Wild C. oustaleti (n = 11) and C. nigrofumosus (n = 9) were collected at Bahia Mansa (32°14′22″S 71°30′54″W) on the central coast of Chile in the austral winter (June 2021) when these species occur in sympatry. We also collected seven individuals of C. nigrofumosus in summer (January 2022) at the same locality. This study site has a Mediterranean climate (mean annual precipitation = 396 mm) characterized by mild dry summers (mean monthly precipitation = 2 mm, T_min = 13°C; T_max = 19) and cold rainy winters (mean monthly precipitation = 85 mm; T_min = 7°C; T_max = 13) (di Castri and Hajek 1976). We observed no clear signs of reproduction (e.g., brood patch) or active molting in C. nigrofumosus captured during the summer. This species is not sexually dimorphic so we could not determine the sex of individuals we captured. We used mist nets and spring traps to capture birds, which were caged individually in the dark after capture to minimize stress. Biometric parameters, cloacal temperature (Tb), and blood samples were collected from each individual in the field. Blood samples were collected from the humeral vein using heparinized hematocrit capillaries. Blood was maintained in coolers (~4°C) for <2 h and then centrifuged at 10,000 rpm for 10 min to separate plasma from red blood cells. The plasma was then stored frozen until cryogenic distillation followed by oxygen isotope analysis. In addition, a subsample of whole blood was dried on two glass microscope slides and then transferred to microcentrifuge tubes and stored for δ¹³C and δ¹⁵N analysis.

Immediately after capture, birds were transported to Algarrobo, Chile ~15 min from the capture site for captive physiological measurements. While in captivity, birds consumed mealworms and water, which were available ad libitum. We measured BMR (mL O₂ h⁻¹) and total evaporative water loss (TEWL) in post-absorptive (fasted for 4-h), resting birds, during their inactive period between 21:00 and 07:00 h using standard flow-through respirometry (Tapia-Monsalve et al., 2018). We removed mealworms from the cages ~4 h before BMR measurements started to ensure a post-absorptive state. Respirometry measurements for BMR were performed on up to three birds per night. Measurements were made at an ambient temperature (T_a) of 30.0 ± 0.5°C, which is within the thermoneutral zone (TNZ), using an infrared O₂-CO₂ analyzer equipped with a hygrometer (FMS, Sable Systems®). All trials were conducted in metallic metabolic chambers (volume 2,000 mL) that received air free of water and CO₂ removed via Drierite and CO₂ absorbent, respectively, at a flow of 800 mL/min (±1%). Inside these darkened metabolic chambers, birds perched on a wire-mesh grid that allowed excreta to fall into a tray containing mineral oil, thus trapping the water from this source. Oxygen consumption was calculated according to the equation (Lighton, 2008): VO₂ = FR × 60 × (F_i O₂ − F_e O₂)/ (1 − F_i O₂), where FR is the flow rate in mL min⁻¹, and F_iO₂ and F_e O₂ are the fractional concentrations of inflow and outflow O₂ in the metabolic chamber, respectively. We calculated absolute humidity (kg/m³) of air entering and leaving the chamber as ρ = P/(T × R_w), where P is water vapor pressure of the air in Pascal, T is the dewpoint temperature in Kelvin and R_w is the gas constant for water vapor (461.5 J/kg K, Lide, 2001). P was determined using the average value of the vapor pressure of the air entering the empty chamber during a baseline period of 15 min before and after each experiment with a dew-point hygrometer located in the FMS. TEWL was calculated as TEWL = (V_e × ρ_out – V_i × ρ_in), where TEWL is in mg/mL, ρ_in and ρ_out are the absolute humidity in kg/m³ of the inlet air and the outlet air respectively, V_i is the flow rate of the air entering the chamber as given by the mass flow controller (800 mL min-1), and V_e is the flow of exiting air. V_e was calculated following as: V_e = V_i – [VO₂ × (1–RQ)] + V_H2O. V_in and VO₂ (mL min-1) are known, and we assumed a respiratory quotient (RQ) of 0.71 (Sabat et al., 2006a,b). Output from the H₂O (kPa) analyzer, the oxygen analyzer (%), and the flow meter were digitalized using a Universal Interface II (Sable Systems, Nevada, United States) and recorded on a personal computer using EXPEDATA data acquisition software (Sable Systems, Nevada, United States). To estimate BMR and TEWL, we averaged O₂ concentrations and water vapor pressures of the excurrent air stream over a 20 min period after steady state was reached, which occurs after 3 h in Cinclodes (Sabat et al., 2021). We estimated the metabolic water production (MWP) using the equivalence of 0.567 mL H₂O per liter O₂ consumed (Schmidt-Nielsen, 1997) and calculated the ratio between metabolic water production and water losses (MWP/ TEWL) for the 20 min period during which steady state was reached. The ratio MWP/TEWL is interpreted as the ability of birds to rely on metabolic water to maintain water balance.

To estimate wet thermal conductance (C_w), we also measured metabolic rates of birds at a T_a below the TNZ, on the day of capture, between 8:00 and 17:00 h, as described above for BMR. Because the “wet” thermal conductance is roughly constant in endotherms below thermoneutrality (Nicol and Andersen, 2007; Rezende and Bacigalupe, 2015; Andreasson et al., 2020), for simplicity and logistic restrictions we measured C_w at 15.0 ± 0.5°C. C_w was calculated as metabolic rate (MR₁₅) measured at 15°C using the equation MR/(Tb-Ta). In this case water and food was available for birds until they were placed in the metabolic chamber. Body mass was measured before the metabolic measurements using an electronic balance (± 0.1 g) and cloacal body temperature (Tb) was recorded with a thin Cole-Palmer copper-constantan thermocouple attached to a Digisense thermometer (Model 92,800–15) within a minute after the birds were removed from metabolic chamber to minimize the effect of manipulation on the temperature measurement (Nord and Folkow, 2019) We considered body temperatures of ≤36°C as hypothermic (Swanson et al., 2012) and all birds were normothermic at the end of 15°C or 30°C exposure trials. After each respirometry measurement, the birds were provided food and water ad libitum until their release.

To estimate the contribution of metabolic water to the body water, we used a method based on the measurement of Δ¹⁷O, which is the positive or negative deviation from the tight correlation that naturally exists between values of δ¹⁷O and δ¹⁸O (Whiteman et al., 2019). The premise of this method is that metabolic water and drinking/food water together provide 80–99% of the body water of most animals (Bryant and Froelich, 1995; Kohn, 1996), with the remaining contribution (1–20%) resulting from condensation reactions that use bound oxygen from dietary nutrients. Here, we ignore this latter contribution and acknowledge that this induces uncertainty (Whiteman et al., 2019). Of the two other sources, metabolic water is assumed to have a Δ¹⁷O value of −0.44‰ reflecting that of inhaled atmospheric oxygen (Whiteman et al., 2019; Wostbrock et al., 2020), and drinking/food water is assumed to have a Δ¹⁷O value of meteoric water with a mean Δ¹⁷O value of ~0.03‰ across a wide variety of potential sources (e.g., lakes, rivers, precipitation; Sharp et al., 2018). While extensive evaporation lowers the Δ¹⁷O values of the remaining water (Aron et al., 2021), recent studies have suggested that in many biological applications assuming a fixed value of ~0.03‰ for meteoric water is reasonable (Whiteman et al., 2019; Sabat et al., 2021). Under this assumption, a linear mixing model can be used to calculate the proportional contribution from each source (Whiteman et al., 2019). For example, an animal body water sample with a Δ¹⁷O value of −0.44‰ represents pure metabolic water and a sample with a Δ¹⁷O value of 0.03‰ represents pure meteoric water. This mixing model is in the form: Δ¹⁷O_{Body Water} = F_M × (−0.44‰) + (1 – F_M) × (0.030‰) where F_M represents the fractional contribution to body water from metabolic water, and (1 – Fm) represents the contribution from drinking/food water. In previous studies of captive sparrows (Zonotrichia capensis) and mice (Peromyscus maniculatus) this equation accurately predicted relative changes in Δ¹⁷O based on metabolic rate and drinking water intake (Whiteman et al., 2019; Sabat et al., 2021).

We cryogenically distilled body water from blood plasma on a vacuum line. We fluorinated 1.5 μL of the distilled water with BrF5 at 450°C for 15 min under a vacuum to evolve O₂, which was purified via a 6 ft. × 1/8″ (1.83 m × 3.2 mm) 60/80 Mol Sieve 13X gas chromatograph (GC) column and analyzed via dual-inlet on a Thermo Scientific™ 253 Plus isotope ratio mass spectrometer (Sharp et al., 2018) against a working reference O₂ gas at the University of New Mexico Center for Stable Isotopes (UNM–CSI; Albuquerque). The measured values of δ¹⁷O and δ¹⁸O were linearized (δ′^xO = 1,000 × ln((δ^xO/1000) + 1); x = ¹⁷O or ¹⁸O) and then used to calculate Δ′¹⁷O (δ′¹⁷O – (0.528 × δ′¹⁸O); Whiteman et al., 2019). Deviations from the normal linear mass-dependent fractionation between δ′¹⁷O and δ′¹⁸O that is defined by an arbitrary reference line with a slope (λ) of 0.528, are expressed as Δ′¹⁷O. Samples were corrected using a one-point calibration based on intermittent measurements of an in-house standard (NM2) calibrated against VSMOW-2.

In addition to using Δ¹⁷O values to understand reliance upon metabolic water, we used the combination of F_M values and δ¹⁸O values of body water to calculate estimated δ¹⁸O values of ingested pre-formed drinking/food water (δ¹⁸O_D + PF) with the equation δ¹⁸O_DFW = (δ¹⁸O_BW - (F_M) × (δ¹⁸O_Air)) / (1-F_M) and assumed δ¹⁸O_Air incorporated via respiration was 19.4‰ due to the fractionation that occurs during absorption of inhaled atmospheric oxygen. This fractionation depends on the efficiency of oxygen absorption (EO₂; Epstein and Zeiri, 1988). Although this efficiency was not measured in our study species, previous research suggests that an EO₂ of 0.4 is reasonable for small passerines (Clemens, 1988; Arens and Cooper, 2005), which in humans produces a fractionation of ~4.4‰ (Epstein and Zeiri, 1988). The estimated δ¹⁸O of ingested water generally changes by <3‰ if you apply the plausible range of fractionation values for absorbed oxygen (2–6‰) to equation 3, which is smaller than much of the naturally-occurring variation in δ¹⁸O of potential water and, therefore, unlikely to affect our conclusions.

We used δ¹³C and δ¹⁵N analysis to estimate the relative contribution of marine and terrestrial prey to the diet of Cinclodes species (Martínez del Rio et al., 2009). In general, baseline δ¹³C and δ¹⁵N values are higher in marine than terrestrial food webs (Martínez del Rio et al., 2009). For C. nigrofumosus, seasonal comparisons were made via analysis of whole blood collected in summer and winter, a tissue that integrates dietary resources assimilated during ~1–2 months prior to capture (Martínez del Rio et al., 2009). For C. oustaleti, seasonal comparisons were made by comparing the isotopic composition of two tissues: a primary feather (P1) reflecting dietary resources assimilated during the molting period that occurs during the austral summer in Cinclodes (Bertolero and Zavalaga, 2003), and blood representing the winter during which they were captured. We chose not to correct for tissue-specific discrimination between feathers and whole blood, which in a controlled experiment on passerine Dendroica coronata consuming a diet containing 97% insect were ~ 2.1‰ for δ¹³C and ca. 0.8 ‰ for δ¹⁵N (Pearson et al., 2003); these values are small in comparison to observed differences between potential marine and terrestrial food resources available at the study sites (Martínez del Rio et al., 2009).

Approximately 0.5–0.6 mg of dried whole blood or feather was weighed into tin capsules, and carbon (δ¹³C) and nitrogen (δ¹⁵N) isotope values were measured on a Costech 4,010 elemental analyzer coupled to a Thermo Scientific Delta V Plus isotope ratio mass spectrometer at UNM–CSI. Isotope values are reported using standard delta (δ) notation in parts per thousand or per mil (‰) as: δX = (R_sample/R_standard–1) where R_sample and R_standard are the ratios of the heavy to light isotope of the sample (e.g., ¹⁵N/¹⁴N) and the reference, respectively. The internationally accepted references are Vienna PeeDee Belemnite (VPDB) for δ¹³C and atmospheric N₂ (AIR) for δ¹⁵N. Within-run precision (SD) for both δ¹³C and δ¹⁵N was estimated via analysis of three proteinaceous internal reference materials and measured to be ≤0.2‰ for both isotope systems.

Because body mass did not differ between seasons, after checking normality (Shapiro–Wilk) and homoscedasticity (Levene), we used a Student t-test to compare mean BMR, TEWL, MWP/TEWL, and thermal conductance between seasons. Because some isotopic data sets (δ¹⁵N) did not meet the assumptions of normality, we decided to use non-parametric test to compare data. A Mann–Whitney U-test was used to compare δ¹³C and δ¹⁵N values between seasons for the same tissue between species and a Wilcoxon W test to compare δ¹³C and δ¹⁵N values of feathers and blood collected from the same species to provide a seasonal comparison. To compare physiological data between species in winter we used ANCOVA with body mass as a covariate. We used a Pearson product–moment correlation matric to explore the relationship between and diet and body water Δ¹⁷O using all dataset. JAMOVI software (Version 2.3., Jamovi project 2022) was used for all statistical analyses. Following Muff et al. (2021), in this paper we expressed statistical results in the language of evidence instead of the use of arbitrary value of p thresholds (e.g., p = 0.05). This approach translates approximate ranges of p-values into specific language, although the boundaries of such ranges should not be understood as hard thresholds: p < 0.01 as “strong evidence,” 0.01 < p < 0.05 as “moderate evidence,” and 0.05 < p 0.10 as “weak evidence”; for further details see Muff et al. (2021).

For C. nigrofumosus, we found no evidence that m_b differed between seasons (t₈ = −1.22, p = 0.243; all results described here are presented in Table 1). We found strong evidence that its summer whole organism BMR was ~19% higher (t₈ = −3.03, p = 0.009), and moderate evidence that their TEWL was ~29% higher (t₈ = −2.36, p = 0.034), than winter. There was no evidence that estimated water balance at 30°C (MWP/TEWL) differed between seasons in C. nigrofumosus (t₈ = 1.26; p = 0.230), but wet thermal conductance increased ~18% in summer compared with winter in this species (t₈ = −2.48; p = 0.026). A linear regression of all data from both species collected in winter provided strong evidence that log BMR (F_{1, 18} = 57.58; r² = 0.76; p = 0.001) and log TEWL (F_{1, 18} = 20.54; r² = 0.53; p < 0.001) were positively correlated with log body mass. After removing the effect of body mass, there was no evidence of difference between species in whole-organismal BMR (F_{1, 17} = 0.019, p = 0.89) and TEWL (F_{1, 17} = 1.53, p = 0.232). The ratio MWP/TEWL was also similar between species (t₈ = 0.022; p = 0.983). Finally, there was strong evidence that C_w was ~46% higher in C. nigrofumosus than C. oustaleti (t₈ = −7.08, p < 0.001).

There was no evidence of a difference between summer and winter δ¹³C (Mann–Whitney U = 17, p = 1.0) and δ¹⁵N (U = 7.5, p = 0.12) values of C. nigrofumosus (Table 2). In contrast, there was strong to moderate evidence that values of δ¹³C and δ¹⁵N differed between summer and winter in C. oustaleti as feather δ¹³C values representing summer were 4‰ higher (Wilcoxon W = −28, p = 0.015) and δ¹⁵N values were 13‰ higher (W = −28, p = 0.05) than isotope values for blood representing winter (Table 2). There was strong evidence that d¹³C and d¹⁵N values of feathers representing summer foraging differed between species (Mann–Whitney U = 0.0, p < 0.001; Table 2). There was no evidence that δ¹³C values from blood collected in winter differed between species (Mann–Whitney U = 11.0, p = 0.11), and moderate evidence of a difference for δ¹⁵N (U = 6, p = 0.04) (Table 2). There was no evidence that δ¹³C values from blood collected in winter differed between species (Mann–Whitney U = 11.0, p = 0.11), and moderate evidence of a difference for δ¹⁵N (U = 6, p = 0.04) (Table 2).

There was little or no evidence that values of Δ¹⁷O differed between summer and winter for C. nigrofumosus (t₇ = −1.76, p = 0.122, Figure 1). Using equation 2, this difference would translate to seasonal contributions from metabolic water to body water (Fm) of 24.8% in summer versus 20.7% in winter. There was no evidence that values of Δ¹⁷O (and hence Fm) differed between species in winter (t₁₂ = 0.652, p = 0.527) (Figure 1). There was also no evidence that the mean δ¹⁸O values of plasma differed between winter and summer for C. nigrofumosus (−1.4 ± 0.7‰ and −1.6 ± 0.5‰ respectively; t₇ = 0.277, p = 0.790), or between C. oustaleti (−3.5 ± 1.1‰) and C. nigrofumosus (using pooled data from summer and winter: −1.5 ± 1.3‰; t₁₇ = 1.57, p = 0.135). The mean estimated δ¹⁸O value of the combined drinking/food water ingested by C. nigrofumosus did not differ between seasons (winter −6.5 ± 2.4‰, summer −8.1 ± 0.8‰; t₇ = 1.41, p = 0.203). However, we found moderate evidence that the drinking/food water ingested by C. oustaleti (δ¹⁸O = −9.7 ± 2.6‰) was more negative than for C. nigrofumosus in winter (−6.5 ± 2.4‰, t₁₂ = 2.18, p = 0.05) (Figure 2).

Using the whole dataset (i.e., both species and seasons), there was moderate evidence that the estimated isotopic value of drinking water (δ¹⁸O_DW) was different between species (t₁₇ = 2.18, p = 0.03) and that d¹⁸O_DW was positively correlated with the δ¹³C and δ¹⁵N values of tissues (Supplementary Table S1 for statistical details). When data were analyzed separately for each species, there was moderate evidence that δ¹⁸O_DW was positively correlated with blood δ¹⁵N in C. oustaleti (r² = 0.714, p = 0.034) but no evidence of a relationship in C. nigrofumosus (r² = 0.018 p = 0.7330) (Figure 3). There was no evidence of a relationship between F_M and δ¹⁵N in either species (r² = 0.439, p = 0.15 and r² = 0.02, p = 0.73 for C. oustaleti and C. nigrofumosus respectively). Finally, when we analyzed the whole dataset we found very strong evidence that Δ¹⁷O correlated positively with δ¹⁸O values of plasma (r² = 0.54, p < 0.001).”

In passerines, the intake of moderately salty water (~400 mOsm/kg NaCl) tends to increase urine osmolality and BMR (Peña-Villalobos et al., 2014; Sabat et al., 2017). The observed seasonal increase in BMR of C. nigrofumosus, however, does not appear to be associated with an increased osmotic challenge as δ¹³C and δ¹⁵N data show this species consumed a high but similar proportion of marine resources between seasons. Previous studies suggest that the osmoregulatory physiology of Cinclodes is influenced by both ecological (diet composition) and environmental (climate) factors in a complex fashion (Sabat and del Río, 2005; Sabat et al., 2006a,b). For example, isotopic data for C. nigrofumosus from another locality in central Chile (Los Molles, 32°14’22’S 71°30’54’W) suggested a greater consumption of marine resources and an increase in plasma concentration during winter compared to summer; however, urine was more concentrated in the (hot and dry) summer than in the (cold and rainy) winter (Sabat and del Río, 2005). This pattern suggests that the effect of salt intake on osmoregulatory physiology in C. nigrofumosus depends on environmental temperature and availability of meteoric water. Furthermore, the mechanistic link between salt intake, energy expenditure, and the role of metabolic water in maintaining water balance is an intriguing topic that requires further attention.

Several studies have investigated seasonal changes in BMR and other measures of metabolic rate (e.g., RMR or FMR) in response to environment temperature, but results have revealed a noticeable difference in the magnitude of the response to thermal (i.e., seasonal) acclimatization (Arens and Cooper, 2005; Cavieres and Sabat, 2008; Noakes and McKechnie, 2020; Swanson et al., 2020) and the mechanisms for the global pattern of BMR acclimatization is poorly understood. It is generally believed that BMR in free-ranging birds is primarily driven by temperature, although it is possible that other abiotic and biotic factors such as photoperiod, reproduction, and body condition may be important (Daan et al., 1990; Chastel et al., 2003; Vézina and Williams, 2003; Zheng et al., 2008; McNab, 2009; Vezina and Salvante, 2010). Our results reveal considerable flexibility in the thermal physiology of C. nigrofumosus, as we found a ~ 20% decrease in BMR and thermal conductance and a ~ 30% decrease in TEWL in winter relative to summer, but no seasonal change in body mass. Reduced thermal conductance may enable a seasonal decline in BMR because heat is more effectively retained in winter (Speakman and Król, 2010; Rezende and Bacigalupe, 2015; Nord and Nilsson, 2019). This combination of trends suggests that the intake/storage of dietary energy and loss of heat to the environment were reduced in concert (Novoa et al., 1994; Cooper et al., 2019). Under such a scenario, the contribution of metabolic water to the body water pool would decrease. Our results contrast with the typical acclimatization response of birds from higher latitudes (McKechnie et al., 2015; Noakes and McKechnie, 2020; Swanson et al., 2022) and supports the idea that changes in BMR is not related to enhancing cold tolerance in areas where birds face milder winter minimum temperatures and more modest thermoregulatory demands.

Reproduction may also influence BMR and TEWL because behaviors such as nest building, courtship/mating, and parental care in addition to synthesizing eggs are costly and influence energy budgets (Wiersma et al., 2004; Mainwaring and Hartley, 2013; Williams, 2018). The influence of reproduction on resting rates of energy expenditure (including BMR) in free-ranging birds, however, is controversial (Nilsson, 2002; Chastel et al., 2003; Welcker et al., 2015). While seasonal increases in BMR can be explained by reproductive demands, such changes in metabolic activity may also be an adaptive response to increase metabolic water production (MacMillen, 1990; Navarrete et al., 2021). This hypothesis is consistent with results of both field- and lab-based studies that report increases in mass-specific BMR in free-ranging desert birds during the summer (Smit and McKechnie, 2010; McKechnie et al., 2015) and in captive sparrows (Z. capensis) who responded to water restriction by losing mass and increasing their mass-specific BMR. This hypothesis is also supported by the trend reported here showing a seasonal increase in the contribution of metabolic water to the body water pool of C. nigrofumosus in summer (see below). Overall, it is important to note that whole-organism metabolic rate can be affected by essentially any change in morphology or physiology, so changes in traits such as BMR can be consistent with multiple, non-exclusive mechanistic explanations.

Δ¹⁷O results suggest that the contribution of metabolic water to the total water budget in C. nigrofumosus was slightly higher in summer (~25%) than in winter (~21%), in agreement with expectations based on differences in BMR between seasons. Both estimates are slightly lower to previously reported ¹⁷O-based estimates for C. nigrofumosus (~28%) sampled from another locality ~200 km to the north of our field site (Sabat et al., 2021). For C. oustaleti, estimates of the metabolic water contribution (23%) were nearly identical to those reported for this species from the more northern locality (Sabat et al., 2021). Overall, these ¹⁷O-based estimates for the importance of metabolic water in Cinclodes in the field agree with those based on (1) respirometry under controlled conditions in the lab, where MWP/TEWL ratios vary from 28 to 33% (Table 2), and (2) doubly-labeled water (DLW) administered to free-ranging zebra finches (Taeniopygia guttata), an arid-adapted passerine (Cooper et al., 2019).

Given the importance of marine resources for both species in the winter, estimated δ¹⁸O values of preformed (drinking/food) water was expected to be close to that of seawater (0‰); however, mean (±SD) values were lower and significantly differed between C. oustaleti (−9.7 ± 2.6‰) and C. nigrofumosus (−6.5 ± 2.4‰). The mean δ¹⁸O value for C. oustaleti is nearly identical to that measured in local meteoric and tap waters (−9.7 ± 0.5‰, n = 3). Acknowledging that the end-member δ¹⁸O value for local meteoric waters is poorly constrained at present, a two-source mixing model shows that seawater contributes ~0–45% and ~ 24–66% of the total water ingested by C. oustaleti and C. nigrofumosus, respectively. These estimates differ from those obtained from a limited number of C. oustaleti (n = 3) and C. nigrofumosus (n = 3) individuals sampled at a more arid locality 200 km to the north of our study site, where ~48–100% of ingested water was sourced from seawater (Sabat et al., 2021). This difference could reflect lower terrestrial primary productivity at the more northern location, where higher consumption of marine prey would result in increased intake of seawater (Sabat et al., 2006b). Lastly, the positive correlation reported here between tissue δ¹⁵N and δ¹⁸O of the blood plasma (Figure 3; Supplementary Table S1) supports the hypothesis that Cinclodes do not directly drink seawater, but passively ingest it when consuming intertidal invertebrates (Sabat et al., 2021). Future studies that evaluate the importance of seawater in Cinclodes along a latitudinal gradient in aridity are crucial to establish the relative importance of ecological (resource use) and environmental (temperature and/or humidity) factors that influence water balance in this unique group of passerines.

A recent meta-analysis of field metabolic rate (FMR) and field water flux (FWF) collected from a diverse set of birds reported that seabirds had a higher FMR than terrestrial species, and granivores had a lower FMR than other functional groups (Song and Beissinger, 2020). Similarly, seabirds and terrestrial birds inhabiting regions with higher rainfall had higher FWF. Because the proportion of metabolic water in the total body water pool is dependent on both metabolic rate and water intake, both variables must be considered to understand water balance. Using data for species with both FMR and FWF data (n = 59) and assuming 0.567 mL of metabolic water is produced per liter O₂ consumed (Sabat et al., 2021), the average proportion of metabolic water to the body water pool for terrestrial birds and seabirds is 22 and 16%, respectively. The estimated proportion of the total water pool derived from metabolic water for Cinclodes (~21–25%) is slightly higher than terrestrial birds and in the upper range for seabirds. These results suggest that Cinclodes is more dependent on metabolic water than other terrestrial birds but cannot solely rely on seawater to maintain water balance, which likely is the result of limitations imposed by renal function. The maximum concentrating capacity of Cinclodes urine rarely exceeds 100 mOsm/kg, which is the concentration of seawater (Sabat et al., 2004), while seabird salt gland secretions can be more than twice this concentration (Sabat, 2000). Finally, it is important to note that the DLW-based estimates of FMR and FWF for free-ranging birds are typically lower than those for captive birds studied in the laboratory under controlled conditions (Bartholomew and Cade, 1963; MacMillen, 1990). Overall, the 17O-based method agrees with more direct methods using DLW (field) or respirometry (lab), which confirms the usefulness of using triple oxygen isotope measurements to estimate the water balance in the field.

Although our Δ¹⁷O-based estimates of the contribution of metabolic water to the total body water pool of Cinclodes are consistent with patterns in other physiological measurements (BMR, TEWL) and previous studies using other methods (DLW), it is important to recognize that Equation 2 is a simplification and includes assumptions that have not yet been fully explored. While meteoric waters collected in a wide variety of environmental contexts have a mean Δ¹⁷O (±SD) value of 0.03 ± 0.02‰ (Sharp et al., 2018), a more comprehensive understanding of how Δ¹⁷O values of drinking and food water available to animals in different environmental contexts is needed to refine this approach in field-based studies of water balance. For instance, extensive evaporation reduces the Δ’¹⁷O values of the residual water (Aron et al., 2021; Passey and Levin, 2021), a process that could impact meteoric waters or organism body water via evaporative water loss in animals living in arid environments. The possibility of fractionation effects on Δ¹⁷O could be captured in a flux-based model, rather than the mixing-model approach described by Equation 2. Future studies should consider these complexities and build on the nascent applications of this method (Pack et al., 2013; Whiteman et al., 2019; Sabat et al., 2021).