Body size and condition, not allochrony, affect temporal reproductive patterns in a prolonged breeding anuran

Abstract

Individual phenological life-history variations in the context of seasonal conditions are well documented in fishes and birds. However, amphibians, a group heavily affected by habitat loss and fragmentation, have received relatively little attention regarding research on life-history adaptations. Here we present 3 years of data on the timing of reproductive activity in a suburban European green toad (Bufotes viridis) population. We found annually consistent patterns of reproductive activity and investigated whether these were caused by allochrony or individual attributes. Body size (a proxy for age), body condition, and sex significantly affected the timing of reproductive activity. However, most individuals showed considerable overlap in their reproductive timeframe, refuting the existence of allochronic subpopulations. Our findings may indicate life-history adaptations in the direction of a faster lifestyle in response to hazardous environments. We propose to focus further research efforts on phenological variations in the context of environmental conditions, and that phenological variations should be considered more strongly in amphibian conservation efforts.

Life-history adaptations develop as evolutionary or plastic responses to biotic and abiotic environmental conditions (Roff 2001; Ricklefs and Wikelski 2002). In seasonal environments, life-history strategies must account for periodically changing conditions (e.g., mortality risk, availability of food, hydroperiod, etc.) throughout the year (Bernhardt et al. 2020). In temperate environments, for many animal species the year is divided into a productive season during which energy gain is possible and an unproductive season defined by energy loss (usually winter; Varpe 2017). The latter is often outlasted in physiologically dormant states (i.e., embryonic diapause, hibernation, and brumation) (Wilsterman et al. 2021), whereas activities during the productive season (i.e., growth, reproduction, and energy storage) have to be allocated and timed to maximize the number of offspring over an individual’s lifespan, also referred to as reproductive value (Grafen 2006; Varpe 2017). As benefits and risks of investing resources in growth, energy storage, and reproduction might vary across the productive season depending on the individual animal’s attributes (e.g., size, body condition, age or life stage) (Varpe 2017; Van de Walle et al. 2023) we expect the optimal timing for each of these activities to vary accordingly. For instance, an individual who starts into the productive season with high body condition might benefit more from an early breeding start. Offspring of early-breeding individuals might have more time to acquire resources to grow and store, therefore having an advantage over competitors and a higher survival rate in winter. Individuals with lower body conditions on the other hand might instead make a tradeoff of postponing reproduction until later in the season to obtain more nutrients, thereby decreasing the survival chances of their offspring but increasing their own survival chances and chances for another season of reproduction. The interplay between individual attributes (changeable over time; e.g., age and size) and timing of reproduction which affects an individual’s reproductive success was investigated in various vertebrate groups, (i.e., in fishes [Danylchuk and Fox 1994; Divino and Tonn 2007], birds [Martin 1995; Verhulst and Nilsson 2007; Saino et al. 2012; Dunn and Møller 2014] and mammals [Bieber et al. 2012]). In amphibians, however, effects of individual states (attributes changeable over time) on the timing of reproduction have hardly been explored.

At the interspecific level in amphibians, reproductive timing can be broadly categorized as continuous or seasonal, strongly modulated by local climatic conditions (i.e., annually constant conditions vs. seasonal patterns in temperature and the hydroperiod) (Juarez and O’Connell 2023; Vági and Székely 2023). Seasonal breeding strategies are again commonly subdivided into explosive and prolonged breeding strategies (Hartel et al. 2007; Ulloa et al. 2019). Explosive breeders apply an extreme strategy in the sense that they focus their reproductive effort on short, synchronized breeding assemblages that can last from a few days to a few weeks (Saenz et al. 2006; Hartel et al. 2007; Ulloa et al. 2019). Thus, individuals show relatively little variation in the timing of reproduction. Prolonged breeders stretch their reproductive activity over longer periods. In Central Europe, this timespan usually lasts from April to August (Hartel et al. 2007) with variations often affected by the local environment (Saenz et al. 2006). However, not all individuals utilize the entire reproductive timeframe of their population and can show high individual variability in the timing of reproduction. It is often observed that prolonged breeders are not continuously active at the breeding sites but that spawning occurs in waves (Flindt and Hemmer 1968; Andrén and Nilson 1985; Sinsch 1988, 1992; Denton and Beebee 1993; Sinsch and Seidel 1995; Paton and Crouch 2002).

When analyzing capture-recapture data on a population of the European green toad (Bufotes viridis), a prolonged breeding anuran species, we noticed temporally distinct peaks of reproductive activity. These peaks, one during early May and one during mid-June were consistent across 3 sampling years. Our data were the years 2021 to 2023 from 19, 21, and 26 sampling nights, respectively, and entails capture–recapture data as well as morphometric data. The studied population is located in a suburban part of south-east Vienna, Austria, best described as a patchwork of greenhouses, fields, and detached houses. While the abundance of toads is likely very high in this population (>1,000 individuals) there are also reports of high road mortality of several hundred individuals per year (Staufer 2022).

In former studies, the natterjack toad (Epidalea calamita), a phylogenetically closely related and ecologically similar anuran species (Sinsch et al. 1999; Stöck et al. 2008; Sinsch and Leskovar 2011), was reported to also exhibit several activity peaks during the reproductive period (Andrén and Nilson 1985; Denton and Beebee 1993). While earlier studies found a correlation between body size and time of arrival at the breeding site (Flindt and Hemmer 1968), this effect, however, was not found in other populations (Denton and Beebee 1993). Furthermore, Sinsch (1988, 1992) and Sinsch and Seidel (1995) proposed a different cause for observed waves of reproductive activity. They presented evidence of temporal breeding separation, termed allochrony, in a natterjack toad population. Allochrony describes genetic isolation between subpopulations, not in space, but by consistently separate times of breeding. This requires, however, that the timing of breeding has a heritable component and is consistent across the relevant periods (day, season, or year) (Taylor and Friesen 2017). The phenomenon of multiple waves of breeding activity was also observed in mammals. Bieber et al. (2012) found two2 annual peaks in reproductive activity in the common dormice (Muscardinus avellanarius). They, however, interpreted their findings as a consequence of high winter survival and high mortality during the rest of the year. They found that high female body condition affected the reproductive timing. Females with high body condition would tend to start reproduction earlier in the season. Thus, individual differences in reproductive timing in species with prolonged breeding periods can have various causes.

To explain our own observation of 2 peaks of reproductive activity consistent across years in a population of the European green toad, we propose 2 hypotheses. First, the pattern might arise because the timing of reproduction is influenced by individual states such as age and/or body condition because different individual states, might call for different breeding strategies to maximize an individual’s reproductive value. Second, the pattern might be caused by allochrony, 2 subpopulations that occupy the same space but are isolated by temporally distinct reproduction activity, which was already described for the natterjack toad (Sinsch 1988, 1992; Sinsch and Seidel 1995).

To test these hypotheses, we identified the first and last appearances each year and analyzed the consistency of individual breeding periods across years, as well as the overlap of individual breeding periods between individuals. Furthermore, we used generalized linear mixed models to test for the potential influence of age (body size as a proxy) and body condition on reproductive timing taking into account individual identity as a random factor (Bolker et al. 2009).

In the case that age and/or body condition cause the observed activity pattern, we expected significant effects of body size and/or body condition on the individual timing of breeding. On the other hand, if the second hypothesis were correct, and there exist allochronic subpopulations, the timing of the individuals’ reproductive activity should be mainly explained by the individuals’ identity. Furthermore, to make an argument for allochrony, the overlap between breeding timeframes should be minimal or non-existent. In this study, we provide new insights into the mechanisms underlying the occurrence of temporal patterns of reproductive activity within the European green toad and possibly other prolonged seasonal breeding amphibians. We discuss the significance of these mechanisms in the context of life-history trade-offs and local adaptations.

Materials and Methods

Study species

The European green toad (B. viridis) is a medium-sized toad found throughout Central and Eastern Europe. In natural environments, this species occupies steppes and wild river floodplains (Stöck et al. 2008; Indermaur et al. 2009; Frank et al. 2018). Therefore, it is adapted to dynamic, early-succession habitats and can cope well with high temperatures and salinity (Degani 1985; Schmidt and Loman 2019). As often reported, European green toads can also be found in various anthropogenic habitats, such as extraction sites (i.e., gravel and clay pits), agricultural areas, and parks (Ensabella et al. 2003; Kovács and Sas 2010; Konowalik et al. 2020; Landler et al. 2023a). The species has a prolonged breeding phase, which usually lasts from early April to July in Central Europe (Stöck et al. 2008). Usually, individuals first reproduce after the second (males) or third (females) winter. Only in rare cases, they might reproduce already after the first winter (Sinsch et al. 2007; Kutrup et al. 2011).

Study site

The study site was located in Simmeringer Haide, a suburban area in southeastern Vienna, Austria (48.169°N, 16.444°E; 155 m asl.) of which large areas are farmland (predominantly greenhouses). Toads were captured in or up to 3 m around a large, rectangular (area: 50 m × 40 m; depth: 2 m), artificial, and permanent pond, which is used as a rainwater reservoir. The rectangular pond had a 2.5 m high embankment covered in polymer sheets. Although to reach the pond toads had to ascend the 2.5 m high embankment first, we observed that the toads managed to do so without major problems. Two meters to the south of the pond there is a fence of sheet iron mounted on metal posts leaving a 5 cm gap at the bottom, thus small animals can easily pass under the wall. Directly south, a small road borders the fence. On all other sides, the pond is surrounded by greenhouses.

Capture procedure, sexing, and individual identification

We sampled toads over 3 consecutive breeding seasons on 19 (31 March 2021–16 July 2021), 21 (27 March 2022–28 July 2022) and on 26 (13 March 2023–-10 August 2023) occasions. Sampling took place at irregular intervals, but usually once or twice a week. We captured the toads by hand or by using a dipnet. Immediately afterward we measured the toads’ size (snout-vent-length [SVL]) using a manual plastic or metal, electronic calipers (Huture, ASIN: B07TJRMK7P, precision: 0.05 mm; Strend Pro DVC75, precision: 0.01 mm), and weighed them with a pocket scale (Brifit Model KA25/UF200H; precision: 0.01 g).

In accordance with previous studies (Sinsch et al. 2007), we classified all individuals < 50 mm (SVL) as immature. Size at maturity can strongly vary between and within European green toad populations, 50 mm is at the lower end of the spectrum (Sinsch et al. 2007; Landler et al. 2023b). We chose 50 mm as the classification cutoff because previous studies already showed that toads in our study population are on average exceptionally small (Landler et al. 2023b). Additionally, we only sampled in and around the spawning pond, where immatures are rarely captured, but still found a considerable number of individuals in the range of 50–55 mm. We sexed adult individuals by identifying secondary sexual characteristics, such as nuptial pads and thickened forelimbs indicative of males. Another sex indicator was the release calls that males often produced while being handled. Although on rare occasions females also produce calls, these calls have different properties and are thus clearly discernable from proper male release calls (Stöck et al. 2008). Furthermore, we looked at the coloration and overall body shape as additional indicators. Females tend to have a more contrasting pattern and a plumper body shape. In the rare cases, we were not able to identify the sex through these indicators conclusively we classified the captures as “unsexed.” To be able to subsequently identify the captures individually we took photos of the dorsal side of each animal for subsequent individual identification using the semi-automatic pattern recognition software IBEIS (Crall 2021). This software uses the same algorithm as Hotspotter (Crall 2020), which has been evaluated for European green toads and has an absolute success rate (match ranked at rank 1) of 94%–97% and a relative success rate (ranked within top 20 matches) of 97%–99% (Burgstaller et al. 2021).

Estimation of body condition

To estimate the toads’ body condition, we calculated the European green toad specific scaled mass index (SMI) using the following formula (Peig and Green 2009; Landler et al. 2023b):

$${\displaystyle \begin{array}{l}{\displaystyle \mathrm{S}\mathrm{M}\mathrm{I}=m_i\ast {\left(\frac{60\mathrm{m}\mathrm{m}}{l_i}\right)}^3}\end{array}}$$

The formula uses the individual mass (m_i) and SVL (l_i) to calculate the SMI. As proposed by Landler et al. (2023b) we used a reference length of 60 mm and a scaling exponent value of three in these calculations. By utilizing an exponent fitting our species’ growth curve, we accounted for length-related biases that may arise with other body condition indices (Gerow et al. 2004).

Statistical analysis and descriptive visualization

To show the overall seasonal patterns, we first plotted a histogram using the gghistogram function from the R package ggpubr 0.6.0 (Kassambara 2023). We set the function’s attribute add_density = TRUE to add a density curve which is a kernel density estimate for the probability function. To draw this curve gghistogram uses the function geom_density from the R package ggplot2 3.4.4 (Wickham et al. 2023). For this first histogram we included recaptures however excluded double captures (individuals found more than once per night). We then plotted the first and last capture dates for all individuals and years in a second histogram. If an individual was captured only once a year, the capture date was counted as both the first and last capture.

Because we wanted to visualize variation in arrival and departure times for individuals between years, but also between individuals, we only considered individuals that were observed at least twice in at least 2 of the 3 sampling years. These data were used in a horizontal boxplot generated using the ggboxplot function from the R package ggpubr 0.6.0 (Kassambara 2023) to show the potential overlap of residencies of different individuals.

To test for correlations between body condition or size and phenology, we performed linear mixed-effects models using the function glmmTMB from the R package glmmTMB 1.1.9 (Brooks et al. 2017). For this analysis, we used the full dataset. Although including annual single captures as the first and last dates might introduce random variation, this approach cannot bias our data in either way. In these models, we used the day of the year (of the first or last capture event) as the response variable, we added the European green toad-specific SMI (a suggested correlate of body condition [Landler et al. 2023b]), SVL, and sex as fixed factors, and the individual ID and year of capture as random factors. We performed this analysis twice, once for the first capture and another time for the last capture events. To calculate and plot model predictions, the ggpredict function from the R package ggeffects 1.4.0 was used (Lüdecke et al. 2023a). We generated the tables using the function model_table from the R package sjPlot 2.8.15 (Lüdecke et al. 2023b). To combine single plots into a figure, we used the R package patchwork 1.2.0 (Pedersen 2024).

Because we were interested in breeding activity, we excluded all data from immature toads (< 50 mm) from the statistical analysis. To test if our model results might have been affected by this cut-off size (50 mm) we built 2 more sets of models with higher cutoff sizes. One set with 55 mm and one set with 60 mm cutoff size and compared the results. The 55 mm cutoff reduced the sampling data by ~17% while the 60 mm cutoff reduced it by ~38%.

Although first and last capture dates are probably often inaccurate due to imperfect detection, we do not think this introduces a bias to our results. Imperfect detection could lead to discrepancies between dates of first detection and true arrival and between last detection and true departure. However, these inaccuracies should only increase variation in the data, thus weakening a potential effect, but would not introduce a bias in either way.

All statistical analyses and plots were created using self-written scripts within the R environment version 4.3.2 (R Core Team 2023). For data preparation and formatting aside from base R functions we used the following R packages: data.table 1.15.0 (Barrett et al. 2024), tidyverse 2.0.0 (Wickham and RStudio 2023), lubridate 1.9.3 (Spinu et al. 2023). All R scripts and corresponding raw data files are available as Supplementary Material ZIP-file (archive.zip).

Results

Over the 3 breeding seasons, we captured and released 1924 toads. Among these captures, we identified 926 individuals (152 females, 772 males, and 2 unsexed). The annual distribution of encounters showed 2 peaks for the sampled population, with the first peak around day 124 (May 4th) and the second one, less pronounced, around day 164 (June 13th) (Figure 1A). Most individuals arrived by day 150 (May 30th). The proportion of first captures was high until the end of the breeding season, meaning that the timing of the start of breeding is highly variable among individuals (Figure 1). The distribution of the last captures is similar to the distribution of the first captures but lags behind slightly, which implies that individuals generally only stay for a short time at the breeding site (Figure 1B).

Figure 1.

Histograms of green toad captures throughout the breeding seasons. (A) Stacked histogram and density functions (curves) of the total number of captures (n = 1924) over the breeding seasons 2021–2023; colors of the curves and bar subdivisions correspond to female (f; red), male (m; green), and unsexed (u; blue) animals; x-axis limits are at day 72 (March 13th) and day 222 (August 10th). (b) Overlapping histograms of annual first (light green) and last (gray) capture dates of all individuals (n = 926 each); if captured only once a year first capture equals las capture; x-axis limits are at day 81 (March 22nd) and day 222 (August 10th).

Focusing on individuals who were captured more than twice for at least 2 out of the 3 years shows that there is a large overlap between the residence times of individuals (Figure 2). Hence, these data do not provide evidence for a consistent temporal separation between individuals.

Figure 2.

Box-plot diagrams visualizing arrival times at (green) and departure times from (gray) the breeding pond of the 16 individual toads (x-axis), captured at least twice in two of the 3 sampling years.

Linear mixed effects models showed significant negative effects of SVL (P_first,SVL < 0.001; P_{last, SVL} < 0.001), and SMI (P_first,SMI < 0.001; P_last,SMI = 0.006) on day of first capture and day of last capture (Tables 1 and 2, and Figure 3). These results suggest that size, a correlation for age, and body condition affect individual breeding phenology in European green toads. In addition, sex significantly effected day of first (P_{first,sex(males)} < 0.001) and last capture (P_{last,sex(males)} < 0.015), with males arriving and leaving earlier than females (Tables 1 and 2, and Figure 3).

Table 1.

Model summary of SMI (scaled mass index) and SVL (snout-vent length) effects on day of first capture. For each predictor, we show the model estimates, their 95% confidence intervals, the critical value (z) and the significance level (P). Base factor level for the predictor sex is the level “females.” For random effects we provide the residual variance (σ²), the random intercept variance (τ00) for individuals (ID) and year, the number of IDs (N_ID) and years (N_year), the total number of observations, and marginal and conditional R-squared (R²) statistics.

	Day of the year
Predictors	Estimates	CI	z	P
Scaled mass index	−1.86	−2.52 to −1.20	–5.52	< 0.001
Snout–vent length	−1.69	−2.08 to −1.30	–8.46	< 0.001
Sex (males)	−16.39	−22.43 to −10.35	–5.32	< 0.001
Sex (unsexed)	−24.78	−56.07 to −6.51	–1.55	0.121
Random effects
σ2	451.63
τ00 ID	174.35
τ00 year	8.19
NID	788
Nyear	3
Observations	880
Marginal R2/conditionalR2	0.111 / 0.367

Table 2.

Model summary of SMI (scaled mass index) and SVL (snout–vent length) effects on day of last capture. For each predictor, we show the model estimates, their 95% confidence intervals, the critical value (z) and the significance level (P). Base factor level for the predictor sex is the level “females.” For random effects we provide the residual variance (σ²), the random intercept variance (τ00) for individuals (ID) and year, the number of IDs (N_ID) and years (N_year), the total number of observations, and marginal and conditional R-squared (R²) statistics

	Day of the year
Predictors	Estimates	CI	Z	P
Scaled mass index	−0.95	−1.63 to−0.28	−2.78	0.006
Snout–vent length	−1.23	−1.62 to −0.83	−6.07	< 0.001
Sex (males)	−7.47	−13.50 to−1.43	−2.43	0.015
Sex (unsexed)	−9.64	−39.86 to −20.58	−0.63	0.532
Random effects
σ2	515.54
τ00 ID	102.64
τ00 year	28.60
NID	793
Nyear	3
Observations	889
Marginal R2/conditional R2	0.046/0.239

Figure 3.

Effects of body size (SVL), body condition (SMI), and sex on arrival at (A, C, E) and departure from (B, D, F) the breeding site. Confidence intervals (95%) are shown as ribbons (A–D) or whiskers (E, F). The factor sex had 3 levels: female (f), male (m), and unsexed, although the latter was excluded from the figure due to small sample size (n = 2). A figure including estimates for unsexed individuals can be found in the supplements (Supplementary Figure S5).

Models with reduced data (only animals >55 mm and >60 mm) showed similar results. Only the models with animals >60 mm did not show a significant effect of sex and of those the model for last capture dates did not show a significant effect of SMI. As with reduced data loss of significance is expected, we only discuss the models with the most sampling data (animals >50 mm). We included the result tables of the models with reduced sampling data in the supplements (Supplementary Tables S1–S4).

Discussion

Our results indicate that size and body condition influence the timing of breeding, whereby toads with smaller sizes and lower body conditions tend to arrive later at the breeding site. Despite 2 activity maxima and high individual variation, there is a continuous overlap between individual reproductive timeframes. Therefore, we rule out that the 2 activity maxima were caused by allochronous subpopulations, as suggested by Sinsch (1988) and Sinsch and Seidel (1995) for natterjack toads.

The timing of breeding in our study population is clearly dependent on body size, which has been found to correlate with age in the European green toad and other bufonid species (Castellano et al. 1999; Leskovar et al. 2006; Kumbar and Lad 2017). Although there is a size overlap between age classes, especially in higher age classes, an individual cannot be clearly assigned to any specific age, and a generally positive relationship between size and age is evident. Therefore, later-breeding individuals are generally younger than early-breeding individuals within the same population. The observed second activity peak may have been caused by individuals who reached maturity only during the season. As winter survival of juvenile toads (aged <1 year) is heavily dependent on body size (Sinsch and Schäfer 2016), breeding late in the season should lower offspring survival, and thus the parents’ fitness, by shortening the time they can spend on growing before brumation. However, life-history theory predicts that if adult mortality rates are high, faster life-history strategies tend to be more successful. Reznick et al. (1990) confirmed this mechanism in an experimental setting. At our sampling site, toads suffer severely from high traffic mortality (Staufer 2022), whereby the road section crossing the sampling site is one of the most dangerous hotspots with up to hundreds of roadkills (only European green toads) documented per season (Staufer 2022). Additionally, urban areas often act as heat islands, which may increase the survival rate of small juveniles (Sinsch and Schäfer 2016). In our case, the high abundance of greenhouses at the sampling site may have caused such an effect. Thus, as adult mortality and juvenile winter survival increase, early maturation combined with late-season reproduction may yield higher fitness compared with saving resources for growth and reproduction in the following year. Hence, differences in breeding phenology may have emerged as a response to local environmental conditions. Since the last decade, life-history adaptations to anthropogenic habitats have been reported for other amphibian species (Zamora-Camacho and Comas 2017; Cayuela et al. 2022). However, these studies focused on age as annually discrete and did not investigate seasonal patterns, disregarding phenology as a key element in life-history strategies.

Our results showed a clear negative effect of body condition on the time of arrival at and departure from the breeding site. The timing of reproduction not only affects the survival probabilities of the offspring but also may influence the amount of energy an individual must spend to reproduce successfully. Competition among male toads is the highest in the early season when female density is also the highest (Figure 1A). Thus, males with high body conditions can maximize their reproductive success by spending a lot of energy and breeding early. Conversely, less competitive males might have a higher chance of finding a mate later in the season, despite low female abundance at the breeding site. Similarly, for females with low body condition, it might be more advantageous to avoid high male densities. Although more pronounced in explosive breeding scrambles, male harassment also occurs in prolonged breeders and might be energy-consuming for females (Trauth et al. 2000; Grayson et al. 2012; Dittrich and Rödel 2023). Despite temporal differences in energy effort, reproduction always has a minimum energy cost (Castellano et al. 2004). Thus, individuals with very low body conditions might not be able to spend energy at all on reproduction right away after brumation. Individuals might decide to skip a reproductive season completely (Bull and Shine 1979; Cayuela et al. 2014) or they might instead spend the early season to “bulk up” to a minimum body condition before trying to reproduce later in the year (Bieber et al. 2012; Lamarre and Franke 2017).

The existence of multiple reproductive waves in prolonged breeders confirms the key role of phenology for amphibian conservation considerations in a rapidly urbanizing world (Deng et al. 2009; Jenerette and Potere 2010; Song et al. 2018), especially for a highly mobile species, such as the European green toad (Blab 1991; Leskovar and Sinsch 2005). Many amphibians migrate annually to and from aquatic breeding sites, which entails crossing roads with the risk of death by traffic. At hotspots, amphibian migration is sometimes protected by temporary measures such as drift fences combined with pitfall traps or volunteers carrying amphibians across the roads. However, temporary measures are usually feasible only for species with short breeding periods, for example, the common toad (Bufo bufo). Prolonged breeding species, such as the European green toad, are rarely considered, and suitable long-lasting measures, such as permanent road tunnels, are still rarely realized. Our study emphasizes the importance of continual protection measures across the entire reproductive cycle of amphibians. In particular, young individuals, reproducing for the first time, arrive at the breeding ponds late in the season being exposed to higher dangers, as protection efforts usually decrease at that time. Higher mortality rates in young adults may negatively affect populations in anthropogenic habitats that are already under heavy pressure.

Asynchronous arrival between the sexes at the breeding ponds as we found it in our population (Staufer et al. 2023) is common in anurans (Lodé et al. 2005). There is likely a selection pressure for males to arrive sooner at the breeding site in order to engage in multiple matings, and thereby increase fitness. It is interesting, however, that males are also leaving the pond earlier than females, hence some females might visit the site without (or only few) males being present. Further research is clearly needed to better understand these dynamics at amphibian breeding sites.

For future studies, we recommend investigating individual variations in reproductive timing in multiple populations under varying environmental conditions. We propose that age- and size-dependent reproductive timing in European green toads and other prolonged breeding amphibians may be mediated by local habitat conditions, specifically seasonal mortality rates. To conclusively identify the causes of individual variations in the timing of breeding in amphibians we suggest future studies to investigate phenology from multiple prolonged breeder populations varying in age-dependent and seasonal survival. Furthermore, in our models, the individuals’ identity explained a large proportion of the variation in the timing of reproduction. Therefore, we conclude that there might be more unknown individual attributes affecting the reproductive timing of European green toads in need of investigation. The relatively short time span females tend to stay at the ponds resulting in unequal capture probability between the sexes. Although our large sampling effort yielded enough captured females to test for sex-specific differences, this might not be the case in studies with less sampling effort or in low-density populations. Thus, we propose implementing automated recording methods to detect more individuals, such as camera trapping (Leeb et al. 2013; Hobbs and Brehme 2017; Boynton et al. 2021).

Supplementary Material

Supplementary material can be found at https://academic.oup.com/cz.

Funding

The authors did not receive any funding for this research.

Conflict of Interest

The authors declare that they have no conflict of interest.

Authors' Contributions

S.B. and L.L. conceived the idea and concept for this study, conducted the statistical analysis, and created tables and figures. A.H. and B.K. conducted most of the data collection during 2021. M.-L.A. conducted most of the data collection in 2023. M.Sp., M.St., and S.B. were involved in the field work and organization over the whole timespan of the study. S.B. and L.L. led the writing of the manuscript. All authors contributed critically to the drafts and gave final approval for publication.

Ethics Statement

All animal captures were conducted under the permit MA22 - 230917-2020 issued by the municipal department for environmental protection of Vienna (MA 22).