital rates of natural populations and determined whether they vary with population size. To date, our understanding of terrestrial stage demographics is based on short-term, caged field studies whose relevance to natural populations has yet to be demonstrated. Theoretical studies have relied on a compilation of vital rates derived from a variety of different species (Biek et al., 2002). Long-term population studies of the terrestrial phase for any amphibian species are remarkably scarce. As a consequence, the relative importance of density dependence in the terrestrial stage remains largely unknown. Here, I present the results of a 21-year population study (1985–2005) of the Wood Frog, Rana sylvatica, from a single population in Michigan. First, I summarize the observed yearly variation in terrestrial vital rates, including juvenile survival, age and size at first reproduction, reproductive traits (egg size and number), and adult survival. I then relate variation in vital rates to variation in population size to test for density dependence. I also examine the potential impact of density-dependent variation in metamorphic traits on terrestrial stage vital rates. Finally, I fit the observed variation in juvenile population size to a logistic growth model to determine the functional relationship between juvenile production and adult population size, and determine if density-dependent processes operating in the terrestrial stage are compensatory. MATERIALS AND METHODS Study site.—I studied the population dynamics of Wood Frogs from 1985–2005 in the Saginaw Forest of the University of Michigan. Saginaw Forest is a natural preserve located approximately 1.5 km west of Ann Arbor, MI (Sec. 26 T25 R5E, Washtenaw County, MI; 42u409N, 83u139W). The preserve is composed of 32 hectares (400 3 800 m) that is bisected at its northern boundary by a large lake (320 3 160 m). Immediately adjoining the preserve to the north is an additional 14 hectares of largely undeveloped woodland habitat. The two areas together constitute a forested island
329
surrounded on four sides by grassy fields, and both commercial and housing developments. Wood Frog breeding activity on the preserve occurs primarily in a semi-permanent pond, roughly elliptical in shape with an average surface area in spring of 2912 m2 and a maximum depth of 1.5 m (Berven and Boltz, 2001). The pond is located 75 m south of the lake and is surrounded by a mesic oak–hickory forest remnant with interspersed historical plantings of other species including Norway maple, spruce, and red pine. The majority (.90%) of adult and juvenile frogs enters and leaves the pond along its southern edge. In addition to the study pond, there are two additional small ponds north of the lake on the adjoining property within 500 and 650 m of my study pond. These ponds maintained relatively small breeding populations of Wood Frogs. The lake and surrounding altered landscapes appear to be significant barriers to dispersal. A previous study of Wood Frogs conducted on the Saginaw Forest between 1980–1983 demonstrated little dispersal (0.15%) of juveniles and adults between my study pond and these areas (Howard and Kluge, 1985). Field procedures.—The pond was completely surrounded by a 75-cm tall fence constructed of aluminum window screening buried to a depth of 20 cm and supported by wooden stakes. The fence was placed approximately 10 m from the pond’s average high water mark. The top of the fence had a baffle (aluminum screening) that extended 20 cm on either side of the fence. I buried buckets (12 L and 9 L) along the inside and outside of the fence to facilitate capture of adults and juveniles. I also inserted plywood boards (0.75 m2) every 10 m along the fence. I removed the boards and covered the buckets at the end of the adult breeding migrations, and again at the end of the juvenile emergence. This allowed all amphibian species to move freely in and out of the pond, and eliminated the need to monitor the fence during periods of low Wood Frog activity. The boards were replaced in late November in anticipation of the spring breeding migrations. I monitored the fence daily during the adult breeding migrations (1 March–30 April) and juvenile emergence (1 June–28 July). All adult frogs collected in the field were transported to my laboratory at Oakland University. I sorted the frogs by sex and placed them in pans filled with tap water. The pans were stored in a cold room maintained at 3uC. I changed the water in each pan every other day. The frogs were typically returned to the pond within one week of their arrival in roughly the order and proportion in which they arrived. Juveniles captured at the fence in the summer were counted and released outside the fence. Marking procedure.—I marked all adult frogs and a subsample of juveniles as cohorts by toe clipping. I marked adult frogs by removing the distal phalange (approximately 2 mm in length) from either the left or right rear third, fourth, or fifth toe. Only one distal phalange was removed each year. The distal phalange removed alternated between the right (even years) and left (odd years) foot beginning with the fourth, followed by the fifth, and finally the third toe. This procedure allowed me to follow a cohort of marked frogs for up to seven years (which exceeded Wood Frog longevity in this population). The procedure was performed on frogs that were maintained in ice water (approx. 2uC),
330
and the frogs displayed no visible response to the procedure. I kept the frogs in the lab for one day following the marking procedure before returning them to the pond. I marked juvenile frogs by removing a single toe from the left front foot (excluding the thumb). The toe removed identified the year of metamorphosis and alternated among years until repeating every fourth year. Juvenile frogs were marked in the field throughout the emergence period and immediately released. With the exception of the first year when I marked 57.6% of the juveniles emerging from the pond (18,138 individuals), I marked an average of 3627 6 572 or 17.9 6 2.9% of the juvenile population each year. I assumed a 1:1 sex ratio (Berven, 1990). Juvenile frogs subsequently recaptured as adults, were ‘cohort’ marked as described above, but because of the ‘juvenile’ clip they could be distinguished from frogs initially marked as adults. Caged studies of marked and unmarked adults and juveniles demonstrated that the frogs did not regenerate phalanges and that the toe-clipping procedure did not affect growth or survival relative to unmarked frogs (unpubl. data). Continuous monitoring and marking of adults and juveniles provided measures of juvenile survival, age and size at first reproduction, and age-specific adult survival. Juvenile and adult survival was estimated from recaptures of 16 marked juvenile cohorts (1985–1990; 1993–1999; 2001–2003) and 21 marked adult cohorts (1985–2005) in subsequent years. I measured body length of adults from the tip of the snout to the caudal end of the ischium (SI) to the nearest 0.5 mm using dial vernier calipers, and wet mass (to the nearest 0.1 g) using an electronic balance. Juvenile wet weight (to the nearest 0.01 g) was determined weekly (at the pond) during the juvenile emergence period using a portable electronic balance. Juvenile metamorphic size was based on an average sample of approximately 100 juveniles each year. Each year I determined the reproductive characteristics of different aged females. I measured egg diameter (to the nearest 0.001 mm) from a sample of ten unfertilized eggs (obtained by gently squeezing the abdomen) using a dissecting scope fitted with an ocular micrometer. Each female was then paired with a male, placed in a pan of water, and allowed to deposit their eggs. After egg deposition, I reweighed each female. I determined total egg weight as the difference between initial weight and final weight (after egg deposition). I determined clutch size by direct egg count. After the eggs hatched, they were returned to the pond. Because reproductive rates in this species correlate with female body size (Howard and Kluge, 1985; Berven, 1988), I used ANCOVA with body size as a covariate to determine body size adjusted mean clutch size, egg size, and total egg weight for different aged females, and for 2-yr-old females among years. Statistical analysis.—I used simple (r) correlations to test the hypotheses that juvenile population size negatively affected male and female survival, age and size at first reproduction, and size adjusted mean reproductive traits. Simple correlations were also used to test the hypothesis that adult population size negatively affected adult male and female annual survival. I used multiple (R) correlations to control for the effects of other variables including metamorphic juvenile body size, larval period length, precipitation, and population sizes of the non-target group on fitness traits. Adult and juvenile population size was the initial number of juveniles or adults the preceding year. Total precipitation
Copeia 2009, No. 2
Fig. 1. Number of breeding adult males (open) and females (closed) in Saginaw Pond for the period 1985–2005.
was determined as the amount of rainfall for the intervening period. Precipitation during the months of November– March was excluded. Precipitation and temperature data were obtained from the United States weather station in Ann Arbor, MI, USA. All correlations used mean cohort values between 1985 and 2005. Annual adult male and female survival was based on the annual survival of pooled age cohorts. I fit the relationship between initial juvenile number and the number of surviving adult males and females, and total biomass (from each juvenile cohort) to both linear (y 5 a*x + b) and theta logistic (y 5 x 2 (x2/a) b) models using Oakleaf 8.2 software. A previous study indicated that this model provided the best fit for similar data (Harper and Semlitsch, 2007). Total biomass for each juvenile cohort was estimated as the sum of the product of the total number of surviving males and females times their average body mass, respectively. Observations with standardized residuals greater than 6 2 SD were treated as outliers and were not included in the model. I performed all analyses using SPSS 11.5. All means are reported 6 SE unless otherwise noted. Proportional data and body weights were arcsine and log-transformed, respectively. Where the assumptions of parametric tests could not be met, I used nonparametric Kruskal-Wallis and Wilcoxon Sign Ranked statistical tests. RESULTS Breeding population size.—The number of breeding adult ¯ 5 3030 6 325; range: 594–6196; CV 5 49%) and males (X ¯ 5 1082 6 206; range: 169–4553; CV 5 87.3%) females (X varied dramatically during the 21 years of this study (Fig. 1). Adult males outnumbered females each year, although the ¯ 5 4.6 6 magnitude varied considerably (range: 2.11–7.7; X 1.02; median 5 2.7; Fig. 1). Larval period length and juvenile body size.—Juveniles were produced in 17 out of the 21 years during this study. The average length of the aquatic larval period (day eggs deposited to day 50% of the juveniles left the pond) ranged
Berven—Terrestrial stage density dependence
Fig. 2. Body size (g) of breeding adult males (closed) and females (open) aged 1–5 yrs. Each bar represents the mean body size (6 1 SE) for the period 1985–2005. Sample sizes are shown above each bar.
¯ 5 78.6 6 1.8 days; CV 5 from 68 to 94 days among years (X 9.5%; n 5 17). For years in which larval metamorphosis occurred, the number of juveniles leaving the pond varied ¯ 5 36,162 6 7865; median 5 19,931; range: dramatically (X 1,210–113,686). Average metamorphic juvenile body size ¯ 5 0.62 6 0.04 g; also varied significantly among years (X range: 0.25–0.83 g; CV 5 25.2%; Kruskal-Wallis, n 5 17, P 5 0.0001). Juvenile survival.—Juvenile survival to first reproduction varied considerably among juvenile cohorts. Male juvenile survival averaged three times higher than female survival ¯ 5 17.2 6 2.4%, range: 4.3–33.2%; female: X ¯ 5 5.5 (male: X 6 0.8%, range: 1.8–11.5%; n 5 16). The majority of male juveniles matured as either 1-yr olds (reproduced their first ¯ 5 57.5 6 7.2%) or as 2spring following metamorphosis: X yr olds (reproduced their second spring following metamor¯ 5 40.2 6 7.1%). In 8/16 years . 70% of male phosis: X juveniles matured as 1-yr olds, while in 5/16 years , 30% of male juveniles bred as 1-yr olds. In most years a small but consistent group of males did not breed until their third year ¯ 5 2.4 6 0.44%). (X Each year without exception, the majority of female ¯ 5 87.2 6 3.6%; range: 40.0– juveniles bred as 2-yr olds (X 98.0%; n 5 16). However, each year a small proportion of ¯ 5 6.7 6 female juveniles were able to mature as 1-yr olds (X ¯ 5 6.8 6 1.7%; 2.9%; range: 0.0–48%) or later as 3-yr olds (X range: 1.1–20%). Male body size (range: 4.6–6.2 g; CV 5 18.8%; F15,1408 5 26.3; P 5 0.0003) and female body size (range: 7.3–10.6 g; CV 5 24%; F15,632 5 15.7; P 5 0.0006) at first reproduction varied considerably among juvenile cohorts. In addition, average body size at first reproduction of both male and females increased with age (Fig. 2). However, 2-yr-old males breeding for the first time were smaller than similar aged ¯ 5 5.8 6 0.15 vs. 6.5 6 males breeding for the second time (X 0.23 g, Wilcoxon Sign Rank Test, P 5 0.007, n 5 16). Similarly, females breeding for the first time as 3-yr olds were smaller than 3-yr olds breeding for the second time; ¯ 5 9.5 6 however, the differences were not significant (X
331
Fig. 3. Annual adult survival (6 1 SE) for males (closed) and females (open) age 1–4 years for the period 1985–2005. Sample sizes are shown above each bar and represent the number of years for which data were available.
0.61 vs. 10.3 6 0.42 g, Wilcoxon Sign Rank Test, P 5 0.14, n 5 10). Adult survival.—Male and female annual survival did not differ (F1,97 5 0.78; P 5 0.38; Fig. 3) and annual survival did not differ among age classes (F2,97 5 0.93; P 5 0.43; Fig. 3). In addition, there was no interaction between sex and age (F2,97 5 0.05; P 5 0.95). The majority of male and females bred only once, and the proportion of males and females breeding multiple times (2–5 times) declined with age (Fig. 4). The maximum longevity observed in this study ¯ 5 was six years for both males and females. Adult male (X ¯ 5 17.8 19.4 6 2.3%; range: 2.0–44.9%; n 5 21) and female (X 6 2.6%; range: 4.4–50.9%; n 5 21) annual survival (pooling age cohorts) differed greatly among years. Reproductive traits.—Older females produced progressively larger clutches (F2,776 5 19.1, P 5 0.0008) of larger egg sizes (F2,767 5 1449.7, P 5 0.0003), which also had a larger total egg weight (F2,767 5 104.4, P 5 0.0003) than younger females (Fig. 5). This relationship was not surprising since reproductive traits in this population, as in others studied (Berven, 1988), are positively correlated with female body size (P , 0.001). However, when body size was treated as a covariate, the observed adjusted mean clutch size and egg size had an inverse relationship to each other across the age classes of females (ANCOVA: egg number: F2,772 5 77.8, P 5 0.0002; egg diameter: F2,764 5 1017.8, P 5 0.0004; total egg weight: F2,766 5 38.4, P 5 0.0001; Fig. 5). One-year-old females produced a greater number of smaller eggs, whereas 3-yr-old females produced fewer, larger eggs. Two-year-old females were intermediate in both egg size and egg number (Fig. 5). Two-year-old females produced the greatest adjusted mean total egg weight relative to 1-yr old and 3-yr olds, which did not differ (Fig. 5). Not surprisingly, year-to-year variation in adult female body size at first reproduction among juvenile cohorts resulted in significant differences in clutch size (F15,339 5 12.1, P 5 0.0007; range: 431–836; Fig. 6), egg size (F15,333 5 12.4, P 5 0.0002; range: 1.56–1.69 mm; Fig. 6), and total
332
Copeia 2009, No. 2
Fig. 4. Proportion (6 1 SE) of male (closed) and female (open) frogs breeding 1–5 times. Proportions are based on the recapture of 10,006 males (22%) and 2466 females (16%) for the period 1985–2005. Sample sizes are shown above each bar and represent the number of years for which data were available.
clutch weight (F15,350 5 10.9, P 5 0.0001; range: 2.14– 4.24 g; Fig. 6). However, when body size differences among female cohorts were adjusted, the differences in reproductive traits persisted (ANCOVA: egg number: F15,338 5 4.4, P 5 0.0001; range: 561–698; egg diameter: F15,332 5 16.7, P 5 0.0003; range: 1.56–1.69 mm; total egg weight: F15,349 5 4.6, P 5 0.0004; range: 2.63–3.44 g; Fig. 6), indicating that some of the variation in reproductive traits among juvenile cohorts resulted from factors other than body size. Correlates of adult survival.—Adult male and female annual survival was negatively correlated with the number of males and females, respectively (Table 1). Correlation coefficients between adult survival and population size of the opposite sex were low and not significant (Table 1). In addition, the survival of both males and females was not significantly correlated with total precipitation (Table 1). Partial correlation coefficients between adult population size and adult male and female annual survival (holding population size of the opposite sex, number of juveniles, and precipitation levels constant) were in close agreement with the zero order correlation coefficients (males: R 5 20.56, P 5 0.016, n 5 16; females: R 5 20.78, P 5 0.0008, n 5 16; Table 1). This indicates that variation in adult male and female survival was largely explained by variation in the number of males and females, respectively. Correlates of juvenile survival, age and size at first reproduction, and reproductive traits.—Both male and female juvenile survival was strongly negatively correlated with juvenile population size but not with juvenile body size, adult population size, or total precipitation (Table 2). Partial correlation coefficients (removing the effects of juvenile body size, larval period length, adult population size, and total precipitation) between juvenile population size and male and female juvenile survival were similar to zero order correlation coefficients (males: R 5 20.69, P 5 0.013, n 5 10; females: R 5 20.62, P 5 0.013, n 5 10; Table 2). Thus juvenile body size, adult population size, and precipitation
Fig. 5. Comparison of mean (6 1 SE) clutch size, egg size (mm), and total egg weight (g) of females breeding at 1–3 years of age. Solid bars represent unadjusted means; open bars represent means adjusted for body size (SI). See text for results of analysis of ANCOVA.
had little effect on male and female juvenile survival to adulthood. The proportion of females breeding as 1-yr olds was significantly negatively correlated with juvenile population size, but not with juvenile body size, larval period length, adult population size, or total precipitation (Table 2). The partial correlation coefficient between the proportion of females breeding at age one and juvenile population size was similar when the effects of juvenile body size, adult population size, and precipitation were held constant (R 5 20.71, P 5 0.01, n 5 10; Table 2). The proportion of males breeding as 1-yr olds, in contrast, was not significantly correlated with any of the variables (Table 2), and partial correlations did not produce any significant correlations with male reproductive age. Male and female body sizes at first reproduction were strongly negatively correlated with juvenile population size and juvenile body size, but not adult population size, larval period length, or total precipitation (Table 2). Holding the
Berven—Terrestrial stage density dependence
333
Fig. 6. Comparison of mean (6 1 SE) clutch size, egg size (mm), and total egg weight (g) of 2-yr-old females (breeding for the first time) from juvenile cohorts for the period 1985–1990; 1993–1999; 2001–2003. Solid bars represent unadjusted means; open bars represent means adjusted for body size (SI). See text for results of ANCOVA.
effects of juvenile body size, larval period length, adult population size, and precipitation levels constant resulted in a slightly lower partial correlation coefficient between juvenile population size and female adult body size (female: R 5 20.71, P 5 0.009, n 5 10) demonstrating that juvenile population size accounted for most of the variation in female size at first reproduction (Table 2). In contrast, partial correlation coefficients between juvenile population size and male reproductive body size (R 5 20.68, P 5 0.016, n 5 10; holding juvenile body size, larval period length, and adult population size constant), and juvenile body size and male reproductive body size (R 5 0.70, P 5 0 .012, n 5 10; holding juvenile population size, larval period length, and adult population size constant) were similar, indicating that both juvenile population size and body size contributed to male reproductive body size. The correlation between the final number of adult males and females from each juvenile cohort and body size at maturity were similar to the correlations between initial juvenile population size and size at maturity (male: r 5 20.56 vs. 20.57; females: 20.69 vs. 20.78, respectively; Table 2). Thus the effects of initial juvenile population size persisted throughout development. Adjusted mean clutch size was negatively correlated with juvenile population size, while egg size was positively correlated with both juvenile and adult population size (Table 2). Partial correlation coefficients (holding the effects of juvenile body size, larval period length, adult population size, and total precipitation constant) were similar (adjusted clutch size: R 5 20.61, P 5 0.035, n 5 10; adjusted mean egg diameter: R 5 0.71, P 5 0.01, n 5 10) to the zero order correlation coefficients, indicating that juvenile population size was the variable most closely associated with variation in these traits (Table 2). Adjusted total clutch weight was negatively correlated with juvenile population size and positively correlated with juvenile body size (Table 2). However, partial correlation coefficients (holding juvenile and adult population size, larval period length, and total rainfall constant) indicated that the majority of the variation in adjusted total clutch weight was due to variation in juvenile body size and not juvenile population size (R 5 0.74, P 5 0.006, n 5 10). Functional relationship between juvenile and adult population size.—The relationship between juvenile population size and the number of surviving males, females, and total adult biomass was best described by the nonlinear theta logistic model (males: y 5 x2(x2/1.78)0.530, F1,13 5 20.1, P , 0.001; Fig. 7; females: y 5 x2(x2/1.20) 0.509, F1,13 5 53.6, P , 0.001;
Table 1. Zero Order Correlation Coefficients (r) for the Relationship among Adult Survival (AS), Number of Adult Males (MN), Females (FN), Juveniles (JN), and Total Precipitation (TP). Analysis performed on arcsine-transformed annual survival (all ages) and log-transformed number. Upper values for males (n 5 19); lower values for females (n 5 19).
MN
AS
AS MN FN JN TP
—
FN
JN
TP
r
P
r
P
r
P
r
P
2.60 .03
(.004) (.895)
2.33 2.71 .38
(.143) (.000) (.089)
2.14 2.06 2.03 2.11
(.536) (.812) (.91) (.62)
.15 .21 2.21 2.31 2.06
(.531) (.37) (.362) (.166) (.783)
—
—
—
—
334
Copeia 2009, No. 2
Table 2. Zero Order Correlation Coefficients for the Relationships among Number of Juveniles (JN), Juvenile Size (JM), Duration of Larval Period (LP), Adult Population Size (AN), Total Precipitation (TP), and Juvenile Survival (JS), Age at First Reproduction (RA), Size at First Reproduction (RM), Body Size Adjusted Mean Clutch Size (EN), Body Size Adjusted Mean Egg Diameter (ED), and Body Size Adjusted Mean Total Clutch Weight (CM). Analysis was performed on log-transformed number of juveniles, adults, and body mass and clutch mass, and arcsine-transformed juvenile survival. Upper values for male juveniles (n 5 14); lower values for female juveniles (n 5 14). Correlation matrix of zero order coefficients (above) are shown in lower section.
JN
JS RA RM EN ED CM JN JM LP AN TP
JM
r
P
2.75 2.64 2.36 2.64 2.58 2.78 .73 2.54 2.52
(.001) (.008) (.17) (.007) (.017) (.000) (.001) (.03) (.04) —
r .39 .29 .44 .21 .73 .68 .29 .02 .69 2.45
LP
AN
TP
P
r
P
r
P
r
P
(.13) (.27) (.09) (.45) (.001) (.004) (.28) (.95) (.003)
2.27 2.21 2.44 2.23 2.25 2.34 2.16 .16 2.23
(.31) (.44) (.09) (.40) (.36) (.20) (.56) (.56) (.39)
2.29 2.29 2.24 2.41 2.16 2.24 2.14 2.51 .09
(.28) (.28) (.38) (.12) (.56) (.37) (.60) (.04) (.74)
2.12 2.05 2.10 2.09 .40 .03 2.34 .25 2.07
(.66) (.86) (.71) (.73) (.13) (.92) (.19) (.34) (.80)
(.078)
.20 2.61
(.46) (.01)
.05 2.06 .40
(.85) (.81) (.12)
.20 .39 2.26 .18
(.46) (.13) (.33) (.51)
—
Fig. 8; total adult biomass: y 5 x2(x2/25.22) 0.727, F1,13 5 70.9, P , 0.001; Fig. 9). Although linear models were also significant, the theta logistic model explained more variation in final adult population size and total biomass (coefficient of determination–males: 0.62 vs. 0.56; females: 0.81 vs. 0.72; total biomass: 0.85 vs. 0.65, respectively) than a linear model. As number of juveniles produced increased, the number of surviving adult males (Fig. 7), females (Fig. 8), and total biomass (Fig. 9) plateaued. This was further evidenced by the deviation from the expected number of adults at higher juvenile population sizes (assuming an average cohort survival of male and females during the study) at all juvenile population sizes (Figs. 7, 8). The theta logistic model predicted a maximum adult population size of 3263 males and 1151 females.
—
—
—
habitats, particularly summer and winter refugia, food availability, or both. Several findings suggest that habitat availability may be important in limiting population size. DeMaynadier and Hunter (1998, 1999) found that both juvenile and adult Wood Frogs displayed strong habitat preferences and were
DISCUSSION The results of this study indicate that density-dependent effects operating in the terrestrial phase may be important in regulating Wood Frog population size. Juvenile and adult population size was the single most important factor explaining year-to-year variation in juvenile and adult vital rates. Higher juvenile population size resulted in lower juvenile survival, delayed maturation, smaller adult body size, and reduced fecundity. While the majority of adults bred only once in their life, in years of low adult population sizes, both males and females lived longer and reproduced multiple times. Terrestrial stage resource limitations.—The fitted logistic growth model describing the relationship between juvenile population size and subsequent number of breeding adult males and females demonstrated that density dependence is compensatory in Wood Frogs and suggests that the terrestrial environment limits adult Wood Frog population size. Density-dependent resource limitations imposed by the environment could include the availability of suitable
Fig. 7. Relationship between the number of male juveniles produced each year and the number of surviving adult males in subsequent years (1–3 yrs). Data points represent the number of adult males surviving to reproduce for each juvenile cohort. The solid line is the fitted theta logistic function describing the relationship (see text for details). The dashed line represents expected number of surviving adult males assuming the average survival at all juvenile population sizes. Value in parenthesis was not included in the model.
Berven—Terrestrial stage density dependence
Fig. 8. Relationship between the number of female juveniles produced each year and the number of surviving adult females in subsequent years (1–3 yrs). Data points represent the number of adult females surviving to reproduce for each juvenile cohort. The fitted theta logistic function (solid line) describing the relationship is shown. The dashed line represents expected number of surviving adult females assuming the average survival at all juvenile population sizes. Value in parenthesis was not included in the model.
more commonly found in mature, closed-canopy forests characterized by dense foliage in both the understory and canopy layers. Within these closed canopy habitats Wood Frog abundance was positively correlated with the structural complexity of forest microhabitat, including the quantity and quality of coarse woody debris, litter depth, and soil moisture (Wyman, 1988; deMaynadier and Hunter, 1998). Recent radio-telemetry studies of adult Wood Frogs have revealed that the availability of preferred habitat may be limiting. Rittenhouse and Semlitsch (2007a) found Wood Frog adults within an oak–hickory forest were clumped within drainages, while in a similar study, Wood Frogs were found in habitats that represented less than 10% of the available forested habitat (Baldwin et al., 2006). Several studies have shown that once established in the nonbreeding habitats, individual adult frogs moved very little, and spent most of the summer in areas comprising a few square meters (Regosin et al., 2005; Baldwin et al., 2006; Rittenhouse and Semlitsch, 2007a). The availability of suitable winter hibernacula may also limit Wood Frog population size. Wood frogs spend the winter near the soil surface and tolerate freezing (Storey and Storey, 1986). However, the ability to tolerate low temperatures (, 210uC), extended periods of freezing, and/or multiple freeze/thaw cycles is largely a function of the quality of the winter hibernacula (forest-, litter-, soil-, and snow-cover) and not Wood Frog physiological tolerances (Layne and Layne, 1987). In the laboratory, Wood Frogs are known to avoid saturated soils when exposed to near freezing temperatures (Licht, 1991) and prefer to winter in drier, upland forest habitats near (approx. 50 m) breeding ponds, as opposed to the moist lowland forest habitats
335
Fig. 9. Relationship between the number of juveniles produced each year and the total adult biomass (g) of adult males and females. The fitted theta logistic function describing the relationship is shown. Value in parenthesis was not included in the model.
preferred during the summer months (Regosin et al., 2003). Indeed, Wood Frog occurrence near suitable ponds has been shown to be particularly sensitive to forest cover within 30 m of the pond’s edge (Homan et al., 2004). Given Wood Frog habitat preferences and limited availability of both summer and winter refugia, it seems likely that in years with high juvenile and adult population sizes, increased competition for limited summer and winter microhabitats could force individuals to occupy habitats that do not provide the conditions required for proper hydration and temperature regulation. Indeed, Wood Frogs held in non-preferred microhabitats during the summer months experienced substantially higher mortality than frogs held in preferred habitats (Rittenhouse and Semlitsch, 2007a). This suggests that microhabitat availability, particularly during periods of high juvenile and adult population sizes, could be important in limiting recruitment and adult annual survivorship. Alternatively, high juvenile population sizes may increase population densities within suitable microhabitats and increase competition for available food resources. Two findings in this study suggest that food may be limited in the terrestrial environment at high population sizes. First, the relationship between total adult biomass and juvenile population size was compensatory. Secondly, juvenile growth rates were density-dependent, as evidenced by smaller adult male and female body size at first reproduction. Food availability has previously been demonstrated to affect postmetamorphic growth in salamanders (Scott and Fore, 1995), and was offered as the likely explanation for lower juvenile growth and maturation rates of juvenile Wood Frogs reared at high and low densities in terrestrial enclosures (Harper and Semlitsch, 2007). Reproductive body size (and its correlated effect on fecundity) and age at first
336
reproduction are critical traits that influence reproductive success and strongly affect the rate of population increase (Cole, 1954). Food limitations may also contribute to lower juvenile survival and adult recruitment. The mass of available prey has been linked to anuran body condition (Pope and Matthews, 2002; Sztatecsny and Schabetsberger, 2005), which in turn correlates with survival (Pope and Mathews, 2002). At higher juvenile and adult densities intraspecific competition for fewer per capita prey items may lower energy reserves and increase the likelihood of starvation (Scott, 1994; Reading and Clarke, 1995) and/or reduce overwintering survival (Pope and Matthews, 2002). This may be particularly true for Wood Frogs that rely on glucose, metabolized from glycogen reserves, as a cryoprotectant during winter (Storey and Storey, 1986). Given the importance of food resources to Wood Frog growth and survival, it seems likely that that lower per capita food availability, caused by high juvenile densities, could also be an important factor limiting adult population size. In 1990–91 Wood Frog population sizes irrupted (Fig. 1) due to large juvenile population sizes combined with uncharacteristically high juvenile survival rates. Population sizes that greatly exceeded the normal carrying capacity (as predicted by the logistic model; Figs. 7, 8) may have resulted from transient factors including favorable weather conditions and/or increases in prey availability. Given the importance of moisture content of both substrate and air in determining Wood Frog habitat preferences (Roberts and Lewin, 1979; Wyman, 1988), it is possible that during particularly wet years (for example, 1989 when precipitation was 21% above normal), marginal habitats may become more suitable, temporarily increasing the carrying capacity of the terrestrial environment. Alternatively, extreme weather conditions may have induced high levels of density-independent mortality of adults and sub-adult frogs and contributed to increased juvenile survival and total adult biomass by indirectly increasing habitat availability and/or per capita food levels in the following years. In 1988, the study site (along with much of the upper Midwest) experienced drought conditions that exceeded the Dust Bowl years of the mid-1930s. In addition to dry conditions, periodic heat waves broke many long-standing records. During one 46-day period in June and July, temperatures at the study site exceeded 32uC on 25 days (normally the area experiences five days above 32uC during the entire summer) and it rained less than 5 mm. During this period (1988–1989) annual adult survival was the lowest observed during the study (3.2%) and juvenile survival (1987 and 1988 juvenile cohorts) was also uncharacteristically low. The large reduction in adult and juvenile frogs in the terrestrial habitat no doubt contributed to the unusually high juvenile survival and total biomass of the 1989 juvenile cohort (Figs. 7–9). Competition among age classes and between sexes.—It was somewhat surprising to find that juvenile fitness traits were more closely related to variation in juvenile population sizes than with adult population sizes. Similarly, male and female fitness traits were more affected by the population size of their own sex. The reasons are not clear, but could be related to competition for prey or habitat segregation. Like most frogs, Wood Frogs are likely opportunistic feeders. In other ranid frogs, invertebrates comprise the majority of prey
Copeia 2009, No. 2
consumed, and similar sized frogs eat similar sized prey (Hedeen, 1972). In addition, the size of prey consumed increases with body size (Forstner et al., 1998). Since juveniles, breeding adult males, and females differ in body size, there may be limited overlap in the prey eaten, reducing competition for prey items among groups. Alternatively, each group may occupy different habitats. Regosin et al. (2003) reported that Wood Frogs were absent from the upland forest adjacent to the breeding pond during the spring and summer, compared to juveniles. They also found that male and female adult frogs overwintered in different areas. Based on differential recapture rates, two separate studies concluded that adults and juveniles established different summer home ranges (Vasconcelos and Calhoun, 2004; Patrick et al., 2006). Female Wood Frogs, and female anurans in general, move greater distances from the breeding habitat in the summer than males (Rittenhouse and Semlitsch, 2007b). It is clear that further work documenting the ambient densities, habitat use, and prey consumption of adults and juveniles is necessary. However, regardless of the causes, my results suggest that intraspecific competition between males and females, and between adults and juveniles may not be occurring. Density effects on reproductive traits.—The finding that clutch size and egg size varied with juvenile population size independent of body size is not surprising and reflects the trade off between maximizing fecundity (egg number) vs. offspring fitness (egg size). Previous studies (Berven and Chadra, 1988) have shown that optimum egg size in Wood Frogs is correlated with environmental factors, including food level and population density. In general, tadpoles derived from smaller eggs were more sensitive to population density compared to larvae derived from larger eggs, irrespective of food level. Adult females appear to be maximizing their fitness by responding with the ‘best’ combination of egg size and clutch size to maximize individual fitness. In years of high juvenile density, females produced fewer larger eggs, reflecting the increased competitive conditions of the larval habitat due to higher larval densities, while in years of low population size, females responded by maximizing fecundity (clutch size) at the expense of offspring size. While the proximal mechanisms are not known, female juveniles could be responding to differences in growth rate. Since juvenile growth rate is negatively correlated with juvenile population size, growth rates may accurately reflect future breeding (and larval) population sizes. It should also be noted in years of low juvenile population size, females were more likely to breed as 1-yr olds. These females also produce greater numbers of smaller eggs relative to 2-yr-old females. Thus, Wood Frog reproductive traits appear to be highly sensitive to population size. Effects of larval traits on adult fitness traits.—Several correlative field studies have shown that the timing and size at metamorphosis (aquatic larval traits) influence adult fitness traits (Berven and Gill, 1983; Smith, 1987; Semlitsch et al., 1988). Experimental studies have directly linked variation in larval density and its effects on larval metamorphic traits to postmetamorphic fitness (Goater, 1994; Scott, 1994; Morey and Reznick, 2001; Altwegg, 2003). The implication of these findings is that density dependence in the larval stage could continue to exert strong effects in the terrestrial stage and
Berven—Terrestrial stage density dependence
stabilize population growth by reducing juvenile survival as well as age and size at first reproduction. In the present study, negative correlations between juvenile body size and juvenile population size, and juvenile body size and length of larval period (Table 2) resulted from density-dependent larval growth rates during the aquatic stage (unpubl. data). However, the effects of larval density on metamorphic body size and the timing of metamorphosis did not affect postmetamorphic male and female juvenile survival in the terrestrial stage. Partial correlations revealed that most of the variation in juvenile survival was due to variation in juvenile population size. This likely reflected increased competition for suitable microhabitats in the terrestrial environment. In contrast, partial correlations indicated that juvenile metamorphic body size (along with juvenile density) positively affected both male reproductive size and, to a lesser degree, female reproductive size. Juvenile body size also affected total clutch weight (adjusted for differences in adult body size), indicating that larger female juveniles tended to devote more energy to reproduction than smaller female juveniles. Fecundity in wood frogs is positively correlated with adult body size (Berven, 1988; this study). Consequently, phenotypic variation in Wood Frog metamorphic body size (resulting from density-dependent larval growth rates) could carry over into the terrestrial stage and affect population growth through its effects on adult body size and fecundity. These results are in agreement with a recent experimental study (Altwegg, 2003) of the Pool Frog (Rana lessonae). In that study, both the aquatic and terrestrial habitat affected juvenile growth rates in the terrestrial stage, although the effect of larval density on juvenile growth rate was greater than the effect of juvenile density. In addition, as in the present study, juvenile body size did not affect juvenile survival in the terrestrial phase. Summary.—The traditional view of population regulation in amphibians has centered on single-stage regulation in the larval stage. However, my results, along with other experimental studies (Pechmann, 1994; Altwegg, 2003; Harper and Semlitsch, 2007), while not demonstrating the direct effects of population regulation, suggest that terrestrial stage regulation may be more common among pond-breeding amphibians than previously thought. Theoretical studies (Hellriegel, 2000; Biek et al., 2002; Vonesh and De la Cruz, 2002) also confirm the potential importance of terrestrial stage vital rates to population stability. While the approach taken in the present (long term population) study documents realistic ranges of vital rates, the inability to control parameters does limit, to some extent, the ability to determine causation. However, although experimental studies are able to determine what is possible under a specified and often-limited set of variables, they are unable to account for the importance of multiple interactions among biotic and abiotic factors. This is particularly true for organisms like Wood Frogs that use different terrestrial habitats depending on sex, stage of development, and season of year, and that regularly move between them. As a consequence, caged experimental studies alone are unlikely to accurately assess the importance of juvenile density and cautions against relying on experimental approaches alone to assay density dependence in the terrestrial environment. While neither approach is ideal, both are necessary and crucial to our understanding of
337
amphibian ecology. Clearly the importance of the terrestrial phase in amphibian research, particularly anuran research, can no longer be ignored. ACKNOWLEDGMENTS I thank the University of Michigan Department of Natural Resources for generously providing access to the research site. Over the years numerous colleagues and students helped in various aspects of this study and I express my profound appreciation. E. Werner and M. Benard provided comments on an earlier draft of this manuscript. The Oakland University Institution Animal Care and Use Committee approved the research, and collection permits were issued by the Michigan Department of Natural Resources. LITERATURE CITED Altwegg, R. 2003. Multistage density dependence in an amphibian. Oecologia 136:46–50. Baldwin, R. F., J. K. Calhoun, and P. G. deMaynadier. 2006. Conservation planning for amphibian species with complex habitat requirements: a case study using movements and habitat selection of the wood frog Rana sylvatica. Journal of Herpetology 40:442–453. Beebee, T. J. C., J. S. Denton, and J. Buckley. 1996. Factors affecting population densities of adult natterjack toads Bufo calamita in Britain. Journal of Applied Ecology 33:263–268. Berven, K. A. 1988. Factors affecting variation in reproductive traits within a population of wood frogs (Rana sylvatica). Copeia 1988:605–615. Berven, K. A. 1990. Factors affecting population fluctuations in larval and adult stages of the wood frog (Rana sylvatica). Ecology 71:1599–1608. Berven, K. A. 1995. Population regulation in the wood frog, Rana sylvatica, from three diverse geographic localities. Australian Journal of Ecology 20:385–392. Berven, K. A., and R. S. Boltz. 2001. Interactive effects of leech (Desserobdella picta) infection on wood frog (Rana sylvatica) tadpole fitness traits. Copeia 2001:907–915. Berven, K. A., and B. Chadra. 1988. The relationship among egg size, density and food level on larval development in the wood frog Rana sylvatica. Oecologia 75:67–72. Berven, K. A., and D. E. Gill. 1983. Interpreting geographic variation in life-history traits. American Zoologist 23: 85–97. Biek, R., W. C. Funk, B. A. Maxell, and L. S. Mills. 2002. What is missing in amphibian decline research: insights from ecological sensitivity analysis. Conservation Biology 16:728–734. Cole, L. C. 1954. The population consequences of life history phenomena. Quarterly Review of Biology 29: 103–137. deMaynadier, P. G., and M. L. Hunter, Jr. 1998. Effects of silvicultural edges on the distribution and abundance of amphibians in Maine. Conservation Biology 12:340–352. deMaynadier, P. G., and M. L. Hunter, Jr. 1999. Forest canopy closure and juvenile emigration by pool-breeding amphibians in Maine. Journal of Wildlife Management 63:441–450. Forstner, J. M., M. R. J. Forstner, and J. R. Dixon. 1998. Ontogenetic effects on prey selection and food habits of
338
two sympatric east Texas ranids: the southern leopard frog, Rana sphenocephala, and the bronze frog, Rana clamitans clamitans. Herpetological Review 29:208–211. Goater, C. P. 1994. Growth and survival of postmetamorphic toads: interactions among larval history, density, and parasitism. Ecology 75:2264–2274. Halpern, B. S., S. D. Gaines, and R. R. Warner. 2005. Habitat size, recruitment, and longevity as factors limiting population size in stage-structured species. American Naturalist 165:82–94. Harper, E. B., and R. D. Semlitsch. 2007. Density dependence in the terrestrial life history stage of two anurans. Oecologia 153:879–889. Hedeen, S. E. 1972. Food and feeding behavior of the mink frog, Rana septentriolalis Baird, in Minnesota. American Midland Naturalist 88:291–300. Hellriegel, B. 2000. Single- or multistage regulation in complex life cycles: does it make a difference? Oikos 88:239–249. Homan, R. N., B. S. Windmiller, and J. M. Reed. 2004. Critical thresholds associated with habitat loss for two vernal pool-breeding amphibians. Ecological Applications 14:1547–1553. Howard, R. D., and A. G. Kluge. 1985. Proximate mechanisms of sexual selection in wood frogs. Evolution 39:260–277. Layne, J. R., and J. R. Layne. 1987. Freeze tolerance and the dynamics of ice formation in wood frogs (Rana sylvatica). Canadian Journal of Zoology 65:2062–2065. Licht, L. E. 1991. Habitat selection of Rana pipiens and Rana sylvatica during exposure to warm and cold temperatures. American Midland Naturalist 125:259–268. Loman, J. 2002. Rana temporaria metamorph production and population dynamics in the field. Effects of tadpole density, predation and pond drying. Journal for Nature Conservation 10:95–107. Morey, S., and D. Reznick. 2001. Effects of larval density on postmetamorphic spadefoot toads (Spea hammondii). Ecology 82:510–522. Patrick, D. A., M. L. Hunter, Jr., and A. J. K. Calhoun. 2006. Effects of experimental forestry treatments on a Maine amphibian community. Forest Ecology and Management 234:323–332. Pechmann, J. 1994. Population regulation in complex life cycles: aquatic and terrestrial density-dependence in pond-breeding amphibians. Unpubl. Ph.D. diss., Duke University, Durham, North Carolina. Pope, K. L., and K. R. Matthews. 2002. Influence of anuran prey on the condition and distribution of Rana muscosa in the Sierra Nevada. Herpetologica 58:354–363. Prout, T., and F. McChesney. 1985. Competition among immatures affects their adult fertility: population dynamics. American Naturalist 126:521–558. Reading, C. J., and R. T. Clarke. 1995. The effects of density, rainfall and environmental temperature on body condition and fecundity in the common toad, Bufo bufo. Oecologia 102:453–459. Regosin, J. V., B. S. Windmiller, R. N. Homan, and J. M. Reed. 2005. Variation in terrestrial habitat use by four pool-breeding amphibian species. Journal of Wildlife Management 69:1481–1493.
Copeia 2009, No. 2
Regosin, J. V., B. S. Windmiller, and J. M. Reed. 2003. Terrestrial habitat use and winter densities of the wood frog (Rana sylvatica). Journal of Herpetology 37:390–394. Rittenhouse, T. A. G., and R. D. Semlitsch. 2007a. Postbreeding habitat use of wood frogs in Missouri oak– hickory forest. Journal of Herpetology 41:645–653. Rittenhouse, T. A. G., and R. D. Semlitsch. 2007b. Distribution of amphibians in terrestrial habitat surrounding wetlands. Wetlands 27:1531–1561. Roberts, W., and V. Lewin. 1979. Habitat utilization and population densities of the amphibians of northeastern Alberta. Canadian Field-Naturalist 93:144–154. Scott, D. E. 1994. The effect of larval density on adult demographic traits in Ambystoma opacum. Ecology 75: 1383–1396. Scott, D. E., and M. R. Fore. 1995. The effect of food limitation on lipid levels, growth, and reproduction in the marbled salamander, Ambystoma opacum. Herpetologica 51:462–471. Semlitsch, R. D., D. E. Scott, and J. H. K. Pechmann. 1988. Time and size at metamorphosis related to adult fitness in Ambystoma talpoideum. Ecology 69:184–192. Skelly, D. K., and J. M. Kiesecker. 2001. Venue and outcome in ecological experiments: manipulations of larval anurans. Oikos 94:198–208. Smith, D. C. 1987. Adult recruitment in chorus frogs: effects of size and date at metamorphosis. Ecology 68:344–350. Smith, M. A., and D. M. Green. 2005. Dispersal and the metapopulation paradigm in amphibian ecology and conservation: Are all amphibian populations metapopulations? Ecography 28:110–128. Storey, K. B., and J. M. Storey. 1986. Freeze tolerance and intolerance as strategies of winter survival in terrestriallyhibernating amphibians. Comparative Biochemistry and Physiology A 83:613–617. Sztatecsny, M., and R. Schabetsberger. 2005. Into thin air: vertical migration, body condition, and quality of terrestrial habitats of alpine common toads, Bufo bufo. Canadian Journal of Zoology 83:788–796. Turchin, P. 1999. Population regulation: a synthetic view. Oikos 84:153–159. Van Buskirk, J., and D. C. Smith. 1991. Density-dependent population regulation in a salamander. Ecology 72: 747–756. Vasconcelos, D., and A. J. K. Calhoun. 2004. Movement patterns of adult and juvenile Rana sylvatica (LeConte) and Ambystoma maculatum (Shaw) in three restored seasonal pools in Maine. Journal of Herpetology 38: 551–561. Vonesh, J. R., and O. De la Cruz. 2002. Complex life cycles and density dependence: assessing the contribution of egg mortality to amphibian declines. Oecologia 133:325–333. Wilbur, H. M. 1980. Complex life cycles. Annual Review of Ecology and Systematics 11:67–93. Wilbur, H. M. 1996. Multistage life cycles, p. 75–108. In: Population Dynamics in Ecological Space and Time. O. E. Rhodes, Jr., R. K. Chesser, and M. H. Smith (eds.). University of Chicago, Chicago. Wyman, R. L. 1988. Soil acidity and moisture and the distribution of amphibians in five forests of southcentral New York. Copeia 1988:394–399.
