Does spatial sorting occur in the invasive Asian toad in Madagascar? Insights into the invasion unveiled by morphological analyses

1 INTRODUCTION

Individuals at the edges of an expanding population may be subjected to differential selective pressures compared to conspecifics from long-colonized areas, which can possibly lead to the emergence of distinctive phenotypic traits (Chuang & Peterson, 2016). This evolutionary mechanism, named spatial sorting, is driven by intrinsic differences of dispersal capabilities among individuals and inevitably brings to the accumulation of dispersal-relevant traits at the leading-edge populations (Phillips & Perkins, 2019; Shine et al., 2011). Spatial sorting can positively select a broad range of traits, including life history, behavioral, physiological, and morphological features (Chuang & Peterson, 2016).

Although spatial sorting has been documented in naturally expanding populations, where range shifting can be induced by climate change (e.g., Hassall et al., 2009; Sanford et al., 2006), biological invasions represent the best study system for investigating this phenomenon, as invasion speeds can be manifold faster than natural expansions of species ranges. Invasive populations of the cane toad Rhinella marina (Linnaeus, 1758) in Australia are an impressive example of such a phenomenon. In this invasive species, spatial sorting has occurred on a whole set of dispersal-related traits. Leading-edge populations expanding in monsoonal floodplain and savannah areas of northern Australia show enhanced dispersal abilities due to several behavioral (Alford et al., 2009; Brown et al., 2007), locomotor, and morphological changes (Hudson et al., 2016, 2020; Phillips et al., 2006). As a result of spatial sorting, cane toads from these invasion forefronts have proportionally narrower heads, longer tibiofibulas and radioulnas, and better dispersal ability (Hudson et al., 2018; Hudson, McCurry, et al., 2016; Phillips et al., 2006). Furthermore, larger and heavier adult cane toads are leading the invasion forefront, as they have longer dispersal than smaller individuals and juveniles, which arrive months later, revealing clear spatiotemporal patterns of invasion (Brown et al., 2013). Limb proportions show a complex pattern across the whole invasive range, initially decreasing from the central to the intermediate populations, where natural selection promotes short-legged individuals. The trend sharply reverses near the invasion front, where individuals have significantly longer limbs and better dispersal ability (Hudson et al., 2016; Stuart et al., 2019). This reveals how trade-offs between contrasting evolutionary forces (i.e., natural selection vs. spatial sorting) determine complex patterns of phenotypic variation across populations (Clarke et al., 2019; Hudson, McCurry, et al., 2016). In addition, environmental features of the recipient ecosystems (Urban et al., 2008) can promote or hinder species dispersal, determining the spatial sorting onset.

Spatial sorting can strongly influence multiple features and the dispersal of invasive populations. Therefore, ascertaining the occurrence of this pattern helps to gather important information on the invasion dynamics and to identify better management responses.

The Asian common toad Duttaphrynus melanostictus (Schneider, 1799) is a smaller relative of the cane toad. Native to Southeast Asia, it was accidentally introduced to Madagascar around 2010 (Moore et al., 2015). The recent invasion of the Asian common toad in Madagascar is posing unprecedented threats to Malagasy biodiversity. First, these toads are toxic and can cause the poisoning of native predators (Marshall et al., 2018). Furthermore, due to its high fecundity, the toad could attain high abundances with multiple potential impacts on native species, especially amphibians, including competition and the transmission of pathogens (Licata et al., 2020).

The invaded area is highly suitable for this species (Vences et al., 2017), characterized by the widespread presence of human settlements associated with rice paddies, which are favorable foraging and breeding habitats for the invasive toad (Jørgensen et al., 1986; Licata et al., 2019), that have granted a successful establishment and are likely to ensure optimal conditions for its dispersal, and the consequent onset of spatial sorting. This invasion provides a unique opportunity of analyzing spatial sorting since the earliest phases of the invasion, thus greatly improving our understanding of the drivers and consequences of this process.

Here, we performed the first broad-scale analysis of morphological variation of the Asian toad across the whole invasive range in Madagascar, to characterize dispersal-relevant morphological traits and test the spatial sorting hypothesis. In so doing, we assessed whether there is a relationship between these traits and the distance from the introduction point. Furthermore, we explored invasion dynamics by testing whether size and body condition of adult toads changed with the distance from the introduction point.

2 MATERIAL AND METHODS

2.1 Study system

The invasive population of D. melanostictus in Madagascar belongs to a species complex of Asian medium-sized toads (Wogan et al., 2016), which have been accidentally introduced by humans to several regions worldwide, and, in particular, in the Wallacea (Othman et al., 2020; Reilly et al., 2017). This terrestrial, lentic-breeding amphibian is an ecological generalist and thrives in human-dominated landscapes (van Dijk et al., 2004). Females reproduce seasonally, producing over 9500 eggs per clutch (Ngo & Ngo, 2013), and, as in most anurans, are larger than males, but size and sexual dimorphism can vary considerably across populations (range of average body size across populations: females = 58.0–107.1 mm; males = 50.8–90.6 mm; Guo et al., 2019). The invasive population established in Madagascar originates from an area between Vietnam and Cambodia (Vences et al., 2017) and was likely introduced via accidental stowaway in commercial containers around 2010 (Moore et al., 2015). The putative introduction point of the toad was identified close to the SW suburbs of Toamasina (Moore et al., 2015). From here, the species expanded in all landward directions with a spreading rate of 1.6–2.5 km/year and, in 2017, it already invaded an area of 549 km², with average abundances of 184 toads/ha (Figure 1; Licata et al., 2019). In 2014, when the invasive range was estimated at only 108 km² (Moore et al., 2015), the total population abundance was estimated to be already >7 million post-metamorphic individuals (Reardon et al., 2018).

The invasive range is located at low-mid elevations (0–500 m a.s.l.) and is intersected by the Ivondro and Ivoloina rivers, on the southern and northern edges of Toamasina, respectively (Figure 1). Apparently, the Ivoloina river could have acted as a partial barrier, temporarily slowing down the northern front of the invasion (Licata et al., 2019). The invaded area is characterized by a human-modified landscape, with more than 1500 villages or small groups of huts scattered throughout the area, and is largely occupied by croplands (Zhang et al., 2020). Uncultivated land is mainly covered by grasslands and shrubland, with only a few remnant patches of forest (Zhang et al., 2020). Climatic conditions in the invasive range are highly suitable for this species (Pearson, 2015; Vences et al., 2017), being characterized by a tropical rainforest climate (Peel et al., 2007), alternating dry (August–November) and wet season (December–July), and monthly temperatures ranging from 21°C (July) to 27°C (February) (Merkel, 2021).

3 RESULTS

We measured 228 adult toads (89 males and 139 females), captured at distances of 3.7–13 km from the putative introduction point (Moore et al., 2015; Figure 1). Morphology of adult toads varied significantly between sexes (Figure 2), with females larger than males for both total length and all the other morphometric parameters (Figure 2, Table S1). Males showed proportionally longer radioulnas and tibiofibulas than females, which, however, showed slightly bigger head proportions (Figure 2; Table S1). Overall, the sexual dimorphic index SDI was 0.120.

All the relationships between SVL and the other morphometric parameters showed a significant non-linear pattern (Figure 3; Table S2). The relationship between SVL and TBL was best described by a second-order polynomial curve (adjusted R² = 0.893), indicating a deceleration of the relationship between SVL and TBL in larger individuals (Figure 3). The relationships between SVL and the other morphological traits were best fitted by a third-order polynomial curve (adjusted R²: HW =0.891; RUL =0.863), which indicate complex changes in trait proportions during toad ontogenesis (Figure 3).

The distance from the introduction point was unrelated to any of the morphological traits taken into consideration, with variation mostly attributable to the sex and/or the size of individuals (see Figure 4; Table 1). RI showed a significant linear relationship with the day of the year, indicating an increase of body conditions toward the end of the dry season.

We did not detect spatial autocorrelation in the residuals of the models with SVL or relative size of head, radioulna and tibiofibula (for all the models, Moran's I < 0.1; p > 0.1). Conversely, the residuals of the model with RI as a dependent variable showed significant autocorrelation at lags up to 1.3 km (Moran's I = 0.13; p = 0.01). This indicates that, after taking into account the effect of sex and SVL, toads living in nearby areas have similar body condition values.

4 DISCUSSION

Dispersal-related morphological traits of the Asian common toad in Madagascar apparently did not change with distance from the introduction point, but were mostly related to the sex and/or ontogenetic phase of individuals, suggesting that spatial sorting is currently not affecting the features of this invasive population. Average body size of adult toads and body conditions did not vary significantly nearby the invasion forefront (see Table 1), unlike what has been documented at the cane toad northern invasion forefront in Australia, where larger adults can be found during the first months following toad arrival (Brown et al., 2013; Phillips & Shine, 2005). In fact, juvenile toads were also found to be extremely abundant at the invasion front in Madagascar, often outnumbering adults (F.L., pers. obs.), suggesting that they are currently contributing to the range expansion. Further data collection at invasion front localities is required, as well as studies on toad abundances and dispersal abilities. In particular, dispersal ability at different ontogenetic stages needs to be ascertained, to identify the life history stages that contribute most to the range expansion.

After taking into account the distance from the introduction point, the body condition index showed a significant spatial autocorrelation at lags up to 1.3 km, suggesting that it is affected by spatially structured environmental features (e.g., food resources) not included in our analyses. Future studies should also include the effect of habitat type on the body condition of the invasive toad, as native populations exhibit lower body condition in human-modified habitats compared with forested habitats (Karraker et al., 2018).

Furthermore, toads were significantly heavier toward the end of the year (see Table 1), namely at the end of the dry season. In its native range in central Vietnam, D. melanostictus shows reproductive seasonality, breeding mainly between April and July (i.e., during the auxiliary rainy season; Ngo & Ngo, 2013), when its body condition decreases (Church, 1960). The variation of body condition index could indicate a better foraging activity before the breeding season and/or gonadal maturation, but more analyses are required to clarify the breeding phenology of this species in Madagascar.

In the study population, females were significantly larger than males which, instead, had proportionally longer limbs (Figure 2), a common pattern in toads and other amphibians (Wells, 2010). The marked female-biased sexual dimorphism in body size is well known in this and several other toad species (Guo et al., 2019; Ngo & Ngo, 2013; Wells, 2010). The observed sexual dimorphism index of the invasive toad of Madagascar (SDI = 0.12) stands among the lowest recorded for this species, where the median is 0.201 (range = −0.033–0.353; Guo et al., 2019). This is due to the larger average size of males in Madagascar (SVL = 72.5 mm; Table S1) with respect to native populations (SVL = 66.4 mm; Guo et al., 2019), whereas females are only slightly bigger than the native ones (SVL = 81.1 vs. 79.9 mm; Table S1; Guo et al., 2019). Variation of sexual size dimorphism could be determined by age differences between sexes (Liao et al., 2013). Further studies will be needed to verify whether the low SDI could be linked to altered selective pressures, which could determine an increased survival rate of adult males, reducing the size gap between sexes and, consequently, sexual size dimorphism.

The lack of spatial sorting in the dispersal-relevant traits considered is likely attributable to the short invasion history of the Asian common toad in Madagascar. Duttaphrynus melanostictus was introduced approx. 10 years before the beginning of our field activities. In its native range, the Asian common toad apparently attains sexual maturity during the second year of life (Nayak et al., 2007). In Madagascar, this process could be faster, as preliminary skeletochronological analysis suggested a continuous physiological growth of individuals (F.M. Guarino, pers. Comm.), contrary to native populations where lines of arrested growth are observed (Nayak et al., 2007). Preliminary estimates reported a rate of invasion spread ranging between 1.6 and 2.5 km/year (Licata et al., 2019), which is coherent with the annual dispersal rate obtained from radiotracking data on adult toads (approx. 2 km/year; F. Licata, unpublished data), therefore, invasion front toads could be at most from the fifth or sixth generation. Simulation studies predict that conditions at the leading edges of an expanding population can maximize the allelic expression of genes which enhance the spatiotemporal fitness of individuals in approx. 10–15 generations (Phillips & Perkins, 2019). In fact, rapid morphological changes resulting from spatial sorting have still not been documented in other invasive species with similar short invasion histories (see Brownscombe & Fox, 2012; Merwin, 2019), while studies of different species indicate that rapid changes are more likely to emerge in dispersal behavior rather than external morphology (see Lombaert et al., 2014; Merwin, 2019; Ochocki & Miller, 2017). Now, it will be important to test the occurrence of spatial sorting for behavioral or life history traits, for instance through telemetric studies, laboratory experiments, or common-garden breeding experiments, which allow investigating evolutionary changes by reducing the potential environmental influence of maternal effect. Furthermore, morphological analysis following a systematic random sampling design (e.g., SECR; Royle et al., 2013), targeting discrete and independent localities, could complement such studies and allow ascertaining the lack of spatial patterns in the morphometric traits of the invasive population.

The study of biological invasions is a key source of information on contemporary evolution (Melotto et al., 2020; Reznick et al., 2019). The implementation of long-term studies of the Asian common toad in Madagascar will allow monitoring of variation of its populations during the invasion, providing a great opportunity to understand the evolutionary dynamics of invasive species. In turn, understanding multiple processes occurring during invasion dynamics is essential for informed management strategies.