Echidnas (Tachyglossus aculeatus) are amongst the largest deep hibernators, but it is difficult to get them to hibernate normally under laboratory conditions. We measured body temperature (Tb)in 14 free-ranging echidnas using implanted data-loggers. Cooling during entry into hibernation bouts followed a Newtonian cooling curve, and conductances calculated from cooling curves were identical to those observed in cold exposed euthermic echidnas. Comparison with a reference soil temperature demonstrated that echidnas showed behavioural thermoregulation during hibernation; early in the hibernation season echidnas preferred to hibernate in cool areas, while during the coldest months they moved to warmer hibernacula, giving a preferred Tb in the range 8–10°C. Thermal buffering against excessive variation in Tb may be as important as maintaining a low Tb.
Although hibernation was once believed to represent an abandonment of control of body temperature (Tb), it is now known to be a precisely controlled physiological process(Lyman, 1982). When an endotherm enters hibernation there is a drop in metabolic rate (MR) and Tb, and as long as ambient temperature(Ta) is not too low, Tb will fall to a value typically 1–3°C above ambient. If Tb falls below a critical value, MR will increase, but when a hibernating endotherm is undisturbed and its Tb is above the critical value, Tb will passively follow Ta, i.e. it will thermo-conform (Geiser,2004).
When first described in the scientific literature, hibernation in the egg-laying echidna (Tachyglossus aculeatus) was interpreted as a confirmation that the echidna was an incomplete homeotherm(Martin, 1902). Since that time it has been demonstrated that hibernation in the echidna closely resembles that of other mammals (Grigg et al., 2004; Nicol and Andersen,2002), although with an adult body mass of 2–7 kg the echidna is one of the largest species to enter deep hibernation(Geiser, 2004; Nicol and Andersen, 2007a). Over the last 10 years we have studied hibernation in free-ranging echidnas in the field using implanted Tb data-loggers, and have collected a very large data set of more than 39 echidna years of data from 14 animals. We have previously analysed parts of this data set data to examine the relationship between Tb and hibernation bout length(Nicol and Andersen, 2000),the timing and patterns of hibernation(Nicol and Andersen, 2002),and rewarming rates and thermogenesis(Nicol and Andersen, 2007b). These data show that during the hibernation season Tbappears to follow Ta quite closely. In the laboratory, we noted a slight increase in Tb, MR and ventilation when Ta fell below 5°C(Nicol and Andersen, 2003; Nicol et al., 1992), and if the low Ta was sustained, or if cooling was too rapid, the echidna would arouse from hibernation. In this paper we make a detailed analysis of Tb in hibernating echidnas. Is there any evidence that echidnas under natural conditions regulate Tb when Ta falls below a certain level?
Our data set includes a large number of cooling episodes as echidnas enter bouts of hibernation. The physiological changes associated with entry into hibernation and torpor have been the subject of considerable debate,particularly with respect to the causal relationship between the reduction in Tb and MR, and the relative timing of these changes(Geiser, 2004; Heldmaier and Elvert, 2004). Although our data do not include measurements of MR, we are able to make some estimates of thermal conductance (C), which in turn allows some inferences to be made about MR. It has been argued, for example, that entry into hibernation and torpor is associated with an initial increase in C, which is then reduced to levels below the euthermic minimum(Snyder and Nestler, 1990). Conductance can be calculated from thermal time constants calculated from cooling curves (Robertson and Smith,1981), and this approach has been used extensively in studies on the thermal properties of ectothermic animals(Bakken, 1976; McNab, 2002). Cooling constants are independent of Ta and Tb, and are preferable to measures such as the time required to reach a specific temperature, or rates of change of temperature,as these will vary with the difference between Tb and Ta. This approach has had only limited use with endotherms; it has been used to calculate C of torpid dormice(Wilz and Heldmaier, 2000) and to characterise the entry of humming birds into torpor(Lasiewski and Lasiewski,1967). Because echidnas normally bury themselves in the substrate before entering hibernation they are sheltered from convective and radiant heat exchange with the environment. Under these circumstances Newton's Law of Cooling should be applicable, and in the second part of this paper we investigate the use of this model to characterise cooling in echidnas entering hibernation, and to estimate C.
Materials and methods
Fourteen echidnas Tachyglossus aculeatus Shaw (nine adult female,four adult male and one juvenile male, mass range 2.3–5.5 kg) were caught by hand at our field study site in the southern Tasmanian midlands, and brought to the University of Tasmania. Temperature data-loggers (Stowaway Tidbit, Onset Computer Corporation, Wilmington, DE, USA) were implanted intraperitoneally under halothane or isoflurane anaesthesia, and a tracking transmitter was glued to the spines on the lower back. After a recovery period the echidnas were returned to their site of capture. The loggers have a measurement range of –5 to –37°C, a stated accuracy of±0.2°C, and resolution of 0.16°C. Loggers were coated in a wax-polymer compound (Elvax, Dupont, Pocasset, MA, USA) to improve water resistance and prevent tissue reactions. The final package (mass approximately 25 g) was slightly larger than the 30×41×17 mm logger, and the 95%response time of the packaged logger to a step change of 10°C was about 7 min in a stirred water bath and 8 min in an unstirred bath. Initially 8K loggers were used but for later measurements we used 32K loggers capable of recording Tb hourly for 3.7 years. To extend the recording period for the 8K loggers the sampling interval for field recording was set to 96 min in some cases, but for the majority of the recordings the sampling interval was 1 h. Loggers were calibrated before and after implantation. In only one case was there a significant deviation from the specified calibration that required correction.
Meteorological data including soil temperature measured at depths of 10, 20 and 50 cm, were obtained from an Australian Bureau of Meteorology (BoM)observation station (Melton Mowbray, 147°11′44″E,42°29′35″S, elevation 218 m) approximately 4.5 km from the centre of the field site. These data were recorded daily to the nearest 1°C at 09:00 h and in some years at 15:00 h. To check the applicability of the BoM data to the field site itself, soil temperatures were measured at several points in the field site for periods of up to 8 months.
Data analyses were carried out using the statistical packages statistiXL(statistiXL, Kalamunda, Western Australia) and Statistica 6.1 (Statsoft Inc,Tulsa, OK, USA), and all results are shown as means ± s.d.
Results and data analysis
Body temperature during hibernation
Details of the timing and patterns of hibernation based on some of these data have been published previously (Nicol and Andersen, 2002), as has an analysis of rewarming(Nicol and Andersen, 2007b). A comparison of the Tb records with the various environmental measures obtained from the BoM station (air temperature and soil temperature at 10, 20 and 50 cm) showed that, during hibernation, Tb varied in parallel with soil temperature at 20 cm,although the absolute difference between the records would sometimes change. Air temperature and soil temperatures at 10 cm showed a much greater variance than Tb, while soil temperature at 50 cm had a lower variance than Tb (Table 1). Recordings of soil temperature at several locations in the field site were found to correspond closely with the BoM data: a 4-month recording of soil temperature in what we judged to be the coldest part of the field site, when plotted against the BoM data, yielded a regression line with a slope of 1.04±0.04, and an intercept of –4. Thus although the absolute values of soil temperature varied around the field area, as would be expected, the BoM data accurately reflected the pattern of temperature change. Hence the BoM soil temperature measurements at 20 cm have been used as a reference substrate temperature (Tref). In the nine winters over which we collected Tb data (1997–2005) Tref fell below 5°C on only 23 days, and only once fell below 4°C. Fig. 1shows the Tb and Tref records for an echidna (male 4572) over the 2000 hibernation period, demonstrating the remarkably close correspondence between the two temperatures. Fig. 2 shows Tb records from early in hibernation for echidna 4572, as well as three other echidnas (females 2753 and 006F and male 1E7C), all hibernating at different locations, up to about 2 km apart. Each animal appears to track the changes in Tref although the absolute temperature varies between echidnas.
Fig. 1 shows that for an individual echidna the relationship between Tb and Tref may change during the course of a hibernation season,and that these changes occur during arousals. In the initial bout of hibernation, before arousal 1, Tb is lower than Tref. Between arousals 1 and 4, Tb is very close to Tref, and between arousals 4 and 6, Tb was about 4°C above Tref. Between arousals 6 and 8, Tb was again very close to Tref, and then from arousal 8 until the final arousal, Tb was about 1°C above Tref.
In order to analyse the relationship between Tb and Tref further, Fig. 3 shows plots of hibernating Tb against Tref for two echidnas for which we have several years of data. ANOVA showed the relationships between Tb and Tref to be essentially identical for the two animals(comparison of slopes: F=0.17, P=0.68, common slope=0.63;comparison of intercepts: F=0.99, P=0.32). In Fig. 4 the frequency distribution of Tb for six echidnas (five female and one male) is compared with the distribution of Tref for the months of May and July. In May the majority of Trefreadings (70%) were in the range 9–12°C, as were 70% of Tb readings, but Tb values less than 9°C are under represented (4%) compared with Tref(17%). In July the distribution of Tb values is offset significantly compared with Tref. 58% of recorded Tref values were between 6 and 8°C, but 62% of Tb values were between 8 and 10°C.
Entry into hibernation and calculation of cooling constants
At each entry into hibernation, Tb fell until it was in equilibrium with the substrate temperature. Cooling constants were calculated for cooling intervals in which there was minimal change in Tref during the cooling period, such that Tb fell smoothly and continuously towards the equilibrium Tb. Twenty eight cooling periods from 11 echidnas (seven female, four male) met these criteria. The starting point for each cooling curve (Tb(0)) was the highest Tb from which Tb fell continuously. The final point in the cooling curve, or equilibrium Tb (Tb(eq)) was the lowest Tb. If the same Tb was recorded for more than 7 h before falling further, this value was taken to be Tb(eq). Each Tb value was then adjusted by subtracting Tb(eq), so that the adjusted Tb(eq) value became 0. A linear regression was calculated for the natural logarithm of the adjusted Tb values against time in hours since Tb(0), and the slope of this regression line is the cooling constant. Fig. 5 shows as examples arousal events for two echidnas with subsequent re-entry into hibernation,along with semi-log plots of the cooling curves. The mean Tb interval(Tb(0)–Tb(eq)) over which the 28 regressions were calculated was 21.4±2.0°C, and mean time interval was 61±13 h. The average r2 value for the regressions was 0.97±0.03. Analysis of variance showed that the between-individual variance was less than the within-individual variance(F=1.0, P=0.5), and there was no discernable relationship between the slope of the regression line and mass or sex. The mean slope of the 28 regression lines was –0.079±0.010 h–1.
Tb during hibernation bouts
Echidnas do not construct or dig a hibernaculum, but before entering hibernation they normally bury themselves in the substrate. We have found echidnas hibernating in loose soil, under sandstone slabs, in piles of leaves and bark, in grass tussocks, or in dry decaying wood in hollow logs. Most hibernating sites appear to provide shelter from convective and radiant heat exchange with the environment and from rain, and the principal means of heat loss of a hibernating echidna is likely to be conduction to the substrate in which it is buried. Figs 1 and 2 and Table 1 indicate that echidnas mostly hibernate in conditions that provide an equivalent amount of thermal buffering to approximately 20 cm of soil, and, because they are buried in the substrate, they track substrate temperature quite closely. Unlike the majority of hibernators, echidnas often move to another location during an arousal, and these moves may result in a change in Tb relative to the reference soil temperature (Tref, Fig. 1). In a number of cases,although there was an abrupt change in Tb relative to Tref at an arousal, the echidna did not actually change locations. Field observations showed that in these cases the echidna had changed its position, or the depth at which it was buried in the substrate. Thus the changes in the relationship between Tb and Tref, which occur at arousals are due to movement of the echidna to a warmer or cooler position. Figs 3 and 4 show that although Tb may track Tref closely during an individual hibernation bout, at high ambient temperatures echidnas seek out cooler places to hibernate, and at low ambient temperatures they move to warmer places. This behavioural thermoregulation is remarkably consistent from year to year and between echidnas. Mountain pygmy possums (Burramys parvus) may also move during arousals, with a resultant change in hibernating Tb, and males seek out warmer hibernacula than females, allowing them to maintain higher a Tb than females (Körtner and Geiser,1998). For echidnas, a comparison of corresponding Tb and Tref values for the last hibernation bout before arousal of reproductively active males and females showed that at any given Tref there was no difference in Tb between females and males. There were consistent differences between hibernating sites: echidnas hibernating in the middle of piles of logs were extremely well buffered from environmental variation, while the lowest Tb values were observed in echidnas hibernating in grass tussocks.
The lowest Tb recorded in this study was 4.7°C,observed in three echidnas, which is close to the minimum of 4.5°C seen in the laboratory (Nicol and Andersen,1993). A minimum Tb of 3.7°C was reported at Mount Kosciuszko at a Ta of 0.8°C(Grigg et al., 1992), but this was calculated from the pulse rate of a Tb transmitter recorded on a tape recorder, which is subject to a number of significant potential errors, particularly at low temperatures. Our field data give no indication that echidnas will increase heat production and thermoregulate if Tb falls below some critical minimum value. Nowhere in the 39+ years of Tb records from 14 echidnas is there any indication of a change in the relationship between Tb and Tref occurring anywhere other than during an arousal. Attempts to measure Ta near hibernating echidnas have suggested a higher Tb–Tadifference at low Ta, but the data are inconclusive(Grigg et al., 1992). Similarly laboratory data that showed increased metabolic rate and ventilation at low Ta (Nicol and Andersen, 1993; Nicol et al.,1992) may be the result of transient disturbances in response to the lowered Ta, or the beginning of arousal. In the field,when Tb falls too far, or too fast, echidnas will arouse and move to a warmer location to maintain Tb in the preferred range. Fig. 4 shows that in the coldest month of the year the preferred Tb of hibernating echidnas is 8–10°C, with 62% of recorded values being in that range. In our study area echidnas appear always to select hibernating conditions that allow them to be thermoconformers.
Cooling during entry into hibernation
The average value of a calculated from 28 cooling curves was–0.079±0.010 h–1. The reciprocal of the cooling constant is the time constant, and the average time constant was 12.9±1.9 h. An exponential curve falls by 1/e (≈0.37)during one time constant, and three time constants represent a 95% response,providing a useful measure of the time required to equilibrate. The 95%response time for echidna cooling was 39±6 h (N=28). Because cooling takes several days, daily torpor with a stable torpid Tb is clearly not an option for echidnas. The low rate of cooling contrasts strongly with what would be expected from an ectotherm: for an alligator of a similar mass (approx. 4 kg) the 95% cooling time in air would be about 3.25 h (Smith,1976), demonstrating that a hibernating echidna has a very low conductance.
The cooling constant, a, for an animal is a function of its body mass (Mb), specific heat (h) and its thermal conductance (C) (Bakken,1976), i.e. a=C/hMb. With a body fat composition of 25–35% (Green et al., 1992), specific heat of echidna tissues would average 3.35 J g–1 °C–1, and the average Cduring the cooling episodes would be 0.024±0.003 J g–1° h–1 C–1. Assuming a heat equivalent of oxygen of 20.1 J ml–1, this can be expressed in terms of oxygen consumption as 0.013±0.002 ml O2 g–1h–1 °C–1, identical to the value of 0.013±0.0005 measured in cold-exposed, non-hibernating Tasmanian echidnas (McNab, 1984).
Similarly, when C of dormice (Glis glis) was calculated from cooling curves, no change was found in C during entry into torpor and during hibernation, and daily torpor C was identical to the value measured in cold exposed euthermic dormice(Wilz and Heldmaier, 2000). Our results for the echidna provide no support for the suggestion(Snyder and Nestler, 1990)that entry into torpor is facilitated by changes in C.
The very close fit of a Newtonian cooling curve to our data(Fig. 5), supports the comment(Lasiewski and Lasiewski,1967) for humming birds entering torpor: “The lowering of body temperature during entry into torpor approximates a Newtonian cooling curve, suggesting that once entry into torpor is initiated, the rate of entry is determined by physical phenomena.” Although the slope of the line, and thus the calculated C, will not be affected by a constant rate of heat production, it will be affected if MR varies with Tb (Bakken and Gates,1974), but in previous laboratory experiments, we were unable to find any relationship between Tb and MR in hibernating echidnas (Nicol et al., 1992). A similar independence of MR from Tb has been observed in a variety of hibernating rodents during thermoconforming hibernation at low Tb (Buck and Barnes,2000; Heldmaier and Elvert,2004; Heldmaier et al.,1993; Ortmann and Heldmaier,2000). Although some small hibernators show a temperature effect on hibernating MR, particularly at high Tb(Geiser, 2004), the linearity of the semi-log plots of the echidna cooling curves(Fig. 5) implies that there is no significant change in echidna heat production or conductance during the cooling period.
These results help us to understand how echidnas use hibernation. It has been well documented that echidnas do not hibernate to avoid periods of extreme cold or absolute food shortage(Grigg and Beard, 2000; Nicol and Andersen, 1996; Nicol and Andersen, 2002). Grigg and Beard (Grigg and Beard,2000) have described echidnas as “using cold as a resource”; they use hibernation to conserve energy, although food is available, and conditions may be mild, when there is no necessity for them to be active. Early in the hibernation season when soil temperatures are high,echidnas will seek out cooler places to hibernate, but in the coldest months they prefer hibernacula where they can maintain Tb in the range 8–10°C. Thermal buffering against excessive variation in Tb may be just as important as maintaining a low Tb.
List of abbreviations
We thank Megan Richards, Christina Vedel-Smith, Anna Reye and David Lovell for their invaluable assistance in the field, and David Lovell for his assistance with the anaesthetic procedures. Thanks to Ian Barnes-Keoghan of the Climate Services Section of the Bureau of Meteorology for his help with the meteorological data. We are grateful to the McShane family of Stockman Stud for allowing us to carry out this work on their property. Fritz Geiser provided helpful comments on the manuscript. Financial support was provided by the Australian Research Council, the University of Tasmania Internal Research Grant Scheme, and the National Geographic Committee for Research and Exploration. This work was carried out under permit from the Tasmanian Department of Primary Industries, Water & Environment, and the University of Tasmania Animal Ethics Committee, and complies with the Tasmanian Animal Welfare Act (1993) and the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes (2004).