Cold acclimation enhances cutaneous resistance

Functional Ecology 2003 Cold acclimation enhances cutaneous resistance Blackwell Science, Ltd to inoculative freezing in hatchling painted turtles, Chrysemys picta G. C. PACKARD* and M. J. PACKARD Department of Biology, Colorado State University, Fort Collins, CO 80523-1878, USA Summary 1. Neonatal Painted Turtles, Chrysemys picta (Schneider 1783), typically spend the first winter of their life inside the shallow, subterranean nest where they completed incubation the preceding summer. This behaviour commonly causes hatchlings in northerly populations to be exposed in mid-winter to life-threatening conditions of ice and cold. Neonates apparently withstand such exposure by remaining unfrozen and becoming supercooled. 2. Recently hatched turtles are unable to resist freezing by inoculation (i.e. freezing that is induced by the penetration of ice crystals through the integument and into body compartments from frozen soil). Consequently, newly hatched animals that make contact with ice are highly susceptible to freezing when temperature goes below the equilibrium freezing point for their body fluids (c. 0 7 C). Freezing is fatal, even when temperature does not go below 2 C, so newly hatched turtles have only a very limited tolerance for exposure to ice and cold. 3. Acclimation (or acclimatization) by neonatal turtles to low temperatures in the interval between hatching and the start of winter elicits an increase in effectiveness of the cutaneous barrier to the penetration of ice into body compartments from the environment. Acclimated hatchlings coming into contact with ice in frozen soil usually remain unfrozen and supercooled, and survival by supercooled turtles is uniformly high. This thermally induced enhancement of the cutaneous barrier to inoculation is key to an overwintering strategy based on supercooling. Key-words: Freezing, inoculation, integument, supercooling, turtle Functional Ecology (2003) Ecological Society Introduction The Painted Turtle (Family Emydidae: Chrysemys picta (Schneider 1783)) has a natural-history unlike that of most other chelonians occurring in the northern USA and southern Canada. Whereas neonates of other freshwater turtles usually emerge from their subterranean nest in late summer or autumn and move to a nearby marsh, lake or stream to spend their first winter in the unfrozen depths (Ultsch 1989; Ernst, Barbour & Lovich 1994), hatchling Painted Turtles typically remain inside their shallow (8 14 cm) nest throughout their first winter and do not emerge above ground until the following spring (Ultsch 1989; Ernst et al. 1994). This behaviour commonly causes neonatal Painted Turtles to be exposed in mid-winter to ice and cold, with temperatures in some nests falling below 10 C *Author to whom correspondence should be addressed. E-mail: packard@lamar.colostate.edu (DePari 1996; Packard 1997; Packard et al. 1997a; Weisrock & Janzen 1999). Hatchling Painted Turtles are small (4 5 g), ectothermic animals with a limited heat capacity, so their body temperature presumably tracks very closely the temperature of the nest (see Claussen & Zani 1991). Despite the apparent threat to their survival, most turtles withstand the exposure and emerge from their nest after the ground thaws in the spring (DePari 1996; Packard 1997; Packard et al. 1997a; Weisrock & Janzen 1999). Neonatal Painted Turtles withstand exposure to ice and cold in mid-winter by remaining unfrozen and supercooled (Packard et al. 1997b, 1999a; Packard & Packard 2001). The key to this adaptive strategy is an integument that resists the penetration of ice into body compartments from frozen soil, because turtles would otherwise be penetrated by ice and caused to freeze (i.e. they would be inoculated) soon after their temperature went below the equilibrium freezing point for body fluids (see Layne, Lee & Huang 1990; Layne 1991; 94

95 Cold tolerance in hatchling turtles Costanzo, Bayuk & Lee 1999). The skin of newly hatched animals, however, seems not to be particularly effective at resisting the inward penetration of ice (Costanzo et al. 2000), so profound changes in permeability of the integument seemingly occur in the interval between hatching of turtles in late summer and the onset of wintery weather to enable them to remain unfrozen and supercooled under conditions encountered in nature. Cutaneous resistance to the penetration of ice into body compartments from frozen soil clearly is fundamental to an adaptive strategy based on supercooling, so knowledge of the origin and nature of the cutaneous barrier is critical to understanding the natural history of neonatal Painted Turtles. Accordingly, we examined the cutaneous barrier to inoculative freezing in recently hatched Painted Turtles and in animals that had been acclimated for a month at high or low temperatures. The goals of the experiment were (i) to verify that cutaneous resistance to penetration of ice does, indeed, increase in neonatal turtles as winter approaches; and (ii) to determine whether any increase in the resistance to inoculation is an outcome of normal developmental processes occurring in the period after hatching or whether it results from an adaptive physiological response to increasing cold. Materials and methods Eggs of Painted Turtles were collected in mid-june 2001 from 24 newly constructed nests located in and around the headquarters compound at the Valentine National Wildlife Refuge, Cherry County, Nebraska, USA. The eggs were packed in coolers containing damp vermiculite and held at ambient temperature until they could be transported to Colorado State University on 23 June. On arrival in the laboratory, eggs were assigned to covered refrigerator boxes containing damp vermiculite (water potential = 100 kpa) and immediately placed into incubation at 27 C. The boxes were re-weighed once each week, at which time water was added to the substrata to replace that which had been lost by evaporation. Hatching occurred mid- to late August. One turtle per clutch was prepared for study shortly after the last egg hatched (i.e. on 28 or 29 August); one turtle per clutch was assigned at the same time to a box of moistened vermiculite for acclimation to a high temperature (24 C); and one turtle per clutch was assigned to a box of moistened vermiculite for acclimation to a low temperature (3 C). Boxes with turtles for the high-temperature acclimation were moved immediately into an environmental chamber set at 24 C and then held there until early October. Boxes with turtles for the low-temperature acclimation, on the other hand, were placed into a second environmental chamber, and the temperature in this second chamber was reduced by approximately 2 C every third day until the desired temperature was reached in early October. Although we were unable to replicate the three treatments in different environmental chambers, the treatments themselves were so very different that they and not random variation among chambers are likely to underlie any differences among the putative treatment groups (Packard & Packard 1993a). Hatchlings in all three treatments were dried, cleaned (by brushing them with a small paintbrush), and then weighed on an electronic balance. A copper/constantan thermocouple (28 gauge) next was attached to the carapace of each animal with epoxy resin, after which the turtles were placed individually into artificial nests constructed in pint-volume canning jars filled to 80% of their capacity with moist soil. The soil was a loamy sand collected at the Valentine NWR, and it was hydrated to field capacity (i.e. water potential was high). Additional quantities of soil were tamped into spaces around each animal to ensure that its carapace, plastron, and exposed extremities made extensive contact with the substratum, thereby to maximize the probability that the turtle would later be caused to freeze by inoculation (Salt 1963). The jars were closed and placed into an environmental chamber set at approximately 2 C. Only one chamber was available in late August to apply a cold stress to unacclimated turtles, but two environmental chambers were available in early October to apply a cold stress to acclimated animals. Consequently, the exposure of acclimated animals to subzero temperatures was replicated for environmental chamber whereas that for unacclimated turtles was not. Ends of the thermocouples were attached via a multiplexor to a Campbell Scientific datalogger (model CR-10, Logan, Utah, USA) that was programmed to monitor temperatures every 10 min. The controls for the environmental chamber(s) were then re-set to bring temperature in the jars to approximately 0 4 C, which is below the equilibrium freezing point for water in moist soils (Bodman & Day 1943) but above that for extracellular fluids of hatchling painted turtles (approximately 0 7 C; Storey et al. 1991; Packard & Packard 1995; Costanzo et al. 2000). Each of the jars next was opened; a few chips of shaved ice were placed onto the soil; and the jar was closed and returned to the environmental chamber. Supercooled water in the soil immediately began to freeze, as indicated by the production of a long-lasting exotherm as water in the soil changed phase from liquid to solid (Fig. 1). Temperature in the chamber remained at the nominal level ( 0 4 C) for the next 4 days so that water in the soil could freeze to an equilibrium. Controls on the environmental chamber(s) were then re-set to lower temperature by 1 C day 1 to a minimum near 2 0 C. This minimum is too high to cause turtles hatching on vermiculite to freeze spontaneously by heterogeneous nucleation (Packard, Packard & McDaniel 2001), so we assume that all cases in which turtles froze were caused by inoculation (see below). Temperatures actually recorded in the jars departed

96 G. C. Packard & M. J. Packard Fig. 1. Temperature profiles for three acclimated Painted Turtles in an experiment assessing the susceptibility of neonatal animals to freezing by inoculation. The dashed line in each panel is an approximation to the equilibrium freezing point ( 0 7 C) for body fluids of hatchlings (Storey et al. 1991; Packard & Packard 1995; Costanzo et al. 2000). The middle and bottom panels reveal spikes in temperature (animal exotherms) resulting from the release of latent heat of fusion by water in the turtles changing phase from liquid to solid. somewhat from the nominal value owing to the spatial variation that is commonplace inside environmental chambers (Measures, Weinberger & Baer 1973). Minima were maintained for 7 days, at the end of which temperature in the chamber(s) was raised rapidly to 2 0 C. After the soil had thawed completely, turtles were removed and their condition (alive or dead) was ascertained. Data were downloaded from the datalogger(s) to a PC, and a temperature profile was generated for each animal. These temperature profiles were searched for spikes in temperature indicative of freezing by water in bodies of the turtles (Fig. 1). Animals were scored for the presence or absence of an exotherm in their temperature profile and for whether or not they survived the exposure. Data for the body mass of hatchlings, for the minimum temperature to which animals were exposed, and for the temperature at which freezing occurred were examined by ANOVA in an incomplete, mixed-model design (Proc Mixed in SAS version 8 1). Treatment was a fixed effect in the analyses, and environmental chamber for applying the cold-exposure was a random effect. Scores for whether turtles froze during their exposures to ice and cold were studied by logistic regression in an ANOVA-like design (Proc Genmod in SAS version 8 1). A preliminary analysis treated only the data for the two groups of acclimated turtles so that we could assess possible effects of the environmental chamber used to apply the cold stress (see Magnusson 2001). When we discovered that chamber had no apparent influence on the outcome, we pooled data across the chambers and performed a one-way analysis on data for the several treatments. Finally, we also studied data for survival by turtles in relation to the presence of a distinct freezing exotherm in their temperature profile. The analysis again was performed by logistic regression (Proc Genmod). Results Acclimated turtles weighed several hundred milligrams less than unacclimated animals (Table 1), owing at least in part to catabolic breakdown of nutrients in the yolk in the month-long interval between the two sets of tests. The mass of animals acclimated at high temperature cannot be distinguished from that of turtles acclimated at low temperature, however. Acclimated turtles were exposed to slightly higher temperatures during cold-exposure than were unacclimated hatchlings (Table 1). The difference is small, and it is unlikely to underlie any difference in the incidence of freezing or in survival by turtles in the several treatments. Table 1. Summary of data for hatchling Painted Turtles in an experiment assessing the susceptibility of animals to freezing by inoculation. Water in the soil did not freeze in one jar containing a turtle acclimated to low temperature, so sample size for that treatment was 23 instead of 24. Data are expressed as least squares means ± SEM. Means that share the same superscript cannot be distinguished at P = 0 05 Variable Unacclimated High acclimated Low acclimated P n 24 24 23 Body mass (g) 4 42 ± 0 09 a 4 0 ± 0 09 b 4 11 ± 0 09 b 0 006 Minimum temp. ( C) 2 2 ± 0 03 a 2 0 ± 0 03 b 2 0 ± 0 03 b <0 001 Temp. at inoculation ( C) 1 9 ± 0 16 a 1 6 ± 0 14 b 1 8 ± 0 18 a,b 0 022

97 Cold tolerance in hatchling turtles Fig. 2. Proportions of Painted Turtles that froze in an experiment assessing the susceptibility of neonatal animals to freezing by inoculation. Unacclimated animals were studied shortly after hatching, whereas acclimated turtles were studied after they had been acclimated to high (24 C) or low temperature (3 C). Sample size = 24 for unacclimated turtles and for turtles acclimated at the high temperature; n = 23 for hatchlings acclimated at the low temperature (because soil in one jar did not freeze and the data for this animal were excluded). Unhatched columns represent the animals whose temperature profiles yielded clear, freezing exotherms; hatched column represents unacclimated animals that are suspected to have frozen without producing a distinct exotherm (see Discussion). Columns that share the same letter cannot be distinguished statistically at P = 0 05. Fewer freezing exotherms were detected in temperature profiles for turtles that were acclimated to low temperature than for unacclimated turtles or for hatchlings that were acclimated to the high temperature (Fig. 2). The apparent incidence of freezing did not differ between unacclimated animals and those that were acclimated to the high temperature (Fig. 2). Freezing of body fluids generally occurred at higher temperatures, however, for turtles acclimated to high temperature than for animals in the other groups (Table 1). We believe that this difference is important, and will return to this point in the Discussion. Survival by hatchlings in this experiment was linked to whether or not a freezing exotherm was detected in their temperature profile (Table 2). Freezing exotherms tended to be detected in the temperature profiles of animals that died, whereas exotherms tended to be absent from the profiles of surviving turtles (Table 2). The 10 animals that seem to have died without freezing, however, were all in the group of unacclimated turtles in the first test. Again, we believe that this observation is important, and will return to this point later. Comparing the survival of unacclimated turtles with that of acclimated animals obscures a potentially important detail, so we concentrate now on data for just the two groups of acclimated hatchlings. Whereas this comparison confirms that a significant proportion of frozen hatchlings died while unfrozen animals invariably survived (Table 2), it also reveals that the few turtles to survive freezing were animals that had been acclimated to low temperature (Table 2). Thus, a tolerance for limited freezing whatever its significance may be expressed only by animals that have become acclimated to cold. Discussion SUSCEPTIBILITY TO INOCULATION Judging by turtles whose temperature profile contained a distinct freezing exotherm, hatchlings that were acclimated to low temperature were much less susceptible to inoculation than turtles in the other two treatments, and unacclimated animals and turtles acclimated at high temperature were about equally susceptible to the penetration of ice into body compartments from the environment (unfilled bars in Fig. 2). Thus, the acquisition of cutaneous resistance to the penetration of ice evidently requires that hatchlings be exposed to low temperatures, but postnatal development per se plays no apparent role in this process. It is important to note, however, that the 10 unacclimated turtles for which no Table 2. Survival by hatchling Painted Turtles in relation to the presence of a freezing exotherm in their temperature profile. Water in the soil did not freeze in one jar containing a turtle acclimated to low temperature, so sample size for that treatment was 23 instead of 24. Likelihood ratio χ 2 = 28 26, d.f. = 1, P < 0 001, in an analysis comparing survival by frozen and unfrozen turtles across the full experiment. Likelihood ratio χ 2 = 13 99, d.f. = 1, P < 0 001, in an analysis comparing survival by frozen and unfrozen turtles that were acclimated at high or low temperature. Likelihood ratio χ 2 = 26 99, d.f. = 1, P < 0 001, in an analysis comparing survival by frozen and unfrozen turtles that were acclimated to high temperature. Likelihood ratio χ 2 = 5 95, d.f. = 1, P = 0 015, in an analysis comparing survival by frozen and unfrozen turtles that were acclimated to low temperature Condition of hatchling Unacclimated High acclimated Low acclimated Exotherm No exotherm Exotherm No exotherm Exotherm No exotherm Alive 0 0 0 6 4 17 Dead 14 10 18 0 2 0

98 G. C. Packard & M. J. Packard freezing exotherm was detected nevertheless died during their exposure to ice and cold whereas large numbers of unfrozen animals in the other treatments withstood such exposure (Table 2). Consequently, a corollary of the preceding interpretation is that recently hatched Painted Turtles have essentially no tolerance for cold, even in those instances where they seemingly remain unfrozen. On the other hand, temperature records for most of the unacclimated hatchlings in the first phase of the investigation were unusually noisy, owing to the malfunction of a baffle directing the flow of air inside the one environmental chamber available to us at that time and to a consequent increase in variability in temperatures maintained by the chamber. Because of the increase in background noise, only the most pronounced of exotherms could be identified with certainty. This observation leads to an alternative interpretation of the evidence specifically, that the high level of background noise in this test masked freezing exotherms for 10 of the 24 animals in the sample, and that all 24 of the unacclimated turtles actually froze during the course of their exposure to ice and cold. Three pieces of evidence support this view. First, when an animal begins to freeze at a temperature that is only slightly below the equilibrium freezing point for body fluids, the resulting exotherm is relatively small (Fig. 1, bottom). A rebound in temperature of less than 0 2 C could easily have been obscured by the background noise in the temperature recordings for unacclimated hatchlings. Second, the temperature at which inoculation occurred in those animals yielding identifiable exotherms was lower for unacclimated turtles than for hatchlings that were acclimated at a high temperature (Table 1). We can think of no reason why turtles acclimated at the high temperature would begin to freeze at temperatures 0 3 C higher, on average, than those at which unacclimated animals would begin to freeze, and consequently suggest that the mean for unacclimated turtles was biased by the systematic exclusion of values from the upper half of the distribution. Only the biggest exotherms could stand out from the background noise, and the biggest exotherms were produced by animals that were inoculated at the lowest temperatures. Finally, all the turtles in the unacclimated group were dead at the end of their exposure (Table 2) an outcome that we might expect for frozen animals but not necessarily for unfrozen ones exposed only to high subzero temperatures. If the second interpretation is correct, the unacclimated turtles actually were completely lacking in resistance to inoculation, and the cutaneous barrier to penetration of ice was enhanced very slightly in animals that were acclimated at high temperature (Fig. 2). By this view, postnatal development of the integument led to a modest increase in its resistance to inoculation. However, the integumental barrier still was substantially more effective in turtles that were acclimated to low temperature than in those that were acclimated to high temperature (Fig. 2) an observation which continues to point to the likely importance of natural acclimatization to cold in preparing animals for overwintering in the field. What changes might occur in the integument during postnatal development and acclimation to enhance barrier function, and do the changes occur over the entire surface of the body? We speculate that the epidermis of skin overlying the bony shell is essentially impermeable from the time of hatching and that enhancement of the cutaneous barrier involves flexible skin on the head, neck, and extremities as well as tissues surrounding the umbilicus. The keratinized layer of epidermis overlying the shell is essentially a syncytium containing β-keratin, which causes the epidermal plates to be hard and, by our hypothesis, impermeable (Alexander 1970; Alibardi 1999, 2002; Alibardi & Thompson 1999). The keratinized layer of epidermis on flexible skin (and presumably the soft tissues surrounding the umbilicus), on the other hand, is formed of distinct cells containing α-keratin (Alexander 1970; Alibardi 1999). We suspect that intercellular spaces in the α-keratin layer provide channels through which ice crystals can grow to inoculate recently hatched turtles, and that deposition of lipid in those spaces occurs during both maturation of the integument and, especially, acclimation/acclimatization of hatchlings to low temperatures (Willard et al. 2000). TO FREEZE OR NOT TO FREEZE? Most workers believe that supercooling is the key to survival by animals in the field when the temperature in their nest goes below 3 to 4 C, because hatchlings in laboratory tests do not recover from freezing when temperature dips only briefly below this level (Churchill & Storey 1992; Costanzo et al. 1995; Packard et al. 1999b). However, the means by which turtles withstand exposure to higher but nevertheless subzero temperatures is subject to debate. One group of investigators suggests that hatchlings overwintering in wet soils are likely to exploit a tolerance for freezing to withstand short periods of cold weather (e.g. early and/ or late in the winter season; Lee & Costanzo 1998), whereas another group argues that neonates are likely to exploit supercooling under virtually all circumstances (Packard et al. 1997b, 1999a). Neonatal Painted Turtles are able to survive exposure to high subzero temperatures in the laboratory either by supercooling (Packard & Packard 1995; Packard et al. 1997b, 1999a) or by recovering from freezing (Churchill & Storey 1992; Costanzo et al. 1995; Packard et al. 1999b), so the generally high survival by animals overwintering in relatively warm nests in the field could be based on either a tolerance for freezing or an ability to supercool (Packard 1997; Packard et al. 1997a; Weisrock & Janzen 1999; Nagle et al. 2000). The competing explanations for cold-tolerance nevertheless differ in important ways. For example, the

99 Cold tolerance in hatchling turtles concept of ecologically relevant freeze-tolerance in neonatal painted turtles is based on the supposition that hatchlings overwintering in wet soils are at high risk of being inoculated by ice crystals growing into body compartments from the environment (Costanzo et al. 1995, 1998, 2001), whereas the concept of supercooling is based on the contrasting supposition that hatchlings are not at substantial risk of being inoculated even when they make extensive contact with ice in the surrounding soil (Packard et al. 1997b, 1999a). If animals are caused to freeze, they presumably would survive at high subzero temperatures by tolerating freezing. If they are able to avoid freezing, however, they presumably would survive by tolerating supercooling. The question of freeze-tolerance vs. supercooling at high subzero temperatures therefore can be reduced in large measure to the question of whether a turtle overwintering in a wet soil is likely to freeze by inoculation when water in the soil freezes. Evidence from the current investigation has bearing on the question of susceptibility to inoculation at high subzero temperatures. We focus here on data for the incidence of inoculation (and subsequent survival) among hatchling Painted Turtles that were acclimated to the low temperature (Table 2). The soil in which turtles were buried was at a high water potential (i.e. the soil was wet), and soil was packed into spaces around animals to ensure that each of them made extensive contact with the medium. These conditions presumably maximized the probability that turtles would be penetrated by ice and freeze (Salt 1963; Costanzo et al. 1998). None the less, few of the animals actually froze, and those that had the misfortune of freezing were less likely to survive than animals that remained unfrozen (Table 2). More of the turtles certainly would have been inoculated had we extended the exposure beyond 7 days (Packard & Packard 1993b; Packard et al. 1997b), but mortality among the frozen animals would have increased as well (Attaway, Packard & Packard 1998). Mortality among unfrozen (supercooled) turtles, however, probably would have remained low or non-existent for upwards of 3 weeks (Packard & Packard 1997; Hartley, Packard & Packard 2000). Whereas findings from the current investigation and from earlier studies of turtles from North Dakota and Minnesota (Packard et al. 1997b, 1999a) do not rule out the possibility that an overwintering hatchling might recover from freezing, they do lead us to suggest that freezing at high subzero temperatures is likely to be infrequent and that it is inherently disadvantageous. The ultimate test for this prediction will come from observations made on turtles overwintering in natural nests. However, an earlier attempt at such a test was unsuccessful, because minute-to-minute variation in temperature inside nests obviated the detection of exotherms produced by turtles in the process of freezing (Packard 1997). Consequently, new technologies may be needed to resolve this issue unequivocally. Acknowledgements Our work was supported by the National Science Foundation (IBN-0112283). The protocol for the study was considered and approved by the Animal Care and Use Committee at Colorado State University (No. 01-052-01). Work at the Valentine National Wildlife Refuge was performed under authority granted by Special Use Permit 01-006VLT from the US Fish and Wildlife Service and by Collecting Permit 2001-174 from the Nebraska Game and Parks Commission. We thank L.L. McDaniel and M. Lindvall for their assistance in seeing the field work to its successful completion. References Alexander, N.J. (1970) Comparison of α and β keratin in reptiles. Zeitschrift für Zellforschung und Mikroskopische Anatomie 110, 153 165. Alibardi, L. (1999) Differentiation of the epidermis of neck, tail and limbs in the embryo of the turtle Emydura macquarii (Gray, 1830). Belgian Journal of Zoology 129, 391 404. Alibardi, L. (2002) Immunocytochemical observations on the cornification of soft and hard epidermis in the turtle Chrysemys picta. Zoology 105, 31 44. Alibardi, L. & Thompson, M.B. (1999) Epidermal differentiation during carapace and plastron formation in the embryonic turtle Emydura macquarii. Journal of Anatomy 194, 531 545. Attaway, M.B., Packard, G.C. & Packard, M.J. (1998) Hatchling painted turtles (Chrysemys picta) survive only brief freezing of their bodily fluids. Comparative Biochemistry and Physiology A 120, 405 408. Bodman, G.B. & Day, P.R. (1943) Freezing points of a group of California soils and their extracted clays. Soil Science 55, 225 246. Churchill, T.A. & Storey, K.B. (1992) Natural freezing survival by painted turtles Chrysemys picta marginata and C. picta bellii. American Journal of Physiology 262, R530 R537. Claussen, D.L. & Zani, P.A. (1991) Allometry of cooling, supercooling, and freezing in the freeze-tolerant turtle Chrysemys picta. American Journal of Physiology 261, R626 R632. Costanzo, J.P., Bayuk, J.M. & Lee, R.E. Jr (1999) Inoculative freezing by environmental ice nuclei in the freeze-tolerant wood frog, Rana sylvatica. Journal of Experimental Zoology 284, 7 14. Costanzo, J.P., Iverson, J.B., Wright, M.F. & Lee, R.E. Jr (1995) Cold hardiness and overwintering strategies of hatchlings in an assemblage of northern turtles. Ecology 76, 1772 1785. Costanzo, J.P., Litzgus, J.D., Iverson, J.B. & Lee, R.E. Jr (1998) Soil hydric characteristics and environmental ice nuclei influence supercooling capacity of hatchling painted turtles Chrysemys picta. Journal of Experimental Biology 201, 3105 3112. Costanzo, J.P., Litzgus, J.D., Iverson, J.B. & Lee, R.E. Jr (2000) Seasonal changes in physiology and development of cold hardiness in the hatchling painted turtle Chrysemys picta. Journal of Experimental Biology 203, 3459 3470. Costanzo, J.P., Litzgus, J.D., Larson, J.L., Iverson, J.B. & Lee, R.E. Jr (2001) Characteristics of nest soil, but not geographic origin, influence cold hardiness of hatchling painted turtles. Journal of Thermal Biology 26, 65 73. DePari, J.A. (1996) Overwintering in the nest chamber by hatchling painted turtles, Chrysemys picta, in northern New Jersey. Chelonian Conservation and Biology 2, 5 12.

100 G. C. Packard & M. J. Packard Ernst, C.H., Barbour, R.W. & Lovich, J.E. (1994) Turtles of the United States and Canada. Smithsonian Institution Press, Washington D.C. Hartley, L.M., Packard, M.J. & Packard, G.C. (2000) Accumulation of lactate by supercooled hatchlings of the painted turtle (Chrysemys picta): implications for overwinter survival. Journal of Comparative Physiology B 170, 45 50. Layne, J.R. Jr (1991) External ice triggers freezing in freezetolerant frogs at temperatures above their supercooling point. Journal of Herpetology 25, 129 130. Layne, J.R. Jr, Lee, R.E. Jr & Huang, J.L. (1990) Inoculation triggers freezing at high subzero temperatures in a freezetolerant frog (Rana sylvatica) and insect (Eurosta solidaginis). Canadian Journal of Zoology 68, 506 510. Lee, R.E. Jr & Costanzo, J.P. (1998) Biological ice nucleation and ice distribution in cold-hardy ectothermic animals. Annual Review of Physiology 60, 55 72. Magnusson, W.E. (2001) On the presentation of statistical tests of place: the importance of editorial consistency. Physiological and Biochemical Zoology 74, 616 618. Measures, M., Weinberger, P. & Baer, H. (1973) Variability of plant growth within controlled-environment chambers as related to temperature and light distribution. Canadian Journal of Plant Science 53, 215 220. Nagle, R.D., Kinney, O.M., Congdon, J.D. & Beck, C.W. (2000) Winter survivorship of hatchling painted turtles (Chrysemys picta) in Michigan. Canadian Journal of Zoology 78, 226 233. Packard, G.C. (1997) Temperatures during winter in nests with hatchling painted turtles (Chrysemys picta). Herpetologica 53, 89 95. Packard, G.C., Fasano, S.L., Attaway, M.B., Lohmiller, L.D. & Lynch, T.L. (1997a) Thermal environment for overwintering hatchlings of the painted turtle (Chrysemys picta). Canadian Journal of Zoology 75, 401 406. Packard, G.C., Lang, J.W., Lohmiller, L.D. & Packard, M.J. (1997b) Cold tolerance in hatchling painted turtles (Chrysemys picta): supercooling or tolerance for freezing? Physiological Zoology 70, 670 678. Packard, G.C., Lang, J.W., Lohmiller, L.D. & Packard, M.J. (1999a) Resistance to freezing in hatchling painted turtles (Chrysemys picta). Canadian Journal of Zoology 77, 795 801. Packard, G.C. & Packard, M.J. (1993a) Sources of variation in laboratory measurements of water relations of reptilian eggs and embryos. Physiological Zoology 66, 115 127. Packard, G.C. & Packard, M.J. (1993b) Delayed inoculative freezing is fatal to hatchling painted turtles (Chrysemys picta). Cryo-Letters 14, 273 284. Packard, G.C. & Packard, M.J. (1995) The basis for cold tolerance in hatchling painted turtles (Chrysemys picta). Physiological Zoology 68, 129 148. Packard, G.C. & Packard, M.J. (1997) Type of soil affects survival by overwintering hatchlings of the painted turtle. Journal of Thermal Biology 22, 53 58. Packard, G.C. & Packard, M.J. (2001) The overwintering strategy of hatchling painted turtles, or how to survive in the cold without freezing. BioScience 51, 199 207. Packard, G.C., Packard, M.J., Lang, J.W. & Tucker, J.K. (1999b) Tolerance for freezing in hatchling turtles. Journal of Herpetology 33, 536 543. Packard, G.C., Packard, M.J. & McDaniel, L.L. (2001) Seasonal change in the capacity for supercooling by neonatal painted turtles. Journal of Experimental Biology 204, 1667 1672. Salt, R.W. (1963) Delayed inoculative freezing of insects. Canadian Entomologist 95, 1190 1202. Storey, K.B., McDonald, D.G., Duman, J.G. & Storey, J.M. (1991) Blood chemistry and ice nucleating activity in hatchling painted turtles. Cryo-Letters 12, 351 358. Ultsch, G.R. (1989) Ecology and physiology of hibernation and overwintering among freshwater fishes, turtles, and snakes. Biological Reviews 64, 435 516. Weisrock, D.W. & Janzen, F.J. (1999) Thermal and fitnessrelated consequences of nest location in painted turtles (Chrysemys picta). Functional Ecology 13, 94 101. Willard, R., Packard, G.C., Packard, M.J. & Tucker, J.K. (2000) The role of the integument as a barrier to penetration of ice into overwintering hatchlings of the painted turtle (Chrysemys picta). Journal of Morphology 246, 150 159. Received 19 April 2002; revised 4 September 2002; accepted 1 October 2002