The thermal dependency of locomotor performance evolves rapidly within an invasive species

Abstract Biological invasions can stimulate rapid shifts in organismal performance, via both plasticity and adaptation. We can distinguish between these two proximate mechanisms by rearing offspring from populations under identical conditions and measuring their locomotor abilities in standardized trials. We collected adult cane toads (Rhinella marina) from invasive populations that inhabit regions of Australia with different climatic conditions. We bred those toads and raised their offspring under common‐garden conditions before testing their locomotor performance. At high (but not low) temperatures, offspring of individuals from a hotter location (northwestern Australia) outperformed offspring of conspecifics from a cooler location (northeastern Australia). This disparity indicates that, within less than 100 years, thermal performance in cane toads has adapted to the novel abiotic challenges that cane toads have encountered during their invasion of tropical Australia.


Many cases of intraspecific variation in thermal performance
curves likely involve adaptation as well as thermal plasticity (Gilbert et al., 2001), but the issue is especially interesting in the case of populations that have diverged only recently from a common ancestor (Eliason et al., 2011;Gilbert et al., 2001). In such a case, we can explore the potential for populations to adapt rapidly to novel challenges (Huey & Kingsolver, 1989;Maron et al., 2004Maron et al., , 2007McCann et al., 2014;Seiter & Kingsolver, 2013;Tingley et al., 2012;Winwood-Smith et al., 2015). Such an ability might be critical to population viability if conditions change, as may be the case for many species under anthropogenically induced climate change. Invasive species from plants to vertebrates offer robust model systems to examine the rate of adaptive change as they colonize new regions (Brown, Phillips, Dubey, & Shine, 2015;Llewellyn, Thompson, Brown, Phillips, & Shine, 2012;Maron et al., 2004Maron et al., , 2007Phillips et al., 2006;Tingley & Shine, 2011;Tingley, Vallinoto, Sequeira, & Kearney, 2014;Tingley et al., 2012). In many cases, such taxa exhibit rapid shifts in phenotypic traits due to adaptation rather than (or as well as) plasticity (Brown et al., 2015;Llewellyn et al., 2012;Maron et al., 2004Maron et al., , 2007McCann et al., 2014;Seiter & Kingsolver, 2013;Tingley et al., 2012).
The cane toad (Rhinella marina, formerly Bufo marinus) is a large bufonid anuran native to Latin America, but intentionally translocated to Australia in 1935 in a misguided attempt to control insect pests of commercial agriculture (Lever, 2001). The toads have since spread widely across Australia (Urban, Phillips, Skelly, & Shine, 2007), colonizing new environments. In the process, toads have undergone substantial shifts in morphology (Hudson, Brown, & Shine, 2016;Hudson, McCurry, Lundgren, McHenry, & Shine, 2016;Phillips et al., 2006), physiology (Brown & Shine, 2014;Brown et al., 2015;Llewellyn et al., 2012), and behavior (Brown et al., 2015;Gruber, Brown, Whiting, & Shine, 2017;Hudson et al., 2017). Studies on captive-reared progeny of toads collected in different parts of Australia have confirmed that many of those divergences are heritable (Gruber et al., 2017;Hudson, Brown, et al., 2016;Hudson et al., 2017) and, hence, are likely to be the result of adaptation. In recent work, we documented significant differences in thermal and hydric performance curves between wild-caught cane toads from different parts of Australia, as well as between Australian populations and those from the native range in Brazil, and a stepping-stone population in Hawai'i (Kosmala et al., 2017). To clarify the degree to which such divergences are a result of adaptation rather than plasticity, we raised toads from eastern and Western Australia under standard ("common-garden") conditions and measured their locomotor abilities under a range of thermal and hydric conditions.

| Study species and collection localities
Cane toads (Rhinella marina) are large "true toads" of the family Bufonidae (Lever, 2001). The species' native range encompasses parts of Mexico, southern Texas, and Central and South America (Easteal, 1981;Lever, 2001;Zug & Zug, 1979). Commercial sugarcane growers imported toads from French Guiana to control insect pests in plantations in Puerto Rico in the 1920s (Lever, 2001).
F I G U R E 1 Average climatic conditions in sites where adult cane toads were collected (i.e., the parents of the progeny studied in the present paper). The graphs show monthly average (dashed line), maximum and minimum (solid lines) of temperature and precipitation (bars with standard errors) based on data for three sites in Queensland (Townsville, Innisfail, and Tully) and four sites in Western Australia (El Questro, Purnululu, Wyndham, and Oombulgurri) Mean rainfall (mm)

Western Australia
From there, 150 toads were translocated to the island of Honolulu, Hawai'i, in 1932 and released in sugarcane fields (Easteal, 1981;Lever, 2001;Zug & Zug, 1979). Three years later, 101 descendants of the Hawai'ian immigrants were collected and shipped to northeastern Australia, where their progeny were released along the Queensland coast (Easteal, 1981).
In the course of their Australian invasion, toads have spread from thermally mild, high-precipitation regions of Queensland into hotter and seasonally arid regions of Western Australia (Kearney et al., 2008;Shine, 2010 Kearney et al., 2008) and experience hotter and seasonally arid conditions ( Figure 1). Monthly average temperatures are approximately 3°C hotter in our WA sites than in our QLD sites, and the WA sites average <20 mm of rain in 7 months of the year whereas all months fall above this level at our QLD sites ( Figure 1). In summary, our QLD sites provide a moister, cooler, and less seasonal environment than our WA sites.

| Methods of obtaining specimens for study
Toads were collected by hand at night, placed in damp cloth bags, kept cool to reduce stress, and transported to local laboratory facilities. They were allocated to 1-m 2 enclosures fitted with sprinklers and lights for ad libitum water and naturally occurring prey, until mating. As the enclosures were located outdoors at a field station in the Northern Territory (12°37′S 131°18′E; an environment characterized by warm temperatures year-round, but clearly defined wet and dry seasons), toads were exposed to natural thermal condi-

| Methods of experimental testing
For 1 week prior to testing (and between trials), we kept toads in the laboratory, fed them with crickets, and provided ad libitum access to water and shelter. The room was maintained at 25°C, with a 12:12-hr light cycle. Prior to trials, we kept the toads in water-filled containers at the test temperature for a minimum of 2 hr. After emptying the toads' bladders by manual pressure on the abdomen, we encouraged the animals to run along a circular wooden track inside the temperaturecontrolled room for 10 min, stimulating them to keep moving by administering gentle pokes to their urostyles. A trial concluded after 10 min of stimulation, or when the animal refused to move even after 1 min of stimulation. We recorded total distances moved in terms of body lengths travelled during the trial period. After these initial trials, we placed toads in desiccating conditions (exposed to a flow of dry air) and allowed them to dehydrate overnight until they lost 10% of their initial body mass. We then repeated the locomotor test protocol over the following days at 90% and 80% of the initial body mass. This protocol was performed at 15, 25, and 35°C, with a week of rest and rehydration in between trials at different temperatures. This protocol is described in more detail by Kosmala et al. (2017). All procedures were approved by

| Statistical analyses
These procedures generated data on the influence of the three categorical variables (test temperature, hydration level, and population of parental origin) on locomotor performance of toads raised in captivity. Data for all analyses conformed to assumptions of normality and homoscedastic variances.
We used linear mixed models (package lme4) in the open-access software R Studio version 0.99.893 (R Core Development Team, 2013), to evaluate the fixed effects of test conditions (temperature and hydration level) and population of parental origin on the toads' locomotor performance. We included individual toad ID as a random factor in the analyses to accommodate multiple measures taken from individual toads. We also included clutch nested within population as a random effect to model relatedness among siblings. We treated test temperature and hydration level as three-level categorical variables.
Temperature interacted with population (F 2,205 = 10.25, p < .0001; Table S1) in its effect on locomotor performance. Toads from both populations performed equally well at temperatures of 15°C and 25°C, but at the highest test temperature (35°C), offspring of WA parents travelled significantly farther than did the offspring of QLD parents (Figure 2). Hydration level had no significant effect on locomotor performance. The three-way interaction (temperature*hydration*population) was nonsignificant (F 4,200 = 0.0981, p = .9830) and, thus, removed from the analysis (Table S1).

| D ISCUSS I ON
In a previous study, we showed that locomotor performance of wild-caught cane toads differed among populations from Brazil, Hawai'i, and four locations within Australia (Kosmala et al., 2017).
However, we were unable to identify the proximate basis for that divergence; it might have been entirely due to toads acclimating to the conditions they experience during ontogeny. The present study extends the earlier work by showing that the divergence between two of the Australian locations (WA vs. QLD) is underpinned at least partly by adaptation. Even when raised under standard conditions, cane toads whose parents were collected in different parts of Australia differed in the degree to which their locomotor ability was affected by temperature. When tested under hot conditions, toads whose parents were collected from a hotclimate location exhibited better locomotor performance than did toads whose parents were collected from a cooler area. However, locomotor performance of toads from the two populations was near identical at lower test temperatures (Figure 2). This is exactly the pattern that one would predict if toads adapt to the thermal conditions encountered within their local area.
Introduced species exhibit both rapid evolution and phenotypic plasticity to respond to the novel conditions they encounter in the introduced range (Maron et al., 2004(Maron et al., , 2007Seiter & Kingsolver, 2013). An adaptive basis to geographic divergence in thermal performance curves, such that local populations function best under the conditions they are most likely to experience, is consistent with an extensive literature (see Introduction). However, the timescale of that adaptive process is much shorter in the current study than in prior comparisons. Cane toads had been present in Australia for only 80 years at the time we collected the parental stock and had been encountering the hot dry conditions of northwestern Australia for only 8 years at the invasion front (Hudson, Brown, et al., 2016;Kearney et al., 2008). Despite their low genetic diversity (due to founder effects in successive translocations: Slade & Moritz, 1998;Rollins, Richardson, & Shine, 2015), cane toads in Australia have evolved in ways that enhance their ability to thrive in their new range. As the invasion expanded westwards, it moved into habitats that were more extreme (both thermally and hydrically) than occur within the species' native range (Tingley & Shine, 2011;Tingley et al., 2012Tingley et al., , 2014; and in response to those novel challenges, cane toads have evolved to function more effectively at higher temperatures than they were capable of when they first colonized Australia (as inferred from toad performance in long-colonized regions). In combination with evidence of local adjustment of the critical thermal minimum to low temperatures in southern Australia (McCann et al., 2014) and of water balance to arid conditions in central Queensland (Tingley et al., 2012), our study suggests that despite the brief time span involved, cane toads are rapidly changing in ways that enable them to better exploit the novel challenges and opportunities within the Australian continent.

ACK N OWLED G M ENTS
Our work was sponsored by the Capes Foundation within the Ministry of Education, Brazil (Grant # BEX/13734-13-0) and the Australian Research Council (FL120100074).
F I G U R E 2 Effect of test temperature on locomotor performance of cane toads that were raised in captivity, but whose parents were collected from two different regions within Australia (Queensland and Western Australia). The Y-axis shows the total distance moved during a 10-min trial, expressed as body lengths. Graph shows mean values and standard errors