Population persistence is informed by the ability of individuals to cope with local abiotic conditions, which is commonly mediated by physiological traits. Among biphasic amphibians, juveniles—which are infrequently studied but play a key role in amphibian population dynamics—are the first life stage to experience terrestrial conditions following the aquatic larval stage. To illuminate phenotypic variation that may allow juveniles to survive the physiological challenges presented by this transition, we examined respiratory surface area water loss (RSAWL) and standard metabolic rates (SMR) among juveniles reared under common larval conditions for five salamander species (Ambystoma annulatum, A. maculatum, A. opacum, A. talpoideum, and A. texanum) collected across ~200 km of latitude in Missouri, USA. We found that SMR described 34% of variation in RSAWL, suggesting that physiological water conservation may be limited by energetic regulation among these species, and vice versa. On average, species differed in juvenile SMR and residual values of RSAWL (corrected for body size/shape) by 0.04 mL CO₂  and 0.16, respectively, possibly because of distinct species ecologies. For example, A. annulatum had higher SMR and RSAWL compared to broadly distributed study species, potentially associated with a relatively narrow range of environmental conditions experienced across the small geographic distribution of A. annulatum. Latitude correlated negatively with temperature and precipitation, and positively with RSAWL, suggesting that variation in RSAWL may be adaptive to local conditions. We provide evidence that species differences likely have a genetic basis, reflecting selection favoring species divergence to effectively use distinct microhabitats.

Using a flow through Field Metabolic System (FMS; Sable Systems International, Inc.), we collected rates of water loss and CO₂ production from each juvenile salamander 24 July–12 August 2017 (see Table 1 for population-specific sample sizes). A 120 VAC/60 Hz pump pushed ambient air through a drying column to an eight-line external manifold and flow control bar (FB-8; Sable Systems International, Inc.). The FB-8 released the dry air at 60 mL/min to one of eight 5 cm × 16 cm (volume ~ 314 mL) capped cylindrical glass animal chambers (RC-2; Sable Systems International, Inc.) housed within an incubator set to 27°C. Each air stream flowed from one of the chambers, out of the incubator, and into an eight-channel multiplexer (RM-8; Sable Systems International, Inc.). The multiplexer directed the airstream from each chamber in sequence to the FMS, which recorded water vapor pressure (WVP; kPa) and percent CO₂ content of the air. The air stream was scrubbed of water vapor before CO₂ analysis. The multiplexer cycled through the chambers five times, with six-minute intervals of data collection per chamber. We recorded temperature inside the incubator each second using a stationary thermistor probe (SA2; Sable Systems International, Inc.).

We selected juvenile salamanders for respirometry trials haphazardly across species, mesocosms, and date of metamorphosis until all n = 357 individuals had been measured. Due to the moss substrate, we could not assure the absence of food for all salamanders prior to trials, but no salamanders were fed within 48 hours of respirometry. We recorded snout-to-vent length (SVL), total length (TL), and the diameter of the midpoint between the forelimbs and hind limbs (DM) of each juvenile immediately prior to data collection using digital calipers (± 0.01 mm). Before the assigned trial, we soaked each juvenile in spring water for ≥20 minutes to ensure hydration across individuals. We then allowed each individual to walk across a paper towel to remove excess water and collected the mass of each salamander (± 0.0001 g). To limit movement within the animal chamber, we secured each salamander in a 3 cm × 11 cm 0.13-cm mesh plastic envelope closed with binder clips. We subsequently loaded a single juvenile into each animal chamber, with the first chamber remaining empty in all trials as a baseline, and up to seven juveniles being measured per trial. Following the trial period, we again collected the mass of each individual and allowed the salamanders to soak in spring water for ≥20 minutes to facilitate rehydration. Each juvenile was subsequently replaced into its individual container. No salamanders perished as a result of these procedures. Data from animals that we observed to have defecated during the trial period (n = 4) were excluded from further analyses. 

We used batch file processing in ExpeData PRO software v. 1.9.10 (Sable Systems International, Inc.) to transform WVP data to respiratory surface area water loss (RSAWL) for each juvenile by first calculating water vapor density (ρ_v; g ∙ m^-3; Equation S1 in Supporting Information) and then evaporative water loss (EWL; mg ∙ hr^-1; Equation S2 in Supporting Information) as in Riddell and Sears (2015) (details available in Physiological Calculations subsection of Supporting Information).