SUMMARY
Stripe-faced dunnarts (Sminthopsis macroura) and Ooldea dunnarts (S. ooldea) were acclimated for 2 weeks to ambient temperature (Ta) regimes of 12–22°C, 18–28°C and 25–35°C, and then measured for standard, basal (BMR) and maximum (MMR) metabolic rate using flow-through respirometry. Sminthopsis macroura maintained a stable body temperature under all experimental Ta and acclimation regimes. Although its BMR was not statistically different between the three acclimation regimes, the lower end of the thermoneutral zone (TNZ) shifted from 30°C under the 18–28°C and 12–22°C acclimation regimes to 35°C under the 25–35°C acclimation regime. MMR increased significantly at the cooler acclimation regimes. EWL increased at Ta=35°C, compared with lower Ta, in all acclimation regimes, but an increase in evaporative water loss (EWL) at Ta=10°C observed in cool acclimations did not occur at the 25–35°C regime. In contrast, S. ooldea had variable body temperature between experimental Ta in all acclimation regimes, but no acclimational shift in TNZ, which was between 30 and 35°C. Neither BMR nor MMR was affected by exposure to the three acclimation regimes. EWL did not change across Ta or with acclimation regime. Sminthopsis macroura was flexible in many aspects of its thermoregulation (involving energy and water balance) in response to thermal acclimation, presumably allowing it to balance its energy and water requirements over a broad range of climatic conditions. Sminthopsis ooldea seems to have an inflexible energetic and water balance in response to thermal acclimation, but has low nominal expenditure of either resource on thermoregulation because it thermoregulates less precisely than S. macroura. It seems that S. ooldea is adapted to a more narrow, stable climate.
INTRODUCTION
Phenotypic plasticity [or flexibility sensu Piersma and Drent (Piersma and Drent, 2003)] occurs when organisms respond to environmental variability on an ecological scale (Nespolo et al., 2001), whereby aspects of physiology, for example, are altered in a way that optimises the phenotype for certain environmental conditions (Lewontin, 1969; Feder, 1987). For example, birds and mammals enhance heat or cold resistance seasonally (acclimatisation) or in response to short-term, experimental thermal challenges (acclimation) (Nespolo et al., 2001; Soobramoney et al., 2003). Dawson (Dawson, 2003) suggested that higher ambient temperatures (either in the laboratory or as a result of seasonal warming) can lead to downregulation of metabolic capacity in birds, including changes in thermogenic capacity measured as basal metabolic rate (BMR) or maximum metabolic rate (MMR) (Bozinovic et al., 1990; Nespolo et al., 2001). Nespolo et al. (Nespolo et al., 2001) suggested, however, that MMR provides the most accurate assessment of acclimation and thermoregulatory flexibility because it incorporates BMR together with shivering and non-shivering thermogenesis, all of which can be enhanced in response to continued exposure to low temperatures. As such, MMR provides an omnibus measure of thermoregulatory response capacity.
Although it has been shown for several species that metabolic rate is flexible in response to environmental changes, less is known about the flexibility of evaporative water loss (EWL). Although evaporative heat loss (EHL) is a major thermoregulatory effector during high temperature exposure (Angilletta et al., 2010), elevated EHL at low temperature exposures, resulting from poor recuperative heat and water exchange in the respiratory tract, can be a significant contributor to thermal balance when ventilation increases to satisfy increased oxygen demand. Therefore, thermal acclimation may not only influence metabolic rate but also impact the water budget. Recent studies have shown a flexibility of EWL in birds (Williams and Tieleman, 2000) and large mammals (Ostrowski et al., 2006). This has implications for the water budget [e.g. relative water economy (RWE), the amount of water lost relative to levels of metabolic water production, as discussed by Cooper and others (Cooper and Withers, 2008; Cooper and Withers, 2009; Cooper et al., 2009; Withers and Cooper, 2009)], especially in species adapted to arid habitats with variable and unpredictable conditions (Lee and Schmidt-Nielsen, 1971; Menon et al., 1989; Williams and Tieleman, 2000). The current paucity of data on the interaction between water loss and energetics makes prediction of the effects of thermal acclimation upon these aspects of physiology difficult, but the importance of these data to the basic physiology suggests that they should be incorporated into acclimational studies.
Enhancement or contraction of metabolism can impart acclimatory responses in homeothermic mammals, but heterothermic mammals are intrinsically more flexible in their metabolic responses to acute changes in their environment (Geiser and Turbill, 2009). Although there has been interest in the acclimation of heterotherms to chronic changes in temperature, these have most often focused upon biochemical aspects of thermogenesis (Klingenspor et al., 2000; Nespolo et al., 2002) and the increased energy efficiency of re-warming after torpor (Opazo et al., 1999). At a coarser scale, however, energetic plasticity may include changes in the propensity towards torpor, the length and depth of bouts, or the intensity of torpor use in response to environmental conditions, optimising torpor use as an evasive strategy at low ambient temperature (Ta) (Geiser, 1994; Geiser et al., 2003). Although some species enter torpor only under thermal challenges of low temperature, food restriction or both (Hudson, 1978; Wang, 1989; Withers et al., 1990; Malan, 1996), the propensity of a species to enter torpor following acclimation to different environmental conditions has been investigated only recently when Munn et al. (Munn et al., 2010) showed that bouts of torpor occurred more often when food was available stochastically than when the same amount of food was available consistently. Given that torpor is viewed as an adaptation that reduces energy expenditure that is normally associated with thermoregulation (Nicol and Anderson, 1996; Grigg and Beard, 2000), a heterotherm acclimated to low temperatures might not utilise torpor as much if it has acclimatory compensation mechanisms to low temperature, other than torpor, that result in lower overall energy use (i.e. torpor use may be reduced as a response to cold in cold-acclimated heterotherms).
Sminthopsis macroura (Gould 1845) and S. ooldea Troughton 1965 represent two distinct clades of dunnart, the S. macroura and S. psammophila species groups, respectively (Archer, 1981; Blacket et al., 1999). Sminthopsis macroura has an extensive geographical distribution across Australia (McKenzie et al., 2006), whereas S. ooldea is limited to the central Australian arid environments of Western Australia, South Australia and the Northern Territory (Aslin, 1983). The considerable environmental variation across the geographic distribution of S. macroura would suggest that this species is highly adaptable, whereas S. ooldea may be more specifically adapted to, and therefore limited to, the ‘hyper-arid’ environments of the Australian centre (sensuArcher, 1981; Withers and Cooper, 2009). The two species are similar in ecology, habit and even size and mass range (Ewer, 1968; Aslin, 1983; Morton, 1983a; Morton, 1983b), but have substantially different basic energetic patterns. Sminthopsis macroura appears to be a strong thermoregulator and a classical heterotherm (i.e. spontaneous, regulated daily torpor), but S. ooldea is more thermolabile while also using torpor (Tomlinson, 2012). Where two closely related species diverge in their distributions, especially where many aspects of their biology are very similar, differences in their thermogenic plasticity may be an important driver of this divergence.
The focus of the present study was to determine the capacity of S. macroura and S. ooldea to acclimate to changes in ambient temperatures (i.e. to respond over a period of weeks rather than tolerate over a period of hours in a respirometry trial), and establish whether any differences were related to their differing distributions and thermal physiology. ‘Phenotypic plasticity’ is taken to be variability in BMR, metabolic heat production (MHP), MMR and propensity for torpor, as well as evaporative water loss (EWL), EHL and RWE following acclimation to different ambient temperature regimes. We predict that both aspects of metabolism (BMR and MMR) of the two dunnarts studied here will show plasticity resulting from acclimation, but we expect this plasticity to differ between the species. As a contributor to thermoregulation, EWL (through EHL) should be reduced at low Ta following chronic cold acclimation, reducing respiratory heat and energy loss and thus increasing RWE. Sminthopsis macroura, having a broad distribution covering a range of climatic conditions, and being a strong thermoregulator (Tomlinson, 2012), is expected to show phenotypic flexibility, essentially tailoring the thermoregulatory energy budget to chronic Ta regimes. Sminthopsis ooldea is expected to be less variable in response to acclimation to a range of chronic Ta regimes because of the lower climatic variation within its geographic distribution.
MATERIALS AND METHODS
Animal housing and acclimation regimes
Stripe-faced dunnarts (S. macroura) were captured from various locations within the natural distribution of the species in Western Australia, and Ooldea dunnarts (S. ooldea) were captured at Lorna Glen Station (26.227°S, 121.5597°E). The dunnarts were transferred to the University of Western Australia (Crawley campus) within 1 week of capture. The dunnarts were maintained on a per diem diet of approximately 1 g of minced red meat (generally kangaroo), a similar portion of canned cat food and three mealworms (Tenebrio molitor larvae). For the duration of acclimation, the dunnarts were maintained in a social arena in which they could socialise freely, but during the period of respirometry experiments the individuals were housed separately (in containers approximately 40×50×30 cm length×width×depth) so that they could be food deprived for ≥12 h prior to measurement. The dunnarts were acclimated to several Ta regimes for 2 weeks each prior to experimental measurements of their metabolic rates under these different regimes. Temperature regimes consisted of 12 h low nocturnal Ta, followed by 10°C elevation to a higher diurnal Ta. These acclimation regimes were 12–22°C, 18–28°C and 25–35°C. The 18–28°C regime was selected because it overlaps the maintenance conditions described by other studies with which these data can be compared (Geiser and Baudinette, 1985; Song et al., 1995; Song and Geiser, 1997; Song et al., 1998; Cooper et al., 2005; Withers and Cooper, 2009), whereas the higher and lower regimes encompass almost the full breadth of average temperatures experienced within the geographical distributions of the species (Australian Bureau of Meteorology, unpublished data).
All animal procedures conformed to guidelines of the National Health and Medical Research Council and were approved by the Animal Ethics Committee of The University of Western Australia under permits RA/3/100/654, RA/3/100/704 and RA/3/100/868.
Respirometry
Two flow-through respirometry systems (see Withers, 2001) consisted of airflow through a cylindrical PVC metabolic chamber (270×250 mm) that was regulated at approximately 437 ml min–1 (standard temperature and pressure) by an Aalborg GFC-17 (Aalborg, New York, NY, USA) or a Brooks 5871-A (Brooks Instrument, Hatfield, PA, USA) mass flow controller. Relative humidity of the excurrent air stream was measured by Vaisala HMP 35B and HMI 33 probes (Vaisala Oyj, Helsinki, Finland), which also measured Ta within the temperature-controlled cabinet. Excurrent air was then dried by a Drierite column (anhydrous calcium sulfate; W. A. Hammond Drierite, Xenia, OH, USA), and passed through a David Bishop 280 Combo gas analyser (David Bishop Instruments, Warwickshire, UK) which measured both [O2] and [CO2]. The analysers were interfaced to a PC using a PICO ADC-11 A/D converter (Pico Technology, St Neots, Cambridgeshire, UK) and recorded using custom-written Visual Basic v6.0 software (Microsoft, Redmond, WA, USA). Ambient temperatures for the acute 8-h measurement trials were 10, 25, 30 and 35°C (in random order), with one trial being conducted per day. Individual dunnarts were only tested once in any three consecutive days.
Baseline readings of background O2 and CO2 were established for 60 min before and after metabolic trials. Metabolic rate was measured for at least 8 h until a 20 min steady recording was obtained. Body temperature (Tb) was measured to ±0.1°C at the end of the trials using a pre-calibrated Radiospares 611-234 thermocouple reader (Wetherill Park, NSW, Australia) by inserting a thermocouple 1.5 cm into the rectum within 1 min of removing the dunnart from the chamber. These methods precluded the measurement of Tb during normothermia or torpor prior to the end of the experiment. Estimation of basic physiological parameters was made using custom-written Visual Basic v6.0 software that calculated V̇O2, V̇CO2 and EWL according to Withers (Withers, 2001), by averaging the lowest and most stable 20 min period. The point where these rates were minimal for euthermic dunnarts was considered to be BMR, and was otherwise referred to as standard metabolic rate (SMR). Wet thermal conductance (Cwet) was calculated as Cwet=V̇O2/(Tb–Ta), measured in ml O2 g h °C, and converted to J g–1 h–1 °C–1 by multiplying V̇O2 by 20.1 J ml–1 O2. Dry thermal conductance (Cdry) was subsequently calculated as Cdry={(V̇O2×20.1)-[(EWL×10–3)×latent heat of evaporation]}/(Tb–Ta). Metabolic water production (MWP; mg H2O ml–1 O2), calculated using the measured RER by the equation MWP=(0.326×RER)+0.337 after Withers (Withers, 1992) and Cooper and Withers (Cooper and Withers, 2009), was used to estimate the relative water economy (RWE; MWP/EWL) and the point of relative water economy (PRWE), which is the Ta where RWE=1. All allometric corrections for the effects of body mass (M) are made using scaling exponents reported by Withers et al. (Withers et al., 2006) of M0.75 for metabolic rate, M0.68 for EWL and M0.57 for thermal conductance.
Maximum metabolism was measured by flow-through respirometry in a helox atmosphere (19.1% O2 in He; BOC, Perth, WA, Australia) at Ta=10°C (Rosenmann and Morrison, 1974; Smith and Dawson, 1985; Geiser et al., 1996; Thomas et al., 1998; Holloway and Geiser, 2001; Geiser et al., 2003). Baseline readings of air and helox were taken for 20 min before and after the trial. The dunnarts were exposed to helox for 20 min, or until their metabolism declined from a peak, reflecting incipient hypothermia, at which time they were removed and Tb was measured. MMR was estimated using custom-written Visual Basic v6.0 software by measuring the peak V̇O2 and V̇CO2 following the introduction of a dunnart to the system, and also by averaging the metabolic rate over the first two and five minutes of the trial.
Statistical analysis
The effect of Ta on physiological variables was examined by linear regression and ANOVA, with Student–Newman–Keuls (SNK) post hoc tests. Initially a repeated-measures design was intended; however, natural attrition of individuals during the study period precluded this. Given that a repeated-measures design accounts for individual variation and so has a lower error term, the use of full-factorial ANOVA is generally more conservative than repeated-measures ANOVA (see Cohen, 2008). The maintenance of normothermic Tb was tested by linear regression against Ta, and was additionally analysed by testing the equality of variance of the Tb residuals between the Ta treatments using a Bartlett's test. Torpor data were compared with normothermia data using ANOVA. The occurrence of torpor appeared to be random amongst individuals, as not all individuals had the same responses at all experimental Ta values. Comparisons of normothermic metabolism between acclimation regimes were made within species by comparing the slope and intercept of the regressions of metabolic rate (V̇O2 and V̇CO2) below the thermoneutral zone (TNZ). Normothermic responses of each species were examined by comparing regressions below the TNZ. Comparisons of maximum metabolism, propensity towards torpor, and torpid metabolism between acclimation regimes were made by ANOVA. All statistical analyses with the exception of Bartlett's test of variance were conducted using statistiXL v.1.7 (statistiXL, www.statistixl.com). Bartlett's test of variance was conducted by hand, following Zar (Zar, 1999). Following discussion by Felsenstein (Felsenstein, 1985) and Garland and Adolph (Garland and Adolph, 1994) suggesting that two-species comparisons are less informative than phylogenetically informed analyses, these data would benefit from phylogenetic correction. Such strategies were, however, precluded by the paucity of data on metabolic physiology and acclimation responses in the Sminthopsini. Values are presented as means ± s.e.m.; sample sizes are given as n=the number of individuals and N=total sample size (i.e. the total number of measurements).
RESULTS
Body mass and acclimation
Body mass across all respirometry trials and all acclimation regimes was 17.1±0.29 g for S. macroura (n=6, N=24) and 11.1±0.13 g for S. ooldea (n=8, N=32). There was no significant change in body mass of S. macroura at the different acute Ta treatments within the 12–22°C regime (F4,25=0.190, P=0.942), the 18–28°C regime (F4,24=0.500, P=0.735) or the 25–35°C regime (F4,24=1.16, P=0.355). However, body mass was significantly higher during the 25–35°C regime than during the other two acclimation regimes (F2,87=17.8, P=3.36×10–7; Table 1). There were no significant changes in the body mass of S. ooldea within the 18–28°C regime (F3,28=0.390, P=0.761), the 12–22°C regime (F3,28=2.04, P=0.131) or the 25–35°C regime (F3,28=0.480, P=0.698), but they were heavier for the 12–22°C regime than the other two regimes (F2,93=4.87, P=0.00980; Table 1).
Body temperature and basal metabolic rate
The Tb of S. macroura was consistent between acclimation regimes (33.8±0.2°C; two-way ANOVA, Ta, F2,55=0.593, P=0.556; Table 2), and under all acute ambient temperature exposures between acclimation regimes (acclimation, F6,55=1.07, P=0.391). There was no difference in the variance of the Tb residuals at different Ta treatments between acclimation regimes (BC,11=19.5, P=0.053). In contrast, S. ooldea was thermolabile across acclimation regimes, where average Tb for S. ooldea showed significant variation with Ta (two-way ANOVA, Ta, F3,72=8.16, P=9.77×10–5) and acclimation regime (two-way ANOVA, acclimation, F2,79=3.44, P=0.0377). The variance of Tb residuals differed significantly between acute Ta treatments and the acclimation regimes (BC,11=64.9, P=1.12×10–9).
At all acclimation regimes, S. macroura showed decreasing SMR from Ta=10°C to BMR. Under the 12–22°C and 18–28°C acclimation regimes, BMR occurred at Ta=30°C, which is presumed to be within the TNZ because metabolism was higher at Ta=35°C (Table 2). Under the 25–35°C acclimation regime, however, SMR was lowest (i.e. BMR) at Ta=35°C. The mass-corrected (M–0.75) BMR was not statistically different between the three acclimation regimes in respect to V̇O2 (F2,15=2.14, P=0.153) or V̇CO2 (F2,15=1.57, P=0.240). Despite the shift in TNZ under the warmest acclimation regime, the metabolic profile below the TNZ was not statistically different between the three acclimation regimes (slope, F2,46=2.54, P=0.0890; intercept, F2,48=0.691, P=0.506), with a common regression of V̇O2=8.089–0.207Ta.
Sminthopsis ooldea maintained a pattern of decreasing SMR from Ta=10 to 30°C between all acclimation regimes. For the 12–22°C acclimation regime, metabolic rate increased between Ta=30°C (TNZ beginning between Ta=25 and 30°C and ending between Ta=30 and 35°C) and 35°C (V̇O2, t6.64=2.38, P=0.049; V̇CO2, t11.0=2.57, P=0.026; Table 2). For the 18–28°C acclimation regime there was no difference in metabolic rate between Ta=30 and 35°C (V̇O2, t10=0.831, P=0.425; V̇CO2, t5.04=0.811, P=0.454; Table 2), which is assumed to represent BMR in the TNZ (beginning between Ta=25 and 30°C and ending above Ta=35°C). For the 25–35°C acclimation regime, BMR at Ta=30°C was not different from that at Ta=35°C (V̇O2, t11=0.327, P=0.750; V̇CO2, t11=0.181, P=0.860; Table 2), again implying that the TNZ began between Ta=25 and 30°C and ended above Ta=35°C. The mass-corrected BMR (Ta=30°C) was not statistically different between the three acclimation regimes in respect to V̇O2 (F2,21=0.253, P=0.779) or V̇CO2 (F2,21=0.947, P=0.404). The metabolic response to Ta was not statistically different between the three acclimation regimes (slope, F2,55=2.252, P=0.115; intercept, F2,57=0.286, P=0.752), with a common regression of V̇O2=10.602–0.288Ta.
The normothermic respiratory exchange ratio (RER) of S. macroura did not change significantly between ambient temperatures within the 12–22°C (F3,18=0.26, P=0.850), 18–28°C (F3,20=1.07, P=0.385) or 25–35°C regimes (F3,20=0.46, P=0.708; see Tables 1, 2 for values). The RER, pooled by acclimation regime, was not significantly different between acclimation regimes (two-way ANOVA, acclimation, F2,67=0.25, P=0.780) and there was no interaction with Ta (acclimation × Ta, F6,63=0.64, P=0.697), averaging 0.79±0.04 (N=12). Sminthopsis ooldea also showed no significant difference in normothermic RER between ambient temperatures within the 12–22°C (F3,25=0.865, P=0.472), 18–28°C (F3,22=0.441, P=0.726) or 25–35°C regimes (F3,24=0.277, P=0.841; see Tables 1, 2). RER pooled by acclimation regime showed no significant difference between acclimations (two-way ANOVA, acclimation, F2,80=2.61, P=0.081), and no significant interaction with Ta (acclimation × Ta, F6,76=0.440, P=0.850), averaging 0.77±0.02 (N=12).
Evaporative water loss, thermal conductance and water economy
There were significant differences in EWL of S. macroura across the experimental Ta range for the 12–22°C acclimation regime (F3,18=3.76, P=0.029; Table 1) and the 18–28°C regime (F3,20=4.69, P=0.0120), but not the 25–35°C regime (F3,20=0.263, P=0.851). EWL increased at both Ta=10 and 35°C compared with the TNZ in the cool acclimation regimes, but not under the 25–35°C regime (Table 1). A pattern of stable Cwet and Cdry between Ta=10 and 30°C, with an increase at Ta=35°C, was consistent between the acclimation regimes. There was no significant effect of acclimation regime on Cwet (two-way ANOVA, acclimation, F2,58=1.21, P=0.308; Ta, F3,57=20.2, P=1.31×10–8; acclimation × Ta, F6,54=1.59, P=0.169) or Cdry (two-way ANOVA; acclimation, F2,58=1.31, P=0.278; Ta, F3,57=6.89, P=0.001; acclimation × Ta, F6,54=1.37, P=0.244), but there were significant increases in both at Ta=35°C during all acclimation regimes. RWE increased with decreasing Ta under all three acclimation regimes (Tables 1, 2). The RWE profile was not statistically different for the three acclimation regimes (slope F2,64=0.542, P=0.584; intercept F2,66=0.375, P=0.688), with a common regression of RWE=0.85–0.02Ta. The extrapolated PRWE was similar between all three acclimation regimes and averaged –6.8±1.2°C (Table 1).
A constant EWL between Ta=10 and 35°C was evident for S. ooldea across all three acclimation regimes (12–22°C, F3,26=1.759, P=0.180; 18–28°C, F3,24=0.548, P=0.654; 25–35°C, F3,25=1.296, P=0.298; Table 1). The pattern of stable Cwet and Cdry between Ta=10 and 30°C, with an increase at Ta=35°C, was also consistent between the acclimation regimes. There was no significant effect of acclimation regime on Cwet (two-way ANOVA, acclimation, F2,71=0.342, P=0.711; Ta, F3,70=12.9, P=1.13×10–6; acclimation × Ta, F6,67=0.277, P=0.946) or Cdry (two-way ANOVA; acclimation, F2,71=0.368, P=0.694; Ta, F3,70=4.443, P=0.007; acclimation × Ta, F6,67=0.180, P=0.981). RWE of S. ooldea showed the same pattern of increasing economy with decreasing Ta for all three acclimation regimes. The RWE profile was statistically different between the three acclimation regimes (slope F2,83=7.221, P=0.001; Tables 1, 2), and each acclimation regime had a statistically significant RWE profile: 12–22°C (F1,29=65.4, P=6.46×10–9), RWE=1.32±0.12–0.03±0.004Ta; 18–28°C (F1,26=33.9, P=3.93×10–6), RWE=0.75±0.08–0.02±0.003Ta; and 25–35°C (F1,28=72.8, P=2.84×10–9), RWE=1.04±0.08–0.03±0.003Ta.
The extrapolated PRWE was different between all three acclimation regimes, ranging from –15.0°C under the 18–28°C regime to 9.0°C under the 12–22°C regime (Table 1).
Maximum metabolic rate
There was a general decrease in MMR of S. macroura with an increase in the acclimation temperature, from 17.01±3.90 ml O2g–1h–1 and 9.50±1.85 ml CO2g–1h–1 at the 12–22°C acclimation regime to 9.53±1.21 ml O2g–1h–1 and 8.07±0.36 ml CO2g–1h– at the 25–35°C acclimation regime. After MMR was log-transformed to normalise unequal variances, maximum V̇O2 (V̇O2,max showed a significant decrease from the coolest acclimation regime to the warmest (F1,16=4.77, P=0.044; Table 3). There was no difference in the maximum V̇CO2 (V̇CO2,max) of S. macroura across the acclimation regimes (F1,16=0.240, P=0.631). There was no change in MMR (V̇O2 or V̇CO2) across the acclimation regimes for S. ooldea (O2, F2,18=0.796, P=0.466; CO2, F2,18=0.905, P=0.422; Table 3). The RER for S. macroura was 0.78±0.07, with no effect of acclimation regime (F2,15=0.557, P=0.584), but RER was 0.69±0.07 for S. ooldea, with a significant increase in the 25–35°C acclimation regime (F2,18=4.21, P=0.032).
Body temperature of S. macroura in helox at Ta=10°C decreased slightly below normothermic Tb in air at Ta=10°C for all of the acclimation regimes (12–22°C, Tb=27.5±1.15°C, ΔTb=4.7±1.45°C, t7=2.78, P=0.027; 18–28°C, Tb=27.5±1.02°C, ΔTb=6.8±0.86°C, t9=4.80, P=0.001; 25–35°C, Tb=26.8±0.75°C, ΔTb=6.4±1.50°C, t9=4.43, P=0.002). In contrast, the Tb of S. ooldea in helox at Ta=10°C was similar to normothermic Tb values in air at Ta=10°C for all of the acclimation regimes (12–22°C, Tb=30.6±2.08°C, ΔTb=0.1±1.28°C, t8.51=1.22, P=0.252; 18–28°C, Tb=30.3±1.26°C, ΔTb=2.2±1.09°C, t7=1.86, P=0.105; 25–35°C, Tb=28.7±2.67°C, ΔTb=3.0±1.51°C, t8.51=2.24, P=0.052).
Factorial thermogenic scope (i.e. MMR divided by BMR) of S. macroura (Table 3) did not differ between acclimation regimes (O2, F2,15=0.631, P=0.546; CO2, F2,15=0.314, P=0.735), averaging 12.7 (range 10.2–14.3) for V̇O2 and 15.2 (range 13.4–18.2) for V̇CO2. The thermogenic scope of S. ooldea was variable (Table 3), but was not significantly different between acclimation regimes (O2, F2,14=0.240, P=0.789; CO2, F2,16=0.381, P=0.690), averaging 7.23 (range 6.36–7.31) for V̇O2 and 10.08 (range 8.58–10.12) for V̇CO2. The thermogenic scope of S. macroura (15.21; range 13.41–18.16) in helox at Ta=10°C was 50% higher than that of S. ooldea (10.08; range 8.58–10.12), perhaps reflecting a difference in their propensities to thermoregulate in the face of thermal challenge.
Torpor
Torpor (with endogenous arousal) was observed for S. macroura only at Ta=10°C, in two trials (33%) in the 12–22°C acclimation regime and in three trials (50%) in the 18–28°C acclimation regime. In the 25–35°C acclimation regime, however, no torpor was recorded but there was one instance of hypothermia (a slow decline of metabolism compared with the rapid entry into torpor). Sminthopsis ooldea had a much greater propensity for torpor at both Ta=10 and 25°C. The proportion of S. ooldea entering torpor remained fairly constant for all acclimation regimes with four individuals entering torpor (50% of trials) at both Ta=10 and 25°C for the 12–22°C acclimation regime, and with five individuals entering torpor (62.5% of trials) at both Ta=10 and 25°C for the 18–28°C and 25–35°C acclimation regimes. The pattern of torpor for these dunnarts was to enter torpor and arouse again before the completion of a metabolic trial; as such, sample sizes for S. macroura are low and thus not amenable to statistical analysis of the effects of acclimation upon torpor. Significant differences in torpid metabolic rate (TMR) were found for S. ooldea, where there was no effect of Ta (F1,25=0.093, P=0.763), but warm acclimation significantly reduced TMR (averaged between the two acute Ta treatments) from 1.78±0.27 ml O2 g–1 h–1 at the 12–22°C acclimation regime to 0.77±0.20 ml O2 g–1 h–1 at the 25–35°C acclimation regime (acclimation, F2,24=6.64, P=0.006).
DISCUSSION
The broad context of this study was to examine whether S. macroura, a strongly thermoregulating dunnart with a broad geographic distribution, had a more flexible physiological response to chronic Ta acclimation regimes than a congeneric species, S. ooldea, a more thermolabile species with a narrow geographic distribution. The rationale was that a species with a narrow distribution should experience less selection for physiological flexibility in response to climatic variability than a species with a broad distribution, which is likely to have evolved greater phenotypic flexibility to different climatic regimes. It has already been shown that S. macroura is a stronger thermoregulator than S. ooldea across a broad Ta range within a single temperature acclimation regime (18–28°C) (Tomlinson, 2012). We show here that S. macroura (from a distribution spanning continental Australia), as well as thermoregulating well in response to a broad range of acute Ta conditions, was flexible in its thermoregulatory strategies following acclimation to a broad range of chronic temperature regimes. Overall, the propensity to thermoregulate may impart resilience to climatic variability, which is expanded by flexibility in water loss and energetics, while thermolability provides a general reduction of energetic requirement, but potentially at the expense of water economy.
Metabolic acclimation
The basic metabolic responses to acute Ta exposures reported here are similar to previous reports for both species (Geiser and Baudinette, 1987; Hinds et al., 1993; Cooper et al., 2005; Tomlinson, 2012), but S. macroura was more flexible in response to chronic thermal acclimation. The similarity in Tb of S. macroura across acute Ta exposures and between chronic Ta acclimation regimes, and the similar Tb variances over all experimental treatments (Fig. 1), indicates robust thermoregulatory control. The variable Tb of S. ooldea and the different variances between experimental Ta values (Fig. 1) indicate substantial thermolability (see Tomlinson, 2012). At high Ta acclimation, although the BMR of S. macroura had not changed, the TNZ extended to a higher Ta (Fig. 1). Sminthopsis ooldea, however, showed no flexibility in its metabolic responses to chronic Ta regimes (Fig. 1). Although we define the TNZ here based upon metabolic and thermal parameters (as opposed to an increase in EWL), similar patterns can be seen by consideration of EWL, but these were less clear and less robust to statistical analysis. The results presented here suggest that S. macroura can modify their metabolic physiology in response to acclimation to low Ta regimes, but that S. ooldea are tightly constrained by a constancy of physiological responses to chronic Ta conditions.
The metabolic response to thermal acclimation differs between various groups of mammals and birds (Hart, 1971; Heldmaier et al., 1986; Dawson and Olson, 1988; Koteja, 1996; Rose et al., 1999; Russell and Chappell, 2007). Heteromyine rodents, for example, enhance metabolism in response to cold acclimation (Hill, 1983; Russell and Chappell, 2007), as do most bird species (McKechnie, 2008). The results presented here for S. macroura concur with this expectation [and those that have previously found effects of thermal acclimation in small marsupials (e.g. Geiser et al., 2003)], although studies of small marsupials have generally reported a lack of phenotypic change to thermal acclimation (Smith and Dawson, 1984; Dawson and Olson, 1988; Withers and Hulbert, 1988), as reported here for S. ooldea.
The most obvious criticism of the conclusion that S. macroura are more physiologically flexible than S. ooldea (and that S. ooldea are not flexible) relates to the limitations of the acclimation period and intensity. The choice of a 14-day acclimation was based on results from several other studies, suggesting that this is a suitable duration for physiological acclimation (Nespolo et al., 2001; Soobramoney et al., 2003). Other studies, particularly of similar small marsupials, used periods of up to 6 weeks (Reynolds and Hulbert, 1982; Smith and Dawson, 1984; Smith and Dawson, 1985; Dawson and Olson, 1988; Withers and Hulbert, 1988). For species such as the dunnarts studied here from arid environments, seasonal changes occur over approximately 2–4 months (Australian Bureau of Meteorology, unpublished data), so an acclimation period of 2 weeks may not be long enough to elicit a full seasonal acclimatisation. Further, the breadth of the experimental acclimations described here may not be wide enough (from cool acclimation at 12–22°C to warm acclimation at 25–35°C) to trigger the fullest extent of acclimation. This seems unlikely, however, given that the range of acclimations encompasses the full extent of climatic conditions within their natural distributions. Further, at a colder acclimation regime (5–15°C) neither species was able to maintain body mass with food restriction.
There is some ambiguity in the acclimation of MMR in the dunnarts studied here. Although both dunnart species had similar MMR at the 25–35°C acclimation regime, the MMR of S. ooldea was not affected by thermal acclimation, whereas S. macroura showed evidence of acclimation in maximal V̇O2 (Fig. 2). The thermogenic scopes calculated here for both species are quite similar to the scopes of 8–10 in air reported by Smith and Dawson for kowari (Smith and Dawson, 1985), but much greater than the 2.6 in air and the 5.1 in helox previously reported for S. macroura (Geiser et al., 1996). The difference almost certainly reflects the Ta exposure during the trials, where we used helox at Ta=10°C, but Geiser et al. (Geiser et al., 1996) used helox at Ta=18°C. Sminthopsis macroura nearly doubled its thermogenic scope in response to cold acclimation, but S. ooldea showed no change in maximal metabolism, and we conclude that S. macroura adjusted MMR as an acclimatory response, but that S. ooldea did not. The different acclimation patterns of MMR between species concur with the patterns of acclimation for SMR, where S. macroura showed phenotypic flexibility but S. ooldea did not. A lack of acclimation capacity of MMR for S. ooldea is in disagreement with many other findings that peak metabolism is decreased in magnitude (Dawson and Olson, 1988; Withers and Hulbert, 1988; Chappell and Hammond, 2004; Rezende et al., 2004; Russell and Chappell, 2007) or duration (Smith and Dawson, 1985) for warm-acclimated mammals. This may reflect the reliance of S. ooldea on thermolability rather than metabolic heat production in response to cold. We conclude that S. macroura seems to be more reliant on plasticity of MMR as an element of its thermoregulation in the face of chronic exposure to different thermal environments. The thermolability of S. ooldea may contribute to the lower thermogenic scopes and conservative MMRs measured here, and it would be of interest to assess the thermoregulation of the kowari studied by Smith and Dawson (Smith and Dawson, 1985).
Acclimation of EWL and RWE
The EWL measured here for S. macroura within the TNZ was not significantly different from that predicted allometrically by Withers et al. (Withers et al., 2006) under the 12–22°C regime (t5=1.256, P=0.264), the 18–28°C regime (t5=0.636, P=0.553) or the 25–35°C regime (t5=1.523, P=0.188). Similarly, the EWL of S. ooldea within the TNZ was not different from that predicted by Withers et al. (Withers et al., 2006) during any of the acclimation regimes (12–22°C, t7=0.052, P=0.960; 18–28°C, t7=0.339, P=0.745; 25–35°C, t7=0.782, P=0.460).
The EWL of S. macroura at both ends of the Ta range (10 and 35°C) decreased with warm acclimation (Fig. 3). Although this should theoretically result in increased RWE, this was not the case (Fig. 4), but the reason for the failure of this mechanism remains unclear. Sminthopsis ooldea, in contrast, showed no change in EWL in response to acute Ta exposures or acclimation (Fig. 3). Although this inflexibility, coupled with the inflexibility of SMR (implying inflexible MWP), should result in consistent RWE between acclimation regimes, there were significant changes of RWE for S. ooldea. The pattern of these differences does not concur with any expectations of thermal acclimation, in that the warmest and the coolest acclimations had similar RWE patterns and both acclimations had generally higher RWE than the intermediate acclimation regime (Fig. 4). Furthermore, there was a strong correlation of V̇O2 with EWL for both S. macroura (r=0.73) and S. ooldea (r=0.71) at low Ta, implying that the increased EWL was due to the ventilatory demands of increased metabolic activity. The high thermal conductance of S. ooldea at Ta=10°C suggests that, at all acclimation regimes other than 18–28°C, the dunnarts did not rest properly, and this elevated metabolism and EWL could reasonably account for the increased RWE. Finally, the order of increase in conductance and RWE was not consistent with the increasing magnitude of chronic acclimation Ta, but in the temporal order of the experimental process, suggesting that there may be uncontrolled responses of S. ooldea to captivity that interact with the physiological lability of the species to produce this unexpected result (see Geiser and Ferguson, 2001). The difficulty in partitioning avenues of water loss, however, limits our current interpretation of similar levels of EWL flexibility. For example, it is likely for S. macroura that its similar levels of EWL (and hence EHL) at Ta=35 and 10°C represent different thermoregulatory and evaporative challenges: elevated EHL to dissipate heat Ta=35°C, and increased EWL at Ta=10°C resulting largely from increased ventilatory requirements for increased thermoregulatory SMR. Thus, the lack of change in EWL of S. ooldea at low Ta is presumed also to reflect the more broadly thermolabile response of this species to a Ta gradient.
Torpor and acclimation
The propensity towards torpor decreased with increasing acclimation temperatures for S. macroura. Acclimation to higher Ta regimes may shift the torpor regulation point upwards (the Ta where TMR increase to regulate torpid Tb); as such, the dunnarts would no longer have a lower critical point of Ta>10°C [established by Geiser and Baudinette (Geiser and Baudinette, 1985) and Tomlinson (Tomlinson, 2012)]. If Ta=10°C fell below the acclimated lower critical point, then entry into torpor may have become prohibitively dangerous (analogous to lower Ta thresholds in cooler acclimation regimes; see S.T., unpublished data) and the dunnarts may not attempt torpor as often under these conditions. Sminthopsis ooldea, however, seemed to have no differences in their patterns of torpor use between acclimation regimes. They did, however, have a higher TMR for the 12–22°C regime, which may be analogous to the higher BMRs that were expected (but not observed) under the same acclimation regime. Given that their normothermic SMR and BMR did not change with chronic thermal challenge, and their reliance on torpor and heterothermy at Ta that nominally approach the TNZ of most small mammals [i.e. Ta=30°C; Tomlinson et al. (Tomlinson et al., 2007) and references therein], it seems that broad thermolability and torpor are the thermoregulatory responses of S. ooldea to cold challenge, not phenotypic flexibility with chronic acclimation.
Ecophysiological and evolutionary interpretations
The ecophysiological implications of our acclimation experiments can only be generalised to possible acclimatisation because of the distinction between laboratory ‘acclimation’ and field ‘acclimatisation’ discussed by McKechnie (McKechnie, 2008) and Piersma and Drent (Piersma and Drent, 2003). With due consideration of these semantic distinctions, however, species that exhibit phenotypic flexibility in response to chronic acclimation could be expected to inhabit broad, variable geographical distributions. This is the pattern for the dunnarts studied here: S. macroura showed flexibility in BMR, MMR and thermoregulatory EWL and has a broad continental distribution, whereas S. ooldea showed less flexibility, exhibiting the same poor thermoregulatory response at all acclimation regimes, and has a comparatively small distribution that falls within a fairly predictable and restricted climatic envelope.
Given that the Sminthopsini are variable in their morphology on the basis of habitat and climate (where S. ooldea exhibits the most derived morphological characters associated with an arid environment) (Archer, 1981), finding different responses of these two species during acclimation to chronic temperature regimes accords with expectations. Aside from S. macroura and S. crassicaudata (Godfrey, 1968; Morton, 1978; Geiser and Baudinette, 1985; Nagy et al., 1988; Frey, 1991; Holloway and Geiser, 1995; Song and Geiser, 1997; Song et al., 1998; Zosky, 2002; Zosky and O'Shea, 2003; Cooper et al., 2005), very few other sminthopsines have been studied in physiological terms. Sminthopsis macroura and S. ooldea also represent different lineages of the Sminthopsini (Blacket et al., 1999). If responses to the realised ecological niche of these taxa differ along phylogenetic lines, then other members of the S. psammophila species group (which includes S. ooldea) may be similarly thermally labile and as dependent upon torpor as S. ooldea, compared with members of the S. macroura species group (e.g. S. crassicaudata and S. macroura) that are not so dependent upon torpor or thermolability. Tight homeothermy is presumably ancestral, not derived, in the Sminthopsini. This allowed S. macroura and S. crassicaudata to expand into broad distributions, as it was selection for thermolability that allowed S. ooldea to adapt specifically to a hyper-arid distribution. Following the conclusions of Felsenstein (Felsenstein, 1985) and Garland and Adolph (Garland and Adolph, 1994) that two-species comparisons suggest associations that can only be ‘proved’ by multi-specific phylogenetic replication, broader interest in the physiology of a wider range of the Sminthopsini may provide great insight into the costs and benefits of maintaining homeothermy in ecosystems of varying productivity.
LIST OF ABBREVIATIONS
- BC
Bartlett's comparison of variance
- BMR
basal metabolic rate
- Cdry
dry thermal conductance
- Cwet
wet thermal conductance
- EHL
evaporative heat loss
- EWL
evaporative water loss
- MHP
metabolic heat production
- MMR
maximum metabolic rate
- MWP
metabolic water production
- PRWE
point of relative water economy
- RER
respiratory exchange ratio
- RWE
relative water economy
- SMR
standard metabolic rate
- SNK
Student–Newman–Keuls post hoc test
- Ta
ambient temperature
- Tb
body temperature
- TEWL
total evaporative water loss
- TMR
torpid metabolic rate
- TNZ
thermoneutral zone
- V̇CO2
rate of carbon dioxide uptake
- V̇O2
rate of oxygen uptake
Acknowledgements
We acknowledge Dr Karl Brennan (DEC Western Australia, Goldfield Region) and Mr Keith Morris (DEC Western Australia, Science Division) for access to Lorna Glen Station for the collection of S. ooldea, and Drs Scott Thompson, Jessica Oates and Graham Thompson (Coffey Environments), Drs Stewart Ford and Victoria Cartledge (ecologia Environment), Mr Roy Teale (Biota Environmental Scientists), and Mr David Steane and Mr Paul Bolton (Outback Ecology) for the capture and provision of S. macroura.
FOOTNOTES
FUNDING
This research was funded by the School of Animal Biology, University of Western Australia, and the Holsworth Wildlife Research Endowment (ANZ Philanthropy Partners). S.T. was supported by an Australian Postgraduate Award.