High Altitude Frogs (Rana kukonoris) Adopt a Diversified Bethedging Strategy in the Face of Environmental Unpredictability

Wei CHEN, Liqing PENG, Lichun JIANG, David A. PIKE, Christopher R. FRIESEN and Gregory BROWN

1Ecological Security and Protection Key Laboratory of Sichuan Province, Mianyang Normal University, Mianyang 621000, China

2Department of Biology, Rhodes College, Memphis Tennessee 38111, USA

3 School of Life and Environmental Sciences, University of Sydney, New South Wales 2006, Australia

1. Introduction

Given finite resources, mothers generally make an energetic trade-off between current reproductive investment and somatic maintenance (Stearns, 1992).This strategy serves to enhance fitness by ensuring future reproductive potential by producing offspring that vary in size and quality (e.g., many small or fewer larger)based on the environmental conditions of the habitat. The outcome is a balance between the benefits of increased fecundity and high offspring quality (Morrongielloet al.,2012). A mother’s best option should be to produce as many offspring of optimal size as is possible, which is adjusted depending on environmental cues (Einum and Fleming, 2004). Such resource allocation within eggs of a clutch is influenced by the environmental variation(Einum and Fleming, 2004; Morrongielloet al., 2012).Empirical support for these patterns is widespread among diverse taxa including plants (Vaughton and Ramsey,1998), insects (Zoviet al., 2008), fishes (Johnston and Leggett, 2002; Morrongielloet al., 2012; Marshall and Burgess, 2015), amphibians (R?s?nenet al., 2008), and birds (Tremblayet al., 2003).

In addition to trade-offs between the size and number of offspring produced, breeding females often vary offspring size within a clutch based on the environmental cues they experience throughout ontogeny (Kaplan and Cooper, 1984). When environments vary predictably with the seasons, females are likely to lay eggs that are of optimal size, and thus that are similar within and among clutches (Kaplan and Cooper, 1984). When environments vary unpredictably among years through evolutionary time, females should instead evolve a bethedging breeding strategy that increases geometricmean fitness (Philippi and Seger, 1989; Simons, 2011,2014). When using a conservative bet-hedging strategy,females may produce offspring that are larger but less variable in size because large eggs may have greater probability of survival in unpredictable environments, as compared to a stable environment (Shine, 1978; Travis,1980). This strategy can result in greater reproductive success in years when environmental quality is poor,especially if larger offspring have a size-based survival advantage in these conditions (Einum and Fleming,2004). Alternatively, females can also employ a diversified bet-hedging strategy to produce a range of offspring sizes, which spreads the risk of incorrectly predicting future environmental conditions to ensure at least some offspring survive, regardless of future conditions (Capinera, 1979; Kaplan and Cooper, 1984;Parker and Begon, 1986; Einum and Fleming, 2004;Simons, 2011). According to the imperfect information hypothesis, the degree of within-clutch variability in egg size will decrease relative to average egg size (Koopset al., 2003), and environmental quality also influences the degree of egg size variation by desiccation stress and/or temperature and climate variability (Morrongielloet al.,2012; Marshall and Burgess, 2015; Shama, 2015).

High altitude environments are ideal for testing the predictions of bet-hedging models because of the high degree of environmental variation and unpredictability that occurs over small spatial scales. Animals living at higher altitudes usually face harsher and less predictable environments compared to ones living at lower altitudes(Duellman and Trueb, 1986; Morrison and Hero, 2003;Chenet al., 2013). These environmental conditions should result in population divergence of reproductive strategies along altitudinal gradients that are most suited to local conditions. Most amphibians show variation in life history traits among populations at different altitudes, including shorter breeding seasons and activity periods, larger eggs and longer juvenile stages, and larger body sizes at all life stages being associated with high altitudes (reviewed by Morrison and Hero, 2003). The shorter and more variable activity periods of high-altitude environments can directly constrain foraging time and therefore reduce energy accumulation (Morrison and Hero, 2003; Chenet al., 2011). Stored energy will determine breeding investment strategies of explosive breeding species,based on an assessment of environmental quality (Wells,2007). However, it is difficult for selection to optimize any one life history trait in a vicissitudinous environment(Morrison and Hero, 2003), and therefore selection may favor bet-hedging strategies. The Tibetan Plateau,which experiences highly variable seasonal temperature and rainfall patterns, provides an ideal location to test the hypothesis that unpredictable environments select for within clutch variation in egg size as predicted by theoretical work on bet-hedging strategies (McGinleyet al., 1987; Einum and Fleming, 2004; Marshallet al.,2008).

We investigated within clutch variation in egg-size ofRana kukunoris, a frog species distributed along an altitudinal gradient on the Tibetan Plateau, to test the following predictions: within-clutch variation in egg size will be (1) positively associated with altitude; (2)negatively associated with mean egg size and positively associated with clutch size; (3) negatively associated with environmental quality, and (4) negatively associated with environmental predictability. We also carried out an experiment to investigate the influence of body size on desiccation and survival of tadpoles. Such trade-offs could help understand how maternal breeding strategies are adapted to unpredictable environmental conditions,and therefore provide a glimpse of reproductive tactics that may continue to be successful under future climatic changes.

2. Materials and Methods

Study speciesThe plateau brown frog,R. kukunoris, is endemic to high-altitude regions of the eastern Tibetan Plateau, and is one of the few frog species that reaches high elevation (2000–4400 m, Chenet al., 2011). Upon emergence from hibernation in the spring, explosive breeding takes place in shallow temporary ponds, where desiccation and freezing can determine larval survival (Luet al., 2008). Egg clutches are deposited on the margins of water bodies, which often freeze over during the night (Luet al., 2008); this, combined with drying of water bodies during the breeding season can result in larval mortality(Chen, personnel observation). After reproducing, females only have a short activity period during which to store energy for the next reproductive bout, prior to ceasing activity for the next winter (Luet al., 2008; Chenet al.,2011, 2013). Female frogs are the larger sex, and clutch size is positively correlated with female body size, with females at higher altitudes producing larger clutches of eggs (Chenet al., 2013).

Data collectionWe collected data from sixR.kukunorispopulations (breeding ponds) along the eastern Tibetan Plateau spanning a 255-km latitudinal gradient and ranging in altitude from 2035 to 3494 m above sea level(Figure 1, Table 1). During 2010-2015 we searched for freshly laid egg clutches at breeding ponds. We counted the number of eggs in a small sample of each clutch (ca.5 g), which was weighed (to 0.01 g) and used to estimate clutch size based on mass of the entire clutch. We quantified egg size by measuring the diameter of 10–30 eggs from digital photographs of each clutch (Abramoffet al., 2004). To ensure that egg size measurements were comparable within and among populations (R?s?nenet al., 2008), all clutches were sampled and measured prior to eggs reaching Gosner Stage 12 (Gosner, 1960). We adopted coefficient of variation (CV) in egg diameter as a measure of egg size variation that describes the standard deviation as a percentage of the mean (Quinn and Keough, 2002).

Figure 1 Map showing the distribution of our sampling sites for Rana kukunoris in the Tibetan Plateau.

Table 1 Study site details for six high-altitude populations of Rana kukunoris in the Tibetan Plateau, including altitude, latitude and sample sizes and clutch attributes.

In order to explore the effects of environmental conditions on egg size, we used mean monthly temperature (MMT) and mean monthly rainfall (MMR)to describe environmental quality of each site from 1980 – 2010 (from data. cma. cn). These variables can mirror different aspects of environmental quality. For example, in frogs, MMT is related to activity period and metabolism, whereas MMR can determine food availability and thus potential for energy accumulation and storage (Wells, 2007). For three populations in close proximity to local weather stations (＜ 20 km),we also used mean monthly rainfall and mean monthly temperature to calculate Colwell’s predictability indices for rainfall and temperature (Colwell, 1974).

Tadpole Survival under Desiccation Experimental DesignWe captured adult frog pairs in amplexus from ponds at an altitude of 3462 m, and took these animals back to Mianyang Normal University. We obtained fertilized eggs by injecting pairs of frogs with synthetic gonadotropins (10 IU/ml chorionic gonadotropin). We measured the egg size and body size of the females, after the eggs were laid.

We used 20 replicates of 10 fertilized eggs each (one from each clutch, Gosner stage 10; Gosner, 1960) of different sizes and placed them into a plastic box (20 cm×10 cm×5 cm) containing aged water maintained at 20°C. After 7 days, the numbers of hatched eggs (Gosner stage 18–20) were recorded. At this time, all eggs had either hatched or were clearly inviable. To measure the impact of egg size on larval life, we maintained the tadpoles in the same boxes for two weeks post-hatching with a room temperature of 20°C, and we feed the tadpoles in each box with fish food every three days and change the water every three days. In order to investigate the influence of drought on tadpoles of different size, 40 tadpoles with different body lengths were haphazardly selected to be placed onto dry copy paper, and have survival time recorded as a function of desiccation.

Statistical methodsFirst, we explored the relationships between altitude and egg size variation using General linear models (Glms) with altitude and latitude as covariates, and CV as the dependent variable. We then investigated the correlation between CV and clutch size as well as CV and mean egg size in another independent model. To explore the effect environmental quality on egg size, we also explored the relationship between environmental quality (mean monthly temperature and rainfall) and CV using the same models. We explored the relationships between both within-clutch variability of egg size and egg size and environmental predictability using linear regressions using stepwise discriminant function.We explored the correlation between egg size and survival rate, between egg size and tadpole performance,and between body size of tadpoles and their survival rate using the Spearman’s correlation coefficients.

Prior to analyses, we log10-transformed all variables to better attain normality and improve homogeneity of variance. All analyses were conducted using IBM SPSS Statistics version 24.

3. Results

Field DataIn total, we measured 410 clutches from six populations spanning a 255-km latitudinal gradient and ranging in altitude from 2035 to 3494 m above sea level.Egg size showed significant variation within a clutch(CV range 1.72%–12.73%), among clutches (t= 56.967,df= 409,P＜ 0.001), and among populations (F5,404=30.357,P＜ 0.001, Table 1). Body size of females was not correlated with the degree of egg-size variation within a clutch (Spearman’s correlation:rs= -0.416,n= 17,P=0.096).

The results from Glms revealed that altitude (F1,407=5.984,P= 0.015), latitude (F1,407= 12.659,P＜ 0.001)can both influence the egg-size variation (Table 2).The degree of egg-size variation did not increase with increasing clutch size (F1,398= 1.470,P= 0.226), and this was also the case for mean egg size (F1,398= 0.122,P=0.727), despite significant variation in the degree of eggsize variation among populations (F5,398= 31.436,P＜0.001). However, the degree of egg-size variation within a clutch was affected by environmental quality (Table 2), including temperature and rainfall (mean monthly temperatureF1,407= 57.117,P＜ 0.001, rainfallF1,407=56.614,P＜ 0.001). Differences in the estimates of variation were unaffected by sample size variation among populations.

In general, when we considered environmental predictability, only rainfall remained in the final model and rainfall change exerted significant influence on egg size variation in a clutch (B= -5.769,t= -7.267,P＜0.001). Clutch size was not significantly associated with these variables (B= 0.346,t= 0.945,P= 0.345).Females inhabiting more predictable environments laid large eggs (B= 1.341,t= 8.691,P＜ 0.001) but that were less variable in egg size (Figure 2), whereas females inhabiting more unpredictable environments laid small eggs but more variable egg sizes, subbestinga diversified bet-hedging strategy of energy allocation to offspring.

Desiccation ExperimentAt Gosner stage 25, the survival rate of tadpoles was not dependent upon initial egg size (rs= -0.18,n= 20,P= 0.941), but there were obvious differences in the developmental speed of tadpoles; tadpoles from larger eggs grew faster than those from smaller eggs (rs= 0.584,n= 20,P= 0.007). We also found the large tadpoles from large eggs survived longer than the small tadpoles when in a desiccating environment(Pearson’s correlationr= 0.775,n= 40,P＜ 0.001; Figure 3), indicating that the offspring from large eggs are more likely to survive simulated harsh conditions.

4. Discussion

Altitude is an important environmental factor that affects many life history traits of amphibians, including the duration of the breeding season, size and age at sexual maturity, lifespan, egg number and size, and development and growth of offspring (Morrison and Hero, 2003).Therefore, many life history traits show altitudinal clines in these characteristics (Morrison and Hero,2003). Environmental unpredictability can also influence strategies of maternal investment among eggs within a clutch (Duellman and Trueb, 1986; Morrison and Hero,2003; Chenet al., 2013); our study revealed considerable variation in egg size ofR. kukunorisfrom Asia, bothwithin a clutch and among populations differing in altitude and latitude. We also detected an obvious altitudinal cline of egg-size variation within a clutch,but no correlation between egg-size variation within a clutch and mean egg size. Interestingly, our results showed that breeding females produce small, but more variable offspring sizes in unpredictable environments.Similar within-clutch variability of eggs has also been documented in tropical tree frogs form Santa Cecilia,Ecuador (Crump, 1981).

Table 2 Parameter estimates and test statistics from General linear models describing within-clutch egg-size variation. Model 1 and model 2 are the results testing the altitudinal cline and environmental quality, respectively. CI: Confidence Interval for Mean.

The imperfect information hypothesis (Koopset al., 2003) predicts the degree of within-clutch variability of egg size will be negatively related to average egg size. Our results did not support this prediction because the degree of egg-size variation was not significantly correlated with mean egg size inR.kukunoris. We also documented a consistent pattern of females producing a greater range of size variation with eggs, indicating that females are fairly consistent in how they divide resources among eggs. This pattern has also been documented in other animals (Koopset al., 2003;Semmens, 2008).

Our results also provide empirical support for the predictive models developed by Marshallet al.(2008), which stipulate that a diversified bet-hedging strategy will occur as an adaptive strategy when future offspring environments are unpredictable. We found that femaleR. kukunorisin these increasingly unpredictable environments produced smaller eggs that are more variable in size. Similar patterns of within-clutch egg size variation with diversified bethedging strategy are found across taxa including brook troutSalvelinus fontinalis(Koopset al., 2003), common jollytailsGalaxias maculatus(Semmens, 2008), quacking frogsCrinia georgiana(Dziminskiet al., 2009) and tree frogs (Crump, 1981). The Tibetan Plateau is a variable and unpredictable environment (Colwell index values 0.89). In this environment, precipitation is an important factor affecting the persistence of habitat and breeding ponds and the survival of offspring, which likely acts as strong selective forces on the breeding allocation strategies of females. We thus speculate that adopting a diversified bet-hedging strategy may be optimal and adaptive in these low quality high-altitude environments by increasing geometric-mean fitness (Simons, 2011,2014), because the large tadpoles from larger eggs survive longer in desiccating environments (Figure 2).

Figure 2 Relationship between egg size and environmental predictability (a) and between within-clutch egg size variation and environmental predictability (b). Environmental predictability is measured using Colwell’s index. Error bars represent standard deviation.

Figure 3 Relationship between survival time and body size of tadpoles in the desiccation experiment, the broken line is the line of best fit.

Within-clutch variation in offspring size has also been attributed to the physiological or developmental inability of females to precisely allocate resources evenly among individual offspring (McGinleyet al., 1987; Fox and Czesak, 2000; Einum and Fleming, 2004), the influence from mothers on offspring environment (van den Berghe and Gross, 1989), or sibling competition (Einum and Fleming, 2002). Even so, we cannot rule out the effects of phenotypic plasticity on egg-size variability, because the concurrent evolution between predictable and unpredictable components in environmental variance is expected (Shama, 2015). In our study, we lack the morphological data on female body condition before oviposition, and were therefore unable to test whether condition influences egg-size variability. Thus the exact causes of egg-size variation within a clutch require further study.

Abramoff M., Magalhaes P., Ram S.2004. Image processing with Image J. Biophotonics Int 11, 36–42

Capinera J. L. 1979. Qualitative variation in plants and insects:effect of propagule size on ecological plasticity. Am Nat, 114:350–361

Chen W., Tang Z. H., Fan X. G., Wang Y., Pike D. A.2013.Maternal investment increases with altitude in a frog on the Tibetan Plateau. J Evolution Biol, 26: 2710–2715

Chen W., Yu T. L., Lu X. 2011. Age and body size ofRana kukunoris, a high-elevation frog native to the Tibetan Plateau. Herpetol J, 21: 49–151

Colwell R. K.1974. Predictability, constancy and contingency of periodic phenomena. Ecology, 55: 1148–1153

Crump M. L.1981. Variation in propagule size as a function of environmental uncertainty for tree frogs. Am Nat, 117: 724–737

Duellman W. E., Trueb D. L.1986. Biology of Amphibians. New York: McGraw-Hill Inc

Dziminski M. A., Vercoe P. E. Dale R. J.2009. Variable offspring provisioning and fitness: a direct test in the field. Funct Ecol, 23:164–171

Einum S., Fleming I. A.2002. Does within-population variation in fish egg size reflect maternal influences on optimal values? Am Nat, 160: 756–765

Einum S., Fleming I. A.2004. Environmental unpredictability and offspring size: conservative versus diversified bet-hedging. Evol Ecol Res, 6: 443–455

Fox C. W., Czesak M. E.2000. Evolutionary ecology of progeny size in arthropods. Annu Rev Entomol, 45: 341–369

Gosner K. L.1960. A simplified table for staging anuran embryos and larvae with notes on identification. Herpetologica, 16: 183–190

Johnston T. A., Leggett W. C.2002. Maternal and environmental gradients in the egg size of an iteroparous fish. Ecology, 83:1777–1791

Kaplan R. H., Cooper W. S.1984. The evolution of developmental plasticity in reproductive characteristics: an application of the‘‘a(chǎn)daptive coin-flipping’’ principle. Am Nat, 123: 393–410

Koops M. A., Hutchings J. A., Adams B. K.2003. Environmental predictability and the cost of imperfect information: influences on offspring size variability. Evol Ecol Res, 5: 29–42

Lu X., Zeng X. H., Du B., Nie C.2008. Reproductive ecology ofRana kukunorisNikolskii, 1918, a high-elevation frog native to the Tibetan Plateau. Herpetozoa, 21: 67–77

Marshall D. J., Bonduriansky R., Bussière L. F.2008. Offspring size variation within broods as a bet-hedging strategy in unpredictable environments. Ecology, 89: 2506–2517

Marshall D. J., Burgess S. C.2015. Deconstructing environmental predictability: seasonality, environmental color and the biogeography of marine life histories. Ecol. Lett, 18: 174–181

McGinley M. A., Temme D. H., Geber M. A.1987. Parental investment in offspring in variable environments: theoretical and empirical considerations. Am Nat, 130: 370–398

Morrison C., Hero J. M.2003. Geographic variation in life-history characteristics of amphibians: A review. J Anim Ecol, 72: 270–279

Morrongiello J. R., Bond N. R., Crook D. A., Wong B. B. M.2012. Spatial variation in egg size and egg number reflects trade-offs and bet-hedging in a freshwater fish. J Anim Ecol, 81:806–817

Parker G. A., Begon M.1986. Optimal egg size and clutch size:effects of environment and maternal phenotype. Am Nat, 128:573–592

Philippi T., Seger J.1989. Hedging one’s evolutionary bets,revisited. Trends in Ecol Evol, 4: 41–44

Quinn G. P., Keough M. J.2002. Experimental design and data analysis for biologists. Cambridge: Cambridge University Press

R?s?nen K., S?derman F., Laurila A., Meril? J.2008. Geographic variation in maternal investment: acidity affects egg size and fecundity inRana arvalis. Ecology, 89: 2553–2562

Semmens D.2008. Maternal Effects on Offspring Fitness in an Australian Native Fish,Galaxias maculatus. PhD dissertation,University of Melbourne, Melbourne, Australia

Shama L. S.2015. Bet hedging in a warming ocean: predictability of maternal environment shapes offspring size variation in marine sticklebacks. Global Change Biol, 21: 4387–4400

Shine R.1978. Propagule size and parental care: the “safe harbour”hypothesis. J Theor Biol, 75: 417–424

Simons A. M.2011. Modes of response to environmental change and the elusive empirical evidence for bet hedging. P Roy Soc B Bio, 278: 1601–1609

Simons A. M.2014. Playing smart vs. playing safe: the joint expression of phenotypic plasticity and potential bet hedging across and within thermal environments. J Evol Biol, 27: 1047–1056

Strarns S. C.1992. The evolution of life histories. New York:Oxford university press

Travis J.1980. Phenotypic variation and the outcome of interspecific competition in hylid tadpoles. Evolution, 34: 40–50

Tremblay I., Thomas D. W., Lambrechts M. M., Blondel J.,Perret P.2003. Variation in Blue Tit breeding performance across gradients in habitat richness. Ecology, 84: 3033–3043

van den Berghe E. P., Gross M. R.1989. Natural selection resulting from female breeding competition in a Pacific salmon(Coho:Oncorhynchus kisutch). Evolution, 43: 125–140

Vaughton G., Ramsey M.1998.Sources and consequences of seed mass variation inBanksia marginata(Proteaceae). J Ecol, 86:563–573

Wells K. D.2007. The ecology and behavior of amphibians. The University of Chicago: Chicago Press

Zovi D., Stastny M., Battisti A., Larsson S.2008. Ecological costs on local adaptation of an insect herbivore imposed by host plants and enemies. Ecology, 89: 1388–1398