JOURNAL OF EXPERIMENTAL ZOOLOGY 309A:297–379 (2008)

A Journal of Integrative Biology

Physiological Ecology of Overwintering in Hatchling 1Ã 1 2 JON P. COSTANZO , RICHARD E. LEE JR , AND GORDON R. ULTSCH 1Department of Zoology, Miami University, Oxford, Ohio 2Department of Zoology, University of Florida, Gainesville, Florida

ABSTRACT Temperate of turtles hatch from eggs in late summer. The hatchlings of some species leave their natal nest to hibernate elsewhere on land or under water, whereas others usually remain inside the nest until spring; thus, post-hatching behavior strongly influences the hibernation ecology and physiology of this age class. Little is known about the habitats of and environmental conditions affecting aquatic hibernators, although laboratory studies suggest that chronically hypoxic sites are inhospitable to hatchlings. Field biologists have long been intrigued by the environmental conditions survived by hatchlings using terrestrial hibernacula, especially nests that ultimately serve as winter refugia. Hatchlings are unable to feed, although as metabolism is greatly reduced in hibernation, they are not at risk of starvation. Dehydration and injury from cold are more formidable challenges. Differential tolerances to these stressors may explain variation in hatchling overwintering habits among taxa. Much study has been devoted to the cold-hardiness adaptations exhibited by terrestrial hibernators. All tolerate a degree of chilling, but survival of frost exposure depends on either freeze avoidance through supercooling or freeze tolerance. Freeze avoidance is promoted by behavioral, anatomical, and physiological features that minimize risk of inoculation by ice and ice-nucleating agents. Freeze tolerance is promoted by a complex suite of molecular, biochemical, and physiological responses enabling certain organisms to survive the freezing and thawing of extracellular fluids. Some species apparently can switch between freeze avoidance or freeze tolerance, the mode utilized in a particular instance of chilling depending on prevailing physiological and environmental conditions. J. Exp. Zool. 309A:297–379, 2008. r 2008 Wiley-Liss, Inc.

How to cite this article: Costanzo JP, Lee RE Jr, Ultsch GR. 2008. Physiological ecology of overwintering in hatchling turtles. J. Exp. Zool. 309A:297–379.

Turtles are evolutionarily successful, long-lived environments as well as ecophysiological factors whose tolerance of heat, cold, dehydration, influencing species’ historical and current distri- and hypoxia permits them to thrive in diverse bution patterns. Additionally, gaining a more environments. Because mortality rates are highest precise knowledge of the winter biology and among hatchlings (e.g., Wilbur, ’75b; Tinkle et al., habitat needs of turtles could help resource ’81; Christens and Bider, ’87; Brooks et al., ’91; managers develop conservation strategies and Iverson, ’91a), chelonian researchers have devoted predict effects of anticipated climate change considerable attention to this age class. In cold- (Willette et al., 2005; Steen et al., 2007). temperate regions, hatchlings are especially Our principal aim in this article is to summarize vulnerable to winter mortality, a fact that may the literature concerning the ecology and physiol- significantly constrain recruitment and limit ogy of overwintering in hatchling turtles. The population size (e.g., Tinkle et al., ’81; MacCulloch and Secoy, ’83; St. Clair and Gregory, ’90; Rozycki, ’98; Schneeweiss et al., ’98). Inability of eggs and Grant sponsor: National Science Foundation; Grant numbers: IBN9817087; IAB 0416750; and IBN 00765592. hatchlings to cope with environmental extremes ÃCorrespondence to: Jon P. Costanzo, Department of Zoology, may also limit the northward extent of species Miami University, Oxford, OH 45056. Email: [email protected] ranges (e.g., Obbard and Brooks, ’81a; St. Clair Gordon R. Ultsch’s current address is 4324 NW 36th St., Cape Coral, FL 33993. and Gregory, ’90). Understanding the winter Received 6 November 2007; Revised 13 March 2008; Accepted 17 biology of temperate turtles may help elucidate March 2008 Published online 16 May 2008 in Wiley InterScience (www. demographic ramifications of living in extreme interscience.wiley.com). DOI: 10.1002/jez.460 r 2008 WILEY-LISS, INC. 298 J.P. COSTANZO ET AL. seemingly narrow focus of this project may (Ewert, ’85; Packard and Packard, ’88). Neonates surprise the casual reader, but neonates differ apparently remain in the nest for some time after biologically from other age classes in so many hatching, but ultimately they leave, either by respects as to warrant special treatment. Although dislodging the overlying soil, or ‘‘nest plug,’’ and the literature pertaining to overwintering in adult ascending through the opening or by descending turtles has been comprehensively reviewed through the floor into the soil column below. Nest (Ultsch, ’89, 2006), a comparable work focused evacuation can occur before or after winter; thus, on hatchlings is lacking. post-hatching behavior dictates many aspects of A secondary objective is to call attention to the hibernation ecology and physiology of hatchl- weaknesses and gaps in our current knowledge ing turtles. and suggest directions for future study. Much of Physical and biological factors stimulating tur- the information about the winter ecology of tles to leave their nests are of considerable interest hatchling turtles is derived from anecdotal reports to chelonian ecologists, although there is little and casual field observations, as few field studies consensus about which of these are of greatest have been published. On the other hand, our importance. Nest emergence sometimes coincides knowledge of their winter physiology is based with rain events (Hammer, ’69; Moll and Legler, almost exclusively on the results of laboratory ’71; Nagle et al., 2004), suggesting that egress is studies. Although many turtle species have tem- triggered by a rapid increase in soil moisture. perature-dependent sex determination (Ewert Precipitation could also stimulate emergence by et al., ’94), gender-related distinctions in physiology softening the soil (Goode and Russell, ’68; DePari, and cold-hardiness responses are virtually unex- ’96) or by flushing carbon dioxide from the nest plored. A preponderance of the literature concerns and delivering the oxygen needed for locomotor temperate North American turtles, certain species activity (Prange and Ackerman, ’74). of which are disproportionately represented. This Temperature, a particularly labile environmen- article necessarily reflects these biases. tal cue, is another potential inducer. With the Finally, we offer the following comments about advent of spring weather, reversal of the tempera- the taxonomic nomenclature used in this article. ture gradient in the soil column could stimulate Recent advances in molecular-based genetics and emergence in hatchlings (Bleakney, ’63; Tucker, morphometric analyses have prompted taxono- ’99) as it does in adult turtles (Grobman, ’90; mists to reclassify and rename various turtles Crawford, ’91b) and also in squamates (Sexton (e.g., over 50 new terminal taxa have been and Marion, ’81; Etheridge et al., ’83; Costanzo, proposed in the last 15 yr); consequently, turtle ’86). Thermal cues, such as rising spring tempera- nomenclature is perpetually dynamic and some- tures, could help synchronize emergence activities times contentious (Iverson et al., 2007; Turtle of hatchling turtles (Gibbons and Nelson, ’78; Working Group, 2007). Given this state Blouin-Demers et al., 2000). In a laboratory study of flux, and being mindful that the primary utility of hatchling red-eared sliders ( scripta), of taxonomic nomenclature is name recognition longer periods of cold exposure hastened emer- (Smith and Chiszar, 2006), we used groupings and gence from artificial nests on rewarming (Thomas, names that have predominated in the literature ’99). Once a particular temperature threshold is for at least the past 5 yr. reached, further chilling or warming may trigger nest emergence (Moran et al., ’99; Nagle et al., 2004). OVERWINTERING ECOLOGY Environmental cues are mediated by physiolo- gical mechanisms that proximally determine Patterns of nest emergence behavioral responses. Unfortunately, studies ad- A female turtle lays her eggs on land, depositing dressing physiological regulation of nest emer- them in an excavation and covering them with gence behavior (or any other post-hatching behav- substrate, and then moves on. In temperate ior) in hatchling turtles are lacking. Comparing regions, turtles nest one or more times from May the endocrine status of emerging and still- to July, choosing warm, unshaded sites conducive dormant hatchlings, and investigating populations to embryonic development (Christens and Bider, that exhibit both fall emergence and spring ’87; Schwarzkopf and Brooks, ’87; Rozycki, ’98). emergence, might prove rewarding. In his study Eggs usually hatch 2–3 months later, in late of Chrysemys picta, DePari (’96) found that summer or early autumn, although unseasonably hatchlings emerging in fall had shorter plastrons cool weather may delay or even prevent hatching but weighed the same and had similar amounts of

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 299 residual yolk as turtles that overwintered inside spring, even in midwinter (Aresco, 2004). By their nests. Working with the same species, contrast, in cool-temperate regions the emergence Costanzo et al. (2004) found no difference in the pattern is a punctuated continuum, probably blood and somatic characteristics between fall- because nest egress is hampered by seasonal cold. emerging and spring-emerging individuals, save In this case, ‘‘spring emergence’’ is a more for a higher lipid content in the former. However, appropriate idiom than ‘‘delayed emergence.’’ neither study tested hatchlings for hormonal Although it is tempting to categorize turtle changes that might trigger emergence. species with respect to emergence timing, it is Endogenous factors could be important in nest important to note that whereas nest emergence emergence behavior. Lindeman (’91) found behavior apparently varies little in some species, remarkable uniformity in the time interval in others it is quite plastic and either or both (300 d) between oviposition and hatchling emer- modes may be used, depending on the population gence and posited whether an internal ‘‘clock’’ and, perhaps, particular environmental and onto- cues emergence activity. Internal timing mechan- genetic circumstances. Overwintering in the nest isms have been implicated in stimulating emer- purportedly is an adaptation for survival in the gence in turtles that vacate nests soon after northern portions of a species’ range (Carr, ’52; hatching (Moran et al., ’99) and could also trigger Congdon and Gibbons, ’85), although it certainly emergence of hatchlings following hibernation, as occurs in more southern temperate and subtropi- they apparently do in some squamates (Drda, ’68; cal areas (Cagle, ’50; Jackson, ’94; Buhlmann, ’98; Garrick, ’72; Weatherhead, ’89; Lutterschmidt Buhlmann and Coffman, 2001; Morjan and Stuart, et al., 2006). Future research may ultimately show 2001; Aresco, 2004; Swarth, 2004). Ecological and that the driving force for hibernation emergence is evolutionary implications of remaining in the an interaction between extrinsic and intrinsic natal nest during winter is a topic of broad influences (Blouin-Demers et al., 2000). interest (Wilbur, ’75a,b; Gibbons and Nelson, Timing of nest emergence varies among taxa, ’78) and will be addressed below. populations, and even among siblings sharing the Timing and patterns of nest emergence, both same nest. Emergence patterns in temperate among nests at a given locale and among siblings species are typified by fall emergence, wherein sharing the same nest, may have marked survival hatchlings exit the nest before winter and hiber- consequences that are incompletely understood. nate elsewhere, and delayed emergence, wherein Synchronization of life-stage transitions and other hatchlings remain inside the natal nest until the phenological phenomena often represent a pre- following spring. Although this dichotomy is dator-swamping strategy that increases offspring ingrained in the literature, problems arise if it is survival (Sweeney and Vannote, ’82; Rutberg, ’87; taken too literally. First, the idiom ‘‘delayed Ims, ’90; but see Tucker et al., 2008). For a clutch emergence’’ invites ambiguity because even fall- of hatchling turtles, emerging in unison not only emerging turtles do not exit the nest immediately obviates mass predation but also facilitates egress after hatching, but rather remain in situ for at and reduces each individual’s energetic cost least several days, during which time they unfold (e.g., Bra¨nna¨s, ’95; Spencer et al., 2001). Emer- their shells and absorb their external yolk sac gence synchrony may improve survival because (Burger, ’76a; Swingland and Coe, ’78; Christens, hatchlings remaining even briefly within a brea- ’90; Dı´az-Paniagua et al., ’97). Emergence latency ched nest are imperiled if the open chamber can occur even among turtles that emerge in fall attracts parasites and predators (Burger, ’77; and hibernate outside the nest (Moll and Legler, Congdon et al., ’83b, ’87, 2000; Christiansen and ’71; Dı´az-Paniagua et al., ’97; Mitrus and Zema- Gallaway, ’84; Christens and Bider, ’87; McGowan nek, ’98). In a Michigan study, for example, et al., 2001) or heightens the risk of dehydration snapping turtles ( serpentina) hatched and/or cold stress (Nagle et al., 2004; Baker et al., in late August but left the nest about 2 months 2006). Late-emerging hatchlings may lag in later (Sexton, ’57). In addition, the distinction somatic growth and poorly compete for limited between fall emergence and delayed (spring) resources (e.g., Mason and Chapman, ’65; Einum emergence can be ambiguous, especially when and Fleming, 2000) with potential fitness conse- considering species (and populations) of southerly quences manifested later in life (Yearsley et al., climes. In northwestern Florida, for example, 2004). On the other hand, asynchrony in life-stage moderate temperatures permit hatchlings to transitions also can be advantageous (e.g., Clark emerge from summer through the following and Wilson, ’81; The´ron and Combes, ’95) and can

J. Exp. Zool. 300 J.P. COSTANZO ET AL. represent a risk-spreading strategy to increase also emerge late from their nests in autumn fitness in an unpredictable and harsh environ- (Parren and Rice, 2004) or even the following ment (Danforth, ’99; Thumm and Mahoney, spring (Ernst et al., ’94). However, there is little 2002). Asynchronous emergence of turtle siblings empirical support for this relationship (e.g., Dı´az- could, in principle, minimize costs of remaining Paniagua et al., ’97). inside the natal nest longer than necessary (Hays Assuming that emergence requires strenuous et al., ’92; Houghton and Hays, 2001). In addition, physical activity, inter-clutch variation in emer- selection may favor asynchronous emergence if gence timing could reflect differences in hatchling predators are attracted to high densities and large body size and physiological condition, or even size groups of hatchlings (Glen et al., 2005). of the sibling group, if social facilitation is Emergence of turtle hatchlings from nests at a important to egress (Carr and Hirth, ’61; Moran given locale may occur virtually simultaneously or et al., ’99). Stress imposed by dehydration or (for over a surprisingly lengthy term. For hatchlings terrestrial hibernators) extreme cold could impair that overwintered inside the natal nest, the neurobehavioral function, delaying nest emer- emergence period reportedly lasts from a few days gence (see Hartley et al., 2000). Similarly, follow- to several weeks (Hartweg, ’44; Lindeman, ’91; ing especially severe winters, adult painted turtles DePari, ’96; Dı´az-Paniagua et al., ’97; Tucker, ’99; (C. picta) resume behavioral activity relatively late Bjurlin and Bissonette, 2004; Nagle et al., 2004), in spring, presumably because they require an but could be even longer. For example, at one extended recovery period (Crawford, ’91b). This study site in northern Indiana, hatchling C. picta idea has not been experimentally tested; however, that overwintered in their nests emerged from Lindeman (’91) observed that surviving hatchlings early March through late April (Fig. 1). in clutches suffering high winter mortality Proximate causes of variation in emergence (a reasonable proxy for stress) tended to emerge timing are unknown, but probably relate to spatial later than hatchlings in clutches with higher survival. variation in environmental factors (e.g., tempera- Clutchmates do not necessarily emerge from ture, moisture) that serve as cues and/or facilitate their nest en masse, but may exit in small groups escape from the confines of the hibernaculum. In or individually over protracted periods. Relatively turn, these factors are influenced by the physical little study has been devoted to this phenomenon, attributes of individual nests, such as depth within in part owing to the technical challenge of con- the soil column, soil characteristics, drainage, currently monitoring multiple nests (see Doody slope, and aspect. Emergence behavior could also and Georges, 2000). Consequently, some authors be influenced by an endogenous timing mechan- have presumed that clutchmates emerge more or ism (Lindeman, ’91), but this idea has not been less in unison, once the first hatchling has rigorously tested. Some investigators have ques- appeared (Christens and Bider, ’87; Tucker, ’97, tioned whether clutches hatching relatively late in ’99). However, several reports attest that indivi- summer or in early fall, perhaps as a consequence duals emerge sporadically over many days or even of retarded development or delayed oviposition, weeks (Burger, ’77; Hays et al., ’92; Butler and Graham, ’95; Dı´az-Paniagua et al., ’97; Houghton 8 and Hays, 2001; Bjurlin and Bissonette, 2004). In 3 7 the extreme, some individuals emerge in fall, the 5 7 remainder of the clutch overwintering in the open 7 5 nest and, if able to survive, emerging in spring 8 4 (Costanzo et al., 2004; Nagle et al., 2004; Swarth, 2004; Baker et al., 2006; Carroll and Ultsch, 2007). 11110 2010 20 March April May This pattern, reported for several taxa, is probably uncommon, but underscores the plasticity in Fig. 1. Timing of emergence of hatchling painted turtles (Chrysemys picta marginata) from natal nest hibernacula at emergence behavior among hatchling turtles in Mount Zion Mill Pond, Fulton Co., IN, in spring 2000 (for a at least temperate regions. description of the study area, see Costanzo et al., 2004). Each Factors influencing timing and patterns of horizontal line represents the duration of emergence of emergence of siblings are as yet undetermined. clutchmates from an individual nest (range for nine nests, Variation in body size and/or physiological condi- 1–31 d). Nests were protected from predators by a wire cage and were monitored until after all live hatchlings in each nest tion among clutchmates could contribute to (number as shown) had emerged. (J. Costanzo, P. Baker and asynchrony (Glen et al., 2005), although findings J. Iverson, unpublished data.) to the contrary have been published (Hays et al.,

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 301

’92; Dı´az-Paniagua et al., ’97). Given the geo- and diverse, and sometimes distinct from those metric arrangement of hatchlings and the thermal used by other age classes. heterogeneity within the three-dimensional nest, Hatchlings of terrestrial species tend to hiber- it is possible that clutchmates receive emergence nate on land, either inside the nest or outside, cues at different times; indeed, emergence asyn- some burrowing into the soil column to hibernate chrony tends to increase with thermal variation at great depths (Doroff and Keith, ’90; Costanzo inside nests (Houghton and Hays, 2001). Environ- et al., ’95b; Converse et al., 2002). Hatchlings of mental heterogeneity within the nest could also some aquatic species also overwinter in the natal cause siblings to become differentially impacted by nest, although many apparently hibernate under stressors (e.g., cold, freezing, dehydration, and water. Published reports characterizing the aqua- hypoxia) that potentially impair neurobehavioral tic refugia of hatchlings are lacking (but see function. Because motor support systems are Ultsch et al., 2008). Hatchlings and young turtles sensitive to oxygen availability, solute and water seemingly prefer shallow-water habitats for feed- balance, ion gradients, and acid–base status (Mill- ing and predator avoidance (Hart, ’83; Congdon er et al., ’87; Lutz and Storey, ’97; Finkler, ’99; et al., ’92; Pappas and Brecke, ’92), but whether or Jackson, 2002), such stress could delay or even not they also hibernate in these locations remains prevent emergence of afflicted individuals. This to be determined. Conventional wisdom maintains notion has not been formally tested, but it is that hatchlings probably hibernate in proximity to noteworthy that Glen et al. (2005) found that older conspecifics, which are known to use nests with higher numbers of dead hatchlings also burrows or other cavities, take refuge under cover had longer emergence durations and higher levels objects, or settle on or beneath the substratum of asynchrony. surface. However, recent findings that hatchlings are relatively intolerant of hypoxic submergence (Reese et al., 2004b; Dinkelacker et al., 2005b) raise questions about their ability to overwinter in Overwintering habits and habitats certain aquatic microenvironments. This liability In recent years, interest in chelonian conserva- may limit the availability of suitable hibernacula tion and mechanisms of cold hardiness has or could suggest that mortality rates are especially generated a wealth of new information about the high in hatchlings. On the other hand, hatchlings winter ecology of hatchling turtles. Still, there might preferentially use lotic systems or the more remain substantial voids in our knowledge of the oxygenated regions of lentic habitats, such as hibernation habits of even common species, and areas of upwelling or inflowing water, air pockets especially of those that hibernate aquatically beneath the ice cap, or the land/water interface. (Ultsch, 2006). This paucity probably stems from Careful field research is needed to resolve these the difficulty in monitoring large samples of nests, questions. and in locating hatchlings after they disperse from Aquatic species that leave their nest before the nesting site. Dispersal behavior has been winter do not necessarily hibernate in water. In studied by tracking movements of hatchling a radiotelemetry study on Long Island, NY, turtles treated with fluorescent dye (e.g., Butler hatchling C. serpentina wandered through various and Graham, ’93; Tuttle and Carroll, 2005), but habitats and encountered a pond, but instead this method has significant limitations. Improve- chose to overwinter in nearby spring seeps (Ultsch ments in radiotelemetry and other tracking et al., 2008). Blanding’s turtles (Emydoidea techniques would enable investigators to monitor blandingii) typically emerge from their nests in hatchling behavior for sufficiently long periods, late summer and wander considerably, apparently providing a more thorough understanding of their avoiding water (Standing et al., ’97; McNeil et al., winter habits and habitat requirements. The 2000), before overwintering. Their hibernation recent study of hatchling C. serpentina by Ultsch habits are as yet unknown, although some authors et al. (2008) is a case in point. speculate that they overwinter in waterlogged Naturalists have long presumed that the hatchl- soils near the margins of ponds or wetlands ings of terrestrial turtles hibernated on land and (Dinkelacker et al., 2004). In an estuarian species, those of freshwater turtles passed the winter the (Malaclemys terrapin), under water. However, field investigations over some hatchlings vacate the nest, briefly inhabit the the last several decades have revealed that upper intertidal zone, and then return to land to hibernacula used by neonatal turtles are variable hibernate in shallow burrows (Draud et al., 2004).

J. Exp. Zool. 302 J.P. COSTANZO ET AL.

Hibernation in the natal nest facilitated by sibling cooperation, then emerging in autumn may be easier and more consistent in A persistent question pondered by turtle species (e.g., C. serpentina) that produce relatively researchers is why the hatchlings of some species large clutches (DePari, ’96). On the other hand, overwinter in the nest whereas others do not. fall emergence is the norm in some species that A related question is why any hatchling would invariably produce few young (Gibbons and overwinter in the nest at all. Gibbons and Nelson Nelson, ’78). Perhaps for these species social (’78) addressed the latter, reporting that, in South facilitation is not required because their nests Carolina, several aquatic species evidently spend have little overburden (e.g., spp.), their first winter on land, presumably inside the their nests are constructed in friable soils (e.g., nest. Their synthesis consolidated a host of early, Terrapene spp.), or they are especially adept at mostly anecdotal reports, emphasizing that hiber- burrowing into the soil column (e.g., spp.). nation of neonates in the natal nest occurs Contrary to the arguments listed above, some worldwide and in at least a dozen genera and five observations suggest that entrapment is not a families. In addition, this seminal report offered primary cause of overwintering inside the natal an evolutionary explanation for variation in nest. Late summer rains that soften the soil and emergence patterns both within and among even erode the nest plug do not necessarily species. Subsequent authors have posited myriad stimulate emergence (Hartweg, ’44; Sexton, ’57; explanations for why hatchlings overwinter in the Bleakney, ’63; Gibbons and Nelson, ’78; DePari, nest, some supposing that the behavior is a passive ’96). Furthermore, hatchlings of some species response, the ultimate consequence of adverse routinely overwinter inside nests constructed in biological or environmental conditions hampering friable soils from which egress should be relatively fall emergence, and others regarding it as a easy (Costanzo et al., 2004; Baker et al., 2006). facultative strategy that increases offspring fitness. Hatchlings may be forced to overwinter inside their nest at least occasionally, but, as a general Passive response hypotheses explanation, the concept of entrapment is One popular hypothesis maintains that hatchl- unsatisfying. ing turtles pass the winter inside the natal nest Heat is an important resource not only for because they are physically incapable of breaching embryonic development and hatching but also for the hardened nest plug until freezing/thawing and nest emergence activities. Accordingly, overwinter- vernal rains have softened the soil (Cagle, ’44; ing in the nest could result if fall emergence is Hartweg, ’44; Legler, ’54; Ernst, ’66; Tinkle et al., hampered by the onset of cool weather. Spring ’81; DePari, ’96). Support for this idea comes emergence seems to coincide with a seasonal from observations that drought precedes over- reversal of the vertical thermal gradient, suggesting wintering in the nest of species (populations) that turtles are thermotaxic (Bleakney, ’63; Tucker, that typically emerge in autumn, and that emer- ’99). Therefore, if, in autumn, the soil strata above gence events in spring often coincide with or the nest become cooler than those below, emergence closely follow rainfall. For example, emergence of cues would be lacking. In addition, the hindering hatchling C. serpentina is reportedly facilitated by effect of cold on locomotor function could directly autumnal rains (Hammer, ’69), whereas dry, hard- prevent turtles from leaving their nests. Thermal ened soil apparently impedes egress and causes insufficiency may account for the failure of some them to remain inside nests during winter (Ernst, clutches of obligate aquatic hibernators to emerge in ’66). In northern New Jersey, DePari (’96) found a autumn (Sexton, ’57; Breckenridge, ’60; Obbard and strong association between nest emergence timing Brooks, ’81a; Parren and Rice, 2004) or even to of C. picta and characteristics of the nesting soil. hatch (Buech et al., 2004). On the other hand, nest Hatchlings occupying natural and artificial nests temperatures at the time of hatchling emergence are constructed in friable soils (sands) were more commonly even lower in spring than in autumn likely to emerge in autumn than turtles in heavier (Gibbons and Nelson, ’78; Holte, ’88; DePari, ’96) or wetter soils, suggesting that the ability to therefore, the role of cold, per se, in inhibiting fall overcome physical barriers to egress can influence emergence is unclear. emergence timing. Yet another explanation is that hatchlings are Nest entrapment could explain interspecific developmentally immature and unprepared to variation in overwintering behaviors of hatchling leave the nest in autumn. Thermal and hydric turtles. For example, if breaching the nest plug is factors strongly influence hatchling phenotype and

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 303

fitness (Packard and Packard, ’88; Deeming, that increase the probability of winter survival. 2004a), and suboptimal incubation conditions could Overwintering within the nest carries its own produce immature hatchlings. Immaturity at risks, such as death from flooding, predation, hatching also could result from delayed oviposition dehydration, energy depletion, or extreme cold, or, at high latitudes, brevity of the summer. In and may eliminate the opportunity for early northern populations of C. picta, hatching may not feeding and accelerated growth. However, Gib- occur until autumn (Costanzo et al., 2004). Declin- bons and Nelson (’78) argued that, at least in some ing soil temperatures apparently retard embryonic circumstances, these costs could be offset by development and ultimately hamper fall emergence certain gains. For example, overwintering in the of hatchlings of several species (Bleakney, ’63; nest would be advantageous if hatchlings emer- Gibbons and Nelson, ’78; Mitrus and Zemanek, ’98; ging in spring entered an environment in which Waye and Gillies, ’99). In some populations of the thermal and food resources were increasing, ( orbicularis), aquatic rather than decreasing, as would be the case in hibernation is the norm, but late-hatching clutches autumn. According to this argument, natural may remain in the nest throughout winter, perhaps selection should favor deferred emergence from a because they are immature and are unable to ‘‘proven sanctuary’’ until environmental cues respond to environmental emergence cues (Dro- signal that hatchlings will enter a resource-rich benkov, 2000). On the other hand, in temperate environment favoring rapid somatic growth (Gib- regions, even early-hatching clutches sometimes bons and Nelson, ’78; Mitchell, ’88). overwinter inside the nest (Gibbons and Nelson, From this reasoning, one would expect over- ’78; Jackson, ’94; Morjan and Stuart, 2001). wintering inside the nest to be commonplace, if Clearly, developmental immaturity is not a uni- not obligatory (see Carr, ’52), for hatchlings of versal factor influencing emergence timing. northern species. However, there may be more Deviation from the routine hibernation habit species of northern turtles whose hatchlings undoubtedly occurs in hatchlings of virtually emerge in fall (e.g., C. serpentina, E. blandingii, every species. For obligate aquatic hibernators, Clemmys guttata, Clemmys [Glyptemys1] insculp- remaining inside the natal nest during winter ta, spinifera, some E. orbicularis) than could stem from failure of autumnal emergence there are of species whose hatchlings overwinter cues to materialize or to be detected (or acted on) in the nest (e.g., C. picta, geographica, by hatchlings. This scenario could explain the odd T. scripta, some E. orbicularis). Perhaps the thesis case reported by Obbard and Brooks (’81a), where of Gibbons and Nelson (’78) applies generally over between 1976 and 1979 in Algonquin Park, Ont., much of the temperate region, but does not apply 62 of 104 C. serpentina clutches (59.6%) producing in northern regions because there the perils of viable hatchlings overwintered inside the nest remaining inside the nest are too severe. The fact (with lethal consequences), rather than emerging that the facultative nest hibernator, C. picta, in autumn. Conversely, in species whose hatchl- commonly overwinters inside the nest, even in ings usually overwinter terrestrially, fall emer- the northernmost reaches of its distributional gence could be triggered by adverse conditions, range (e.g., Woolverton, ’63; St. Clair and Gregory, such as flooding or degradation of the nest plug or ’90; Rozycki, ’98), suggests that this behavior is chamber. For example, although ornate box advantageous if the associated challenges can be turtles () usually spend their overcome. first winter in the soil column beneath the nest Overwintering inside the nest could constitute a chamber, clutches hatching in particularly shallow means of reducing predation risk during a period nests sometimes emerge in fall and hibernate of limited opportunity for somatic growth. This elsewhere (Converse et al., 2002). In some popula- argument was advanced for C. picta (Wilbur, tions of C. picta, plasticity in nest emergence ’75a,b), but certainly would apply to other tempe- behavior apparently is tied to local weather and rate species. Hatchlings remaining in the nest soil conditions (DePari, ’96). during winter may be relatively safe, as depreda- tion of turtle nests generally diminishes with time after oviposition (Burger, ’77; Tinkle et al., ’81; Adaptive response hypotheses Christens and Bider, ’87; Rozycki, ’98). Remaining A popular alternative hypothesis for why turtles hibernate in the natal nest states that deferring 1Recommended by the Committee on Standard English and emergence until spring confers special benefits Scientific Names (Iverson et al., 2008).

J. Exp. Zool. 304 J.P. COSTANZO ET AL. in situ until spring eliminates the need to disperse 25 overland at the time predator populations have 20 Pond bottom reached a seasonal peak and some migratory predators have concentrated in staging areas. 15

Additionally, after making the trek to water, 10 hatchlings would become exposed to both resident and migratory predators, such as wading birds, 5 which are feeding heavily in preparation for Temperature (˚C) 0 Soil winter. Little is known about predation pressure Pond surface on hatchling turtles in aquatic habitats, in part -5 owing to the difficulty of tracking the fate of this SONDJFMAM cohort (e.g., Brooks et al., ’91). Hatchlings of many Fig. 2. Seasonal dynamics in environmental temperature aquatic turtles exhibit effective antipredator be- potentially encountered by hatchling painted turtles (Chrys- havior and aposematic coloration, and probably emys picta marginata) remaining in natal nests until spring or are not preyed on heavily by fish (Semlitsch and emerging from nests in autumn and overwintering under water. Temperature data were collected during 2004–2005 at a Gibbons, ’89; Britson and Gutzke, ’93). More turtle nesting site at Mount Zion Mill Pond, Fulton Co., IN research in this area would permit comparison of (for a description of the study area, see Costanzo et al., 2004). the cost/benefit relationships of overwintering in Dataloggers recorded temperature each hour in the pond the nest and in aquatic habitats. (water surface and surface of substratum beneath 1 m of Another potential benefit of passing the winter water) and in the soil column 7.5 cm below ground, the typical depth of C. p. marginata nests. Lines connect weekly mean inside the ‘‘proven sanctuary’’ is that protection is values, each based on 168 measurements. (J. Costanzo, P. afforded to turtles that have hatched, but are not Baker and J. Iverson, unpublished data.) morphologically or physiologically mature enough to survive outside the nest. These would Dinkelacker et al., 2005a,b) and none have have additional time to complete development in a compared the energetic costs of overwintering in relatively safe environment where demands for water with those of overwintering on land. performance (feeding and digestion, evading pre- dators, habitat selection, etc.) are relatively few. Ecological and evolutionary implications Neonatal C. picta dive and swim poorly, perhaps owing to high buoyancy imparted by their large As mentioned above, one particularly vexing internalized yolk sac (T. Muir, personal commu- question is why some species hibernate in the nest nication). Turtles consume much of their yolk in a given area, particularly in the northern parts during winter and presumably are more adroit of their ranges, whereas other species do not. If when they finally enter water. Accordingly, species hibernating in the nest confers a fitness advantage whose hatchlings provision their eggs with rela- relative to emerging in autumn for any of the tively large amounts of yolk also tend to defer reasons listed above, then, intuitively, all species emergence until spring (Congdon et al., ’83c; ought to exhibit this behavior. Several hypotheses Congdon and Gibbons, ’85, ’90; Rowe et al., ’95; can be advanced to explain why this is not the Nagle et al., ’98; Costanzo et al., 2000b); however, case. this association may be derived from additional (or First, perhaps only species with strictly north- other) factors. ern distributions have evolved a life history that From a physiological perspective, overwintering includes overwintering of hatchlings inside the within the nest may be an adaptive response that nest. For example, with the exception of the obviates problems associated with submergence in southern subspecies, C. picta is basically a north- cold, hypoxic water. It may be more favorable ern taxon whose hatchlings typically hibernate in energetically if turtles in nests substantially the nest, whereas musk turtles, with the exception reduce their metabolic demands. Intuitively, this of one species (), are basi- may be the case if nesting soil, by virtue of its cally a southern group whose hatchlings over- lower specific heat, cools more quickly than water, winter elsewhere. On the other hand, two species and if prevailing winter temperatures are lower of map turtles ( Graptemys), a predomi- inside nests than within aquatic refugia (Fig. 2). nantly southern group, range into cold climates However, few investigations have addressed the and their hatchlings do overwinter inside the nest. physiology of aquatic hibernation in hatchling A telling argument contradicting this hypothesis is turtles (Finkler et al., 2002; Reese et al., 2004b; that several species (e.g., C. guttata, C. insculpta,

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 305

E. blandingii) that are closely related, basically species whose hatchlings can cope with the threats northern forms, have hatchlings that apparently of freezing and/or desiccation. The alternative is to routinely hibernate outside the nest. hibernate in places where these threats can be A second hypothesis is that overwintering of avoided, but other risks abound. Mortality during hatchlings in the nest is restricted to species the first winter of life may well be higher among of certain phylogenetic groupings. Indeed, most of hatchlings that overwinter aquatically than the species exhibiting this behavior are in the among those that spend the first winter in the family . On the other hand, most of the nest. Taking this point further, if a species has a species inhabiting the United States and Canada high mortality not only during the egg stage, as all are in this family. Moreover, a number of northern species do, but also during the initial winter emydid species do not, and perhaps cannot, because it hibernates aquatically, then it must successfully overwinter inside the natal nest. It compensate for this additional demographic bot- is noteworthy that, among emydids, most of those tleneck. One approach is to live long and/or have a northern hard-shelled species with hatchlings that high annual reproductive output. C. serpentina are do remain inside the nest during winter are in one long-lived, for example, and thus can persist in far clade (Graptemys–Malaclemys–Trachemys–Chrys- northern populations (e.g., Algonquin Park, Ont.) emys), whereas those that hibernate elsewhere in spite of very low annual reproductive success are in another (Clemmys [including ]– (Galbraith and Brooks, ’87; Brooks et al., ’91). Emys–Emydoidea–Terrapene (Stephens and Weins, Because they lay large clutches, an occasional 2003). Therefore, if a turtle is especially cold successful year of reproduction is apparently hardy, it is highly likely to be an emydid, although enough to sustain the population. not all emydids use the same cold-hardiness If overwintering in the nest is more beneficial strategies (see below). than hibernating elsewhere, then, lacking the Another hypothesis is that only species in certain challenges of severe cold and desiccation, hatchl- morphological or physiological groupings have ings universally ought to do so. Accordingly, a hatchlings that overwinter in the nest. In northern species that does not overwinter in the nest in the environs, all such species must be adapted to cope northern portion of its range perhaps could do so with exposure to potentially injurious cold and in the more southern parts of its range, where dehydration. Species whose primary strategy of cryoinjury is unlikely, if the nest environment cold hardiness is freeze tolerance share physiologi- were sufficiently humid. There is some support for cal adaptations that limit osmotic and anoxic this notion, but only in that the hatchlings of damage to frozen cells and tissues, whereas species many species do commonly overwinter in the nest that avoid lethal freezing through supercooling at low latitudes (Carr, ’52; Congdon and Gibbons, (remaining liquid at temperatures below the ’85), although it certainly occurs in more tempe- equilibrium freezing/melting point) probably share rate areas (Cagle, ’50; Jackson, ’94; Buhlmann, behavioral and morphological traits that reduce the ’98; Buhlmann and Coffman, 2001; Morjan and risk of contact inoculation. The integument is a Stuart, 2001; Aresco, 2004; Swarth, 2004). There particularly important mediator of both inoculative is no compelling evidence for a major reversal of freezing and evaporative water loss (EWL), and hibernation habits with latitude for a given various species (e.g., C. serpentina, S. odoratus,and species. For example, hatchlings of A. spinifera, A. spinifera) that do not, and perhaps cannot, C. serpentina, S. odoratus, and T. carolina overwinter in the nest as hatchlings have a routinely emerge from the nest in autumn and relatively large amount of skin exposed to their overwinter outside the nest throughout their surroundings (Costanzo et al., 2001b). The extensive geographical ranges. Additionally, implication of this relationship is that excessive although information for southern populations is dehydration and a greater propensity for inocula- meager, several species, including C. picta (Cagle, tive freezing constrains nest overwintering beha- ’54) and T. scripta (Gibbons and Nelson, ’78; vior. On the other hand, species such as Jackson, ’94), which hibernate in the nest in the C. guttata, C. insculpta,andE. blandingii do not north apparently also do so in the south. Thus, hibernate in the natal nest even though they have a working hypothesis is that if a species can little exposed skin. overwinter in the nest, then it will, and will do According to the aforementioned hypothesis, so throughout its geographic range. Clearly, overwintering of hatchlings in the nest is the most however, such behavior is not a requirement for favorable survival strategy, but is only used by a species to be successful, or even to be competi-

J. Exp. Zool. 306 J.P. COSTANZO ET AL. tive. C. picta and C. serpentina best exemplify this (Waye and Gillies, ’99) or overwinters in situ (St. point: hatchlings of the former overwinter in the Clair and Gregory, ’90), attesting that the hiber- nest and those of the latter do not, yet both are nation habit of this species is plastic, even in a abundant and wide-ranging species that coexist in severe environment. In this regard, M. terrapin is very cold climates. particularly interesting, as up to 30% of the nests on a given beach harbor hatchlings during winter (Roosenberg, ’94; Baker et al., 2006); young Taxonomic and geographic patterns produced in the other nests seek winter quarters Overwintering habits and habitats of hatchling elsewhere (Auger and Giovannone, ’79). turtles are remarkably diverse, even among In the final analysis, there is no definitive syntopic species (Christiansen and Gallaway, ’84; answer as to why hatchlings, especially those of Costanzo et al., ’95b). For example, in the Sand- aquatic species, overwinter inside the natal nest, hills region of westcentral Nebraska, hatchling nor is it clear why some species exhibit this C. picta hibernate within their natal nests, behavior and others do not. The evidence does E10 cm below ground; T. ornata and Kinosternon show, however, that species that can successfully flavescens overwinter in sand dunes Z1 m beneath hibernate in the nest will do so throughout their the nest chamber; and A. spinifera and geographic range, and that the behavior occurs C. serpentina emerge from their nests after only incidentally in species that are not well suited hatching and move to ponds or streams for to terrestrial hibernation. The adaptive signifi- overwintering. As we will discuss below, the cance of overwintering inside the nest remains hibernation habits of a given species reflects its equivocal. particular life-history traits and physiological tolerances to environmental extremes. Species accounts The hatchlings of some temperate species Gibbons and Nelson (’78) earlier compiled the exhibit considerable variability in overwintering available information concerning the hibernation behavior on both regional and local scales. For habits of the hatchlings of various turtle species. example, hatchling E. orbicularis overwinter Much new information has subsequently become under water in central Poland (Mitrus and available, yet the contemporary literature lacks an Zemanek, ’98), inside the natal nest in north- updated account. We here summarize, for mostly eastern Germany (Andreas and Paul, ’98; Schnee- North American species, the current state of weiss et al., ’98), or in self-constructed burrows in knowledge pertaining to the overwintering habits southern Russia (Mazanaeva and Orlova, 2004). In and habitats of hatchling turtles, and briefly Belarussian Polesye, near the northeastern edge comment on the physiological and ecological of this species’ range, and in northeastern implications of habitat selection. Ukraine, they can hibernate inside or outside the nest, depending on whether the eggs began Terrestrial species incubation early or late, respectively (Drobenkov, 2000; Zinenko, 2004). The of North spp. (; Testudinidae): Of the America is also notable in this regard. Although three species of land tortoises in the United States overwintering inside the nest appears to predomi- (G. agassizii, G. polyphemus, and G. berlandieri), nate, autumnal emergence (and presumed aquatic all are southern forms, as the northern limit of hibernation) occurs with variable frequency with- their range is southernmost Utah in the west and in the same populations of all northern subspecies: southernmost South Carolina in the east. Hatchl- C. p. marginata in Indiana (Costanzo et al., 2004), ings of the first two species emerge from nests in C. p. picta in New Jersey (DePari, ’96), Connecti- autumn; unhatched eggs do not survive the winter cut (Finneran, ’48), New Hampshire (Carroll and (G. agassizii—Woodbury and Hardy, ’48; Turner Ultsch, 2007), Pennsylvania (Ernst, ’71) and et al., ’86; Rostal et al., ’94; Averill-Murray et al., Maine (Rozycki, ’98), and C. p. bellii in New 2002; Bjurlin and Bissonette, 2004; G. polyphe- Mexico (C. Morjan, personal communication), mus—Iverson, ’80; Landers et al., ’80; Wright, ’82; Iowa (Christiansen and Gallaway, ’84), Nebraska Martin, ’89; Butler et al., ’95; Butler and Hull, ’96; (Costanzo et al., 2004), and Minnesota (Pappas Epperson and Heise, 2003). Gopherus berlandieri, et al., 2000). In British Columbia, near the limited to southern Texas in the United northwestern edge of its geographic range, States, probably also emerges in autumn, as the C. p. bellii either emerges from the nest in autumn laboratory incubation period was 88–118 d for 13

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 307 eggs, with hatching occurring August–November 2004); however, because these individuals were (Judd and McQueen, ’80). Whether overwintering not observed leaving a nest, it is uncertain where in the nest occurs in any is an open they spent the winter. They could have been waifs question, as it has been reported only once, for stranded on land that managed to survive the horsfieldii in Kazakhstan, Turkmenia, winter by burrowing, or simply relatively small and Uzbekistan (Kuzmin, 2002). turtles that were found on land after hibernating Terrapene spp. (box turtles; Emydidae): These aquatically. are terrestrial emydids with two species in the At high latitudes, hatchling C. serpentina will United States. Terrapene carolina (eastern box perish if they fail to emerge from their nests before turtle) ranges from Florida to Texas north into winter, and pipping may not occur if cold weather New England and Michigan, and T. ornata (ornate arrives early enough to kill developing hatchlings. ) ranges across the Great Plains north to At Algonquin Park, Ont., very close to the northern South Dakota. Eastern box turtle hatchlings limit of their range, Obbard and Brooks (’81a) emerge from nests in autumn throughout their found that of 257 clutches monitored, only 47 range (Cooke, ’10; Cahn, ’33; Ewing, ’33; Allard, successfully emerged in autumn, and only one of ’35, ’48; Conant, ’38; Smith, ’61; Minton, ’72; the remaining nests had any survivors the follow- Iverson, ’77; Dundee and Rossman, ’89; Forsythe ing spring. This nest, which on excavation (May 5) et al., 2004), although little is known about their was found to contain 16 live hatchlings, 11 dead hibernacula. Presumably they burrow below the hatchlings, and three infertile eggs, may have been frost line, as do adults (but see Costanzo and buffered from cold by a snow bank formed by road Claussen, ’90; Claussen et al., ’91). Some may also plowing. Similarly, in North Dakota, Hammer (’69, overwinter in or below the nest cavity (Ernst et al., ’72) found that hatchlings died in the nest if they ’94; Trauth et al., 2004), but such behavior failed to emerge before winter. Moreover, despite appears rare. Hatchlings of T. ornata, however, the usually high hatching success (93.5%), emer- routinely overwinter below the nest cavity at gence rates were typically low (19.8%) and in some depths approaching 1 m (Doroff and Keith, ’90; years no turtles left the nests. Why hatchlings in Costanzo et al., ’95b; Converse et al., 2002). The these northern populations did not leave the nest depth to which hatchings burrow suggests that once they have left the egg remains an open they are attempting to stay below the frost line, question. In Michigan, some were found entangled although they do tolerate somatic freezing in plant roots that prevented them from reaching (Costanzo et al., ’95b, 2006; Dinkelacker et al., the surface (J. Dusseau, personal communication). 2005b). If hatching occurs late in autumn, low body temperatures may hamper their ability to dig out, or the overlying soil may freeze and become Aquatic species impenetrable. C. serpentina (snapping turtle; ): In northern populations, hatchlings remaining This species is wide ranging and common in the in nests during winter are usually doomed, as they central and eastern United States and, with lack a well-developed capacity for freeze tolerance C. picta, has the most northerly range limit of and cannot supercool extensively in the nest North American turtles. Throughout its range, all environment (Packard and Packard, ’90; Packard reports of hatchlings observed leaving the nest or et al., ’93; Costanzo et al., 2001b, 2006; Dinkelack- being collected in numbers are for autumn (Noble er et al., 2005b). However, in climates where and Breslau, ’38; Hamilton, ’40; Pell, ’41; Cagle, freezing is not an issue, they perhaps could survive ’44; Norris-Elye, ’49; Petokas and Alexander, ’60; winter on land. Of particular note are the findings Hammer, ’69; Punzo, ’75; Obbard and Brooks, of J. Gibbons and J. Greene (unpublished data), ’81a,b; Congdon et al., ’87; Mitchell, ’94; Pappas who during 1978–1998 monitored a drift fence et al., 2000; Hulse et al., 2001; Kolbe and Janzen, that encircled a bay in South Carolina. Of 45 2002; Carroll and Ultsch, 2007). It is widely hatchlings collected, four were captured moving accepted that these hatchlings overwinter aqua- toward the bay in February–April, presumably tically, although nothing in the literature confirms after overwintering on land, although not neces- this supposition. Some workers have found occa- sarily inside natal nests. In this population, sional hatchlings in spring that were presumed to autumnal nest emergence and aquatic hibernation have hibernated on land (Toner, ’40; Ernst, ’66; is the rule, but successful overwintering on land Minton, ’72; Congdon et al., ’87; Parren and Rice, may be possible.

J. Exp. Zool. 308 J.P. COSTANZO ET AL.

Hatchling C. serpentina on Long Island, New and Nelson, ’78) and K. flavescens (Christiansen York and in southeastern New Hampshire were and Gallaway, ’84; Long, ’86; Iverson, ’90, ’91b). In fitted with radiotransmitters and followed to their South Carolina, 91.7% of 870 hatchling K. sub- hibernation sites (Ultsch et al., 2008). On Long rubrum captured at a drift fence from 1968 to 1998 Island, all hatchlings initially moved to water, but were collected in March or April, and only two later movements included terrestrial sojourns; were collected in late summer (J. Gibbons and those that could be located left the water and J. Greene, unpublished observations; see also overwintered in spring seeps. In New Hampshire, Gibbons and Nelson, ’78). In the Nebraskan hatchlings moved directly to nearby aquatic Sandhills, hatchling K. flavescens descend over habitats after emergence and hibernated aqua- 0.5 m below the nest chamber to overwinter tically in root masses near banks. (Costanzo et al., ’95b). This behavior presumably S. odoratus (common musk turtle; Kinosterni- is necessary because they are susceptible to dae): This species ranges into southern Maine and inoculative freezing (Costanzo et al., 2001b) and Ontario in the north and to the Gulf Coast in the do not tolerate somatic freezing (Costanzo et al., south. Most reports indicate that, throughout its ’95b). It seems likely that K. subrubrum exhibits range, hatchlings emerge from natal nests in similar behavior in the northern part of its range. autumn and presumably overwinter aquatically The , Kinosternon baurii, (Risley, ’33; Cagle, ’42; Carr, ’52; Minton, ’72; which inhabits the southeastern coastal plain Ernst, ’86; Mitchell, ’88). from Virginia through Florida, has an atypical Generally, it is easier to find hatchling turtles in developmental pattern among turtles of the Uni- fall, when most emerge from nests over 3–4 weeks, ted States and Canada. Generally, embryonic than in spring, when they may emerge over a development within the female proceeds to the period of months. This may be especially true for late gastrula stage, and is arrested in the oviduct musk turtles, whose hatchlings are very small and until the eggs are laid (Ewert, ’85). K. baurii also rarely bask, and hence not readily observed. There exhibits embryonic diapause, a state of arrested are some anecdotal reports of finding hatchling development that usually can be terminated only musk turtles emerging from nests in spring by a period of prolonged chilling followed by (G. Folkerts [Alabama—emerging in spring from rewarming (Ewert and Wilson, ’96). Eggs laid in nests on beaver dams], E. Keiser [Mississippi], autumn, and at lower temperatures, are more J. Weilbacher [Ohio], personal communications). likely to undergo diapause than eggs laid in spring Long-term monitoring of a drift fence encircling a or at relatively high temperatures. Because a South Carolina bay (as described in the section on given female may nest in autumn or spring C. serpentina) showed that of 60 hatchling musk (September–June, in Florida), the diapause is turtles collected between 1969 and 2000, 11 were regulated environmentally, rather than geneti- captured in March–June and 49 were captured in cally (Ewert and Wilson, ’96). Eggs laid in spring August or September (J. Gibbons and J. Greene, probably hatch in summer and fall, but it is unpublished data). Thus, 18% of the hatchlings unknown how many of the hatchlings then over- apparently overwintered on land (Gibbons and winter inside the nest. Iverson (’77) found hatchl- Nelson, ’78). Whether terrestrial hibernation is ings outside the nest most frequently in January possible in more northern climates is unknown; and March, but also in August, September, and however, given that this species often constructs December. Similarly, Mushinsky and Wilson (’92) shallow nests (Risley, ’33; Cagle, ’37; Edgren, ’42) reported predominantly spring emergence from and poorly tolerates freezing (Costanzo et al., nests, but also made some captures in October and 2006), hatchlings undoubtedly would need to December. Thus, there appears to be a minor leave the nest chamber for a more protected emergence in fall and a major one in spring, hibernaculum. perhaps with some individuals overwintering in Kinosternon spp. (mud turtles; ): nests as hatchlings and others as embryos. Two species of mud turtles occur in seasonally cold Two southwestern/Mexican kinosternids (K. so- climates in the United States: the eastern mud noriense and K. integrum) are found at high turtle (K. subrubrum) of the eastern region and altitudes where winters can be cold (Degenhardt the (K. flavescens) of the Great and Christiansen, ’74). Both species exhibit Plains. Terrestrial overwintering is routine for embryonic diapause and, at least at low altitudes, hatchlings of K. subrubrum (Richmond, ’45; late-stage embryos defer hatching until the follow- Skorepa and Ozment, ’68; Lardie, ’75; Gibbons ing summer (Ewert, ’91; Iverson, ’99). Whether

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 309 embryos can survive harsh winters at high alti- the occasional hatchling was caught at drift fences tudes (or else are forced to complete development in spring (Congdon et al., ’83b, 2000), suggesting in autumn) is an interesting question worthy of that they had successfully hibernated terrestrially. study. In Minnesota, 26 of 1,590 hatchlings collected at E. blandingii (Blanding’s turtle; Emydidae): drift fences were captured in spring (Pappas et al., One of the most northerly distributed of all 2000). Because in both cases it is unlikely that chelonians, Blanding’s turtles reportedly leave these hatchlings had found other water in which the natal nest in autumn from their eastern limit to hibernate, it appears that they successfully in Nova Scotia to their western limit in Nebraska overwintered on land, either within the natal nest (Conant, ’38; Sexton, ’57; Bleakney, ’63; Minton, or elsewhere. Terrestrial hibernation would be ’72; Graham and Doyle, ’78; Congdon et al., ’83b; aided by this species’ well-developed capacity for Butler and Graham, ’95; Standing et al., ’97, ’99; freeze tolerance (Packard et al., 2000; Dinkelacker McNeil et al., 2000; Pappas et al., 2000; Kolbe and et al., 2004; Costanzo et al., 2006), which, in turn, Janzen, 2002). In Nova Scotia, Herman et al. (’95) is promoted by its high susceptibility to inocula- found five apparently dormant hatchlings inside a tive freezing (Packard et al., 2000; Dinkelacker nest in late fall, but it is uncertain what their fate et al., 2004). Thus, the available information would have been had they not been disturbed. suggests that autumnal emergence is normally Although hatchlings commonly emerge from followed by movement (albeit not necessarily their nests in autumn, and their post-emergence direct) to water for overwintering, that some behavior has received considerable study (Butler hatchlings may be stranded on land during the and Graham, ’95; Standing et al., ’97; McNeil winter, and some unknown percentage of the et al., 2000; Pappas et al., 2000), it remains latter can survive until the following spring. Given unclear whether they ultimately overwinter in the mild temperatures they may encounter under water or on land. Some evidence suggests that snowpack, overwintering in terrestrial sites can- they could successfully hibernate on land. For not be ruled out (Congdon et al., 2000). example, hatchlings are not especially diligent in C. guttata (; Emydidae): Most heading for water after leaving the nest in studies that mention hatchlings of this species autumn, some wandering on land for days, even concern northern populations, although C. guttata if water was within 5 m (Standing et al., ’97, ’99); ranges southward to Florida. Numerous reports McNeil et al. (2000), who studied hatchlings at the suggest that some hatchlings may emerge from same location, made similar observations. Oddly, nests in spring (Conant, ’38; Nemuras, ’67; Ernst, these hatchlings showed no directionality to open ’75, ’76; Chippindale, ’89). This notion was based water and most avoided a lake near the nests. on findings of turtles that were still hatchling- However, this behavior may be particular to this sized in spring and presumably had just emerged locale because no shallow-water bodies were from their nests. However, an alternative explana- nearby. Apparently hatchling E. blandingii prefer tion is that these hatchlings actually emerged in shallow wetlands for overwintering and can travel autumn and overwintered elsewhere, but did not considerable distances to such sites. For example, grow before spring. Investigators who have in Massachusetts, Butler and Graham (’95) directly monitored nests and/or paid special tracked nine hatchlings that emerged from their attention to hatchlings all report autumnal nest. All went to water, but took up to 9 d and emergence (Belmore, ’80; Joyal, ’99; Carroll and traveled a mean distance of 187 m to reach it. B. Ultsch, 2007). Nevertheless, there are enough Butler and S. Smyers (personal communication) anecdotes of hatchlings captured abroad in spring found that hatchlings typically move about on that at least occasional terrestrial overwintering land for several days before entering aquatic cannot be ruled out; further studies with mon- habitats. M. Jones (personal communication) itored nests are required to make a definitive noted that hatchlings emerging from nests in statement. August or September may move to wetlands, Clemmys [Actinemys2] marmorata (western occupying forms usually near the margins, but pond turtle; Emydidae): This species ranges along whether they overwinter there or elsewhere is the west coast of the United States and Mexico, unknown. from Baja western California into Oregon and Nevertheless, not all such wanderers reach parts of Washington (Ernst et al., ’94). Across its water before cold weather arrives, and overwinter- ing on land may not always be lethal. In Michigan, 2See Footnote 1.

J. Exp. Zool. 310 J.P. COSTANZO ET AL. range, winters vary from mild to cold, although species sometimes does not dig a typical subterra- the hatchlings commonly hibernate inside the nean nest, but instead lays its eggs in moss or natal nest (Buskirk, ’91; Reese and Welsh, ’97; clumps of grass (Holub and Bloomer, ’77); it is Rathbun et al., 2002; Van Leuven et al., 2004). As doubtful that in such cases hatchlings could with most species whose hatchlings overwinter in successfully overwinter in situ. There are some the nest, autumnal emergence does occur occa- reports of finding hatchling-sized turtles in spring, sionally, even as far north as Washington or other suggestions that the turtles remain in (F. Slavens, K. Slavens, and D. Anderson, personal nests during winter (Bloomer and Bloomer, ’73; communications), although it is more common in Mitchell, ’94), but the extent of terrestrial over- the southern portion of the range, where hatchl- wintering by hatchlings, if it occurs, is unknown. ings may emerge in autumn or spring (Holland, C. picta (painted turtle; Emydidae): There are ’94; Hays et al., ’99; Bettelheim, 2004). Although four recognized subspecies of painted turtles, adults reportedly may overwinter on land or three of which occur as far north as Canada: underwater (Holland, ’94), hibernacula used by C. p. picta (eastern painted turtle) in the eastern hatchlings emerging in autumn have not been seaboard states and provinces; C. p. marginata described. (midland painted turtle) in the Great Lakes area; C. [Glyptemys3] insculpta (; Emydi- and C. p. bellii (western painted turtle) from the dae): This turtle primarily inhabits the north- Great Plains westward to the Pacific Ocean. The eastern United States, southeastern Canada, and fourth subspecies, C. p. dorsalis5 (southern the Great Lakes region, and ranges only as far painted turtle), is a form of the lower Mississippi south as northern Virginia. The adults overwinter River drainage and ranges from southern Missouri in rivers and streams and, at the northern limit of to the Gulf Coast. It had been long suspected that their range, may encounter near-freezing tem- overwintering in the nest may occur in the peratures inside hibernacula (Greaves and Litz- northern subspecies (C. p. bellii—Thacker, ’24; gus, 2007). Hatchlings evidently do not overwinter Woolverton, ’63), based largely on observations of in the nest (Harding and Bloomer, ’79; Farrell and what appeared to be numerous newly hatched Graham, ’91; Harding, ’91; Tuttle and Carroll, ’97, turtles abroad in spring, and/or the lack of similar 2005; Ernst, 2001; Buech et al., 2004; Carroll and observations in autumn. Many later studies, in Ultsch, 2007), but rather emerge in autumn and which investigators caged nests to determine seek winter quarters elsewhere. They do not precise emergence dates or used drift fences to always head directly to water after leaving the capture hatchlings as they migrated to water, have nest. In New Hampshire, Tuttle and Carroll (’97, verified that the majority of the northern hatchl- 2005), using fluorescent powder to track move- ings spend their first winter inside the nest cavity ments, found that 12 hatchlings took 1–24 d to (C. p. bellii—Christiansen and Gallaway, ’84; St. enter nearby brooks. In New Jersey, S. Castellano, Clair and Gregory, ’90; Lindeman, ’91; Costanzo J. Behler, and G. Ultsch (unpublished observa- et al., ’95b, 2004; Pappas et al., 2000; tions) used radiotransmitters to follow seven C. p. marginata—Tinkle et al., ’81; Costanzo hatchlings, from 1 August to October 15, as they et al., 2004; C. p. picta—Finneran, ’48; Ernst, dispersed from their nests. The turtles remained ’71; DePari, ’96; Rozycki, ’98; Carroll and Ultsch, in agricultural fields for at least 13–62 d during the 2007). There have not been as many reports on period of observation. Foraging on small slugs was C. p. dorsalis, but overwintering in the nest occurs observed seven times, including once within 24 h frequently as far south as Louisiana and Mis- of nest emergence; therefore, the turtles were sissippi (Cahn, ’37; Cagle, ’54). feeding during this time. The hatchlings gained Trachemys spp. (sliders; Emydidae): Trachemys mass (1.1 g) and increased their length (2.6 mm) scripta elegans, the red-eared slider, is the only before hibernating in presumably aquatic sites. member of this group with a significant northern Clemmys [Glyptemys4] muhlenbergii (; element to its range, which extends from the Gulf Emydidae): Relatively little is known of hatchling coast to northern Illinois and Indiana. In Illinois, behavior in this small and secretive turtle. Fall emergence from the nest occurs almost entirely in emergence seems to be the rule (Barton and Price, spring (Cagle, ’44; Smith, ’61; Tucker, ’97, ’99, ’55; Bloomer and Bloomer, ’73; Mitchell, ’94). This 2000a; Willette et al., 2005). Overwintering in the

3See Footnote 1. 5Considered a full species, chrysemys dorsalis, by the Committee 4See Footnote 1. on Standard English and Scientific Names (Iverson et al., 2008).

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 311 nest also occurs in northern Indiana (Costanzo tions). In Virginia, at least some hatchlings leave et al., 2001c; P. Baker, J. Iverson, P. Meyer, the nest in autumn and hibernate elsewhere personal communications) and possibly in Iowa (Mitchell, ’94). In northern Florida, 13 of 15 (Christiansen and Vandewalle, 2000). Hatchlings hatchlings found crossing roads were encountered appear not to be particularly freeze-tolerant in March and April (Jackson, ’94), although in (Churchill and Storey, ’92b; Packard et al., ’97c; peninsular Florida, nest emergence usually occurs Dinkelacker et al., 2005b; Costanzo et al., 2006). before winter (Jackson, ’88). Aresco (2004), who They also do not supercool appreciably when in excavated and reburied eggs in northern Florida, contact with frozen soil, evidently because the skin found that hatchlings emerged in both autumn offers little resistance to the penetration of ice and spring, although relative numbers were not crystals (Packard et al., ’97c; Costanzo et al., reported. In New Mexico, a nest of the Big 2001b). Hence, overwintering in the nest requires Bend slider (T. gaigeae) contained hatchlings in that the soil about the hatchlings does not freeze, January that were gone in April, suggesting which in turn may limit the northern distribution that the turtles had overwintered successfully of the species, given that the hatchlings do not opt (Morjan and Stuart, 2001). In summary, it appears to overwinter aquatically. that the majority of sliders overwinter inside The hibernation habits of slider hatchlings are the natal nest, although some fall emergence not so clear in the southern portions of the range. occurs in the more southern portions of their Cagle (’44) cited what he considered a reliable range. source, who said that emergence of red-eared spp. (red-bellied turtles and cooters; sliders in Tennessee was approximately evenly Emydidae): The Florida cooter (P. floridana divided between fall and spring. One turtle farmer floridana6) occurs on the coastal plain from in Louisiana (J. Evans) told us that T. s. elegans Virginia to Alabama, exclusive of peninsular emerge in both autumn and spring on his farm, Florida, which is occupied by the peninsular cooter but that in the field, hatchlings apparently emerge (P. f. peninsularis7). Gibbons and Nelson (’78) mostly in spring. Another (H. Kleibert) said that trapped all but one of 32 hatchling P. f. floridana any eggs not harvested in autumn produce during spring at a drift fence in South Carolina. hatchlings that emerge the following spring, and During 1968–1998, at a drift fence encircling a overwintering of hatchlings inside the nest is also South Carolina bay, 84.8% of 119 hatchlings were the case in the field. In contrast, Cagle (’50) found captured in March and April and only two were that 34 natural nests in Louisiana ‘‘hatched’’ in captured in September–November (J. Gibbons and July–September; however, one cannot assume J. Greene, unpublished data; see also Gibbons and that this means that the hatchlings emerged, as Coker, ’77). In Alabama and the Florida panhan- this was not specifically stated, especially as he dle (Walton County), most of the hatchlings also stated that they often overwinter in the nest, collected by Thomas (’72) were taken in spring, and that a farmer in the New Orleans area found leading to the presumption that hatchling many hatchlings in his field each spring. The P. f. floridana overwinter in the nest. In north appearance of hatchlings abroad in spring has Florida, Jackson (’94) observed four hatchlings been noted elsewhere (T. Mann [Mississippi], crossing roads in spring, but found none in D. Nelson [Alabama], D. Haynes, F. Rose [Texas], autumn, and Aresco (2004) reported that hatchl- personal communications); therefore, it appears ings emerged from artificial nests in both autumn that overwintering in the nest occurs throughout and spring. In Alabama, some hatchlings emerge the range, with possibly a trend toward increasing, in fall, although the majority hibernate inside the but still minor, fall emergence at lower latitudes. nest (Thomas, ’72). Trachemys s. scripta: Trachemys s. scripta, the In peninsular Florida, P. f. peninsularis nests yellow-bellied slider, is a southeastern subspecies. from fall through spring and is considered a In South Carolina, hatchlings typically hibernate ‘‘winter nester,’’ as is the inside the natal nest, as indicated by drift fence (Deirochelys; see below) in this area (Ewert, ’91). captures (Gibbons and Nelson, ’78). Monitoring a Hatchlings have been found abroad in autumn drift fence encircling a South Carolina bay showed 6Considered a subspecies of Pseudemys concinna by the that of 320 captures of hatchlings during Committee on Standard English and Scientific Names (Iverson 1968–1998, 84.7% were made in March and April et al., 2008). 7Considered a full species, Pseudemys peninsularis, by the and only 5.9% occurred in August–October Committee on Standard English and Scientific Names (Iverson (J. Gibbons and J. Greene, unpublished observa- et al., 2008).

J. Exp. Zool. 312 J.P. COSTANZO ET AL.

(Iverson, ’77; Jackson, ’88), although there are no ings that hibernated inside the nest, 10 produced data regarding when the majority of nest emer- hatchlings that emerged in autumn and over- gence occurs. wintered elsewhere, and 5 had hatchlings that In Illinois, P. concinna () hatchlings emerged in both autumn and spring. In total, 163 are found abroad in autumn (Cahn, ’37), whereas in hatchlings hibernated inside the nest and 126 northern Florida, of the 11 nests studied by Jackson hibernated elsewhere, with some females produ- (’94), hatchlings emerged from six in autumn and cing hatchlings exhibiting either behavior. Again, fiveinspring;therewasnocleartendencyfor there was no correlation between emergence hatchlings in late-constructed nests to hibernate in timing and date of oviposition. Neither nest situ. In Alabama, hatchlings sometimes overwinter temperature nor rainfall influenced season of outside the nest, but the tendency is to remain emergence, although in spring, rainfall and rising inside (Fahey, ’87). In West Virginia, Buhlmann and temperature did stimulate nestling emergence, as Vaughn (’91) collected five P. concinna they con- was found for red-eared sliders in Illinois sidered to be hatchlings in April–June, which they (J. Tucker, personal communication). In sum- presumed had overwintered in the nest; however, at mary, the timing of hatchling emergence of the 32–42 mm, these individuals were large as compared complex of red-bellied turtles appears to be some- with the 27–36 mm size given for hatchlings by what plastic, perhaps dependent on the micro- Ernst et al. (’94); therefore, more definitive evidence environment within a given nest or even among of overwintering would be needed to affirm their eggs within a single nest. supposition. Graptemys spp. (map turtles; Emydidae): In D. Nelson ([Alabama] personal communication) general, map turtles are a southern group; how- had found hatchling P. alabamensis (Alabama red- ever, several species range into northern areas, bellied turtle) during spring roadkill surveys, especially the common map turtle (G. geographi- suggesting that these turtles hibernate inside the ca), which is found in southern Canada, the Great natal nest. In Florida, the Florida red-bellied Lakes region, and northern Vermont. Overwinter- turtle (P. nelsoni) appears to emerge in fall in ing in the nest in northern areas seems to be the the northern part of the state (Lardie, ’73; more prevalent habit. In Minnesota, Pappas et al. Jackson, ’88), but the situation in the southern (2000) captured 38 of 45 hatchlings moving from part is unclear (Lardie, ’73); in any event, data are land to water at a drift fence in spring, and in scarce for this species. Iowa, the only two hatchlings similarly caught at Mitchell (’74, 2003) reportedly observed red- drift fences by Christiansen and Gallaway (’84) bellied turtle (P. rubriventris) hatchlings emerging were captured in spring. By monitoring individual from nests in April and in September in Virginia, nests, Nagle et al. (2004) found that the incidence but farther north, T. Graham (personal commu- of hatchlings overwintering inside the natal nest nication) has collected many hatchlings abroad, all was 95% (n 5 75 nests); Baker et al. (2003) found in autumn, from an isolated population in Massa- the same percentage (n 5 20 nests) in a northern chusetts. This is a reversal of the situation in most Indiana population. Overwintering inside the nest other species of facultative terrestrial hibernators, has been suspected by others, mostly on the basis wherein overwintering in the nest tends to be of capturing hatchlings abroad in spring and more common in the north than in the south. In finding them scarce in autumn (Newman, ’06; Maryland, Swarth (2004) determined that emer- McCallum, 2003). However, some fall nest emer- gence of P. rubriventris hatchlings from 17 nests gence occurs in northern areas, although it is was split about evenly between fall (n 5 60) and usually minor (Newman, ’06). The only reports spring (n 5 85) and that timing was not correlated stating that overwintering inside the nest does not with oviposition date. For one female that laid occur (or occurs rarely) in this species, as well as in three nests over 2 years, a nesting on 7 June G. ouachitensis and G. pseudogeographica at the produced hatchlings that emerged in spring; same Wisconsin locale, are those of Vogt (’80, ’81), hatchlings from a 29 June nesting emerged in who did not directly monitor nests. Thus, most autumn and overwintered elsewhere; and a second evidence indicates that hatchlings in northern nest laid on 12 July of the same year produced 12 populations remain inside nests during winter. hatchlings that remained in situ, but died during Reports for southern populations are lacking, but, winter. These findings are bolstered by a second near Birmingham, AL, R. Guthrie (personal com- study at the same site (Friebele and Swarth, munication) observed hatchling G. geographica 2005): of 31 monitored nests, 16 produced hatchl- abroad in spring.

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 313

The hibernation habits of hatchlings of other naeva and Orlova, 2004; Novotny et al., 2004); in species of map turtles have been less well studied. the same population, either situation may occur Vogt (’80, ’81) unequivocally stated that both (Kotenko, 2000). Hatchlings may emerge from G. ouachitensis and G. pseudogeographica in nests either in autumn or spring in Austria Wisconsin emerge from nests in autumn. In (Farkas, 2000; Rossler, 2000), Russia (Nikolskii, Indiana, M. Ewert (personal communication) ’15; Bozhansky and Orlova, ’98; Kuzmin, 2002), found hatchling G. ouachitensis abroad in spring, France (Servan, ’98), and Lithuania (Snieshkus, whereas, in Missouri, J. Tucker (personal commu- ’98). However, in the two latter locales, hiberna- nication) found them abroad in autumn. Similarly, tion inside the nest is the far more common habit; Cahn (’37) found hatchling G. pseudogeographica this is also true of populations in Germany in autumn in Illinois. In contrast, in Minnesota, (Andreas and Paul, ’98; Schneeweiss et al., ’98) M. Pappas (personal communication) observed and Slovakia (Novotny et al., 2004). The evidence many hatchlings of this species migrating to water suggests that emergence timing is influenced by in spring, and a drift fence deployed in Iowa climate and year-to-year variation in temperature, captured hatchlings in spring, but not in autumn such that relatively warm weather often results in (Christiansen and Gallaway, ’84). In the southern fall emergence (e.g., in Italy; Zuffi, 2000) and cold United States, hatchlings of G. flavimaculata weather induces overwintering in the nest. There (Anderson, ’58; Horne et al., 2003), G. nigrinoda is no evidence of an embryonic developmental (Lahanas, ’82); G. gibbonsi (R. Jones [who mon- diapause (Ewert, ’91); hence, eggs that do not itored caged nests in Mississippi], personal com- hatch before winter will die. The pattern of nest munication), G. pulchra (Shealy, ’76), G. barbouri emergence in this species appears much less fixed (Wahlquist and Folkerts, ’73), and G. oculifera than in North American species. (Anderson, ’58) typically emerge from nests in Deirochelys reticularia (chicken turtle; Emydi- autumn and presumably hibernate aquatically. dae): This turtle has an unusual reproductive The available information for Graptemys indicates pattern. In Florida, nesting occurs from late that the more northerly species hibernate in the September to mid-March (Iverson, ’77; Jackson, nest in both the northern and southern parts of ’88), whereas, in South Carolina, the winter their range, but principally southern species segment of the period is absent and nesting occurs emerge from nests in autumn; thus, overwintering only in fall and spring (Congdon et al., ’83a; in the nest may be a fixed trait in some species Buhlmann and Gibbons, 2001). In Florida, re- that is independent of latitude. In cold climates, cently emerged hatchlings have been found in terrestrial hibernation is promoted by a well- August (Iverson, ’77), but in South Carolina, most developed capacity to supercool (Baker et al., 2003). hatchlings (394 of 408 over an 18-year period; J. E. orbicularis (European pond turtle; Emydi- Greene and J. Gibbons, unpublished data; see also dae): We include this species because of its wide Gibbons, ’69; Gibbons and Greene, ’78) appear in distribution in Europe and western Asia, the March and April. Hatchlings apparently hibernate abundant information on its biology, and its inside the nest, even in the deep South (Gibbons proximate phylogenetic relationship to North and Nelson, ’78; Buhlmann, ’98). In fact, diapaus- American turtles of the genera Clemmys (includ- ing eggs (Ewert, ’85) and/or hatchlings spend at ing and Glyptemys) and Emydoidea least one winter, and possibly two, inside the nest. (formerly Emys). In Poland, clutches may fail to M. terrapin (diamondback terrapin; Emydidae): develop due to cold, may emerge from nests in This is the only strictly brackish-water turtle in autumn and hibernate elsewhere, may success- the United States, occurring in estuaries and tidal fully overwinter (up to 93% survival) as hatchlings marshes from Massachusetts to Texas. Hatchlings inside nests, or may burrow into the soil below the exhibit marked variability in timing of emergence level of the nest (Mitrus and Zemanek, 2000, 2003; from the natal nest. On Long Island, Wojakowski Mitrus, 2005). Timing of nest emergence can vary et al. (2005) found that of 87 monitored nests, 78 considerably from year to year, apparently de- produced hatchlings that emerged in autumn. In pending on the average summer temperature. an extension of that study, in which 122 nests Perhaps this dependence explains why hatchlings were monitored, all live hatchlings in 114 clutches, in other locales emerge from nests in both autumn and some of the live hatchlings in four clutches, and spring. For example, in Ukraine, overwinter- emerged in autumn and hibernated elsewhere, ing reportedly occurs inside the nest (Szcerbak, whereas only two clutches had at least one ’98) and outside the nest (Kuzmin, 2002; Maza- hatchling that survived to emerge in spring

J. Exp. Zool. 314 J.P. COSTANZO ET AL.

(A. Widrig and R. Burke, personal communica- Winter environment tion). Most other reports have also indicated that The physical environment to which hatchling fall emergence predominates (Burger, ’76a, ’77; turtles are exposed in winter impacts their physio- Auger and Giovannone, ’79; Mitchell, ’94; Butler, logical condition and prospects for survival, and 2004; Draud et al., 2004); however, it is possible even their fitness after emergence from hiberna- that hibernation inside nests has gone unnoticed tion. Natural hibernacula are notoriously difficult at some sites (Roosenburg, 2003). For example, in to locate and study, especially during inclement a 3-yr study in New Jersey, Baker et al. (2006) conditions; therefore, there is a paucity of even found that a significant number of hatchlings fundamental information about them. This is hibernated inside the nest. Overwintering success unfortunate because their physical characteristics was ascribed to a well-developed freeze tolerance govern the exchange of heat, water, and gases by (see also Dinkelacker et al., 2005b; Costanzo et al., overwintering turtles. In turn, these factors drive 2006), which is accompanied by a high suscept- evolution of strategies for coping with cold, dehy- ibility to inoculation from ice and ice-nucleating dration, and hypoxia, the primary environmental agents (INA) in the environment. challenges confronting hatchlings in winter. It is commonly assumed that turtles emerging from Intolerance of environmental extremes may the natal nest in autumn ultimately overwinter in cause significant attrition of hatchling turtles. water. However, hatchling M. terrapin often do not Reports of widespread mortality among hatchling move directly to water (W. Roosenburg, personal turtles are not uncommon (Andreas and Paul, ’98; communication), preferring to hide in the tidal wrack St. Clair and Gregory, ’90; Lindeman, ’91; Pack- (Lovich et al., ’91). Juveniles are found under ard, ’97; Packard et al., ’97a; Nagle et al., 2000) vegetation up to 100 m from water (Pitler, ’85). and demographic studies have shown that winter- Terrestrial overwintering of hatchlings outside of the kill may contribute significantly to mortality of nest also occurs on Long Island, NY (Draud et al., this age class (Tinkle et al., ’81), but few studies 2004). Whether hibernation under water during the have identified the specific mortality factors. first winter is also common remains to be determined. Hypothermia may be such a factor, especially in Apalone spp. (softshell turtles; ): northern populations of terrestrial hibernators Both the smooth softshell (A. mutica) and spiny (St. Clair and Gregory, ’90), as losses often are softshell (A. spinifera) range from the Gulf coast to greatest in years of low temperatures and scarce the northern states and/or Canada, and therefore snowfall (Breitenbach et al., ’84; Andreas and are found in areas with cold winters. Throughout Paul, ’98; Schneeweiss et al., ’98; DePari, ’96; their ranges, both species are known to emerge Nagle et al., 2000; Carroll and Ultsch, 2007). from their nests in autumn, presumably to allow Terrestrial hibernators also succumb to flooding aquatic overwintering (A. mutica—Muller, ’21; (Christens and Bider, ’87; Holte, ’88; DePari, ’96; Anderson, ’58; Fitch and Plummer, ’75; Plummer, Rozycki, ’98). Dehydration may contribute to ’76, ’77; Christiansen and Gallaway, ’84; winter mortality, either directly or indirectly A. spinifera—Surface, 1908; Cahn, ’37; Brecken- (Christiansen and Gallaway, ’84; Schneeweiss ridge, ’60; Smith, ’61; Christiansen and Gallaway, et al., ’98; Packard et al., ’89; Costanzo et al., ’84; Costanzo et al., ’95b). Hatchlings are not 2001b). Environmental stressors may interact in freeze-tolerant (Costanzo et al., ’95b) and owing to ways that have particularly significant effects on the large amount of exposed skin, they presum- overwintering success. For example, cold and wet ably would be highly susceptible to inoculation by winters are particularly challenging to hatchling ice and INA in the nest environment (Costanzo C. picta hibernating within natal nests (Rozycki, et al., 2001b). Softshells are also found in cold ’98; Costanzo et al., 2004), probably owing to climates in Russia (e.g., sinensis), increased risk of freezing mortality. Below, we where they, too, overwinter outside the nest summarize the available information concerning (Kuzmin, 2002), presumably in an aquatic habitat. physical features of turtle hibernacula. Limited cold hardiness could explain why hatchl- ing softshells leave the nest in autumn at high Thermal regime latitudes, but cannot account for similar behavior in more tropical populations. Because A. spinifera Hatchling turtles are small, ectothermic organ- and A. mutica typically nest on sandbars, fall isms whose heat content closely tracks changes in emergence may be a way to escape dehydration ambient temperature. Nevertheless, there is some and winter flooding. evidence for metabolic thermogenesis in turtle

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 315 nests (Bustard, ’71; Burger, ’76b). During emer- Investigators have lavished attention on the gence from their nest, loggerhead turtles (Caretta severity of the thermal conditions within indivi- caretta) reportedly liberate enough heat that dual hibernacula, particularly with the aim of the surrounding sand warms noticeably (Moran linking them to observations of hatchling mortal- et al., ’99), but it is unclear whether this heat ity (e.g., Weisrock and Janzen, ’99). Drawing was endogenously produced or whether the turtles inferences about this association hinges on the simply radiated heat they had gained from assumption (probably true) that freezing is a deeper, warmer places in the soil column. Even major cause of winter mortality; factors influen- hibernating hatchlings release small amounts cing freezing risk, such as seasonal minimum of metabolic heat, but there is no evidence temperature and frost duration, are the most that this heat significantly alters their thermal critical elements of such analyses (for a discussion, environment. Hatchlings liberate a substantial see Lee and Costanzo, ’98). Because freezing risk amount of heat (335 J gÀ1) when water in their can fluctuate with changes in certain environ- tissues freezes, but whether or not this heat mental variables, frequency of frost exposure is could impact the thermal dynamics of other also an important metric. turtles sharing the hibernaculum has not been Some analyses (Costanzo et al., ’95b, 2004; determined. Packard, ’97; Packard et al., ’97a; Baker et al., Investigators have attempted to characterize the 2003, 2006) of the thermal conditions inside thermal conditions affecting hatchling turtles in terrestrial hibernacula have enumerated and natural hibernacula (Breitenbach et al., ’84; characterized individual bouts of frost exposure, DePari, ’88; Schneeweiss et al., ’98; Packard each representing a discrete period of chilling at et al., ’89, ’97a; St. Clair and Gregory, ’90; temperatures that are at or below the equilibrium Costanzo et al., ’95b, 2004; Packard, ’97; Weisrock freezing/melting point of turtle tissues, À0.61C. and Janzen, ’99; Nagle et al., 2000). This task is One assumption implicit in these analyses is that simplified by deploying miniaturized temperature the risk of freezing (and death) increases with the recorders directly inside occupied turtle nests. To frequency and duration of such exposures, and avoid disturbance of the nest chamber, investiga- also with the severity of the cold. The consensus of tors sometimes place these units adjacently at an these studies is that the typical subzero chilling appropriate depth in the soil column; resulting excursion is relatively benign, as turtles commonly data satisfactorily represent nest temperatures encounter temperatures only slightly below (Breitenbach et al., ’84; Nagle et al., 2000; À0.61C. An analysis of the thermal relations of Costanzo et al., 2004). hatchling C. picta in the Nebraskan Sandhills, for Recordings of temperatures within northern example, showed that hibernaculum (natal nest) natal nest hibernacula commonly show that temperature was within 21C of the tissue freezing average daily temperature gradually decreases point for 30–50% of the total period of freezing from 151C in October to near 41C in December risk (Costanzo et al., ’95b). During an exception- and slightly lags the seasonal decrease in air ally cold winter, the typical chilling excursion temperature (see Costanzo et al., ’95b, 2004). exposed hatchlings to temperatures above À3.51C Warming of the nest chamber, following a reversal and lasted only 24 h (Table 1). The analysis also of this pattern in spring, usually begins in early showed that, although hatchlings in deeper nests March. Examined on a finer scale, temperature would have encountered fewer frost exposures, recordings often show daily fluctuations of up to varying the nest depth from 5 to 25 cm had little several degrees Celsius. In winter, such oscilla- effect on minimum temperature, duration, or tions, which correspond to diel heating and cooling maximum cooling rate. cycles, may expose hatchlings to frost on a nightly There is substantial geographic variation in basis. During the coldest segment of winter, thermal dynamics and severity of cold inside typically from late November through early hibernacula used by hatchling turtles. In regions March, temperatures inside nests may or may where snowfall is abundant, temperatures in natal not descend appreciably below the equilibrium nests often hover around the freezing mark and freezing point of turtle tissues, À0.61C. Despite instances in which hatchlings actually encounter the popular notion that hatchlings encounter a frost are relatively uncommon (Obbard and winter-long period of intense cold, data suggest Brooks, ’81a; Breitenbach et al., ’84; Schneeweiss that frost exposure is transient, even in severe et al., ’98; Paukstis et al., ’89; Rozycki, ’98; Nagle winters (Table 1). et al., 2000; Costanzo et al., 2004). During periods

J. Exp. Zool. 316 J.P. COSTANZO ET AL.

TABLE 1. Characterization of potential freezing episodes at a site in Garden County, NE, where painted turtles (Chrysemys picta bellii) overwinter inside nests

Minimum temperature (1C) Duration (h) Maximum cooling rate (1ChÀ1)

N Median (range) Median (range) Median (range)

Typical winter 16 À1.6 (–9.6 to –0.8) 8 (2–86) 0.33 (0.05–0.85) Cold winter 14 À3.4 (–11.8 to –0.8) 23 (6–320) 0.58 (0.15–0.55)

Characteristics of potential freezing episodes (N), each a discrete sequence of soil temperatures below À0.61C, the approximate equilibrium freezing/melting point of turtle tissues, during two winters of record. Temperatures were measured 10 cm below the surface of a sandhill and are representative of conditions inside nests containing hatchling C. picta. Cooling rate was determined from successive temperature recordings made at 2-h intervals. In the ‘‘typical’’ winter (1993–1994), the average daily air temperature for December and January was similar to the mean of these values (À3.8 and À4.01C, respectively) over the preceding 15 yr; in the ‘‘cold’’ winter (1990–1991), the average daily air temperature for December and January was lower than the respective long-term means by 3.8 and 2.21C, respectively. Adapted from Costanzo et al. (’95b). of severe weather, nest temperature may dip sites that promote development of viable offspring slightly below freezing, but an insulating snow- and potentially influence their sex (Schwarzkopf pack usually renders this peril modest and brief and Brooks, ’87; Ewert et al., ’94; Janzen, ’94; (i.e., hours to days). However, hatchlings may Morjan, 2003), but whether or not they also select encounter extreme cold on the rare occasion that sites that minimize environmental stress on over- snow is lacking (e.g., Rozycki, ’98; Weisrock and wintering hatchlings is not clear (Weisrock and Janzen, ’99; Nagle et al., 2000). In Acadia National Janzen, ’99). Park, ME, temperature in three C. picta nests Developing embryos experience a substantial vacillated between 0 and À81C, reaching minima range in temperatures due to vertical thermal of À10.7 to À12.61C during January and February gradients within the soil column (Thompson, ’88; owing to a lack of snow cover. Snow fell in March Georges, ’92; Tucker, ’99). By the same token, and nest temperatures moderated, remaining near thermal conditions can also vary considerably 01C (Rozycki, ’98). within a single hibernaculum as, for example, Thermal conditions differ considerably in locales hatchlings overwintering near the top of a nest where winters are extreme and snow cover is light could become colder than those remaining near and transient. One such place is the Sandhills the bottom (Costanzo et al., ’95b; Tucker, ’99). region of central Nebraska. There, most chilling Thermal variation probably is stronger in shallow bouts in C. picta nests are brief, such that nests (e.g., Houghton and Hays, 2001) and in hatchlings cool only a few degrees below 01C larger nests (i.e., those containing many hatchl- before rewarming; however, extended chilling can ings), and in soils prone to developing a sharp occur and turtles can reach critically low tem- depth–temperature gradient, and potentially peratures, sometimes, though not always, with could explain the mixed survival occurring in lethal consequences (Costanzo et al., ’95b, 2004; some nests (MacCulloch and Secoy, ’83; DePari, Packard, ’97). Extreme chilling excursions occur ’88; Packard et al., ’89; Lindeman, ’91; Costanzo infrequently and tend to be relatively brief, et al., 2004). although they can profoundly impact cohort survival. Hydric regime Several studies have demonstrated pronounced variability in the intensity and duration of Water is critical to life and, if not obtained subfreezing cold among hibernacula at the same through feeding (likely the case in overwintering locale (Breitenbach et al., ’84; Packard et al., ’89, hatchlings), must be obtained from the environ- ’97a; Paukstis et al., ’89; Packard, ’97; Weisrock ment, if possible, or by catabolism of endogenous and Janzen, ’99; Nagle et al., 2000). For example, fuel reserves. Water availability is obviously not seasonal minimum temperatures recorded in an issue for hatchlings overwintering aquatically, different C. picta nests vary considerably, ranging but may be a significant problem for terrestrial 101C or more (Table 2). Such variation apparently hibernators. Indeed, desiccation and associated reflects differences in nest depth, as well as slope, osmo-ionic perturbations may be a more formid- aspect, patchiness of snow cover, and other able mortality factor than starvation in dormancy physiognomic factors (Weisrock and Janzen, ’99; (Seidel, ’78; Gregory, ’82). On the other hand, Costanzo et al., 2004). Female C. picta choose nest excessive moisture and inundation can cut off

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 317

TABLE 2. Survival and minimum temperatures to which hatchling turtles are exposed during hibernation inside natal nests

Survival (%) Minimum Sample Locale Winter n Range Avg. Nest temp. (1C) Period Reference

Chrysemys picta SE British Columbia 1987–1988 3 0–10 3.3 À6 (1) Jun–Mar 1 1988–1989 8 0 0.0 –5 (1) Jun–Apr 1 St. Louis Co, MN 1961–1962 1 100 – –11 (1) Nov–Mayf 2 Cherry Co., NE 1987–1988 7 22–100 60.1 –6.2 to –0.2 (7) Nov–Febf 3 1994–1995 14a 0–100 81.4 –11.3 to –2.9 (14) Nov–Mar 4 1995–1996 18a 0–100 63.2 –21.0 to –3.0 (18) Nov–Mar 5 Garden Co., NE 2000–2001 9 0–100 70.6 –10.2 to –3.9 (9) Jun–Mar 6 2001–2002 9 67–100 87.4 –15.0 to –6.1 (9) May–Mar 7 Carroll Co., IL 1995–1996 9 0–100 72.3 –12.3 to –2.4 (10) Nov–Mar 8 Morris Co., NJ 1985–1986 1 100 – –10.2 (1) Sep–Apr 9 1986–1987 1 100 – –5.5 (1) Jun–Decf 9 Central Ontario 1987–1988 3 100 100.0 –8, –6 (2) Jan–Febf 10 Livingston Co., MI 1981–1982 9 100 100.0 –3.3 to 0.5 (9) Nov–Apr 11 1995–1996 8 n.r. 58.5c –7.5d Nov–Mar 12 1996–1997 35 n.r. 98.9c –1.0d Nov–Mar 12 1997–1998 16 n.r. 98.9c –4.4d Nov–Mar 12 1998–1999 17 n.r. 100.0c –2.4d Nov–Mar 12 Graptemys geographica Huntingdon Co., PA 2000–2001, 2001–2002 10 100 100.0 –8.4 to –2.4 (10) Jun–Apr 13 Fulton Co., IN 2000–2001 6 100 100.0 –3.5 to –1.4 (6) May–Jul 14 2001–2002 6 90–100 98.3 –5.4 to –3.6 (3) May–Jul 14 Malaclemys terrapin Cape May Co., NJ 2002–2003 6 0–100 80.4c –2.2, –1.6 (2) Jul–Mar 15 2003–2004 43 0–100 85.1c –4.7 to –1.0 (8) Jul–Mar 15 Trachemys scripta Jersey Co., IL 1996–1997 7b 28–100 85.1 –4.6 to –1.1 (7)e Jan–May 16

Survival represents the percentage of fully formed hatchlings in individual nests alive in spring; minimum nest temperature is lowest temperature observed inside or adjacent to individual nests (range is reported, n given in parentheses) during the indicated sample period. n.r., not reported. References: (1) St. Clair and Gregory (’90); (2) Woolverton (’63); (3) Packard et al. (’89); (4) Packard et al. (’97a); (5) Packard (’97); (6) Costanzo et al. (2004); (7) Costanzo et al. (2003); (8) Weisrock and Janzen (’99); (9) DePari (’88); (10) Storey et al. (’88); (11) Breitenbach et al. (’84); (12) Nagle et al. (2000); (13) Nagle et al. (2004); (14) Baker et al. (2003); (15) Baker et al. (2006); (16) Tucker and Packard (’98). aNatural nests were stocked in late October with turtles hatched in the laboratory. bArtificial nests were stocked in mid-October with turtles hatched in the laboratory. cAverage for all hatchlings observed, not individual nests. dTemperature of soil at the average depth of nest bottom, 7.5 cm. eRecorded near the top of the nests. fDiscontinuous sampling regimen probably underestimated true winter minimum temperature. oxygen supply and cause osmoregulatory pertur- encounter static, humid conditions. The few speci- bations. mens examined were found above the water table There is little information about the hydric (Costanzo et al., ’95b; Converse et al., 2002), relations of neonates overwintering in natal nests although in lowland situations inundation by or other terrestrial hibernacula. Hatchlings over- rising ground water could occur. Inundation can wintering inside natal nests probably are exposed also be problematic for hatchlings that hibernate to a wide range of hydric conditions reflecting the inside natal nests. Riverine species, such as diversity of oviposition sites; however, turtles G. geographica and G. pseudogeographica, over- generally nest in mesic sites because adequate wintering inside floodplain nests are susceptible to moisture is required for proper development of flooding, and hatchling M. terrapin hibernating on incubating embryos (Packard and Packard, ’88; tidal flats may be inundated by storm surges Ratterman and Ackerman, ’89; Cagle et al., ’93). (R. Wood, personal communication). Even species Hatchlings overwintering deep in the soil column, that commonly nest in upland sites are occasion- such as K. flavescens and T. ornata, probably ally impacted by flooding or chronic soil saturation

J. Exp. Zool. 318 J.P. COSTANZO ET AL.

(Moll and Legler, ’71; Gibbons and Nelson, ’78; Gases Christens and Bider, ’87; DePari, ’96; Rozycki, ’98; Standing et al., ’99). Holte (’88) observed that Oxygen, carbon dioxide, and water vapor are the nesting soils used by C. marmorata were compact three most labile and influential gases with and dry during nesting and incubation, but many respect to the winter biology of hatchling turtles. nests containing hatchlings were inundated be- Apparently there are no published reports of tween November and April. Higher mortality ambient gas concentrations in aquatic or terres- occurred in nests that remained waterlogged for trial hibernacula used by hatchling turtles. In long periods, but some turtles apparently survived streams, rivers, and other lotic systems, oxygen brief inundation. Live C. picta have been recov- tension near hibernating turtles probably remains ered from nests beneath standing water in early high and there probably is little accumulation of spring (J. Iverson, personal communication). carbon dioxide. In lentic systems, dissolved gas Hydric conditions inside terrestrial hibernacula tensions tend to be more variable, as they are vary spatially and temporally, generally reflecting more strongly influenced by ambient temperature, patterns of precipitation and temperature, drai- rainfall, biochemical oxygen demand, and rates of nage characteristics of the surrounding soil, and exchange with the atmosphere. Notably, oxygen position of individual hatchlings (Ackerman, ’97). tension may fall precipitously after an ice layer The moisture level could increase when water forms on the water surface and in shallow, vapor condenses on the surface of hatchlings that eutrophic systems, persistent snow cover can are cooler than their surroundings (Thompson, virtually eliminate the oxygen supply (e.g., Craw- ’88). On the other hand, freezing of the soil ford, ’91b). These conditions may persist for weeks solution removes liquid water, creating a desiccat- or months, rendering the environment inhospita- ing environment (Spaans and Baker, ’96), and ble to all but highly anoxia-tolerant organisms precipitation falling as snow may not recharge soil (Ultsch, 2006; Bickler and Buck, 2007). moisture levels until melting occurs. No measurements of water vapor pressure Dynamics of water exchange in terrestrial inside terrestrial hibernacula used by hatchling hibernacula can have potentially significant im- turtles have been reported, although humidity plications on the physiology and cold hardiness of undoubtedly varies with temperature and precipi- overwintering hatchling turtles. Percolation of tation. Humidity could vary even within the rainwater through the soil column may flush confines of a single hibernaculum. For example, carbon dioxide from the nest chamber and deliver inside turtle nests, moisture diffuses in and out of oxygen to hatchlings (Packard, ’91). Moisture in different parts of the nest chamber in concert with soil also influences thermal conductivity and thus diel thermal cycles (Packard et al., ’85). Thermal rates at which turtles heat and cool, and the fluctuations necessarily alter saturation vapor frequency and duration of frost events. An pressure and chilling can cause moisture to abundance of environmental moisture can benefit condense on turtles. Continued cooling will lead hatchlings by reducing EWL. However, this con- to the freezing of this water, inoculation of turtle dition also promotes inoculative freezing of tur- tissues being the probable outcome (see below). tles, an outcome that may be advantageous Methods for sampling gas from inside turtle (Dinkelacker et al., 2004; Baker et al., 2006) or burrows (Ultsch and Anderson, ’86) and intact deleterious (Baker et al., 2003), depending on the experimental nests (Wallace et al., 2004), have species. By periodically sampling soil from C. picta been devised but whether or not gas exchange nests, Costanzo et al. (’98, 2004) found that soil poses problems for terrestrially overwintering moisture levels may peak in mid-winter, when hatchlings has yet to be investigated. Ventilation nest temperatures are lowest and the risk of of subterranean hibernacula occurs by diffusion of freezing is therefore the greatest. These studies respiratory gases through spaces among soil also showed that the moisture regime is linked to particles and their ultimate exchange with the edaphics and regional precipitation patterns, free atmosphere. Consequently, gas tensions are but also varies markedly among hibernacula at influenced by soil texture, porosity, particle size, a given nesting locale. Such variation may strongly water content, temperature, rainfall, and other impact turtle physiology, particularly with respect factors affecting diffusion (Congdon and Gibbons, to water balance, and may also govern risk ’90; Booth, ’98; Shams et al., 2005). Gas tensions of freezing and, possibly, mortality (Costanzo inside hibernacula are also influenced by respira- et al., 2004). tion of the hatchlings as well as microbes and

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 319 other resident organisms (Ackerman, ’77). Some concentrations of INA than do deeper strata research has shown that PO2 and PCO2 can vary (Vali, ’91) and therefore these agents probably among regions of a given nest as a function of the are of greatest importance to hatchlings that interplay between diffusion and metabolic activity overwinter within natal nests or other superficial of the occupants (Ralph et al., 2005). hibernacula. Turtle nests harbor INA of at least Gas tensions inside nests containing developing two major classes: inorganic particulates, which embryos often approximate those in ambient air, are active in the range À8toÀ61C, and small although hypoxia and hypercapnia conceivably (o10 nm), water-soluble, organic entities that can could develop under certain circumstances. For express activity at temperatures as high as À31C example, diffusion coefficients are especially low in (Costanzo et al., ’98, 2000a, 2001c, 2003, 2004). soils rich in clay (Shams et al., 2005) and organic Although both types can significantly constrain matter (van der Lee et al., ’99), probably because supercooling capacity of hatchling turtles, work in they have little air space and tend to retain water, this area has focused on the latter, which which has a relatively low gas conductance. Snow apparently have the greater impact on turtle cold deposition and freezing of the soil solution could hardiness. hamper diffusion even in porous soils. Because Water-soluble INA found in turtle nesting ground water usually contains little dissolved substrata are sensitive to autoclaving, heating, oxygen (Malard and Hervant, ’99), environmental and exposure to pH extremes and thus are hypoxia can result from inundation caused by a probably of organic, and possibly biogenic, origin rising water table (Christens and Bider, ’87; Kam, (Costanzo et al., 2000a). Some evidence suggests ’94). Thus, environmental constraints on gas that they are derived from or associated with ice- exchange potentially could influence whether nucleating microbes. Samples of soil collected hatchlings of a given species can successfully from inside C. picta nests exhibited an increase hibernate in terrestrial situations. in ice-nucleating activity when incubated in the cold, and a subsequent loss of activity on rewarm- ing (Costanzo et al., 2000a), possibly reflecting a Ice-nucleating agents thermally sensitive expression of genes encoding A host of terrestrial environments, including ice-nucleating proteins in microbes (e.g., Rogers ones in which hatchling turtles overwinter, harbor et al., ’87; Gurian-Sherman and Lindow, ’95). natural substances that catalyze the freezing of Preliminary studies of microbial colonies cultured supercooled water. These substances, commonly from these samples produced DNA sequences that called INA, act by orienting water molecules in a matched primers for ice-nucleating proteins geometric configuration favoring the formation of synthesized by ice-nucleating bacteria. However, an ice embryo, which in turn precipitates ice microscopic examination of extracts prepared from crystal growth (Lusena, ’55; Langham and Mason, these samples yielded no recognizable cellular ’58; Vali, ’95). INA are important elements of the structures, suggesting that the proteins had been winter microenvironment because, once ingested shed into the environment (e.g, Phelps et al., ’86). or otherwise incorporated into an organism’s INA are ubiquitous in terrestrial environments tissues, they sharply limit supercooling capacity and have been found at turtle nesting locations (Lee and Costanzo, ’98; Zachariassen and throughout North America (Costanzo et al., 2000a, Kristiansen, 2000). With hatchling turtles, 2001c, 2004; Baker et al., 2006). INA activity tends contamination with INA can reduce supercooling to be greater in wetter soils (Costanzo et al., 2004), capacity by 8–101C (Costanzo et al., 2000a, 2003). but apparently is not strongly associated with INA commonly found in soils include mineral abundance of organic matter (Costanzo et al., particulates, such as quartz and silicates (Mason 2000a). Activity levels in soil exhibit a seasonal and Maybank, ’58; Roberts and Hallett, ’68; pattern, such that maximal activity coincides with Kumai, ’76; Shen et al., ’77), organic crystalloids the coldest part of winter, presumably when their and organic/inorganic complexes formed during effects on turtle cold hardiness would be greatest decay (Power and Power, ’62; Fukuta, ’66; Vali, (Costanzo et al., ’98, 2000a, 2004). Soil inside nest ’91), and certain bacteria and fungi (Hirano and hibernacula can vary considerably with respect to Upper, ’95; Upper and Vali, ’95). Although the INA abundance and/or potency, perhaps owing to soils in which deeply burrowing turtles hibernate differences in incident precipitation and relief have not been examined, it is clear that upper soil (Costanzo et al., ’98, 2000a). In an extreme case, horizons generally support more kinds and greater ice-nucleating activity in soil collected from four

J. Exp. Zool. 320 J.P. COSTANZO ET AL.

C. picta nests in Ontario ranged from À11.2 to suggest that the latter substrate was preferred, À3.41C (Costanzo et al., 2000a). The environment but not necessarily that the former is inhospitable; inside these hibernacula can exhibit INA activity elsewhere, this species hibernates in sandy soils, that is high relative to that in adjacent soils, often with good success (Christiansen and Gall- suggesting that INA can become concentrated away, ’84; Costanzo et al., ’95b, 2004; Packard, inside turtle nests; however, this is not always ’97; Packard et al., ’97a). the case (Costanzo et al., 2004). Attributes of a particular soil, such as texture, porosity, and organic matter content, strongly Edaphics influence the abundance and distribution of ice within the soil matrix, thereby affecting the risk of Turtles nest in diverse substrata, both natural inoculative freezing (Costanzo et al., ’98). Inocu- and anthropogenic (e.g., DePari, ’88; Rozycki, ’98). lative freezing of hatchlings occurs readily in The influence of physical characteristics of soil on substrata, such as sands, which attenuate little hatching success and hatchling phenotype has moisture. Additionally, although porous soils may received extensive study (Packard et al., ’87; drain well, they also contain large voids that can Ratterman and Ackerman, ’89; Congdon and fill with ice to which hatchlings are intimately Gibbons, ’90), but relatively little attention has exposed. As a consequence, susceptibility to been paid to their impacts on the physiology and inoculative freezing (and possibly death) is far survival of turtles overwintering in terrestrial greater for animals exposed to sandy soil as situations. Attributes of soil, such as texture (i.e., compared with clayey soil, all else being equal fractional composition of sand, silt, and clay), (Packard and Packard, ’97; Costanzo et al., ’98, friability, and organic matter content, can influ- 2001c). In principle, therefore, variation in soil ence hatchlings through their variable effects on attributes on even a microgeographic scale can thermal, hydric, and gas regimes within hiberna- markedly influence winter survival rates. Costan- cula (Fig. 3). For example, although porous, well- zo et al. (’98) noted in reports of overwinter drained soils accelerate EWL, heavier soils, such mortality (Packard, ’97; Packard et al., ’97a) that as clays and clay loams, retain moisture, and thus viability was in fact lower for C. picta hibernating tend to heat and cool relatively slowly. Friable inside nests constructed in fine sand as compared soils are prone to infiltrating the nest cavity with loamy sand. Direct tests of this association (Congdon and Gibbons, ’90), placing hatchlings further bolster the conclusion that regional and at an increased risk of freezing through INA local variation in soil characteristics can impact contamination and inoculative freezing (see be- overwintering success in hatchling turtles (Cost- low). There is little evidence that certain soils are anzo et al., ’98, 2001c). strictly avoided for overwintering. DePari (’96) reported that New Jersey C. picta hatching in sandy soil tended to emerge from their nests in fall, whereas turtles hatching in clayey soil over- Spatial and temporal variability wintered inside their nests. His observation could Given that turtles are distributed over broad ecological and climatological regimes, environ- 15 mental conditions in hibernacula must vary 12 tremendously. In the north, ponds and marshes remain cold for extended periods with attendant 9 implications for duration of dormancy in aquatic 6 hibernators. In the same regions, the presence and

Moisture (%) 3 persistence of ice cover, which strongly influences dynamics of gas exchange, bears on the physiolo- 0 gical state of turtles. The microenvironment inside 0 5 10 15 20 25 terrestrial refugia also varies with edaphics and a Organic matter (mg/g) host of climatic factors, such as precipitation and Fig. 3. Relationship between organic content and moist- temperature. In cold climates, conditions that ure content (%, w/w) of soil in which hatchling painted turtles elevate freezing risk are principle threats to (Chrysemys picta bellii) overwintered at a nesting area in Garden Co., NE. Soil samples were collected in late January at winter survival, whereas in more temperate a depth of 10 cm, the typical depth of the nest chamber regions, conditions promoting dehydration and (hibernacula). (Adapted from Costanzo et al., ’98.) energy depletion are probably more critical.

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 321

Environmental conditions to which overwinter- in water balance, energy stores, cold hardiness, and ing hatchlings are exposed can vary even among ultimately winter survival (Costanzo et al., 2004). hibernacula at the same locale. For example, Hatchlings overwintering within the nest typi- temperatures inside natal nests can be mild or cally form a cluster, their heads near the top of the severe depending on the depth and structure of nest chamber (Breitenbach et al., ’84; Andreas and the hibernaculum, and on local edaphics, topogra- Paul, ’98; Packard et al., ’89; St. Clair and phy, and physiognomy; such microgeographic Gregory, ’90; DePari, ’96; Tucker, ’97; Nagle heterogeneity can account for inter-clutch varia- et al., 2000). Fragments of eggshell on the bility the in physiological state (Costanzo et al., chamber floor and walls may insulate turtles from 2004) as well as overwintering success (DePari, contact with soil. However, if the clutch is ’88; Packard et al., ’89, ’97a; Packard, ’97; Nagle relatively large, the hatchlings become distributed et al., 2000; Costanzo et al., 2004). This scenario vertically and laterally around the nest center, and raises some interesting questions about the fitness some workers (Costanzo et al., 2000a, 2001c, 2004; implications of nest-site selection. Maternal Nagle et al., 2000) have questioned whether choice of nesting site is an important, albeit little turtles near the periphery of the nest may be studied, factor shaping the life-history evolution of predisposed to inoculative freezing via contact oviparous, precocial organisms (Resetarits, ’96). with ice crystals and INA. This scenario pertains A female provides her offspring a measure of care only to nests that remain relatively free of through her choice of nesting site, which has infiltrating soil. However, as suggested by Cost- portentous implications for embryo survival anzo et al. (2004), regional variation in soil texture (Hughes and Brooks, 2006). Because the nesting and friability determines whether nests retain landscape is a complex, temporal mosaic of varying their form (Breitenbach et al., ’84; DePari, ’96) or physical conditions, choice of the oviposition site become infiltrated (Hartweg, ’44; Packard et al., can strongly impact embryonic development and ’89). Nevertheless, differential freezing risk may nest success (Congdon and Gibbons, ’90; Cagle explain the rather common instances of mixed et al., ’93; Kolbe and Janzen, 2001; Hughes survival of siblings overwintering in the same and Brooks, 2006) as well as hatchling phenotype nest (Packard et al., ’89, ’97a; Lindeman, ’91; (Schwarzkopf and Brooks, ’87; Packard and DePari, ’96; Weisrock and Janzen, ’99; Nagle Packard, ’88). The consequence of nest microen- et al., 2000; Costanzo et al., 2004; Carroll and vironment on hatchling performance or fitness, Ultsch, 2007). including cold tolerance (Du et al., 2006), has received considerable study (Miller et al., ’87; OVERWINTERING PHYSIOLOGY O’Steen, ’98; Finkler, ’99; Rhen and Lang, ’99; Tucker and Paukstis, ’99; Finkler et al., 2000; The diversity of environmental conditions con- Kolbe and Janzen, 2001; Steyermark and Spotila, fronting hatchling turtles raises myriad physiolo- 2001; Janzen and Morjan, 2002); however, the gical challenges to surviving winter. For aquatic influence of nest-site location on winter survival of hibernators, extreme cold and risk of freezing are turtles has received scant attention (Breitenbach not significant issues (Ultsch, ’89, 2006). On the et al., ’84; Weisrock and Janzen, ’99; Nagle et al., other hand, prolonged submergence presents 2000). The question of whether or not female potentially severe problems of gas exchange, turtles choose nest sites that are advantageous for metabolic acidosis, and iono-osmoregulatory im- hibernation within the nest remains open. balance. Terrestrial hibernators must cope with Environmental variability occurs on yet a finer dehydration and the attendant osmotic stress, scale: among regions of a given hibernaculum. and, if gas exchange is limited, hypoxia and Eggs occupying a three-dimensional nest chamber hypercapnia. Hatchlings have no access to food are exposed to a range of thermal (Ratterman and and therefore must endure a prolonged period of Ackerman, ’89; Tucker, ’99) and hydric (Legler, aphagia. In northern regions, cold exposure pre- ’54; Hotaling et al., ’85; Marco and Dı´az-Paniagua, sents a formidable challenge for the winter 2008) conditions, such that position within a survival of hatchlings overwintering terrestrially. clutch may modulate expression of phenotypic Energetics traits that ultimately affect fitness (Packard and Packard, 2001b). For hatchlings that defer emer- Generally, reptiles can endure remarkably long gence from the nest until spring, such hetero- periods of aphagia (Frye, ’73). Hatchling turtles geneity could explain instances of sibling variation are provisioned with large stores of yolk that can

J. Exp. Zool. 322 J.P. COSTANZO ET AL. fuel metabolic activities at times when food is not Ecological considerations available (Ewert, ’85; Congdon and Gibbons, ’90; Deeming, 2004b). This is particularly important In the north, hatchlings must subsist on en- for species whose hatchlings remain inside nests dogenous energy stores for extended periods. during winter or hibernate in the soil column, as Fortunately, their hibernacula usually are quite feeding is not possible. Hatchlings emerging from cold and the depressive effect on metabolism can, their nests in autumn presumably have an to a large extent, offset the effects of prolonged opportunity to feed before winter, although the aphagia. In milder climes, turtles probably go extent to which they do so remains uncertain. without feeding for shorter periods; however, DePari (’88) found that aquatic invertebrate prey owing to higher metabolic rates, it is unclear were as abundant during the period of autumnal whether they fare any better energetically. This emergence of hatchling C. picta as during the interesting conundrum has been considered for period of spring emergence. Hatchling G. insculp- other ectotherms that also exhibit a wide geogra- ta have been observed feeding before hibernation phical range (Costanzo, ’88; Irwin and Lee, 2000), (Tuttle and Carroll, 2005), but it is unknown how but has not been resolved for hatchling turtles. It common this behavior is among other species. If seems clear, however, that energy use is closely feeding occurs, it will cease as the weather cools linked to prevailing ambient temperature. In a before hibernation; indeed, food intake is probably 2-yr study of hatchling C. picta in Indiana and negligible at temperatures below 151C (Cagle, ’50; Nebraska, Costanzo et al. (2004) found that Sexton, ’59; Ernst, ’71). Contrary to the experi- utilization of stored lipids and other energy- ence of one of us (G. Ultsch), neonatal yielding substrates was greater in the warmer C. serpentina did not eat when offered food in winter, apparently reflecting a higher metabolic the laboratory (Finkler et al., 2002, 2004); some demand. In the same study, more energy was authors have surmised that these hatchlings defer expended by hatchlings of the more temperate feeding until the following spring (Sims et al., Indiana population, probably because winter 2001). Thus, as both embryo and neonate, turtles temperatures were milder. In another study, may rely on energy provisioned by the mother for hatchling T. scripta had less residual yolk follow- an extended period (Gibbons and Nelson, ’78). ing a mild winter than after a colder winter, this Is energy limiting in the winter survival of result portending the ill effects of global warming hatchling turtles? Neonates usually have abun- (Willette et al., 2005). Intuitively, it seems ad- dant energy reserves, primarily in the form of yolk vantageous for turtles inhabiting relatively warm lipids and proteins (Wilhoft, ’86; Congdon and climes to provision their eggs with more nutrients Gibbons, ’90; Janzen et al., ’90). The quantity and in order that a large surplus be available for use quality of residual yolk is determined by the after hatching, assuming that feeding before amount of nutrients maternally invested in each winter is not possible. Finkler et al. (2004) particular egg, as well as the thermal and hydric reported that solid matter, a proxy for energetic conditions prevailing during egg incubation reserves, in eggs of C. serpentina exhibits a (Deeming, 2004b). More energy is allocated to climatological cline, with populations experiencing eggs in a female’s first clutch than in her second higher temperatures before hibernation provision- (Harms et al., 2005), but whether hatchlings ing eggs with more solids. Energy provisioning of resulting from the latter are disadvantaged for hatchlings can vary markedly among conspecific winter survival has not been studied. Residual populations (Costanzo et al., 2004; Harms et al., yolk is a critical fuel for post-hatching activities, 2005), perhaps because energy demand varies with such as nest emergence and migration to other local environmental conditions. An alternative habitats. It is also critical to meeting basal strategy of energy conservation in relatively warm metabolic needs during hibernation because most climes, which has yet to be explored in hatchling species probably are unable to (or simply do not) turtles, is to invoke behavioral and physiological feed before winter. If the yolk supply is ample, mechanisms for reducing metabolic demand some may be used for somatic growth, and body (Ewert, ’91). mass will increase during winter (Tucker et al., Another intriguing, albeit unresolved, question ’98; Filoramo and Janzen, ’99). However, if the pertains to the energetic consequences of hiber- supply is inadequate, other endogenous energy nating on land vs. under water. Terrestrial reserves are catabolized and the hatchling will hibernators presumably are exposed to relatively tend to lose body mass (Finkler et al., 2002). low prevailing temperatures, and, if gas exchange

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 323 is limited, low ambient gas tension could enhance 0.8 Carcass the hypometabolic effect of cold (Burggren and a a b Pinder, ’91). Still, avenues for rapid energy loss 0.6 exist. In the event of tissue freezing, liver glycogen is rapidly converted to glucose (Churchill and 0.4 Storey, ’91; Hemmings and Storey, 2000), which, if not reconverted after thawing, could be lost from the body via renal excretion (although we know Dry mass (g) 0.2 little about the amount of renal excretion that occurs during winter). 0 Underwater hibernation presents a different suite of energetic considerations. In some amphi- 80 Liver bians and reptiles, apnea initiates responses that a facilitate energy conservation, including bradycar- 60 a,b dia, peripheral vasoconstriction and oxygen ra- b tioning, and a decrease in metabolic rate (Shelton 40 and Boutilier, ’82; Herbert and Jackson, ’85; Donohoe and Boutilier, ’98; Ultsch et al., 2004). 20 Hatchling turtles might drastically reduce energy use if they also exhibit the so-called ‘‘diving response,’’ although whether or not they do so is 0 apparently unknown (Burggren and Pinder, ’91). 25 Yolk Reese et al. (2004b) provided evidence that a metabolism is reduced in hatchlings submerged 20 a in anoxic water as compared with hatchlings exposed to gaseous anoxia. In any case, a portion 15 of the energetic savings probably is offset by the 10 increased cost of managing salt and water balance a necessitated by chronic submergence. Further- Dry mass (mg)5 Dry mass (mg) more, submergence in hypoxic water would necessitate rapid depletion of glycogen reserves, 0 particularly those of the liver, to mobilize glucose Initial Terrestrial Aquatic to the tissues (Warren et al., 2006). Species whose hatchlings overwinter inside the Fig. 4. Variation in energy consumption in hatchling painted turtles (Chrysemys picta bellii) during simulated hibernation on nest reportedly (Congdon et al., ’83c; Congdon and land and in water, as indicated by changes in dry mass of the Gibbons, ’85, ’90; Long, ’86; Rowe et al., ’95; Nagle carcass, liver, and residual yolk sac. Eggs were collected from et al., ’98; Costanzo et al., 2000b) have greater gravid females at a nesting area in Garden Co., NE, and incubated energy reserves than species whose hatchlings in the laboratory at  291C until hatching in August. Neonates ultimately overwinter outside the nest (presum- (n 5 24) were kept at 211C on damp vermiculite until 1 October and then assigned to one of three groups. One group (initial)was ably aquatically), the favored interpretation being immediately euthanized and assayed; another group (terrestrial) that terrestrial hibernation is energetically more was placed in a simulated nest constructed of soil collected from costly. Unfortunately, there are no published the field site; another group (aquatic) was transferred to a plastic comparisons of energy budgets for hatchlings tank containing aerated water (8 cm deep) and rocks for shelter hibernating terrestrially and aquatically. Preli- and basking. Aquatic turtles were first exposed to 151Cand indirect lighting (12L:12D) provided by a full-spectrum lamp; minary work has shown that, during laboratory ambient conditions were changed to 101C, 10L:14D on 1 hibernation at 41C, hatchling C. picta lost more November, and to 41C, 10L:14D on 1 December. In mid- yolk, dry liver mass, and dry carcass mass if December, ice cover was simulated by reducing light intensity submerged in normoxic water than if kept inside and installing a horizontal screen 3 cm below the water surface, simulated nests (Fig. 4). The accelerated energy thus preventing hatchlings from breathing air; these conditions were maintained thenceforth. Terrestrial turtles were exposed to consumption in the aquatic hibernators, which the same thermal conditions, but kept in darkness. Hatchlings remained quiescent throughout the experiment, from both treatments were euthanized and assayed in mid-April. probably resulted from efforts to maintain osmor- Means7SE (n 5 8) identified with different letters were statisti- egulatory balance, although water gain and cally distinguishable (ANOVA/Bonferroni; Po0.05). (J. Costanzo, reduced plasma osmolality suggested that such B. Dishong, and T. Muir, unpublished data.)

J. Exp. Zool. 324 J.P. COSTANZO ET AL. efforts were incompletely successful. This experi- Much of the change in dry body mass undoubt- ment also showed that hatchling C. picta are edly reflects consumption of the residual yolk, the capable of surviving winter in cold, normoxic primary source of energy for overwintering water without access to air; in contrast, hatchlings hatchlings. In one laboratory study (Costanzo of many (most?) species do not long tolerate anoxia et al., 2003), yolk mass in C. picta on hatching (Reese et al., 2004b; Dinkelacker et al., 2005a,b). decreased by 75–90% during a 5.5-month period of simulated hibernation. The remaining yolk weighed about the same as that measured in Physiological considerations hatchlings collected in the field in spring; thus, Given the length of time during which hatchling energy use in the experimental animals probably turtles remain in negative energy balance, one was reflective of the natural situation. Similarly, might expect them to lose much body mass during little yolk remains in T. scripta on emergence from winter. Such a loss could weaken the and natal nest hibernacula in Jersey County, west- adversely affect its fitness and survival. Field central Illinois (Tucker et al., ’98). On the other investigations suggest that success in migrating hand, decrements in body mass and nutrient from the nest chamber to the relative safety of reserves can vary considerably among individual aquatic habitats is greater in heavier hatchlings hibernacula and can be negligible at least occa- (Janzen, ’93; Tucker and Paukstis, ’99). Changes sionally (Costanzo et al., 2004). in body mass of hibernating hatchlings obviously Lipids are an important energy source in many are linked to conditions prevailing during winter, hibernating reptiles (Derickson, ’76; Gregory, ’82). although some evidence suggests that conditions This may also be the case for the hatchlings of during embryonic development are also impor- some turtle species (Wilhoft et al., ’83), although tant. For example, T. scripta developing in caution is needed in making this generalization. relatively dry nests tend to lose more mass in During a relatively mild winter, hatchling C. picta hibernation than turtles hatching in wetter nests catabolized only 25% of their total lipid reserves (Tucker and Paukstis, ’99). and the resulting energy (1.22 kJ) accounted for Little is known about the dynamics of body mass less than one-half of the total energy (2.70 kJ) during winter, except that limited information is produced from all substrates in the average-sized available from laboratory studies. Only 15% of turtle (Costanzo et al., 2004). Furthermore, in the body mass is lost during winter in hatchling G. same investigation, turtles used lipids sparingly, if geographica (Nagle et al., 2004) and T. scripta at all, during a subsequent, colder winter. This (Congdon and Gibbons, ’90). This amount gener- result could reflect severe hypometabolism and/or ally concurs with the findings of Tucker and an inability to catabolize lipid at low temperature, Paukstis (’99), who reported that T. scripta particularly if cells become hypoxic during somatic hatchlings lost 14–23% of their initial body mass freezing (Storey, ’90b) and supercooling (Hartley during simulated terrestrial hibernation. Simi- et al., 2000; Costanzo et al., 2001a). Mechanisms of larly, body mass was reduced 15–25% in hatchling lipid sparing probably help to ensure the avail- C. picta kept under comparable conditions (Cost- ability of a rich energy reserve to support post- anzo et al., 2003), but in neither study did the emergence activities (Parker and Brown, ’80; authors discover the extent to which the change Gregory, ’82; Costanzo, ’85). reflected water loss, rather than nutrient con- Other energy-yielding substrates, such as glyco- sumption. In contrast to these findings, one field gen and proteins, apparently are important in the study, in which workers tracked the mean body energy budget of overwintering hatchlings. Eggs mass of hatchling C. picta over a 6-month winter, of C. picta are provisioned with 2–3 times as much showed that the decrease was slight, suggesting protein as lipid (Harms et al., 2005), and blood that energy use was minimal (Costanzo et al., urea levels in hatchlings increase during fall and 2004). Similarly, Baker et al. (2006) reported that winter, suggesting that they actively catabolize the mean body mass of hatchling M. terrapin proteins (Costanzo et al., 2004). Indeed, during emerging from nests in spring was nearly identical simulated hibernation, hatchling C. picta (but not to that of hatchlings emerging in autumn. In some C. serpentina) maintain a rich pool of amino acids cases, hatchling turtles overwintering in the (Costanzo et al., 2000b), which can be metabolized laboratory lose weight sparingly, if at all (Tucker even by hypoxic tissues (van Waarde, ’88), the et al., ’98; Filoramo and Janzen, ’99; Costanzo resulting carbon being converted to glucose via et al., 2000b). gluconeogenesis. This pathway may also be active

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 325

15 during desiccation, as evidenced by the strong c correlation between urea and glucose levels in some turtles (Costanzo et al., 2004). Amino acids can serve as ‘‘compatible osmolytes,’’ benign 10 Q =7.5 solutes that protect cells against dehydration, 10 and some, such as proline, offer protection against both freezing and desiccation stresses (Takagi 5 b

et al., 2000). consumption (µl/g/h)

2 a

Turtles must limit energy expenditure during O hibernation if they are to adequately fuel beha- 0 41015 vioral activities in spring. Little is known about Temperature (˚C) mechanisms of metabolic regulation in hatchling turtles, although both physiologic and genetic Fig. 5. Metabolism, estimated from oxygen consumption factors are probably important. Resting metabo- rate, as a function of thermal acclimation and temperature in hatchling painted turtles (Chrysemys picta bellii). Eggs were lism in hatchlings of some species is influenced by collected from gravid females at a nesting area in Garden Co., environmental conditions prevailing during em- NE, and incubated in the laboratory at 291C until hatching bryonic development (O’Steen, ’98; O’Steen and in August. Neonates (n 5 24) were kept in darkness at 211Con Janzen, ’99; Steyermark and Spotila, 2000). In damp vermiculite until 1 October, when ambient temperature northern regions, hatchlings undoubtedly benefit was reduced to 151C. One group (n 5 8) remained at 151C thereafter; other hatchlings (n 5 16) were transferred on 1 from extended periods of cold and its effect on November to an incubator set at 101C. Half of these turtles biochemical processes; indeed, metabolism at remained at 101C; the others were transferred on 1 December subzero body temperatures is barely measurable to an incubator set at 41C. Closed-system respirometry trials (Halpern and Lowe, ’68; Voituron et al., 2002b). were conducted in mid-winter using an oxygen analyzer (S-3A/ Additionally, in hatchling C. picta, chilling to near- I, AEI Technologies, Pittsburgh, PA); data were corrected to STPD. Means7SE (n 5 8) identified with different letters freezing temperatures cause a marked slowing of were statistically distinguishable (ANOVA/Bonferroni; energetically expensive functions, such as cardiac Po0.05). (T. Muir, B. Dishong, and J. Costanzo, unpublished contraction (Birchard and Packard, ’97) and data.) pulmonary ventilation (Larson, 2004), perhaps owing to an increased reliance on anaerobiosis or from 0.48 mLgÀ1 hÀ1 at 41C to 0.15 mLgÀ1 hÀ1 at metabolic reprogramming. Even a modest de- 01C(Q10 5 2.6). When exposed to subzero tem- crease in ventilation frequency would be beneficial peratures, hatchlings had extremely low levels of because, at least in older individuals, the majority aerobic metabolism; for example, the rate of CO2 of energy expended by inactive turtles is used for release at À81C was only 0.02 mLgÀ1 hÀ1. This pulmonary respiration (Kinney and White, ’77). severe drop undoubtedly stems from the down- Metabolic processes in overwintering hatchlings regulation of aerobic metabolic function, but may have not been studied in great detail. Larson also reflect an increased reliance on anaerobic (2004) investigated effects of thermal history and energy production with cooling into the subzero acute exposure to various temperatures on temperature range (see Voituron et al., 2002b). C. picta, using CO2 release as a proxy for metabolic The thermal sensitivity of metabolism in hatchl- rate in a group of hatchlings that he successively ing C. picta has also been investigated using acclimated to 15, 10, and 41C. Metabolic rates at closed-system respirometry (T. Muir, B. Dishong, 101C were essentially the same regardless of and J. Costanzo, unpublished data). In this whether the hatchlings had been acclimated to project, metabolism of hatchlings acclimated and 15 or 101C. However, the rate measured at 41C tested at 15, 10, and 41C varied exponentially with was significantly lower after turtles were accli- body temperature, although the Q10 for this effect mated to 41C as compared with 151C, suggesting was considerably higher than that reported by that inverse compensation is a strategy of energy Larson (2004); indeed, oxygen use at 41C was only conservation in dormant hatchlings. Additional 9% of that measured at 151C (Fig. 5). A second experimentation with 41C-acclimated hatchlings aim of this study was to determine whether elucidated the effect of varying temperature on supercooling or freeze tolerance was the more metabolism. The mean rate of CO2 release energetically favorable strategy of cold tolerance declined with temperature in an exponential in hatchling C. picta. For hatchlings exposed manner, decreasing from 1.20 mLgÀ1 hÀ1 at 151C to À2.51C, the mean rate of oxygen consumption À1 À1 to 0.61 mLg h at 101C(Q10 5 3.9), but only was about three times greater in supercooled

J. Exp. Zool. 326 J.P. COSTANZO ET AL.

(1.33 mLgÀ1 hÀ1) vs. frozen (0.42 mLgÀ1 hÀ1) ani- ’77; Ralph et al., 2005). Although there are far mals. Interestingly, during recovery at 41C, pre- fewer propagules in the nests of freshwater viously frozen C. picta consumed oxygen at a turtles, these conditions might develop inside greater rate (3.50 mLgÀ1 hÀ1) than turtles that had nests harboring overwintering hatchlings, espe- been supercooled (2.11 mLgÀ1 hÀ1), perhaps be- cially if gas conductance is impeded by high clay cause there is a latent energetic cost to freezing. In content, water saturation, or freezing of the soil. other freeze-tolerant , repayment of an Low temperatures in winter obviously help reduce ‘‘oxygen debt’’ after thawing is sometimes ob- oxygen demand and CO2 production; however, served (Layne, 2000), but may be absent, depend- metabolism is relatively high in autumn and again ing on severity of the freezing exposure (Voituron in spring, just before emergence from hibernacula. et al., 2002b). Hypometabolism at subzero tem- Field studies are needed to determine whether or peratures could provide significant energetic sav- not gas exchange is a challenge to terrestrially ings. This is especially true of frozen animals in hibernating hatchlings. which the effect could result from enzymatic It is widely appreciated that, among vertebrates, perturbation caused by the attendant rise in ion anoxia tolerance is particularly well developed in concentrations (Kanwisher, ’59). Moreover, a fall turtles, although it certainly varies taxonomically, in tissue pH brought on by chilling, and perhaps tending to be greater in northerly distributed accelerated by lactate accumulation, could create a forms that naturally encounter periods of hypoxia ‘‘metabolic context’’ that favors development of during long winters and persistent ice cover metabolic depression (Storey and Storey, ’90). (Ultsch, ’89; Reese et al., 2004a; Ultsch and Reese, 2008). Research on this phenomenon has largely focused on survival and physiological responses in Respiration adult animals, which, except for terrestrial and Exchange of respiratory gases is not problematic estivating species, invariably hibernate aquati- for hatchling turtles overwintering in ponds, cally. Physiological adaptations promoting survi- streams, and other aquatic systems so long as val during hypoxia include a depression of they can breathe air. However, pulmonary venti- metabolism brought on by low ambient tempera- lation could be difficult for cold, lethargic turtles, ture and reduced oxygen availability; reliance on and, in temperate regions, is often prevented by anaerobic pathways for energy production and ice formation. In these situations, diffusion across tolerance of the associated lactacidosis; and an epithelia of the skin, buccopharynx, and cloaca extensive mineral buffering system to manage becomes the only avenue for gas exchange (for a acid–base disturbance (Lutz and Storey, ’97; review, see Ultsch, ’89). Physiological mechanisms Jackson, 2000a; Bickler and Buck, 2007). Anoxia- sustaining aerobic metabolism during submer- tolerant turtles also exhibit well-developed anti- gence in oxygenated water include hypometabo- oxidant defenses that minimize or prevent damage lism, a high efficacy of cutaneous gas exchange, by reactive oxygen species (ROS) during reoxy- and adaptations of the blood oxygen transport genation (Willmore and Storey, ’97; Jackson, 2002; system (Ultsch and Jackson, ’82; Maginniss et al., Milton et al., 2007). ’83, 2004; Saunders and Patel, ’98). If the Survival and physiological responses to anoxia in refugium becomes anoxic, which can occur in hatchling turtles have been examined only re- regions where ice cover persists, hatchling survi- cently (Reese et al., 2004b; Dinkelacker et al., val will then depend on a host of biochemical and 2005a,b). The consensus of these studies is that physiological responses to low tissue PO2 (Jack- anoxia tolerance varies taxonomically, as it does son, 2002, 2004). with adults, and is severely limited relative to that Whether or not gas exchange is problematic for of conspecific adults. For instance, although adult hatchlings overwintering on land is unknown, C. picta can survive continuous submergence in although this scenario remains a possibility, cold, anoxic water for periods lasting 4–5 months particularly for animals hibernating deep in the (Ultsch and Jackson, ’82), survival of hatchling soil column or in close proximity to other hatchl- C. picta under similar conditions is limited to 40 d ings, as is the case with species that hibernate (Reese et al., 2004b). Survival in gaseous anoxia inside their nest. Hypoxia and hypercapnia can (N2 exposure) is even more restricted (Reese et al., develop in nests when respiratory 2004b), suggesting that submergence invokes demands of late-stage embryos exceed the diffu- responses that delay onset of physiological pertur- sive capacity of even highly porous soil (Ackerman, bation (although greater activity of hatchlings in

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 327

N2 atmosphere is another possibility). Neverthe- ’96). Because the vapor pressure of ice is lower less, for the few species that have been examined, than that of unfrozen body fluids, even super- survival times in cold, anoxic submergence are cooled organisms will lose water to a frozen only on the order of 1–6 weeks (Reese et al., 2004b; environment (Forge and MacGuidwin, ’92; Lund- Dinkelacker et al., 2005a). The comparatively low heim and Zachariassen, ’93; Holmstrup et al., tolerance relative to adults apparently reflects an 2002). Behavioral, physiological, and morphologi- incomplete ossification of the shell and an inade- cal mechanisms of hydro-osmotic homeostasis are quate buffering system (Reese et al., 2004b; paramount to winter survival of hatchling turtles, Dinkelacker et al., 2005a,b; Ultsch, 2006). Varia- as in other ectotherms (Ring and Danks, ’94; tion in shell composition and buffering capacity Block, ’96; Danks, 2000; Williams et al., 2002). explains some of the taxonomic differences in anoxia tolerance among freshwater turtles (Jack- Water balance son et al., 2007). Hypercapnia, which potentially accompanies The water budget of hatchling turtles in terres- hypoxia, is a lesser-studied problem that could trial hibernation is fairly simplistic. They pre- beset hatchlings crowded inside communal hiber- sumably neither drink nor obtain water through nacula with barriers to gas exchange. Lower feeding during winter, although this supposition vertebrates commonly respond to increased requires empirical validation. Unlike amphibians, PCO2 with a reduction in aerobic metabolism that hatchling turtles reportedly cannot reabsorb may be precipitated by increased acidification moisture from even a saturated substratum (Storey and Storey, ’90). Hypercapnia stimulates through nonoral routes (Costanzo et al., 2001b). ventilation in C. picta, although the effect is Therefore, production of metabolic water may be largely negated if the inspired gas is also hypoxic the sole hydration source (Costanzo et al., 2004). (Milsom and Chan, ’86). Adult aquatic turtles can On the other hand, water readily leaves the body become hypercapnic during the initial weeks of via multiple routes. Some is lost through normal cold submergence, but this condition does not functions of the excretory and digestive systems, persist in the long term in either normoxic or although these debits probably are unimportant in anoxic water. If there is an acidosis associated with the overall water budget. Turtles apparently submergence, it is initially a combined respiratory defecate within a week or two after hatching. and metabolic acidosis, but eventually becomes a Elimination of fecal matter rids the body of potent pure metabolic acidosis, often partially offset by a INA derived from resorbed yolk (Costanzo et al., slight respiratory alkalosis resulting from a fall in 2003), as well as bits of eggshell and soil that are PCO2 relative to air-breathing turtles (Jackson consumed during or shortly after hatching (for et al., 2000; Reese et al., 2001, 2002, 2003). reasons that remain unclear; see Packard et al., Retention of CO2 by hatchlings overwintering 2001; Costanzo et al., 2003; Packard and Packard, under water has not been investigated; however, 2006). Micturition is a potentially significant given their small size and favorable geometry for avenue for water loss during hibernation, although cutaneous exchange, hypercapnia is unlikely to little is known about excretory function in hatchl- develop unless the PCO2 of the surrounding water ing turtles. Changes in blood urea levels during becomes elevated. simulated terrestrial hibernation imply that hatchl- ing C. picta are anuric or at least refrain from urinating, whereas hatchling C. serpentina, which Osmoregulation typically overwinter under water, apparently do Water availability obviously is not a problem for void urine (Costanzo et al., 2000b). Generally, urine turtles that hibernate in streams and ponds. production ceases in turtles during dehydration Rather, excessive water influx is the primary (Dantzler and Schmidt-Nielsen, ’66). osmoregulatory challenge confronting aquatic hi- Respiration accounts for a small proportion of bernators (Trobec and Stanley, ’71). Water deficit water efflux when body surfaces are exposed to the may pose serious difficulties for hatchlings over- atmosphere, but this mode has greater signifi- wintering in brackish water or on land. Terrestrial cance when the animal is buried in soil. Manage- hibernators are subjected to dehydration when ment of respiratory water loss in hatchling turtles environmental water potential is low, such as has received little study. Ventilation rates gener- when precipitation falls as snow rather than rain, ally are sensitive to body temperature, although or the soil solution freezes (Spaans and Baker, some evidence suggests that these variables can

J. Exp. Zool. 328 J.P. COSTANZO ET AL. become uncoupled during cooling. For instance, EWL is also governed by permeability character- hatchling C. picta show a disproportionately large istics of the integument (Spotila and Berman, ’76); reduction in ventilatory frequency during cooling these largely reflect epidermal keratinization at subzero temperatures (Larson, 2004), perhaps (Seidel and Reynolds, ’80) and/or lipid content reflecting an overall reduction in cellular metabo- (Roberts and Lillywhite, ’80; Lillywhite and lism as well as increased reliance on anaerobiosis Maderson, ’88). Cutaneous permeability in turtles (Hartley et al., 2000; Costanzo et al., 2001a). In has not been extensively studied but probably addition, the ventilatory pattern is transformed as varies by taxon, life stage, and even acclimatiza- body temperature falls and, importantly, breath- tion. Hatchling C. picta are highly resistant to holding can be an effective means to reduce water EWL (Costanzo et al., 2001b), this trait deriving loss (Schmidt-Nielsen and Bentley, ’66; Schmidt- from a dense epidermal layer of lipid deposits that Nielsen et al., ’66). According to Larson (2004), apparently is lacking in hatchlings of several other hatchling C. picta exhibited relatively cyclic and species (Willard et al., 2000). During dormancy, rhythmic ventilation at 151C, but as the tempera- K. flavescens exhibits reduced rates of EWL, ture was reduced to –81C, turtles began breathing apparently owing to an increase in skin keratini- in a discontinuous, noncyclic manner. This shift zation (Chilian, ’76). Among turtles, dehydration presumably allowed metabolic adjustments as tends to suppress EWL (Chessman, ’84), as does oxygen demand decreased with cooling; probably cold (Olson, ’89). it also conserved both water and energy (Seidel, Given the strong influence of morphological ’78; Glass et al., ’79). attributes and physiology on susceptibility to Evaporation from cutaneous (as opposed to water loss, it is perhaps not surprising that respiratory) surfaces typically accounts for the chelonians vary considerably in this characteristic majority of EWL in reptiles (Bentley and Schmidt- (Bogert and Cowles, ’47; Bentley and Schmidt- Nielsen, ’66; Chessman, ’84). Among turtles, Nielsen, ’66; Ernst, ’68; Chessman, ’84; Olson, carapacial and plastral scutes of the shell better ’89). Mechanisms of water conservation probably resist water loss as compared with the softer are of great significance to species that overwinter epidermis of the head, neck, and limbs (Rose, ’69; terrestrially or in other desiccating habitats. Spotila and Berman, ’76). Consequently, EWL is These include behaviors such as retracting the influenced by the cutaneous surface area, which, head and limbs into the shell, and in communal in turn, is governed by shell morphology and body hibernacula, forming a cluster with other hatchl- size (Wygoda and Chmura, ’90; Stone et al., ’92; ings. Measurements of EWL in hatchling turtles Stone and Iverson ’99). All else being equal, larger are available for few species (Finkler, ’99; Costan- turtles have lower mass-adjusted rates of EWL zo et al., 2001b; Dinkelacker et al., 2004), although owing to a lower surface-to-volume ratio that the data indicate that terrestrial hibernators favors water conservation. However, no significant generally exhibit lower (mass-standardized) rates association between body size and hibernation of EWL than aquatic hibernators (Table 3). Of habitat has been found among hatchling turtles particular interest is E. blandingii, which exhibits (Costanzo et al., 2001b). an intermediate susceptibility to water loss.

TABLE 3. Rates of evaporative water loss (EWL) at 41C in winter-acclimatized hatchling turtles

Taxon Hibernation habitat EWL (mg gÀ1 dÀ1) Body mass (g) Reference

Terrapene ornata Terrestrial 0.9 10.7 1 Chrysemys picta bellii Terrestrial 1.9 4.0 1 Graptemys geographica Terrestrial 2.0 5.4 2 Kinosternon flavescens Terrestrial 2.4 3.1 1 Trachemys scripta Terrestrial 2.6 7.7 1 C. p. dorsalis Terrestrial 3.6 5.0 1 Emydoidea blandingii Unknown, possibly terrestrial 4.1 6.8 3 Chelydra serpentina Aquatic 6.3 8.7 1 Sternotherus odoratus Aquatic 9.6 3.2 1 S. carinatus Aquatic 11.4 2.8 1

Values are means (N 5 3–10) reported in the source document. References: (1) Costanzo et al. (2001b); (2) Baker et al. (2003); (3) Dinkelacker et al. (2004).

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 329

Although the precise hibernation habits of its universally, and K1 also increased in C. serpenti- hatchlings are unknown, it is possible that they na. These changes apparently reflected a mobili- avoid excessive dehydration by hibernating in wet zation of mineral from the shell, and, together soil in marshes or pond margins (Dinkelacker with a marked accumulation of lactate, contrib- et al., 2004). Overall, however, the association uted to an increase in plasma osmolality. between EWL and hibernation habitat suggests Some evidence suggests that osmoregulation that dehydration resistance may be an important during continuous submergence is stressful, trait in terrestrial hibernation (Costanzo et al., even if the water remains normoxic. Hatchling 2001b). C. serpentina, which commonly hibernate under water, survived winter (6.5 months) while held at 41C in water-filled jars (Sims et al., 2001). Over Hydric and osmotic stress the course of the experiment, plasma osmolal- Water influx is the salient osmoregulatory ity dropped precipitously, from 340 to 290 challenge besetting turtles that spend long periods mosmol kgÀ1, and the hatchlings lost 5% of their submerged in cold water (Trobec and Stanley, initial body mass. Lacking data on body water ’71). This phenomenon stems from a continuous concentration, the authors surmised that in- osmotic uptake of water and is compounded by an creased hydration caused the drop in osmolality; uncompensated salt loss and marked reduction in however, an uncompensated salt loss could also cardiac output, which allows fluid to pool in have been involved. In contrast, hatchling C. picta underperfused capillary beds (Semple et al., ’70; succumbed when kept in water-filled jars at 41C Stitt and Semple, ’71). Another complicating for 70 d (Packard and Packard, 2001c). Unfortu- factor is the inhibiting effect of cold on ion uptake nately, there was no attempt to control or measure (Dunson, ’67) and urine production (Crawford, the ambient PO2, although the authors suspected ’91a). Although limited renal function in sub- iono-osmoregulatory imbalance was a factor in the merged hatchlings probably curbs excretion of turtles’ demise. In a more elaborate experiment, salts, it also contributes to water retention. hatchling C. picta (n 5 8) were kept over winter in Few studies of hydro-osmotic balance in sub- aquaria containing cold, normoxic water, and merged hatchling turtles have been published. shelters (J. Costanzo, T. Muir, and B. Dishong, Turtles actively exchange water and the major unpublished data). Denied access to air from ions (except potassium) with their environment, December–February (simulating ice cover), none although fluxes in hatchlings apparently differ in died during the 5.5-mo experiment; however, in some respects from those in adults (Dunson, ’67; spring all hatchlings appeared edematous and in Dunson and Seidel, ’86). Limited study shows that poor health, and three were unable to right the hatchlings of at least some species actively themselves. Body water concentration in these osmoregulate during hibernation in water. During animals (84.170.6% wet mass) exceeded that in normoxic submergence of hatchling C. picta, shell control turtles (80.670.9%; n 5 8) held at the concentrations of Na1 and K1 rose two-fold to same temperature and for the same duration in three-fold, but this change was not observed in simulated terrestrial hibernation; accordingly, hatchling G. geographica or C. serpentina (Reese their plasma osmolality was also lower (23774 et al., 2004b); the significance of these findings vs. 35175 mosmol kgÀ1, respectively). Differences remain unclear. Dinkelacker et al. (2005a) studied in dry carcass mass between the groups indicated hydro-osmotic responses in hatchlings of three that the submerged turtles had consumed more species of turtles during simulated hibernation in energy than the terrestrial hibernators, probably normoxic and severely hypoxic water. E. blandin- owing to higher costs of managing salts and water. gii, C. serpentina, and A. spinifera varied in their Little study has been devoted to water relations responses, although some generalized patterns during the months following hatching; therefore, emerged. During 75–77 d of normoxic submer- the extent to which neonates must cope with gence, plasma ion concentrations generally re- hydric and osmotic stress in terrestrial hiberna- mained static (Na1,K1, and Ca11) or declined tion is uncertain. Indeed, some evidence indicates (Mg11) and, although body water content did not that terrestrial hibernators can remain fully change, plasma osmolality tended to decrease. hydrated and in osmotic homeostasis throughout Responses of hatchlings in hypoxic submergence, winter. In fact, under favorable environmental which was tolerated for only 6–19 d, were more conditions, some species actually gain weight that dramatic. Plasma Ca11 and Mg11 increased presumably reflects increased hydration (Filoramo

J. Exp. Zool. 330 J.P. COSTANZO ET AL. and Janzen, ’99; Costanzo et al., 2000b). Net water hatchlings embarking on a nest-to-water journey gain in terrestrial hibernators could be promoted in a partially dehydrated state are less successful by a combination of low EWL and a favorable in reaching their permanent habitat (Tucker and morphology (Costanzo et al., 2001b), the effect of Paukstis, ’99). cold on vapor pressure (Olson, ’89), and produc- tion of metabolic water via b-oxidation of fatty Thermal physiology acids. The relatively large lipid and yolk reserves in hatchlings that hibernate terrestrially, as Hatchlings of temperate-region turtles are in- compared with those in aquatic hibernators herently tolerant of cold, irrespective of whether (Congdon et al., ’83c; Congdon and Gibbons, ’85, they hibernate on land or under water, and many ’90; Rowe et al., ’95; Nagle et al., ’98; Costanzo revive even after chilling to subzero temperatures. et al., 2000b), seemingly is adaptive in this regard. Myriad biophysical, physiological, and behavioral On the other hand, it is clear that turtles can responses enable hatchlings to survive tempera- dehydrate while hibernating on land. Christiansen tures encountered in their hibernacula, although et al. (’85) noted that K. flavescens emerging from cold tolerance is limited, even in the hardiest hibernacula in sand dunes appeared dehydrated, species. Behavioral avoidance of extreme cold is and hatchling C. picta and T. ornata collected from the most fundamental and reliable defense against hibernacula are often—though not always— chilling or freezing injury. Avoidance is easily thirsty (Costanzo et al., 2001b, 2004). The latter achieved in aquatic hibernators, but hatchlings observation may be taken as evidence of dehydra- that overwinter terrestrially, particularly inside tion because, when placed in water, experimen- nests, have limited ability to evade frost. They tally dehydrated hatchlings imbibe the quantity therefore depend on either of two mechanisms for needed to restore hydration (Costanzo et al., surviving transient exposure to subzero tempera- 2001b). A field study by Costanzo et al. (2004) tures: freeze tolerance or supercooling. Each coping showed that hydro-osmotic balance in hibernating mechanism has its own advantages and disadvan- turtles may be influenced by environmental con- tages. For example, although freeze tolerance ditions, especially patterns of precipitation and permits revival after the tissues have frozen and soil moisture abundance. During an unusually dry thawed, it offers protection only at modest winter, hatchling C. picta tended to lose body temperatures. Indeed, cooling of frozen hatchlings water and increase plasma levels of urea and other below À3toÀ41C apparently is lethal (Storey osmolytes, whereas during a wet winter turtles et al., ’88; Churchill and Storey, ’92a; Costanzo remained in positive water balance and accumu- et al., ’95b, 2006; Attaway et al., ’98). In contrast, lated little urea. supercooling permits survival over a much broad- Most reptiles tolerate a considerable degree of er temperature range (to À121C; Packard and desiccation. Turtles, generally, can survive the Packard, ’99); however, the supercooled state is loss of about one-third of their body water (Hall, inherently precarious because turtles are suscep- ’22; Bogert and Cowles, ’47; Ernst, ’68). Never- tible to inoculation by endogenous and exogenous theless, physiological functions are sensitive to the INA (e.g., Costanzo et al., 2003; Packard and attendant increase in osmotic pressure and ion Packard, 2006) and the rapid, uncontrolled freez- concentrations, which perturb metabolic and ing of deeply supercooled tissues usually is lethal acid–base regulation. In addition, hypovolemia (Storey and Storey, ’88). Freeze tolerance and and increased blood viscosity can impair circula- supercooling usually are dichotomous strategies, tory and gas exchange efficiency. Aside from these but at least one species, C. picta, apparently can direct effects, water loss can impose subtle, benefit from either mechanism, the one invoked deleterious consequences for hatchling fitness during a particular chilling episode depending on and survival, for example, by altering resistance physiological and environmental conditions (Lee adaptations, such as thermal tolerance (Plummer and Costanzo, ’98). et al., 2003). Additionally, because locomotor The following sections summarize the literature performance of hatchlings is influenced by body on cold-hardiness mechanisms of hatchling hydration (Miller, ’93; Miller et al., ’93; Finkler, turtles, highlighting advantages and limitations ’99; Finkler et al., 2000), excessive water loss can of each. We caution that our synthesis suffers from contribute to mortality during overland migration ambiguities and uncertainties in the literature. (Kolbe and Janzen, 2001, 2002). Some work A primary shortcoming is that virtually all such suggests that, unless water is discovered en route, information comes from laboratory experiments

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 331 that may not adequately represent field condi- supercools only if it lacks potent INA that organize tions. In several studies, design flaws stemming water molecules into ice embryos, which then from investigators’ failure to observe key aspects foster rapid, widespread freezing of a solution of the animal’s physiology or natural history have (Vali, ’95). Furthermore, injury or death from obfuscated the tenets of cold hardiness, leading to extensive chilling can occur in the absence of incremental progress in the field. Yet another freezing. Freeze avoidance through supercooling limitation is that virtually all of the information is an important survival mechanism in a host can be traced to only three research laboratories. of ectothermic animals, but is not without Therefore, the literature not only reflects the potentially deleterious consequences (Lee and biases of relatively few investigators, but it also Costanzo, ’98). lacks validation through extensive peer scrutiny. Finally, there is virtually no information available Chilling responses and cold tolerance for the majority of turtle families, and most of the pertinent literature concerns only one How do hatchling turtles behave when exposed species, C. picta. This species is of especial interest to extreme cold? Responses to progressive cooling because it is broadly distributed over the con- have been examined in only one study (Costanzo tinent, it (along with C. serpentina) occurs at et al., ’99). In both C. picta and C. serpentina, higher latitudes than other North American locomotor activity increased as body temperature turtles, and it regularly overwinters within the approached 01C, a response also seen in other shallow natal nest, even near the northern limits ectotherms (Costanzo, ’88; Cook, 2004), but of its geographic range (Nichols, ’33; Bleakney, declined as temperature was further reduced. ’63; St. Clair and Gregory, ’90; Lindeman, ’91; Ultimately, all turtles became immobile, but there Koonz, ’98; Rozycki, ’98). Furthermore, it is the were some important differences between the subject of a spate of studies attempting to resolve species. The terrestrial hibernator, C. picta, disparate hypotheses about the nature of its retained some neuromotor functions at tempera- cold hardiness (e.g., see commentary by Packard, tures as low as À101C; C. serpentina, which occurs 2004a). Nevertheless, studies of other turtle in the north but usually hibernates in thermally taxa, though much less abundant, have greatly buffered aquatic sites, entered cold rigor at À51C. helped to elucidate the diversity of environ- This work also revealed that C. serpentina from a mental conditions and associated survival southern population became immobilized at tem- mechanisms. peratures higher than those impacting northern conspecifics, suggesting that natural selection has favored increased cold tolerance in the latter Supercooling (Costanzo et al., ’99). A cold-hardiness strategy predicated on super- Hatchling turtles are remarkably tolerant of cooling permits survival with relatively little being chilled in the unfrozen state to very low physiological stress over a broad range of body temperatures. Critical thermal minima have not temperatures. Reliance on supercooling for freeze been determined for many species, although avoidance does, however, carry a significant C. picta reportedly can recover from acute chilling mortality risk. In fact, supercooled solutions are to temperatures as low as À10 to near À151C inherently metastable and, although the probabil- (Packard and Packard, ’93b, ’99; Costanzo et al., ity of ice nucleation increases with passage of time ’99, 2000a) and C. serpentina to at least À4orÀ51C and decreasing temperature, freezing of a super- (Packard et al., ’93; Costanzo et al., ’99). Hatchling cooled solution can occur spontaneously at any C. picta reportedly recover from supercooling for temperature below its equilibrium freezing/melt- 11 d at À101C (Packard et al., ’99a), for 25 d at –81C ing point (Salt, ’66; Sømme, ’82). Spontaneous (Hartley et al., 2000), and for 30 d at –51C (Packard freezing of a supercooled organism causes ice to and Packard, ’97). In both C. picta and form haphazardly and quickly, often with a lethal C. serpentina, cardiac contractions continue at body consequence (Storey and Storey, ’88). temperatures approaching À81C, although the Most organisms can be supercooled to some heart rate is greatly reduced (Birchard and Pack- extent (Andjus, ’55; Lowe et al., ’71; Spellerberg, ard, ’97). Even C. picta from the southern limit of ’76), although not all organisms can effectively use its geographical distribution can tolerate 8 d of supercooling as a means to avoid freezing under continuous supercooling, with a critical minimum natural conditions. Indeed, an aqueous medium temperature near À91C (Packard et al., 2002).

J. Exp. Zool. 332 J.P. COSTANZO ET AL.

It was long presumed that supercooling was determine chill tolerance. In one study, hatchling benign (Lowe et al., ’71; Spellerberg, ’72; Halpern, C. picta recovered after being held at À81C for 5 or ’79; Costanzo and Lee, ’95), but recent studies 15 d, but high mortality resulted after 25 d have shown that extended hypothermia and acute (Hartley et al., 2000). exposure to severe cold can impair physiological The literature maintains that hatchling turtles functions. In hatchling turtles, diminished cardiac can survive being supercooled for prolonged output and tissue perfusion at subzero tempera- periods so long as they avoid extreme cold. Thus, tures (Birchard and Packard, ’97) results in it is noteworthy that mortality has occurred in lactate accumulation and, presumably, reduced hatchling C. picta exposed only to À4orÀ71C intracellular pH (Hartley et al., 2000; Costanzo (Packard and Packard, ’97), and even À21C et al., 2001a; Baker et al., 2003). Consequently, (Packard and Packard, ’95a). Unfortunately, the species that poorly tolerate hypoxia may be unable experimental methodology in these studies does to use supercooling as a winter survival strategy. not preclude the possibility that the succumbing It is conceivable that supercooled tissues generate turtles actually froze, unbeknownst to the ROS and suffer oxidative stress. However, recent authors. In fact, Packard and Packard (2003b) work (Baker et al., 2007) suggests that tissue suspected that they failed to detect the freezing of levels of protein carbonyls and thiobarbituric acid C. picta that died after being exposed to À21C reactive substances (TBARS), products of oxida- for 7 d. tive damage to proteins and lipids, respectively, do The aforementioned studies focused on survival not increase during 24-h recovery for hatchling in response to subzero temperature exposure. C. picta held supercooled at À61C for 48 h. Some information suggests that physiological Oxidative stress resulting from supercooling and impairment can occur at temperatures above subsequent rewarming has not been examined in those that prove lethal. Hatchling C. picta hatchlings of other species, but apparently is promptly recovered from being held supercooled absent in the cold-hardy common lizard (Lacerta at À41C for 5 d; on rewarming, they responded vivipara), perhaps because they increase activities normally to tactile stimulation and burrowed into of certain antioxidant enzymes (Voituron et al., the substratum (Costanzo et al., 2001a). In 2005b). The hatchlings of some turtle species contrast, hatchlings held supercooled at À61C for maintain high constitutive levels of antioxidants the same period required 3 d to fully recover and (Dinkelacker et al., 2005b; Baker et al., 2007) and were initially lethargic, unresponsive to stimula- thus are in a constant state of preparedness. tion, and remained on the soil surface. Packard Extended chilling and exposure to severe cold and Packard (’93b) reported that hatchling can result in mortality, even if tissues never C. picta supercooled to À121C, a near-lethal freeze. In one study (Burke et al., 2002), Italian temperature, were slow to recover normal beha- wall lizards (Podarcis sicula) reanimated quickly vioral functions, but ultimately became active following experimental chilling to between À2.4 and fed. and À3.71C, but an individual supercooled to What are the proximal cause(s) of chilling À161C did not revive. Extreme chilling can also injury? Damage might stem from adverse lipid be injurious or lethal to hatchling turtles. Mortal- phase transitions and their deleterious effects on ity in hatchling C. picta occurs when turtles are membrane structure and function (Drobnis et al., cooled to between À11 and À131C (Packard and ’93; Kostal et al., 2003; Tomcala et al., 2006). Packard, ’93b), and none revive after being Consistent with this hypothesis, recent investiga- supercooled even briefly to À151C (Packard and tions with hatchling turtles have shown that Packard, ’99). Some evidence suggests that chill injury is induced over a narrow thermal range tolerance depends on season or stage of develop- (J. Costanzo and T. Muir, unpublished data). In ment. For example, winter-conditioned C. picta these experiments, hatchling C. picta (n 5 20) hatchlings survived exposure to À101C for at least tolerated exposure to À10.51C for 10 h, but no 24 h (Packard and Packard, ’93b), whereas ani- turtles (n 5 20) recovered from a 24-h exposure to mals tested in spring had poor survival when held À131C. Turtles in the latter group had beating at À101C for as little as 5 h (Packard and Packard, hearts and patent circulation, but did not respond ’99). Furthermore, recently hatched, unacclima- to tactile stimulation; thus, the nervous system tized C. picta died when chilled at À21C for 7 d may be a primary site of chilling injury (see (Packard and Packard, 2003a). Apparently, tem- Costanzo et al., ’99). Because nervous tissues perature and exposure duration interactively respond to thermal adaptation over both brief

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 333 and evolutionary time periods (Macdonald, ’81), perature at which body fluids spontaneously one would expect to find variation in chilling freeze. The ensuing phase transition is marked tolerance among hatchlings of species with diverse by an abrupt release of the latent heat of fusion, life histories. Notably, although C. picta tolerated which is detected by a thermocouple placed on, or exposure to À10.51C, neither M. terrapin (n 5 4) immediately adjacent to, the specimen. Note that, nor T. scripta (n 5 5) recovered, suggesting that although internal placement of the probe gives a the former is particularly well adapted to survive more accurate measure of core body temperature, extreme cold. The basis of its enhanced cold this approach should be avoided because it can hardiness remains to be elucidated. In hatchling initiate freezing. The lowest temperature recorded C. picta, cell membranes reportedly (DePari, ’88) before the exotherm appears, the temperature of have a higher percentage of unsaturated fatty crystallization (Tc), represents the limit of super- acids than in hatchling C. serpentina, a feature cooling under the particular experimental condi- that presumably maintains membrane fluidity and tions. The Tc is equivalent to the supercooling prevents deleterious phase transitions at lower point; however, the latter term, an unfortunate temperatures (Guschina and Harwood, 2006). literature convention, erroneously implies that the One of the most instructive findings of Costanzo measured value is an invariable, phenotypic and Muir (unpublished data) was that chilling attribute (sensu Lowe et al., ’71). tolerance was sensitive to thermal history. In one If the experimenter’s aim is to determine innate experiment, hatchling C. picta were precondi- supercooling capacity, then experimental subjects tioned by cooling them at À21CdÀ1 until they must be isolated from dust, soil, and other potent had reached À71C, holding them at this tempera- INA, both before and during testing. Subjects ture for 2 d, and then passively warming them to must also be free of surface moisture, which can 01C. Transient exposure to this intermediate freeze and inoculate tissues. For hatchling turtles, temperature markedly improved their chill toler- this requirement can be met by keeping them on ance, as 10 of 20 hatchlings survived an ensuing clean vermiculite, a common incubation medium 24-h exposure to À131C, an otherwise lethal that is sterilized during its manufacture and lacks treatment. This result is reminiscent of the rapid significant ice nucleation activity (Costanzo et al., cold-hardening phenomenon exhibited by various 2000a). Before testing, hatchlings are cleaned and arthropods, and could derive from remodeling of permitted to air dry before being chilled in a clean, cell membranes (Kostal et al., 2003; Overgaard dry environment. Cooling rates used in super- et al., 2005, 2006; Lee et al., 2006a) and myriad cooling trials have not been standardized (Cost- changes in the metabolite profile (Bohn et al., anzo and Lee, ’95), although unless extreme (e.g., 2007). Stress-induced thermoprotection can also 4101ChÀ1), they probably do not materially affect stem from upregulation of stress proteins, such as Tc (Costanzo et al., 2000a). Hsp 70, which protect against cytoskeletal pertur- The Tc values determined for specimens main- bation, among other functions (Klose and Robert- tained and tested in isolation from ice and INA son, 2004; Robertson, 2004). Cold preconditioning represents the maximal limit of supercooling may also improve the organism’s ability to main- under idealized conditions. Packard and coauthors tain transmembrane ion gradients (Kostal et al., (’95a) stated that refraining from thoroughly 2004), perhaps by suppressing ion channels. cleaning and drying hatchling C. picta rendered Whether or not chilling injury is an important them ‘‘extraordinarily prone to freezing at high natural cause of winter mortality remains to be subzero temperatures’’ in supercooling trials. determined. Also unknown is whether individuals Even brief contact with nesting soil harboring surviving periods of supercooling can later exhibit native INA can markedly constrain supercooling diminished performance and fitness, as has been capacity. Not surprisingly, the limit of super- demonstrated for other organisms (Layne and cooling for C. picta hatched and reared on natural Kuharsky, 2001). substrata typically is 5–61C higher than that determined for vermiculite-reared hatchlings (Costanzo et al., 2000a; Packard et al., 2001). Methodological considerations Consequently, inter-study comparisons are con- The usual laboratory method for determining founded by inconsistencies in experimental proto- supercooling capacity is to monitor the body col, which can lead to erroneous conclusions. In temperature of individuals during progressive particular, the observation that supercooling chilling in a refrigerated bath, noting the tem- capacity is greater in hatchlings near the northern

J. Exp. Zool. 334 J.P. COSTANZO ET AL. limit of their range (Packard and Packard, ’95a; anzo et al., 2000a, 2001c, 2003; Packard et al., Packard et al., ’97b; Costanzo et al., 2000a) than in 2001; Packard and Packard, 2003a, 2006). hatchlings from several more temperate locales in Moreover, hatchlings collected directly from the eastern United States (DePari, ’88; Claussen natural hibernacula supercool on average to and Zani, ’91; Packard and Janzen, ’96) and only À7orÀ81C (Costanzo et al., 2001c, 2003; southern Canada (Storey et al., ’88, ’91; Churchill Packard et al., 2001). and Storey, ’92a) has promulgated the idea that Organisms obey physical laws governing solu- this species exhibits adaptive geographic differ- tions; therefore, it is not surprising that many ences in supercooling capacity. However, despite ectotherms, including reptiles, are capable of its attractiveness from an evolutionary perspec- supercooling deeply (Lowe et al., ’71; Spellerberg, tive, the variation simply reflects different meth- ’72). Because the various INA in a solution are ods of rearing hatchlings, which may or may not randomly distributed, and the probability that a have become contaminated with INA (for a more given solution contains an efficient, high-tempera- extensive discussion, see Costanzo et al., 2000a). ture nucleator increases with volume (Vali, ’95), Results of laboratory supercooling trials must be small organisms generally can be supercooled carefully interpreted in the context of the experi- more extensively than large ones (Costanzo and mental protocols used by the investigators. Lee, ’95; Lee and Costanzo, ’98; Wharton, 2002; Ansart and Vernon, 2003). Accordingly, hatchling turtles (and perhaps some small lizards—Halpern, Morphological and physiological ’79; Costanzo et al., ’95a) are among the few influences vertebrates that can reliably use supercooling as a With the approach of winter, many temperate winter survival mechanism. Claussen and Zani ectotherms undergo physiological changes that (’91), in an investigation of allometric variation in improve cold hardiness. Biochemical and physio- adult and hatchling C. picta, found no correlation logical mechanisms promoting development of between supercooling capacity and body size; cold hardiness are best known among inverte- however, the relatively high Tc values reported brates (Lee, ’91; Loomis, ’91; Oswood et al., ’91; for all individuals suggest that nucleation was Duman et al., ’95). With the exception of certain triggered by a thermocouple probe placed in the coldwater fishes (DeVries, ’82), evidence for intestine. In a comparative study of hatchling seasonal variation in cold hardiness among turtles, Costanzo et al. (2000a) found that, after ectothermic vertebrates is scanty (Costanzo and excluding the exceptionally low Tc values obtained Lee, ’95). Nevertheless, several laboratory studies for C. picta, smaller turtles supercooled more (Costanzo et al., 2000b, 2003; Packard et al., 2001) extensively than larger individuals, irrespective of of hatchling turtles have documented a profound taxon. However, in a subsequent study (Costanzo increase in supercooling capacity coincident with et al., 2001b), variation in body size (range in winter acclimatization. In these investigations, C. mean body mass, 3–11 g) did not appreciably picta reared on (INA-free) vermiculite and tested influence supercooling capacity in hatchlings of soon after hatching supercooled modestly (Tc , À6 eight taxa. All of the specimens used in that study to À111C); however, following cold acclimation were relatively small, however. The authors they supercooled to near À171C. This change, concluded that body size, per se, probably does induced by elimination or masking of endogen- not preclude the use of supercooling as a winter ously produced INA (see below), is striking in survival mechanism in hatchling turtles. vermiculite-reared animals and gives the illusion The amount and distribution of water within an that these hatchlings could avoid freezing in even organism may strongly influence supercooling the coldest winter. To the contrary, C. picta reared capacity, and some ectotherms seasonally undergo on native nesting soil (rather than vermiculite) a partial dehydration that enhances their cold exhibit a relatively modest gain in supercooling hardiness (Zachariassen, ’91; Block, ’96; Lee and capacity (e.g., 2–51C) with cold acclimation (Pack- Costanzo, ’98; Danks, 2000). This effect at least ard et al., 2001; Costanzo et al., 2003; Packard and partly reflects the lowered equilibrium freezing/ Packard, 2003a, 2006). These turtles, which are melting point (owing to solute concentration) and irreversibly contaminated with natural INA, ex- reduced body fluid volume. Experimental dehy- hibit mean Tc values that range from À10 to À71C, dration apparently enhances supercooling capa- even when fully cold-hardened and cooled in city in hatchling C. picta (Costanzo et al., ’95b). isolation from ice and environmental INA (Cost- These turtles can dehydrate considerably during

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 335 dry winters (Costanzo et al., 2004) and this markedly (by 60–70 mosmol LÀ1), much of the change, in principle, could enhance their capacity increase being attributed to a two-fold to three-fold to supercool (Storey et al., ’88; Claussen and Zani, increase in urea, which in fully cold-acclimated ’91; Churchill and Storey, ’92a; Costanzo et al., animals comprised 20% of the osmotically active ’95b). Whether or not this effect is manifested solutes. In contrast, hatchling C. serpentina, which under field conditions remains to be determined. do not supercool as extensively as hatchling In a comparative study, Costanzo et al. (2001b) C. picta, showed a decrease in plasma osmolality determined that marked variation in the body with a concomitant reduction in tissue urea. In water content (range, 76–83% of fresh body mass another study, urea was found to be a major in eight taxa) did not explain interspecific varia- organic osmolyte in winter conditioned hatchlings tion in supercooling capacity, but neither did of various turtle taxa (Costanzo et al., 2006). having high body water content preclude the use In some organisms, certain amino acids may of supercooling as a winter survival mechanism. function as compatible osmolytes (Somero and Antifreeze proteins (AFPs), also known as Yancey, ’97). The amino acid pool in hatchling ‘‘thermal hysteresis factors’’ and ‘‘ice-structuring C. picta is maintained during cold acclimation, and proteins,’’ occur in various organisms relying on slight increases are seen only in serine and tyrosine supercooling to survive exposure to subfreezing (Costanzo et al., 2000b). Similarly, cold acclimation temperatures (DeVries, ’82; Duman et al., ’91, has little effect on the quantity of glucose or ’95). In blood samples, these highly specialized carbohydrates that in many insects and other peptides and glycopeptides arrest the growth of ice ectotherms contribute to increased osmotic pres- embryos at temperatures as low as 2–61C below sure. Glycerol, mannose, sorbitol, erythritol, threi- the equilibrium freezing/melting point. However, tol, and fructose are absent or are present in low functioning in vivo, they may prevent freezing concentrations in cold-acclimated hatchling over a much larger range of temperatures C. picta (Storey et al., ’88; Churchill and Storey, (Zachariassen and Husby, ’82). Antifreeze com- ’92a; Costanzo et al., 2000b). High-performance pounds play a critical role in the winter survival of liquid chromatography analysis of plasma indicated some polar fishes (DeVries, ’82), but their im- that cold acclimation is associated with a 50–70% portance in cold hardiness of hatchling turtles increase in plasma concentration of inositol and de (and other vertebrates) is equivocal. They appar- novo production of six unidentified carbohydrates ently are absent from the blood of freeze-tolerant (Costanzo et al., 2000b); however, given their very wood frogs, Rana sylvatica (Wolanczyk et al., ’90), low concentrations, it seems unlikely that they hatchling C. picta (Storey et al., ’91; Costanzo could enhance supercooling capacity appreciably. et al., 2000b), and hatchling C. serpentina (Cost- In fact, it is doubtful that the colligative effect of anzo et al., 2000b). It is unknown whether they this change can account for the exceptional super- occur in turtles, such as hatchling G. geographica, cooling capacity of winter C. picta (Zachariassen which survive extreme chilling exclusively by and Hammel, ’76; Lee et al., ’81). freeze avoidance (Baker et al., 2003). Further- more, it is possible that AFPs are absent from the Taxonomic and ecological associations blood but nevertheless occur in extravascular compartments or intracellular spaces (Olsen Although the available literature suggests that et al., ’98). supercooling capacity in hatchling turtles varies As a rule, increasing the concentration of osmoti- taxonomically (Table 4), comparing data among cally active solutes in a solution colligatively studies is risky because methods of incubation, enhances its capacity to supercool. Many freeze- housing, acclimation, and testing strongly influ- avoiding invertebrates exploit this principle by ence the outcome of supercooling trials. Two accumulating compatible osmolytes, such as polyols, investigations examined taxonomic variation in sugars, and certain amino acids, and/or by shedding this character, both ensuring that all specimens a portion of their body water (Storey and Storey, ’88; were reared on (INA-free) vermiculite, acclimated Lee, ’91; Ansart and Vernon, 2003). The question of to 41C, and cooled in the absence of external ice whether freeze-avoiding turtles (and other ectother- and INA (Costanzo et al., 2000a, 2001b). In the mic vertebrates) also use this mechanism has earlier report, a terrestrial hibernator, C. picta, received little study. Costanzo et al. (2000b) showed froze at À17.81C, 7–121C below the Tc recorded that acclimation of hatchling C. picta to simulated for three aquatic hibernators, S. odoratus, winter conditions increased plasma osmolality A. spinifera, and C. serpentina. In the latter

J. Exp. Zool. 336 J.P. COSTANZO ET AL.

TABLE 4. Supercooling capacity in hatchlings of North American turtles, as indicated by the temperature of crystallization (Tc) of turtles hatched, reared, and tested in the absence of ice nucleating agents (INA)

Tc (1C) Taxon Hibernation habitat Origin Mean7SE Min n Reference

Apalone spinifera Aquatic Jennings Co., IN À7.570.4 n.r. 6 1 Chelydra serpentina Aquatic Kosciusko Co., IN –5.770.8 n.r. 12 1 Garden Co., NE –8.570.5 n.r. 8 2 Chrysemys picta bellii Natal nest Cherry Co., NE –13.370.8 n.r. 3 1 Garden Co., NE –16.570.2 n.r. 6 3 Garden Co., NE –14.671.9 n.r. 5 4 Garden Co., NE –16.270.8 n.r. 10 5 Garden Co., NE –15.571.2 n.r. 8 2 Clearwater Co., MN o–10.0 n.r. 10 6 Bottineau Co., ND o–8.9 o–9.1 12 7 C. p. marginata Natal nest Central Ontario –15.971.5 n.r. 7 1 Kosciusko Co., IN –17.870.3 –20.0 14 1 Emydoidea blandingii Unknown; possibly terrestrial Grant Co., NE –14.371.2 –17.0 7 8 Graptemys geographica Natal nest Fulton Co., IN –14.870.4 n.r. 10 9 Kinosternon flavescens Below natal nest Garden Co., NE –9.371.5 n.r. 5 2 Malaclemys terrapin Natal nest, other terrestrial, or aquatic Cape May Co., NJ – 15.171.1 –18.2 8 10 Sternotherus odoratus Aquatic Kosciusko Co., IN –11.470.1 n.r. 10 1 Lonoke Co., AR –8.171.1 n.r. 8 2 S. carinatus Aquatic McCurtain Co., OK –10.670.7 n.r. 5 2 Terrapene ornate Below natal nest Garden Co., NE –12.071.6 n.r. 3 2 Trachemys scripta Natal nest Lonoke Co., AR –13.971.4 n.r. 8 2

Data are reported only from experiments on turtles hatched from eggs collected directly from females and hatched on (INA-free) vermiculite. n.r., not reported. References: (1) Costanzo et al. (2000a); (2) Costanzo et al. (2001b); (3) Costanzo et al. (2000b); (4) Costanzo et al. (’98); (5) Costanzo et al. (2003); (6) Packard et al. (’99a); (7) Packard et al. (’97b); (8) Dinkelacker et al. (2004); (9) Baker et al. (2003); (10) Baker et al. (2006).

project, mean Tc varied among eight taxa, ranging cooling capacity of hatchling C. picta remains from À8.1 to À15.51C. These studies showed that unknown. the supercooling limit for each species more or less Although turtle species clearly vary in their reflected the thermal extremes it encounters in its intrinsic capacity to supercool, the underlying particular winter microenvironment. For example, causes of such variation are incompletely under- the terrestrial hibernators, T. ornata, T. scripta, stood. Body size and body water content influence and C. p. bellii, supercooled extensively, whereas supercooling limits in some organisms, but these aquatic hibernators, C. serpentina, S. odoratus, factors probably explain little of the observed and S. carinatus, supercooled only modestly. variation among hatchling turtles, as neither Supercooling capacity was also less developed in varies appreciably among species (Costanzo K. flavescens, a species that hibernates deep in the et al., 2001b). Certain species apparently are more soil, below the reach of frost (Costanzo et al., ’95b). proficient than others at masking or eliminating Other studies, conducted in the same laboratory internal INA. For example, although the blood using identical methods, documented extensive plasma of hatchling C. serpentina can be super- supercooling (mean Tc, À14 to À151C) in hatchling cooled to À121C (Costanzo et al., 2000b), intact G. geographica (Baker et al., 2003), M. terrapin animals invariably freeze at much higher tem- (Baker et al., 2006), and E. blandingii (Dinkelack- peratures (Costanzo et al., 2000a, 2001b), perhaps er et al., 2004), species known or suspected to because this species retains INA in extravascular hibernate terrestrially. Supercooling capacity in spaces, even after cold acclimation. By contrast, C. picta exceeds that of other turtle species and the limit of supercooling for plasma of winter- most other similarly sized ectotherms (Lee and acclimatized C. picta more or less matches that of Costanzo, ’98). One hatchling froze only after the intact organism. Perhaps some species en- reaching À201C; this value is the lowest hance supercooling capacity through behavioral Tc reported for any and is on par with mechanisms, such as remaining quiescent during that of a water droplet (Costanzo et al., 2000a). cooling (Costanzo et al., ’99), although this idea The basis for the exceptional intrinsic super- remains conjectural.

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 337

Regulation of ice nucleation removing or masking endogenous INA (Costanzo et al., 2000b, 2003; Packard et al., 2001). Costanzo Principles of biological ice nucleation, including et al. (2000b) determined that the blood plasma of those germane to overwintering in hatchling neonates lacks ice-nucleating activity and deduced turtles, have been summarized (Costanzo and that the INA are restricted to intracellular and/or Lee, ’95, ’96; Lee and Costanzo, ’98). Organisms interstitial compartments. Further study (Costan- depending on supercooling for winter survival zo et al., 2003) showed that the neonatal gut must manage the risk of freezing by minimizing harbors INA that, derived from internalized yolk, the time they spend at subzero temperatures and are eliminated in feces during the weeks following by limiting their exposure to excessive cold. More hatching. Factors initiating gut evacuation appar- critically, they must remain isolated from all ently include maturational changes, given that entities that can trigger ice nucleation in body some individuals defecated within a few weeks of fluids at high subzero temperatures. These ha- hatching, even before ambient temperature was zards include ice itself and various INA of reduced. Collectively, these studies suggest that endogenous and environmental origins. To safely seasonal development of cold hardiness depends exploit the physical phenomenon of supercooling, on elimination of endogenous INA and an un- organisms must be capable of masking or elim- masking of innate supercooling capacity (Fig. 6). inating internal INA and mounting barriers to One perplexing discovery in this line of investi- ingress of external ice and INA. gation was that C. picta commonly ingested Regulation of ice nucleation is also of great substratum and bits of eggshell during, or shortly importance to freeze-tolerant organisms, but for a after, hatching (Packard et al., 2001; Costanzo very different reason. Survival of any given et al., 2003; Packard and Packard, 2006). This freezing episode requires that inoculation be matter is not retained throughout hibernation, controlled in a manner that prevents intracellular but is voided with feces during the weeks following ice formation and permits ice to propagate slowly hatching. The significance of this behavior is (Storey and Storey, ’92; Duman et al., ’95; unknown, although presumably it could aid gut Zachariassen and Kristiansen, 2000). Accordingly, elimination (e.g., roughage), inoculate the tract many species employ one or more mechanisms with normal flora, or otherwise promote proper that initiate freezing at a temperature close to gut function. the equilibrium freezing/melting point of body On the other hand, ingesting external matter tissues. Three potential strategies for achieving seemingly predisposes hatchlings to freezing risk. nucleation control, in increasing order of efficacy Nesting soil contains at least two classes of INA, and reliability, include endogenous synthesis of INA, active incorporation of environmental INA, and inoculative freezing through intimacy with 0 environmental ice and INA. The diversity of nucleation-regulation mechanisms employed by -5 freeze-avoiding and freeze-tolerant hatchlings is C) ˚ an interesting and multi-dimensional feature of ( -10 No Soil their winter biology. c T Frozen Soil Endogenous INA -15 Seasonal cold hardening in freeze-avoiding -20 ectotherms commonly involves elimination of August November January ingested food and other matter that could trigger ice nucleation (Sømme, ’82; Zachariassen, ’85; Fig. 6. Effect of exposure to environmental ice and ice- nucleating agents (INA) on temperature of crystallization (Tc) Bale et al., ’89; Duman et al., ’95). Similarly, for in hatchling painted turtles (Chrysemys picta). Hatchlings hatchling turtles, achieving the full capacity to were cooled in isolation from INA (no soil) or in contact with supercool requires purging the gut of potent INA. INA (submerged in frozen soil) until each produced a freezing C. picta hatched and reared on vermiculite, an exotherm. Trials were conducted in mid-August with 221C- essentially INA-free substratum, initially freeze at acclimated hatchlings, late November with 101C-acclimated 1 1 hatchlings, and late January with 4 C-acclimated hatchlings. relatively high temperatures (À6toÀ11 C), but Turtles were hatched and reared on (INA-free) vermiculite develop a more extensive supercooling capacity before testing. Means7SE (n 5 6). (Modified from Costanzo during winter acclimatization, coincident with et al., 2000b.)

J. Exp. Zool. 338 J.P. COSTANZO ET AL.

0 temperatures spanning the range recorded in natural nests are at virtually no risk of freezing -5 when they are removed from contact with crystals

C) b of ice’’ (Packard and Packard, ’95a). This conten- ˚ a,b ( a c -10 tion may be true of specimens kept in a sterile T environment, but seems untenable for turtles in -15 nature. Indeed, studies of field-collected animals

c c confirm that fully cold-hardened hatchlings can -20 supercool to only À7orÀ81C (Costanzo et al., Soil Soil 2001c, 2003; Packard et al., 2001). Soil Washing Washing, +33 d Water Sham For most freeze-tolerant organisms, it is critical Treatment to survival that ice nucleation occurs at a Fig. 7. Effect of exposure to internal ice-nucleating agents temperature near the tissue equilibrium freez- (INA) on temperature of crystallization (Tc) in hatchling ing/melting point. Many such species use extra- painted turtles (Chrysemys picta). Hatchlings were cooled until each produced a freezing exotherm 24 h (or 33 d) after cellular INA, often synthesized during seasonal being force-fed a small quantity of autoclaved nesting soil cold hardening, to control ice nucleation (Duman (organic INA absent; particulate INA present) or a washing et al., ’95; Lee and Costanzo, ’98; Zachariassen and prepared from nesting soil (organic INA present; particulate Kristiansen, 2000). These agents include both INA absent). Control turtles received ultrapurified water inorganic crystals and specialized proteins, the or nothing (sham). Turtles were hatched and reared on (INA- free) vermiculite, progressively acclimated from 22 to 41C, latter being more common. Ice-nucleation pro- and tested in winter. Means7SE (n 5 5) identified with teins, which have been characterized for several different letters were statistically distinguishable (ANOVA/ invertebrates, apparently organize water mole- Student–Newman–Keuls Multiple Comparisons test; Po0.05). cules through hydrogen bonding to form an ice (From Costanzo et al., 2000a.) embryo. Agents exhibiting similar characteristics have including various mineral particulates (insoluble, been found in the blood of freeze-tolerant verte- largely unaffected by high temperature and brates, including C. picta (Wolanczyk et al., ’90; pressure) and organic entities (water-soluble, Storey et al., ’91), but their adaptive significance autoclave-sensitive) that apparently derive from has been challenged for several reasons (Costanzo certain microorganisms (Costanzo et al., 2000a). and Lee, ’96). One relevant issue is that they Feeding a small quantity of either type to hatchl- nucleate blood plasma in vitro at À7toÀ81C, ing C. picta severely limits supercooling, as mean several degrees below the lower limit for freezing Tc rises from À171CtoÀ7toÀ81C (Fig. 7). survival. In addition, under idealized laboratory Dissecting the intestine attests that turtles conditions, hatchling C. picta can supercool well hatched on nesting soil ingest, but ultimately below À81C, so the activity of these agents is not eliminate, particulate INA (Costanzo et al., 2003; expressed in vivo. Finally, similar ice-nucleating Packard and Packard, 2006). However, turtles activity is also found in the blood plasma of various long retain organic INA, which, in the case of freeze-intolerant species (Costanzo and Lee, ’95, microbes, may have colonized the gut. Conse- ’96). Costanzo et al. (2000b) determined that the quently, any gain in supercooling capacity asso- activity was an artifact of sample contamination ciated with purging the gut is slim, as, for with exogenous INA and confirmed that the plasma example, the limit of supercooling measured in of hatchling C. picta (and C. serpentina)lacksice- winter C. picta was only 21C lower than that nucleation proteins. Additional investigation is measured in autumn, À61C (Packard et al., 2001). needed to determine whether specialized nucleat- Collectively, these studies indicate that contam- ing proteins are found in the blood (or perhaps on ination by soil INA, either by ingestion or some cell membranes; see Izumi et al., 2006) of any turtle other route, is a serious constraint on supercooling taxa. Such studies might profitably focus on freeze- capacity and, further, raise pointed questions tolerant species, such as M. terrapin, whose ability about the efficacy of supercooling as a cold- to supercool is relatively poor. hardiness strategy in terrestrially hibernating turtles. Experiments with C. picta reared on Inoculative freezing (INA-free) vermiculite and meticulously groomed and dried before testing have led authors to Eliminating endogenously produced and in- conclude that ‘‘supercooled turtles exposed to gested INA from the body provides no guarantee

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 339 that an organism will remain supercooled; freez- fluids. This scenario could explain why, in some ing can still occur through seeding by environ- species, inoculation is not necessarily instanta- mental ice or INA (Salt, ’63; Ansart and Vernon, neous, but may be delayed hours or days after 2003; Holmstrup, 2003). Among terrestrially-hi- initial exposure to environmental ice and INA bernating turtles, inoculative freezing often is (Packard and Packard, ’93a, ’95a, ’97; Costanzo triggered by contact with frozen soil or other et al., ’95a; Packard et al., ’97b, ’99a, 2002). substrata at relatively high temperatures (Pack- Much of the literature concerning inoculative ard and Packard, ’93a, b, ’95a, ’97, 2003b; Packard freezing in hatchling turtles has addressed the et al., ’93, ’97b; Costanzo et al., ’95b, ’98, 2000b, mediating role of the skin, perhaps suggesting that 2001b, c; Willard et al., 2000; Baker et al., 2003, the integument is the sole avenue for INA 2006; Dinkelacker et al., 2004). Ice is a particularly transmission (e.g., Packard and Packard, ’93a,b, potent catalyst, but inoculation can result from ’95a, 2003b; Packard et al., 2000; Willard et al., contact with various inorganic and organic INA 2000; Packard, 2004a). However, it is important to ubiquitously found in terrestrial hibernacula realize that ice and/or INA can also access body (Costanzo et al., 2000a). Susceptibility to inocu- fluids through ingestion and by way of the ocular lative freezing is governed by temperature and a and otic openings, nares, cloaca, and umbilicus host of environmental factors that mediate an (Kalabukhov, ’58; Lowe et al., ’71; Spellerberg, organism’s intimacy with these agents (Costanzo ’72; Costanzo et al., 2000a, 2001b, 2003). Empiri- and Lee, ’96; Lee and Costanzo, ’98). It is also cal evidence for inoculation through these avenues influenced by behavior, morphology, and physiol- is scanty, although manual application of ice to the ogy, and thus varies markedly among taxa and, umbilicus or skin of a supercooled turtle promptly potentially, within populations of conspecifics triggers freezing (Packard and Packard, ’93a; (Costanzo et al., 2001b). Packard et al., ’93, ’97c). That freezing can be Physical events precipitating inoculative freez- initiated by contact with inorganic INA (e.g., sand ing in ectothermic animals are incompletely grains and other particulates) too large to pene- understood. In some invertebrates, ice apparently trate skin pores (Costanzo et al., 2000a) implies infiltrates the body through respiratory pores, that INA can invade through nonintegumentary exoskeletal joints, gill lamellae, and body orifices routes. Several different routes may admit INA, (Salt, ’63; Lee et al., ’96; Frisbie and Lee, ’97; as activity is found in some, but not all, body Holmstrup, 2003; Issartel et al., 2006). Amphi- compartments (Costanzo et al., 2003). Whether bians are highly susceptible to inoculative freezing INA ultimately access and nucleate the tissues owing to the ease at which ice propagates across depends, in part, on the relative sizes of the moist, permeable skin (Layne et al., ’90; Layne, agents and the external openings (Gehrken, ’92; ’91). The less permeable integument of reptiles Valerio et al., ’92). and fishes better resists ice transfer (Valerio et al., Some authors ardently maintain that hatchling ’92; Costanzo and Lee, ’96), although turtle skin is turtles can fully ‘‘resist’’ inoculative freezing, not a perfect barrier, as experimentally wetted being essentially immune to this peril (Packard hatchlings are readily inoculated when the moist- and Packard, 2001a; Packard, 2004a). A compet- ure on their surface freezes (Attaway et al., ’98; ing view is that any hatchling achieving sufficient Packard et al., ’99b; Packard and Packard, 2004; intimacy with environmental ice or INA is Costanzo et al., 2006). This fact has biological destined to freeze, and, furthermore, that ice relevance when, owing to differential patterns of nucleation via this process can occur at any heating and cooling between turtles and the temperature equal to or below the tissue equili- surrounding soil, water vapor condenses on the brium freezing/melting point (Costanzo et al., surfaces of terrestrially hibernating turtles ’95b). Indeed, the fact that hatchlings are highly (Thompson, ’88). susceptible to direct physical contact with ice (but Some work indicates that inoculative freezing see Packard and Packard, ’93a) underlies the can occur even if a hatchling avoids physical common method used to control ice nucleation in contact with ice. Interpreting the seminal work freeze-tolerance trials (Packard et al., ’99b; Pack- by Salt (’63), Packard and Packard (’93a) proposed ard and Packard, 2004; Costanzo et al., 2006). that minute ice crystals grow in a plume of water However, as we discuss below, inoculative freezing vapor emanating from integumentary ‘‘pores,’’ is not necessarily inevitable, and whether or not ultimately transmitting across the skin and it occurs during a given chilling episode depends triggering global freezing of supercooled body on exposure duration, various physiological and

J. Exp. Zool. 340 J.P. COSTANZO ET AL. environmental factors, and, to some extent, results of such trials should be interpreted stochasticity. cautiously. It is important to realize that experi- mental outcomes are highly specific to the testing conditions used, and, unless a standardized pro- Methodological considerations tocol is followed, results of different studies should Generally, inoculative freezing is studied in the not be directly compared. laboratory by progressively cooling specimens in Another important consideration is that erro- the presence of ice and/or INA until their body neous conclusions may be drawn if conditions of fluids freeze. The resultant Tc, measured with a inoculation trials are inappropriate to the biology thermocouple placed near each specimen, is taken of the subjects under investigation. For example, as a measure of the capacity for freeze avoidance measures of inoculation resistance for C. picta achieved under the specific conditions of the trial. hatched and reared on a (INA-free) vermiculite, Although the procedure is simple, designing meticulously groomed to remove moisture and meaningful experiments presents several techni- debris, and tested in a non-indigenous clayey soil cal challenges. First, it is important to configure (e.g., Packard and Packard, ’93a,b, ’95a; Packard the experiment so that the study subjects are and Janzen, ’96; Packard et al., ’97c) undoubtedly exposed in a naturalistic manner to the kinds and overestimate the ability of these hatchlings to amounts of INA prevalent in the hibernaculum supercool in nature. Overestimates also result microenvironment. It is also important to ensure from the practice of ‘‘head-starting’’—i.e., cooling that the exposure occurs over ecologically relevant hatchlings in a substratum that for a time remains time periods and temperature ranges. supercooled and, therefore, lacks ice crystals Studies of inoculation resistance in hatchling (Packard and Packard, ’90, ’93a,b, ’95a, ’97; turtles commonly involve cooling hatchlings in a Packard et al., ’93, ’97c, 2000; Costanzo et al., simplified representation of a terrestrial hiberna- ’95b; Packard and Janzen, ’96; Willard et al., culum. In practice, a turtle is placed in a vessel 2000). Failure to freeze the substratum before and fully immersed in an appropriate substratum. initiating the trial also creates a serious technical A thermocouple, used for tracking temperature problem in that the large, protracted exotherm during cooling and registering the freezing produced when the substratum eventually freezes exotherm, is positioned adjacent to the hatchling. can wholly eclipse the relatively minor signal from If the objective concerns inoculative freezing by the hatchling, thus making it impossible to environmental ice, the substratum should be determine whether the animal has actually frozen. frozen before the subject is allowed to become Although this difficulty is easily avoided by using supercooled. This is achieved simply by inoculat- the smallest volume of soil that fully envelops the ing the soil with a few ice crystals and allowing it hatchling (e.g., Costanzo et al., ’98), exotherm to freeze at a temperature slightly above the ‘‘masking’’ is a recurring, confounding issue in equilibrium freezing/melting point of turtle tis- many reports (Packard and Packard, ’90, ’93a,b, sues. Once the soil solution has thoroughly frozen, ’97, 2003b; Packard et al., ’93, ’97c, 2000; Willard the entire system is progressively cooled until the et al., 2000). thermocouple registers a freezing exotherm, yield- ing the T of the turtle within. c Morphological and physiological The methodology described above, by virtue of influences its simplicity, can be standardized so that data for different experimental groups, taxa, etc. are Freeze-avoiding organisms commonly exhibit a comparable. It can determine fairly precisely the seasonal development or enhancement of inocula- lowest temperature each specimen attained before tion resistance that is timed to provide maximum freezing, and this metric is a useful gauge of protection against (lethal) freezing during the freezing resistance capacity under the conditions coldest weather (Fields and McNeil, ’86; Lee, ’91; of the trial. On the other hand, it is a poor Gehrken, ’92; Rojas et al., ’92). This response is facsimile of the actual hibernaculum microenvir- also exhibited by hatchlings of some turtle species onment, which is far more complex, both physi- (Fig. 8). In laboratory studies, recently hatched cally and thermally. Natal nest hibernacula, for C. picta were highly susceptible to inoculation, example, commonly include other hatchlings, whereas cold-acclimated hatchlings supercooled broken eggshells, various organic debris, and more extensively in the presence of environmental possibly an air space. Given these limitations, ice and INA (Costanzo et al., 2000b; Packard and

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 341

Packard, 2003b). Data for C. picta collected from Hatchling Age (d) natural hibernacula at intervals from fall through 30 60 90 120 early spring corroborate these laboratory findings (Costanzo et al., 2004). Such variation implies that 125 Skin inoculation resistance is under morphological and/ or physiological control, although the underlying mechanisms have not been elucidated. 100 Ultrastructural characteristics of the integu- ment have important influences on the skin’s role as a barrier between the internal and external 75 milieus. In dormant turtles, for example, extensive 125 keratinization may reduce cutaneous water ex- change (Chilian, ’76) and potentially curb trans- Carcass Lipid density (mg/g) mission of ice and INA. Willard et al. (2000) 100 suggested that C. picta skin is both heavily keratinized and rich in lipid deposits, perhaps accounting for the profound inoculation resistance 75 exhibited by this species. Other investigators have found that hatchling C. picta maintain concentra- 0 tions of lipids in the skin (though not in carcass) n.d.

throughout winter, but failed to link lipid density (˚C) -5 c with the distinct seasonal/ontogenetic develop- T ment of inoculation resistance (Fig. 8). Although -10 these results do not preclude an important role for skin lipid in inoculation resistance, they seem -15 contrary to the hypothesis that fat deposition 22 15 10 4 accounts for the observed seasonal/maturational Acclimation temperature (˚C) development of this character (Packard and Pack- Fig. 8. Thermal acclimation effects on concentration of ard, 2003b). A particularly interesting finding in lipid in skin and carcass, and inoculation resistance in this line of research was that transmission proper- hatchling painted turtles (Chrysemys picta bellii). Means7SE ties of the integument varied anatomically, the (n 5 8). Turtles were hatched and reared on (INA-free) skin within axillary and inguinal pouches posing vermiculite, kept in darkness, and progressively acclimated less of a barrier to ice and INA than that on from 22 to 41C. Lipid concentrations in dried skin samples and dried carcasses were determined using a chloroform/methanol exposed appendages (Packard and Packard, ’95a). extraction procedure (J. Costanzo and J. Jack, unpublished Physiological regulation of inoculative freezing data). Capacity for inoculation resistance was represented by in hatchling turtles has received scant attention in the temperature of crystallization (Tc) of individual hatchlings the literature. In principle, organic osmolytes cooled while submerged in a frozen matrix of sandy soil accumulated during winter acclimatization (e.g., collected from near actual C. p. bellii nests in Garden Co., NE, Costanzo et al., 2000b, 2006) influence the process the source of the animals. (From Costanzo et al., 2000b.) because INA activity is sensitive to the solute concentration in the nucleating environment epidermis or other structures (Olsen et al., ’98; (Zachariassen and Hammel, ’76). Some osmolytes Duman, 2001). apparently work cooperatively with specialized proteins to inhibit internal INA (Duman, 2002). Temporal and environmental influences Furthermore, in some organisms, AFPs not only protect against spontaneous freezing but also Inoculative freezing is influenced not only by an block inoculative freezing (DeVries, ’82; Zachar- organism’s morphology and physiology, but also iassen and Husby, ’82; Duman et al., ’91). It is by factors determining duration of exposure to believed that turtles lack such proteins, as their and intimacy with ice and INA. The process blood exhibits no thermal hysteresis (Storey et al., becomes increasingly efficacious as temperature ’91; Costanzo et al., 2000b). However, it is possible falls below the tissue equilibrium freezing/melting that hatchlings harbor these proteins in concen- point (Salt, ’63; Packard and Packard, ’93a; trations too low to induce hysteresis (Ramlov Packard, ’97; Packard et al., ’97b). For example, et al., ’96) or solely in close association with the inoculation frequency for hatchling C. picta in

J. Exp. Zool. 342 J.P. COSTANZO ET AL.

100 0

75 -5 50 ND NE 25 (˚C) -10 NM c T MN Inoculation frequency (%) 0 -15 Indiana -12 -10 -8 -6 -4 -2 0 Nebraska Temperature (˚C) Fig. 9. Incidence of inoculative freezing in hatchling painted turtles (Chrysemys picta) as a function of ambient -20 temperature. Hatchlings were cooled individually inside 0 5 10 15 20 25 artificial nests consisting of a 473-mL glass jar containing Soil moisture content (%) frozen loamy sand and thus were exposed to ice and ice- nucleating agents (INA). Each point represents the percentage Fig. 10. Temperatures of crystallization (Tc) of individual of turtles in the sample that froze. The number of turtles in hatchling painted turtles (Chrysemys picta) cooled in contact each sample was as follows: North Dakota, n 5 32 (Packard with frozen nest soil as a function of soil moisture content and et al., ’97b); Nebraska, n 5 30 (Packard and Packard, ’97) or soil texture. Indiana turtles (C. p. marginata), hatched inside n 5 23 (Packard and Packard, 2003a); New Mexico, n 5 10 natural nests, were tested in a clayey soil collected from one of (Packard et al., 2002); and Minnesota, n 5 18 (Packard et al., the nests at Mount Zion Mill Pond, Fulton Co., IN; hence, the ’99a). The equation describing the line of best fit: y 5 À7.56x1 lower limit of supercooling is constrained by ice-nucleating 6.75 (r2 5 0.77). Exposure to temperatures below the equili- agents (INA). Nebraska turtles (C. p. bellii) hatched on brium melting/freezing point of turtle tissues lasted 9–11 d; vermiculite in the laboratory were tested in loamy sand exposure to the indicated target temperature lasted 1.5–9 d. collected from a nesting site in Garden Co., NE. All hatchlings Where reported, soil moisture was 16.4–25% (w/w) and water were cooled in artificial nests in which the soil enveloping the potential was À75 to À50 kPa. turtle was frozen. (Adapted from Costanzo et al., 2001c.) artificial nests increased more than 10-fold as but often peak when temperatures are lowest, temperature was reduced from À2.5 to À10.51C thus compounding freezing risk; consequently, (Fig. 9). The number of hatchlings inoculated mortality tends to be greater in winters that are inside thermostatic artificial nests increases over both cold and wet (Costanzo et al., ’98, 2004). The time, some individuals freezing many days after heightened inoculation susceptibility in wetter initial exposure to ice and INA (Packard and soils stems from a greater abundance of ice, and Packard, ’93a; Packard et al., ’99a). Apparently, hence the higher probability of contacting crystals; freezing risk varies with seemingly small differ- moisture also facilitates transmission of soil INA ences in exposure time, as the inoculation fre- to turtle tissues (Costanzo et al., 2000a, 2004). quency of hatchling C. picta held at À2.51Cina Laboratory studies with C. p. bellii (Costanzo frozen, loamy soil was 6% after 11 d (Packard et al., et al., ’98) and C. p. marginata (Costanzo et al., ’97b), but 39% after 14 d (Packard et al., ’99a). 2001c) cooled in native nesting soils showed that Notwithstanding the importance of temperature the relationship between soil moisture content and time, inoculative freezing is exquisitely and hatchling Tc is logarithmic. The nature of this sensitive to complex, indeterminable factors med- response permits inoculation under wet conditions iating intimacy between an organism and ice and at temperatures conducive to freezing survival INA in its immediate surroundings. Ambient (i.e., ZÀ41C); at lower ambient moisture levels, moisture has a potent influence on freezing risk the probability of inoculative freezing (and death in diverse organisms (Salt, ’63, ’66; Layne et al., due to uncontrolled ice formation) is greatly ’90; Lundheim and Zachariassen, ’93; Costanzo diminished (Fig. 10). Recall, however, that super- et al., ’95a, ’97, ’98). For hatchling turtles, cooling is ultimately constrained by the action of inoculation frequencies of specimens in simulated endogenous INA. hibernacula climb sharply with soil moisture Moisture abundance in terrestrial hibernacula content (Costanzo et al., ’95b, ’98, 2000a, 2001c; is governed by various environmental factors, Baker et al., 2003). Moisture levels in natural including climatic and seasonal precipitation hibernacula, such as C. picta nests, vary seasonally patterns, topography, and edaphics. Composition

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 343 and texture (particle-size distribution) of soil 80 strongly affect its drainage and moisture-retention 60 A. characteristics. Soils comprised of large particu- lates drain readily, whereas soils containing 40 organic matter or many fine particulates and content (%) 20 colloids tend to hold moisture (Fig. 11). However, Saturation water 0 these entities also tend to adsorb water, lowering its activity and limiting ice formation (Forge and 0 MacGuidwin, ’92; Costanzo et al., ’97). Moisture abundance in natural hibernacula has been -250

examined in the context of inoculative freezing (kPa) (Costanzo et al., ’98, 2001c, 2004; Baker et al.,

2003) and varies geographically as well as among Water potential B. hibernacula in the same general locale. Indeed, -500 soil moisture content—and thus freezing risk— vary such that, during any given chilling episode, 0 turtles in adjacent nests may be either frozen or -4

supercooled (Costanzo et al., 2004). Differential (˚C) -8 freezing risk may account for gross differences in c T C. survival rates of hatchlings overwintering in -12 neighboring hibernacula (Packard et al., ’89, -16 ’97a; DePari, ’96; Packard and Packard, ’97). Soil Soil + Soil + Findings of several laboratory studies underscore Clay Organic the importance of soil texture as an environmental Fig. 11. Relationship between substratum composition determinant of freezing risk. Inoculation resistance and (A) moisture-holding capacity, (B) water potential, and in hatchling C. picta cooledinamatrixofnative (C) resistance to inoculative freezing in hatchling painted turtles (Chrysemys picta bellii). Experimental substrata were sandy soil was markedly improved after adding a unadulterated soil collected at a nesting area in Garden Co., small quantity of either clay or organic matter (peat) NE (sand, 91%; silt, 1%; clay, 8% w/w); nesting soil augmented to the soil (Fig. 11). Packard and Packard (’97) with 10% (w/w) clay; and nesting soil augmented with 10% found that inoculation frequencies of hatchling C. (organic) peat. All substrata were hydrated to 7.5% (w/w). picta chilled to À71C in artificial nest hibernacula Inoculative freezing resistance was determined as the mean (7SE) temperature of crystalization (Tc) of hatchling turtles were markedly higher with ‘‘loamy sand’’ (80%) as (n 5 4À5) cooled in contact with the frozen substratum. compared with ‘‘clayey soil’’ (33%), even though Turtles were hatched and reared on (INA-free) vermiculite, water potential of the two substrata was similar. In progressively acclimated from 22 to 41C, and tested in winter. other studies of hatchling C. picta,thetemperature Dashed line indicates the lower limit of supercooling (–14.61C) at which half of an experimental group was determined for turtles cooled in the absence of substratum. inoculated was À6.51C if the substratum was (Adapted from Costanzo et al., ’98.) ‘‘loamy sand’’ (Packard et al., ’97b), but was À91C if turtles were exposed to ‘‘clayey soil’’ (Packard and 1995–1996 (90% vs. 36%). Because the soils in this Packard, ’93a). Studies by Costanzo et al. (2001c, area contain little clay or organic matter, indigen- 2004) revealed that marked variation in inoculation ous turtles may freeze more readily than turtles resistance of C. picta hatchlings from different endemic to areas with different soil types, other populations was due to differences in characteristics factors being equal (Costanzo et al., 2004). of their respective indigenous soils (Fig. 10). Although the relative importance of soil charac- The contention that soil composition influences teristics to survivorship in field studies is inoculation susceptibility, and hence winter survi- confounded by variation in hibernaculum val, of hatchling turtles draws support from field temperature, these field observations, together studies. Analyzing the winter mortality data with available laboratory data, suggest that winter reported by Packard (’97) and Packard et al. survival rates are linked to edaphics on both local (’97a) for hatchling C. picta, Costanzo et al. (’98) and regional scales. This situation raises interest- noted that the survival of hatchlings occupying ing questions about the maternal role of choosing nests constructed in ‘‘loamy sand’’ exceeded that the nest site. Selecting a clayey soil over a sandy of turtles hibernating in ‘‘fine sand’’ in two one is not only beneficial to developing embryos consecutive years: 1994–1995 (94% vs. 65%) and (Legler, ’54; Ratterman and Ackerman, ’89), but

J. Exp. Zool. 344 J.P. COSTANZO ET AL. in principal also promotes freeze avoidance. et al., ’95; Lee et al., ’95; Vernon and Vannier, Perhaps fortuitously, in some populations, C. picta 2002). As we earlier discussed, efficacy by which emerge in fall from nests constructed in sand, but environmental ice and INA seed the freezing of overwinter in nests constructed in organic or body fluids depends on a given organism’s anato- loamy soils (DePari, ’96). mical and morphological constitution, ambient Investigations of inoculative freezing in hatchling temperature, and physico-chemical characteristics turtles most commonly focus on ice, the supreme of the nucleating environment. Assuming that nucleator. Nevertheless, various INA, including inoculation resistance is a heritable trait on which both inorganic and organic entities, naturally occur natural selection can act, one might suppose that in the winter microenvironment and can also freeze avoidance varies among species, being trigger freezing. Two studies have examined the particularly well developed in freeze-avoiding inoculation susceptibility of C. picta by cooling terrestrial hibernators. In testing this idea, hatchlings in a matrix of nesting soil that contained Costanzo et al. (2001b) compared inoculative INA but no ice (Costanzo et al., ’98, 2000a). The freezing resistance among hatchlings of eight taxa INA sharply reduced supercooling in some turtles. of terrestrial and aquatic turtles collected from Moistening the soil slightly induced inoculation in a both northern and southern locales. Using stan- higher percentage of the sample, apparently by dardized trials, in which hatchlings were cooled in causing more soil to become lodged in axillary and a matrix of frozen soil, mean Tc values ranged inguinal pockets and nuchal skin folds. Thus, from À0.8 to À13.61C and tended to be lower environmental moisture not only provides a direct among terrestrial hibernators. Other studies, source of ice crystals, but also facilitates commu- which examined inoculation resistance in indivi- nion of soil INA with body surfaces. dual species, including C. picta (Packard The literature hints of an important, as yet and Packard, ’93a; Costanzo et al., ’98, 2000a), unexplored, interaction between environmental C. serpentina (Packard and Packard, ’90; Packard INA and inoculation efficacy in hatchling turtles. et al., ’93), G. geographica (Baker et al., Most studies of inoculative freezing have, for the 2003), E. blandingii (Packard et al., 2000; sake of convenience, used specimens produced by Dinkelacker et al., 2004), and T. scripta (Packard hatching eggs in an incubation medium of (INA- et al., ’97c), generally support the maxim that free) vermiculite. However, hatchlings exposed to this trait is better developed in terrestrial natural INA are substantially more susceptible to hibernators. inoculation. For example, Packard and Janzen The foregoing conclusion notwithstanding, an (’96) found that C. picta hatched inside natural especially poor inoculation resistance is found in nests incurred a 44% inoculation frequency when some terrestrial hibernators whose primary cold- cooled to À4.61C inside artificial nests containing hardiness strategy seems to be freeze tolerance. clayey soil. The impact of INA contamination is Lacking endogenous INA or specialized ice-nucle- evident even in hatchlings produced from eggs laid ating proteins, inoculation by environmental ice or in natural nests but subsequently hatched on INA ensures that freezing commences at high vermiculite. Using hatchling C. picta produced in temperatures and that ice growth is controlled and this manner, the inoculation frequency for speci- proceeds slowly, a condition requisite to freezing mens kept in loamy sand for 7 d at À21C was 26% survival (Lee and Costanzo, ’98). This system in one study (Packard and Packard, 2003b) and would benefit T. ornata (Costanzo et al., 2001b) 57% in another (Willard et al., 2000). The and M. terrapin (Baker et al., 2006), freeze- propensity for inoculation at such a high tempera- tolerant species that commonly overwinter in ture illustrates the precarious nature of super- sandy soil, potentially within reach of frost. cooling as a freeze-avoidance strategy, even in Hatchling E. blandingii probably also belong to species (such as C. picta) that have a well- this group, as they tolerate somatic freezing developed capacity to resist inoculative freezing. (Packard et al., ’99b; Dinkelacker et al., 2004; Costanzo et al., 2006) and poorly resist inoculative freezing (Packard et al., 2000; Dinkelacker et al., Taxonomic and ecological associations 2004). The case with G. geographica is sharply Abundance, activity, and inoculation efficacy of contrasting. These hatchlings hibernate terrest- various endogenous and exogenous INA are rially but are freeze intolerant; they exhibit an influential factors in the evolution of cold-hardi- extraordinary ability to resist inoculative freezing ness strategies of ectothermic animals (Duman (Baker et al., 2003).

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 345

Although inoculation resistance varies taxono- turtles, occurs primarily via transcutaneous mically, it remains equivocal as to whether this routes (Bentley and Schmidt-Nielsen, ’66). Indeed, trait varies among populations of conspecifics. One Costanzo et al. (2001b) found that the rate of study found marked differences in inoculation water loss was an excellent predictor of inocula- resistance in C. p. bellii from Nebraska and tion resistance among several taxa of terrestrial C. p. marginata from Indiana; however, the and aquatic hibernators, data from the exceptional disparity could be explained by characteristics of C. picta excluded. This finding implies that both the soil indigenous to each locale, rather than processes are mediated by the same morphological phenotypic variation (Costanzo et al., 2001c). and/or physiological characters; it also explains Another study indicated that resistance exhibited why inoculation resistance is greater among by Nebraskan C. p. bellii far exceeded that of terrestrial hibernators. Although the authors did Arkansan C. p. dorsalis, a primarily southern not make direct measurements of cutaneous sur- subspecies (Costanzo et al., 2001b); however, this face area, they suggested that the extensive finding remains tentative because the sample plastron found in hatchlings of species that over- contained relatively few hatchlings. Directly com- winter terrestrially limits exposure of the skin and paring results from different studies is not appro- mucous membranes of the cloaca to ice and INA. priate unless identical experimental protocols Taxonomic variation in inoculation resistance were used. That caveat notwithstanding, the may also derive from differences in the qualities of literature suggests that inoculation resistance in skin that influence its barrier properties. Willard C. picta varies little among populations in Nebras- et al. (2000) found a discrete lipid layer in the ka (Packard and Packard, ’93a,b, ’95a), Minnesota epidermis of hatchling C. picta, but no such layer (Packard et al., ’99a), North Dakota (Packard in skin of T. scripta, E. blandingii and C. et al., ’97b), Illinois (Packard and Janzen, ’96), serpentina, species that exhibit poor resistance. A and even New Mexico, near the southern limit of recent study (J. Costanzo and J. Jack, unpublished the species’’ distribution (Packard et al., 2002). data) determined that inoculation resistance and Referencing findings for the New Mexico popu- skin lipid density were positively correlated lation, Packard and Packard (2003c) argued that (r2 5 0.62; P 5 0.035) in hatchlings of seven taxa inoculation resistance (and profound supercooling of North American turtles, affirming the general capacity) are not maintained by selection for cold importance of lipids in inoculation resistance hardiness because these traits would have no (Fig. 12). Lipid density could underlie the strong value in such a mild environment. However, link between inoculation resistance and dehydra- climatological records from a U.S. National tion resistance in hatchling turtles (Costanzo Weather Service station at Alamogordo, NM, et al., 2001b). 200 km SSW of the source of the animals How any hatchling could avoid freezing while (Soccorro County), indicate that frost descends in overwintering in intimate contact ice and INA, if the soil column to 51 cm, and certainly would indeed they do, remains something of an enigma. penetrate turtle nests. Furthermore, contrary to One possible answer is that, in nature, transmis- the claim (Packard and Packard, 2003c) that this sion of ice and INA to the body fluids is imperfectly population was derived from ancestors that had efficient. This notion is supported by the skewed never confronted cold stress, these turtles appar- distributions of Tc values sometimes obtained in ently are descended from northern populations inoculation resistance trials (Costanzo et al., ’98, (Starkey et al., 2003); thus, their cold-hardiness 2000a). Another possibility is that turtles are to traits likely predate glacial expansion (Holman some extent buffered from ice and INA by egg- and Andrews, ’94). Carefully controlled studies are shells or other debris within the nest cavity. needed to examine populational variation in this Behavioral avoidance of these agents may also be important trait, especially in terrestrial hiberna- important. Given that limb movements apparently tors other than C. picta. permit soil to infiltrate the axillary and inguinal What is the basis for taxonomic variation in pockets and thereby contact skin that is suscep- inoculation resistance? If skin is a primary barrier tible to transmitting ice and INA (Packard and to ice and INA transmission, then, all else being Packard, ’97), a hatchling might limit its contact equal, species (and individuals) having more with these agents simply by retracting its head exposed skin should be at greater risk of freezing and limbs within the shell (Packard and Packard, through contact inoculation. This argument is also ’95a). Forming a tight cluster with one’s siblings germane to water loss, which, in hibernating (e.g., Breitenbach et al., ’84; Tucker, ’97; Andreas

J. Exp. Zool. 346 J.P. COSTANZO ET AL.

0 Eb -2 Cs Mt -4 Ts (˚C) -6 c T Gg -8 Cpm Loamy sand Clay loam

-10 Cpb Position: Peripheral Central Peripheral Central -12 Survival: 0/9 0/4 0/9 4/4 Adherent soil (mg): 576 ± 94 548 ± 121 142 ± 25 55 ± 17 25 50 75 100 Lipid density (mg/g) Fig. 13. Relationship between hatchling position in the hibernaculum, exposure to environmental ice and INA, and Fig. 12. Relationship between skin lipid density and freezing-induced mortality. A group of hatchling painted capacity to resist inoculative freezing in hatchlings of seven turtles (Chrysemys picta) was exposed to À51C inside an taxa of North American turtles: Chrysemys picta bellii (Cpb), artificial nest constructed of 1.5 L damp (5%, w/w) loamy sand Chrysemys picta marginata (Cpm), Chelydra serpentina (Cs), or damp (10%, w/w) clay loam; each group contained nine Emydoidea blandingii (Eb), Graptemys geographica (Gg), ‘‘peripheral’’ turtles arranged around a core of four ‘‘central’’ Trachemys scripta (Ts), and Malaclemys terrapin (Mt). The turtles. Soil was permitted to freeze and nests were held at equation describing the line of best fit: y 5 À0.17614.40 À51C for 7 d. After nests were thawed at 41C, hatchlings were 2 (r 5 0.73). Turtles were hatched and reared on (INA-free) examined for viability and the amount of soil adhering to each vermiculite, progressively acclimated from 22 to 41C, and turtle was determined. In the clay loam nest, central turtles tested in winter. Inoculative freezing resistance was deter- had less adherent soil (mean7SE) and higher survivorship 7 mined as the mean ( SE) temperature of crystalization (Tc) than peripheral turtles. This relationship was not evident in of hatchling turtles (n 5 4À5) cooled in contact with frozen the nest constructed of loamy sand, possibly because soil substratum (loamy sand; water content, 4% w/w). Lipid infiltrated this nest. (J. Costanzo and J. Larson, unpublished 7 concentrations in dried skin samples (mean SE; n 5 5À8) data.) were determined using a chloroform/methanol extraction procedure. (J. Costanzo and J. Jack, unpublished data.) and Paul, ’98) would further reduce exposure, with life in an extreme environment (Tinkle, ’61; particularly for individuals in the group’s center Gemmell, ’70; Moll, ’73; Iverson et al., ’93; Iverson (Costanzo et al., 2000a, 2001c, 2004; Nagle et al., and Smith, ’93). This ‘‘protected siblings’’ hypoth- 2000). esis suggests a mechanism on which such selection The hypothesis that winter mortality depends might operate, at least for freeze-avoiding, terres- on such factors as location within the hibernacu- trial hibernators. We note, however, that it does lum poses some interesting questions about not preclude the possibility that large clutches also certain life-history traits of northern turtles. If are beneficial, albeit through some other mechan- one assumes the primary cause of death is freezing ism, in the more northern populations of species initiated by contact with ice and INA, and that a whose hatchlings hibernate aquatically. hatchling gains the greatest protection by being Challenges imposed by the harsh winter envir- shielded by its siblings, then, all else being equal, a onment may shape other life-history traits, such relatively large clutch should have a survival as propagule size. Contrary to the paradigm that advantage. Costanzo et al. (2004) found that larger individuals have a survival advantage, ice- C. picta nests containing at least nine hatchlings nucleation theory predicts that a small body had markedly higher survival and lower incidence should favor survival in freeze-avoiding species. of total winterkill compared with nests with fewer Ice nucleation occurs once sufficient water mole- siblings. Laboratory experiments also support the cules have organized to form an ice embryo idea that a hatchling’s location within the hiber- (Rasmussen and MacKenzie, ’73) and, because naculum and proximity to siblings moderates its the probability of embryo formation increases exposure to ice and INA, and hence survival of with fluid volume (Bigg, ’53; Langham and Mason, chilling episodes (Fig. 13). Additional research is ’58; Vali, ’95), small organisms tend to supercool needed to test these relationships rigorously, but more extensively than large ones. Furthermore, the concept is interesting. The hypothesis argues small animals, by virtue of their limited body for a strong directional selection for large clutches surface, are less likely to incur inoculative freezing in turtles living at high latitudes and in cold (Costanzo et al., 2001b). habitats, and, in fact, northern populations of Theory aside, the ‘‘smaller is better’’ concept turtles do produce relatively large clutches, pur- advanced by Costanzo et al. (2004) has not as portedly to offset a higher mortality associated yet received rigorous empirical testing. One

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 347 laboratory study determined that incidence of on C. picta; evidence for freeze tolerance in this (apparent) freezing mortality was in fact lower in species abounds (Storey et al., ’88; Paukstis et al., lighter (smaller?) hatchling C. picta (Costanzo ’89; Claussen and Zani, ’91; Rubinsky et al., ’94; et al., 2001c); similar results have been reported Costanzo et al., ’95b; Packard et al., ’99b; Hem- for invertebrates (Holmstrup et al., 2007). On the mings and Storey, 2000; Willard et al., 2000; other hand, in a Michigan field study of the same Packard and Packard, 2004). species, Nagle et al. (2000) found no association In common usage, freeze tolerance denotes an between mean body size and the proportion of organism’s ability to survive the freezing and hatchlings within individual nests surviving win- thawing of a biologically significant quantity of ter. However, this analysis included data from body water under thermal and temporal regimes only eight nests in a single winter, and the authors relevant to the organism’s life history (Baust, ’91). used body mass, a labile metric, as proxy for body Unfortunately, the literature is fraught with size for both live and dead turtles. A large-scale reports that have failed to distinguish responses field study would be needed to thoroughly address that are adaptive in the sense of evolutionary this question, but some ancillary evidence sug- process (e.g., Bennett, ’97) from those of recovery gests that such an effort may prove rewarding. from superficial injury (Storey and Storey, ’96). Some northern turtles potentially can lay eggs Indeed, some reports of freeze tolerance in larger than those actually produced, so selection invertebrates (Paukstis et al., ’96; Ansart et al., apparently favors production of hatchlings of 2001; Cook, 2004), amphibians (Pasanen and smaller than maximal size (Tucker, 2000b). Karhapa¨a¨, ’97; Croes and Thomas, 2000; Steiner Costanzo et al. (2004) conjectured that size- et al., 2000), and reptiles (Claussen et al., ’90; specific freezing risk could partly explain why Andersson and Johansson, 2001; Burke et al., C. picta populations in severely cold regions tend 2002) are based on experiments in which subjects to produce relatively small offspring. Annual have survived a measure of internal ice formation, reproductive output of C. picta is more or less but the tolerance is limited to superficial freezing constant (relative to female body size) among over very brief exposures to relatively high populations at all latitudes and is partitioned temperatures. Compounding this problem, some within clutches via a trade-off between offspring authors have endeavored to catalogue various number and size (Iverson, ’92; Iverson and Smith, degrees of cold hardiness, including subcategories ’93). Thus, any female that produces relatively of freeze tolerance (Bale, ’93, ’96; Sinclair, ’99; small hatchlings also produces many hatchlings, Nedved, 2000; Vernon and Vannier, 2002; Voitur- this combination compounding the increase in her on et al., 2003; Hawes et al., 2006). These fitness. The possibility that the winter environ- constructs, which have generated a litany of ment is a selective force shaping the life-history confounding idioms, including the oxymoron traits of turtles merits further consideration. ‘‘partial freeze tolerance,’’ have done little to advance our understanding of freeze tolerance as a natural survival adaptation. Some authors hold Freeze tolerance in high regard the ability to revive after attaining Natural freeze tolerance is a complex, multi- an equilibrium ice content, irrespective of tem- faceted adaptation derived from a coordinated perature (Sinclair, ’99; Voituron et al., 2003); this suite of responses to the formation and melting of criterion also lacks relevance to life history. ice in tissues. It has evolved independently many Others have suggested recovery after 24 h of times in diverse organisms. Long known in freezing at –2.51C as the minimum basis for arthropods and other invertebrates, it was first designating freeze tolerance in ectothermic verte- recognized as a winter survival strategy among brates (Storey and Storey, ’92). These conditions, vertebrates only relatively recently (Schmid, ’82). which generally result in the freezing of 450% of These seminal investigations focused on woodland the body water (Claussen and Costanzo, ’90), frogs, but it was soon discovered that certain undoubtedly are lethal to organisms lacking reptiles also tolerated tissue freezing (Costanzo specific adaptations to somatic freezing. Adopting et al., ’88; Storey et al., ’88). Currently, freeze this criterion may provide an objective, discrimi- tolerance is known in anuran and urodele amphi- nating test that can be applied broadly across taxa, bians, both suborders of squamates, and turtles but nevertheless fails to consider life-history (Ramlov, 2000; Baker et al., 2004; Storey, 2006). variation amongst cold-hardy species. In our view, Amongst hatchling turtles, most work has focused it is important to consider the demonstrated

J. Exp. Zool. 348 J.P. COSTANZO ET AL. capacity for tolerating somatic freezing in both whether this reduced ability is simply vestigial or physiological and ecological contexts. We recog- if it enhances survival under certain circum- nize, however, the current paucity of ecological stances. Adult C. picta usually hibernate aqua- information for all but a few species. tically and, thus, are buffered from extreme cold Freeze tolerance is usually regarded as a winter (Ultsch, ’89, 2006); however, they become exposed survival strategy, but in temperate locales some to frost at least occasionally (Christiansen and organisms also are at risk of freezing early and Bickham, ’89; St. Clair and Gregory, ’90). late in the activity season. Instances of mass mortality of reptiles have been attributed to Methodological considerations unseasonable frosts occurring in autumn and spring (Spellerberg, ’76; Gregory, ’82). Freeze Laboratory studies must be carefully designed if tolerance presumably protects organisms from they are to properly assess an organism’s ability to this hazard and conceivably can extend behavioral tolerate freezing/thawing under ecologically rele- activity into the cooler months. For example, in vant conditions. It is essential to control ice the northern part of its range, T. carolina some- nucleation and to ensure that freezing proceeds times remains abroad until early November, well slowly; spontaneous nucleation of deeply super- after the first autumnal frost (Claussen et al., ’91), cooled tissues is deleterious, even in profoundly perhaps owing to its tolerance to somatic freezing freeze-tolerant species, because it promotes (Costanzo and Claussen, ’90; Costanzo et al., ’93; (lethal) intracellular freezing and accelerates ice Storey et al., ’93). Similarly, the garter snake accumulation, leaving cells insufficient time to (Thamnophis sirtalis) encounters frost during adapt to volume change and hastening circulatory spring and fall owing to its habit of emerging arrest (Storey and Storey, ’92). For example, in from hibernation early and retiring late (Vincent the wood frog, rapid freezing is injurious (Costan- and Secoy, ’78). Its ability to survive the freezing zo et al., ’91a) because, among other things, it of at least 35% of its body water for at least 48 h hampers water flux and cryoprotectant mobiliza- (Costanzo et al., ’88) is more important at these tion (Costanzo et al., ’91b, ’92). It is also important times than during winter, as this species usually to thaw the specimens slowly, to hold them at hibernates in thermally buffered sites (Costanzo, appropriate temperatures, and to allow adequate ’86; Lutterschmidt et al., 2006). A tolerance for time for them to revive before assessing survival. somatic freezing, even if limited, could similarly In practice, ice nucleation is readily induced by protect hatchling turtles that encounter frost bringing ice or frozen substratum into contact while abroad (Packard et al., 2000). with the wetted surface of a specimen that has Organisms whose overwintering habitats vary been chilled slightly below its equilibrium freez- with developmental stage commonly exhibit an ing/melting point (Attaway et al., ’98). Once ontogenetic shift in cold-hardiness strategy (Bou- freezing begins, the turtle can be slowly cooled to chard et al., 2006; Rinehart et al., 2006). It is the prescribed temperature. This is easily accom- possible for freeze tolerance to be expressed in plished by immersing the vessel containing the neonatal turtles but lacking in older individuals turtle in a refrigerated bath. It is important to (or conversely), although no such case has been hold the animal at the target temperature long discovered. Hatchling and adult box turtles enough such that ice content attains equilibrium; (T. carolina and T. ornata) are freeze-tolerant, as however, investigators should choose an exposure both life stages overwinter terrestrially, poten- duration that is sensible in the context of the tially within the frost zone (Costanzo and Claus- species’ ecology. Finally, the specimen is thawed sen, ’90; Claussen et al., ’91; Costanzo et al., ’95b; by permitting it to gradually warm to 0–41C. Converse et al., 2002; Bernstein and Black, 2005). Survival criteria used in studies of freeze Adult mud turtles (K. flavescens) pass the winter tolerance have not been standardized. Commonly, deep in the soil column (Tuma, ’93, 2006) and are viability is inferred from the subject’s physical intolerant of freezing, as are hatchlings (Costanzo appearance, passive behaviors, and response to et al., ’95b). Freeze tolerance is well developed in mechanical stimulation. Although convenient, hatchling C. picta (Storey et al., ’88; Willard et al., such endpoints are inherently subjective and could 2000; Packard and Packard, 2004) and, although lead to bias. Costanzo et al. (2006) reported that the adults can survive brief freezing of at least plasma lactate dehydrogenase (LDH) activity was 15% of their body water (Johnson, ’90; Claussen a reasonably objective marker of cryoinjury in and Zani, ’91; Claussen and Kim, ’93), it is unclear hatchling turtles. The premise of this assay is that

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 349 freezing perturbs cell membranes, permitting developmental stage. Critical temperatures have cytosolic contents to leak into the interstitium; not been defined for hatchling turtles, but see- thus, cryoinjury can be indexed by circulating mingly are near À41C. Tolerable ice contents of levels of this easily measured enzyme. 40–55% of total body water have been measured in Another concern regarding survival assessments hatchling emydids frozen at an equilibrium tem- is that few investigators monitor the post-trial perature of À2.51C (Storey et al., ’88; Churchill vitality of frozen/thawed hatchlings for very long. and Storey, ’92a,b). Evaluations made shortly after freezing/thawing As a general rule, freezing mortality increases do not necessarily reveal injury or dysfunction with time spent in the frozen state (e.g., Sømme, manifested later in life (Baust and Rojas, ’85; ’96; Ansart and Vernon, 2003). Although this Ansart et al., 2002; Layne and Peffer, 2006). On principle has been amply documented in hatchling the other hand, freeze tolerance is underestimated turtles (Churchill and Storey, ’92a,b; Attaway or entirely overlooked if survival is assessed before et al., ’98), physiological factors governing freeze animals have fully recovered. Indeed, although endurance are not well known. Some evidence some authors have reported heavy freezing mor- suggests that lactate accumulation and/or energy tality in hatchlings they examined only 24 h after depletion are involved. As ice propagates within thawing (Packard et al., ’97c, ’99b; Packard and the body, cardiac activity diminishes and ulti- Packard, 2003a), several studies indicate that mately ceases (Rubinsky et al., ’94), rendering recovery requires from several days to perhaps a tissues ischemic and reliant on anaerobic glyco- week, depending on the severity of the stress lysis to meet metabolic demands. Isolation of cells (Costanzo et al., 2004, 2006; Baker et al., 2006). from nutrient delivery and waste management Because investigators do not always follow appro- systems may contribute to their mortality (see priate experimental practices, discretion should be Belkin, ’63). Some authors have speculated that used when interpreting results of freeze-tolerance accumulated lactate, the main end-product of trials. glycolysis, contributes to freezing mortality (Pack- ard and Packard, 2004). Support for this idea comes from the recent finding that blood lactate Limits of freeze tolerance and injury from experimental freezing were The severity of freezing/thawing stress—and correlated among the hatchlings of various species potential for cryoinjury and death—varies with (Costanzo et al., 2006). Time-dependent mortality factors such as minimum temperature, exposure may also reflect depletion of the metabolic fuels duration, and rates of cooling and warming. necessary to meet cellular demands for ATP Because the outcome of a freezing/thawing trial (Layne et al., ’98). Yet another explanation is depends, in part, on the experimental conditions, that mortality coincides with onset of physical determining an organism’s capacity for freeze damage due to recrystallization, the gradual tolerance is not a simple matter. Furthermore, enlargement of small, unstable ice crystals at freeze tolerance can be considered both in terms of the expense of others (Knight and Duman, ’86). minimum tolerable temperature, which has im- Relatively little study has been devoted to defining plications for the amount of ice being formed and limits of freeze endurance. Hatchling C. picta associated cellular dehydration and osmotic/ionic reportedly can survive freezing for 6–7 d at À21C perturbation tolerated, as well as the maximum (Willard et al., 2000; Packard and Packard, duration, which reflects tolerance of additional 2003b) and at least 11 d at À2.51C (Churchill stresses such as energy depletion and hypoxia and Storey, ’92a). Several hatchling C. picta tolerance. believed to be frozen recovered after spending Generally, freeze-tolerant animals can survive 1 month inside an explanted natural nest kept at the freezing of up to two-thirds of their body À2.51C inside a laboratory incubator (Costanzo water; formation of additional ice causes excessive et al., 2001c). cell dehydration and irreversible damage to cell membranes and organelles (Storey and Storey, Adaptive mechanisms of freeze tolerance ’88). Because ice content increases as an organism cools, the survival limit can be associated with a To survive even a single, mild freezing episode, critical minimum temperature (Claussen and organisms must withstand myriad stresses simul- Costanzo, ’90), although this value may vary taneously manifested at multiple levels of biologi- considerably with the physiological state and the cal organization. Such stresses derive from the

J. Exp. Zool. 350 J.P. COSTANZO ET AL. formation and melting of ice in extracellular isms commonly accumulate cryoprotectants, compartments, osmotic withdrawal of water from organic osmolytes that reduce ice formation and cells, and attendant osmotic and ionic effects on the attendant cell dehydration, and also preserve macromolecules, membranes, and cell homeostasis the structural integrity of proteins, membranes, (Steponkus, ’84; Storey and Storey, ’88; Mazur, and other cell structures (Storey, ’97; Mazur, 2004). As the extracellular water freezes, a growing 2004; Yancey, 2005). Principles of natural freeze osmotic tension withdraws cytoplasmic water, tolerance are reasonably well known, especially as thereby causing cells to shrink. The resulting they pertain to invertebrates and anurans. Freeze hyperconcentration of solutes, together with the tolerance is also well developed in certain reptiles, release of lysosomal proteases, can denature but the underlying molecular and physiological intracellular proteins. mechanisms remain incompletely understood Freezing impacts the physical state of lipids and (Costanzo et al., 2006; Storey, 2006). can alter membrane fluidity and composition, impairing various membrane-mediated processes, Seasonal variation in freeze tolerance such as transport and metabolism (Steponkus, ’84; Capacity for freeze tolerance varies in association Bischof et al., 2002). As tissues freeze, the with cold hardening or winter acclimatization in expanding ice front can force cells apart and many temperate ectotherms (Aarset, ’82; Murphy, damage capillaries. Mechanical and osmotic stres- ’83; Lee, ’91), including hatchling turtles. Field ses can combine to irreversibly disrupt tissue investigations of C. p. bellii demonstrated that, in architecture (Muldrew et al., 2004) and this result, two consecutive years of study, freeze tolerance was if not directly injurious, may ultimately cause weakly expressed in late summer, but had improved ischemic necrosis or apoptotic destruction of markedly during winter and persisted at least into tissues (Hoffman and Bischof, 2005). April (Costanzo et al., 2004). Similarly, Churchill At the organ system level, pervasion of ice and Storey (’92a) reported that hatchling C. p. throughout the body eventually arrests breathing marginata, collected from natural nests at winter’s and cardiac function, rendering tissues ischemic end, were substantially more freeze tolerant than a and hypoxic (Churchill and Storey, ’91, ’92a; November cohort; thus, cold hardening is required Packard and Packard, 2004). Consequently, energy for the full expression of freeze tolerance. Labora- stores are depleted and oxidative stress can occur tory experiments also suggest that a period of cold when oxygen levels are restored on thawing; acclimation improves freeze tolerance in hatchling membrane damage resulting from oxidative stress C. picta (Packard and Packard, 2003b). Factors is a potentially significant cause of freezing-related regulating this change are as yet unknown, but may injury (Murphy, ’83). Nervous tissues are particu- relate to metabolic and osmoregulatory shifts larly sensitive to freezing stress (Cameron, ’30) accompanying seasonal cold hardening. For exam- and, accordingly, cryoinjury manifested as neuro- ple, enhanced freeze tolerance in winter animals behavioral dysfunction (e.g., lethargy, impassive- could derive from accumulationofurea(Costanzo ness to tactile stimulation, absence of the righting et al., 2000b, 2006), an efficacious cryoprotectant in response) varies with severity of the freezing some species (Costanzo and Lee, 2005). Seasonal exposure (Costanzo et al., ’95a; Attaway et al., variation in freeze tolerance could also reflect shifts ’98; Burke et al., 2002; Dinkelacker et al., 2004). in plasma membrane composition, structure, and Freeze-tolerant organisms employ a suite of function that render this sensitive structure more coordinated molecular and physiological responses tolerant to freezing-induced stresses (Orvar et al., that minimize injury caused directly or indirectly 2000). Membrane remodeling probably occurs in by the formation and melting of ice (for reviews, hatchling turtles during preparation for winter. see Storey and Storey, ’88; Storey, ’90a). Some Notably, an increase in the membrane cholesterol of these protective mechanisms probably origi- level, which is known to stabilize membranes during nated from fundamental responses to osmotic cooling (Drobnis et al., ’93), has been observed in and hypoxic stress. For many species, adaptive hatchling C. picta undergoing cold acclimation strategies of freeze tolerance include control of the (Fig. 14). ice-nucleation event, management of the redis- tribution of water and solutes, hypometabolism, Ice-nucleating agents upregulation of antioxidant defense systems, and expression of genes involved in homeostasis and Some freeze-tolerant organisms synthesize INA somatic repair. In addition, freeze-tolerant organ- that initiate and control the freezing process,

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 351

10 Costanzo, ’98). Known classes of endogenously Erythrocyte c produced INA include inorganic crystals and the 8 more common ice-nucleating proteins. An early report of ice-nucleating activity in the range of À7 to À81C in the blood of hatchling C. picta seemed 6 to indicate that endogenous INA could play a role a in turtle cold hardiness (Storey et al., ’91). 4 However, subsequent results argue that the b activity was an artifact of sample contamination with exogenous INA (Costanzo et al., 2000b). 2 In the latter study, C. picta blood exhibited ice-nucleating activity if turtles were hatched 0 and reared on nesting soil containing native INA, but not if they were reared on vermiculite, 10 a substratum lacking soluble INA. The ability of Enterocyte vermiculite-reared C. picta to supercool to nearly 8 À201C (Costanzo et al., 2000a) is perhaps the best a a a evidence that endogenous INA are not used to

Cholesterol (µmol/mg protein) 6 regulate ice nucleation. Rather, inoculation by environmental ice and INA serves this critical function (Costanzo et al., ’95b). 4 Antifreeze agents 2 AFPs and glycoproteins, more aptly termed ‘‘ice- structuring proteins’’ (Clarke et al., 2002), are 0 best known for their role in arresting the growth 20 10 4 of ice and preventing wholesale tissue freezing. Acclimation temperature (˚C) However, in some freeze-tolerant organisms, these Fig. 14. Cholesterol content in plasma membranes of specialized agents help prevent recrystallization, a erythrocytes and enterocytes of hatchling painted turtles thermodynamic phenomenon in which some ice (Chrysemys picta) during seasonal cold hardening. Eggs were crystals gradually enlarge within tissues at the collected from gravid females at a nesting area in Garden Co., expense of others (Knight and Duman, ’86; NE, and incubated in the laboratory at 291C until hatching in August. Turtles were reared on vermiculite, kept in Loomis, ’91; Tursman et al., ’94; Ramlov et al., darkness, and progressively acclimated to 41C; experiments ’96; Duman et al., 2004). If unchecked, growing were conducted on 15 October with 201C-acclimated hatchl- crystals could irreparably damage tissue architec- ings (n 5 9), 15 November with 151C-acclimated hatchlings ture and disrupt intercellular communication (n 5 9), and 15 December with 41C-acclimated hatchlings systems. Turtle blood lacks thermal hysteresis (n 5 10). Erythrocytes and a 10-mm section of small intestine were harvested from euthanized turtles and randomly (Storey et al., ’91; Costanzo et al., 2000b), assigned to one of three separate pools, each containing tissue suggesting a lack of AFPs, but whether or not from 3–4 individuals. Membranes were isolated by homo- these agents are actually present (and effective in genizing and sonicating samples in buffer (0.15 M NaCl, preventing recrystallization) in concentrations too 3 mM MgCl2, and 0.01 M Tris-HCl, pH 7.2) and centrifuging low to be detected is unknown. (500g, 5 min) the suspension. Membranes in the resultant suspension were collected by centrifugation (25,000g, 15 min), washed in buffer containing 0.1 M choline chloride, and again Bound water centrifuged. The membrane pellet was used in spectrophoto- metric assays for cholesterol (No. 234-60, Diagnostic Chemi- The equilibrium freezing/melting point of water cals Limited, Oxford, CN) and protein (Bio-Rad, Hercules, CA) is lowered not only by the colligative effects of assays. Means (7SE; n 5 3) identified by different letters were dissolved solutes but also by the hydration of statistically distinguishable (ANOVA; Student–Newman–- macromolecules, membranes, and subcellular Kuels, Po0.05). (M. Polin and J. Costanzo, unpublished data.) structures. Water molecules proximal to hydro- phylic surfaces are subject to attractive forces that guarding against damage from rapid, uncontrolled greatly reduce their potential energy (Wolfe et al., ice growth in deeply supercooled tissues (Storey 2002). Known as ‘‘bound water,’’ this fraction has and Storey, ’88; Duman et al., ’91; Lee and an exceedingly low freezing point and does not

J. Exp. Zool. 352 J.P. COSTANZO ET AL. crystallize at the temperatures usually encountered calcium carbonate and magnesium carbonate in natural hibernacula; consequently, the quantity from the skeleton (especially the shell), which of bound water in tissues influences the amount of buffer most excess protons, and by sequestering ice that forms at any given temperature (Murphy, lactate in the shell (Jackson, 2000b). Hatchlings ’83; Claussen and Costanzo, ’90). The bound water do not tolerate oxygen deprivation as well as fraction can change seasonally in some temperate adults, perhaps because their lactate buffering organisms (Storey et al., ’81), an increase some- system is incompletely developed (Reese et al., times accompanying accumulation of micro- and 2004b; Dinkelacker et al., 2005a; Ultsch and macromolecules (Ring, ’82; Issartel et al., 2006). Reese, 2008). Hatchling turtles accumulate osmolytes during cold Anoxia tolerance is relevant to freeze tolerance acclimation (Costanzo et al., 2000b, 2006) and because as ice propagates through the body, maintain large deposits of glycogen, a water- circulatory function slows and ultimately ceases binding polymer (Rubinsky et al., ’94; Hemmings (Claussen and Kim, ’93; Costanzo et al., ’93). This and Storey, 2000), but whether or not bound water leaves tissues without oxygen, rich in lactate, and contributes significantly to freeze tolerance has not acidotic (Churchill and Storey, ’91,’92a,b; Packard been tested directly. and Packard, 2004). That freezing induces acute anoxic stress in hatchling turtles is evidenced by Water balance rapid accumulation of lactate in brains of C. picta and T. scripta (Hemmings and Storey, 2000). Many cold-hardy organisms dehydrate partially Anoxia tolerance may be requisite to the evolu- as they physiologically prepare for winter. This tionary development of freeze tolerance (Storey response can benefit freeze-tolerant species by et al., ’88; Greenway and Storey, ’99). Taxonomic concentrating osmolytes and by increasing the variation in anoxia tolerance, which reportedly fraction of bound water in tissues (Ring and Tesar, reflects differences in buffer reserves and meta- ’81; Storey et al., ’81; Aarset, ’82; Murphy, ’83; Lee bolic rates (Reese et al., 2004b; Dinkelacker et al., and Costanzo, ’93; Hayward et al., 2007), both 2005a,b), could explain differential capacities for effects reducing the percentage of body water that freeze tolerance among hatchling turtles. Results freezes. Reducing the hydration of tissues also of one study failed to support a strong association lowers ice content, thereby limiting mechanical between these traits, leading the authors to damage to the microvasculature and other sensi- conclude that, although anoxic stress is certainly tive structures. In principle, freeze-tolerant relevant to freezing survival, it is not necessarily turtles would benefit from partial dehydration, limiting (Dinkelacker et al., 2005b). Nevertheless, although the relationship between body water given that oxygen limitation is a principle deter- content and freeze-tolerance capacity has not been minant of thermal tolerance in ectothermic verte- investigated. It is also possible that freeze toler- brates (Po¨rtner, 2002), and that tissue ischemia ance is better developed in species that maintain underlies some of the important gene-expression relatively low hydration levels; however, although responses to somatic freezing (Storey, 2006), tissue water concentration does vary among taxa anoxia tolerance appears to be a key exaptation (Costanzo et al., 2001b, 2006; Dinkelacker et al., in natural freeze tolerance. 2005b), Costanzo et al. (2006) found no support for the expected correlation. Antioxidant systems Ischemia and anoxia tolerance Oxidative stress occurs whenever production of As a group, turtles exhibit a remarkable ability ROS, such as superoxide anion (O2À), hydrogen to survive oxygen lack (Lutz and Storey, ’97; peroxide (H2O2), and the hydroxyl radical (OH Á ), Bickler and Buck, 2007). This trait permits them in the mitochondrial respiratory chain overtaxes to estivate or overwinter in habitats that other- the antioxidant defense mechanisms. Superoxide wise would be inhospitable or lethal (Ultsch, ’89; and hydrogen peroxide are not particularly reac- Crocker et al., ’99; Jackson, 2002). Turtles are tive, but they can be readily converted into the endowed with a host of mechanisms for coping harmful hydroxyl radical. Ischemia associated with anoxia, including metabolic depression, effi- with somatic freezing can result in oxidative stress cacious management of oxidative stress, and manifested when oxygen returns to the tissues tolerance of ionic and acid–base perturbations. after thawing (Storey, ’96). Maintaining high They contend with lactacidosis by mobilizing constitutive levels of antioxidant enzyme activity

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 353 is a defensive strategy employed by species prone to that minimizes oxidative damage stemming from oxidative stress, although some organisms also (or freezing/thawing. However, whether or not oxida- instead) upregulate antioxidant defenses either in tive damage contributes to cryoinjury of freeze- anticipation of, and/or in direct response to, increas- intolerant species remains to be determined. ing levels of ROS production (Hermes-Lima et al., ’98; Hermes-Lima and Zenteno-Savin, 2002; Voitur- Cryoprotectants on et al., 2005b). In aquatic turtles, the antioxidant complement tracks seasons and generally decreases Certain organic osmolytes accumulated before, with body temperature, but remains relatively high during, or perhaps after freezing substantially in winter, apparently as an adaptation to submerged improve the tolerance of cells and tissues to freeze/ hibernation (Pe´rez-Pinzo´n and Rice, ’95). Freezing thaw stresses (Storey, ’97; Mazur, 2004; Yancey, and/or thawing-induced enhancement of antioxidant 2005). These compounds, cryoprotectants, have defenses, including increases in antioxidant enzyme multiple roles in freeze/thaw protection. When activities, has been reported in freeze-tolerant present in high concentrations, osmolytes, such as vertebrates, including the wood frog (Joanisse and polyhydric alcohols and sugars, colligatively Storey, ’96), garter snake (Hermes-Lima and Storey, reduce ice formation and, if membrane permeable, ’93), and common lizard (Voituron et al., 2005b). cellular shrinkage. Some compounds, including Dinkelacker et al. (2005b) investigated responses certain amino acids and polyols, can function in of the antioxidant enzyme catalase (which combats much lower concentrations to stabilize mem- oxidative stress by decomposing hydrogen perox- branes and macromolecules. Some serve as anti- ide) to freezing/thawing and anoxia exposure in oxidants or metabolic fuels, or increase the bound hatchlings of seven turtle species, both freeze- water fraction. tolerant and freeze-intolerant. All species exhibited Although various freeze-tolerant invertebrates high constitutive levels of the enzyme in liver, and amphibians accumulate cryoprotective solutes although the levels varied taxonomically. Although to high concentrations (e.g., 0.1–1 M or more), this catalase activity increased during anoxia in most seems not to be the case with reptiles, including species, it increased with experimental freezing/ hatchling turtles. In fact, although freezing usual- thawing only in the poorly freeze-tolerant species. ly triggers an osmolyte response in hatchling This outcome could mean that freeze-tolerant C. picta, glucose levels rise only modestly (Church- species maintain adequate constitutive levels of ill and Storey, ’92a) or not at all (Hemmings and catalase, or that they tolerate high levels of free Storey, 2000), and glycerol is virtually undetect- radical damage. Alternatively, perhaps other able (Storey et al., ’88; Churchill and Storey, ’92a; antioxidants are more important in the freezing Costanzo et al., 2000b, 2004). It remains equivocal response, or the experimental freezing regimen whether glucose is synthesized during freezing or (2 d at À2.51C) simply was too mild to cause during thawing. In a recent study, blood glucose significant ROS production in the freeze-tolerant levels in frozen/thawed hatchlings representing species. nine turtle taxa were 5- to 20-fold higher than To help resolve the role of antioxidant defenses in unfrozen counterparts. The finding that glucose in reptilian freeze tolerance, antioxidant capacity accumulated in freeze-intolerant species as well and indices of oxidative damage in tissues of as freeze-tolerant ones suggests that the mobiliza- hatchling C. picta subjected to a 48-h bout of tion is a fundamental response to stress, rather freezing (À2.51C), followed by gradual thawing than an adaptation specific to freeze tolerance and 24 h of recovery were examined by Baker et al. (Costanzo et al., 2006). (2007). Samples of plasma, brain, and liver were The aforementioned results pertain to turtles assayed for Trolox-equivalent antioxidant capa- subjected to experimental freezing/thawing and city, TBARS, markers of oxidatively damaged should be interpreted cautiously. Few studies have lipids, and protein carbonyls, markers of oxida- examined the physiology of turtles hibernating in tively damaged proteins. Finding an increase only nature, although the available data do suggest in plasma TBARS, the authors concluded that they become hyperglycemic in winter (DePari, ’88; freezing/thawing induced little stress. Hatchling Costanzo et al., 2004). Costanzo et al. (2004) found C. picta, like other freeze-tolerant vertebrates that plasma glucose in hatchling C. picta increased (Hermes-Lima and Storey, ’93; Joanisse and up to 15-fold in winter over levels prevailing in Storey, ’96; Voituron et al., 2005b), apparently autumn, the onset of hyperglycemia coinciding have a well-developed antioxidant defense system with the advent of frost in the hibernaculum.

J. Exp. Zool. 354 J.P. COSTANZO ET AL.

Glycemia tended to decrease as winter progressed, (Churchill and Storey, ’92a), and the metabolic albeit remaining above autumn levels, and was end products alanine and succinate (Churchill and higher in the colder of the two winters of record. Storey, ’91), but found that these compounds do not Indeed, turtles encountering periods of chilling to rise with freezing or are present only in minute subzero temperatures became profoundly hyper- amounts. Apparently, glucose and lactate synthesis glycemic, with glucose concentrations in some accounts for essentially all of the measured incre- individuals reaching 60 mmol LÀ1. A strong, direct ment in osmolality (up to 77 mosmol kgÀ1 in hatchl- correlation between plasma glucose level and ing C. picta) associated with freezing (Costanzo plasma osmolality attested that this metabolite is et al., 2006). an important osmolyte in hatchling C. picta. Efforts to elucidate the role of cryoprotectants in The putative role of lactate as a cryoprotectant reptilian freeze tolerance have been hampered by a in turtles and other vertebrates has received scant singular focus on osmolyte accumulation in response attention, although this compound reportedly to freezing. This mindset emanates from studies of (Loomis et al., ’89) has cryoprotective properties. freeze-tolerant frogs, which were found to mobilize Lactate invariably accumulates in the blood and large quantities of glucose and glycerol after freezing organs of freezing-exposed hatchling turtles (Stor- commenced, rather than during cold acclimation ey et al., ’88; Churchill and Storey, ’92a; Dinke- (Schmid, ’82; Storey and Storey, ’85). The anuran lacker et al., 2004, 2005b; Packard and Packard, cryoprotectant system is unusual because most 2004; Costanzo et al., 2006) and the early initia- freeze-tolerant organisms accumulate osmolytes in tion and rapidity of this response suggested to anticipation of freezing (Storey and Storey, ’88; Lee, some authors (Churchill and Storey, ’91; ’91). Only recently, however, was it reported that Hemmings and Storey, 2000) that the rise, at terrestrially hibernating, freeze-tolerant frogs accu- least initially, is not triggered by ischemic hypoxia, mulate a cryoprotectant, urea, in autumn and early but rather represents a cryoprotective response. winter (Layne and Jones, 2001; Costanzo and Lee, On the other hand, it is uncertain that the amount 2005). Whether or not reptiles also show an of lactate accumulated during freezing, typically anticipatory response requires more study, but 15–25 mmol LÀ1 in blood plasma, could have a preliminary findings suggest that they do. Costanzo significant colligative effect in reducing cellular et al. (2000b) reported that seasonal development of dehydration and body ice content. Additionally, cold hardiness in hatchling C. p. bellii was associated the finding that lactate increases with freezing to with an increase in plasma osmolality chiefly caused essentially the same degree in hatchlings of both by accumulation of urea (to 80 mmol LÀ1), the freeze-tolerant and freeze-intolerant turtles principle nitrogenous waste product in most chelo- suggests that the response is not a specific nians. Subsequently, elevated plasma urea adaptation to freezing survival (Dinkelacker (30–50 mmol LÀ1), accounting for 10% of the total et al., 2005b; Costanzo et al., 2006). On the other plasma osmotic potential, has been found in cold- hand, although lactate accumulation invariably acclimated hatchlings of several other species occurs with experimental freezing, one study (Costanzo et al., 2006). Hatchlings do not necessarily (Costanzo et al., 2004), the first to examine turtles remain hyperuremic throughout winter (Costanzo during natural hibernation, showed that blood et al., 2004), but those that do stand to benefit from lactate levels in hatchling C. p. marginata and this solute’s cryoprotective properties. Indeed, urea C. p. bellii hibernating in natural nests remained accumulation could underlie the observed increase low (2–6 mmol LÀ1) throughout winter, even in freeze tolerance associated with winter acclima- though some of the animals were sampled whilst tization (Churchill and Storey, ’92a; Packard and frozen. Packard, 2003b; Costanzo et al., 2004). Additional Investigation of the putative roles of cryoprotec- research is needed to determine whether accumu- tants in reptilian freeze tolerance has focused lated urea is perturbing, as little is known about the mainly on glucose and lactate. However, it is levels of trimethlyamine oxide and other ‘‘counter- important to note that, in sufficient quantity, any acting’’ methylamines in turtle tissues. permeable, compatible solute will colligatively Although hatchling turtles can accumulate moderate ice formation and limit cell shrinkage. organic osmolytes both before and during freez- Investigators have screened the blood and other ing/thawing, the importance of such compounds tissues of freezing-exposed turtles for amino acids in freezing survival remains equivocal. On the (Storey et al., ’88; Churchill and Storey, ’92a), one hand, organic solutes collectively account for sugars such as sorbitol, fructose, and mannose up to one-third of total plasma osmotic pressure

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 355 in frozen/thawed animals; without them, substan- research on vertebrates has focused on freeze- tially more ice would form, mechanical stress and tolerant frogs, but recent investigations have cellular dehydration would increase, and turtles found in organs of hatchling C. picta several would sooner and at higher temperatures attain a classes of freezing-responsive genes, the products critical ice content (Costanzo et al., 2006). On the of which include antioxidant enzymes, inhibitors other hand, it is uncertain that lactate exhibits the of serine protease, and iron-binding proteins most essential quality of any cryoprotective agent, (Storey, 2006). One major finding emanating from and that of cytoprotectants in general: compat- this line of research is that various genes respon- ibility with cellular processes. Packard and Pack- sive to anoxia or desiccation also respond to freeze/ ard (2004) suggested that excessive lactate thaw stress, bolstering the view that freeze accumulation, failure of the buffering system, tolerance developed from fundamental mechan- and attendant acid–base disturbance contributed isms for coping with oxygen deprivation and cell to freezing mortality in hatchling turtles. Working volume regulation. with seven species, Dinkelacker et al. (2005b) found no clear relationship between freeze-toler- Somatic correlates ance capacity and blood lactate levels in frozen/ thawed hatchlings. To the contrary, Costanzo Relationships between somatic and life-history et al. (2006) examined nine turtle taxa, finding traits and cold-hardiness strategies of ectothermic lower rates of freezing survival in those with animals are of potential interest to adaptational copious lactate, as well as a direct association biologists, yet have received scant coverage in the between lactemia and cryoinjury among conspe- literature. The ice nucleation theory predicts that cific individuals. Disparity in experimental out- as body size increases, supercooling becomes less comes reported by Dinkelacker et al. (2005b) and effective as a cold-hardiness strategy. Thus, Costanzo et al. (2006) probably reflects the relatively large organisms may be predisposed to comparatively colder and longer freezing expo- evolving a strategy of freeze tolerance rather than sure, which resulted in two-fold higher lactate freeze avoidance. In addition, because a larger concentrations, in the latter study. Unfortunately, body will cool more slowly, ice growth in relatively those authors could not ascertain whether the link large individuals tends to be more controlled and, between hyperlactemia and freezing injury simply overall, less body water will freeze before the reflected differential lactate clearances among animal rewarms (Claussen and Zani, ’91; Voituron injured and healthy specimens, or whether it et al., 2005a). Body size bears significantly on stemmed from adverse effects of the osmolyte freeze-tolerance capacity in some taxonomic (e.g., acidification) or the stress (e.g., hypoxia) groups, such as mollusks (Loomis, ’95; Ansart triggering its accumulation. In any case, the avail- and Vernon, 2003, 2004). Among hatchling turtles, able literature concurs that the hatchlings of freeze- cryoinjury tends to be less severe among larger tolerant species can recover from freezing/thawing individuals (Costanzo et al., 2006), this finding without the need to accumulate large quantities of being consistent with the idea that freeze toler- colligative cryoprotectants. This also is the case with ance is more easily developed in larger species. a few freeze-tolerant insects (Salt, ’61; Ring and Hatchlings of species that usually overwinter Tesar, ’81), most intertidal invertebrates (Murphy, terrestrially tend to have larger supplies of yolk ’83; Ansart and Vernon, 2003), and reptiles, includ- and storage lipids than species whose hatchlings ing adult turtles, snakes, and lizards. hibernate aquatically (Congdon et al., ’83c; Congdon and Gibbons, ’85, ’90; Rowe et al., ’95; Nagle et al., ’98; Costanzo et al., 2000b). Not only Gene and protein responses is residual yolk important to the energetics of Innovations in molecular biology, such as cDNA aphagic, overwintering hatchlings, but it also library screening, DNA array screening, and real- could enhance their cold hardiness (Bleakney, time PCR, have greatly facilitated efforts to ’63). Freezing reportedly increases the vascularity identify the myriad gene and protein responses and shrinks the internalized yolk sac, presumably supporting natural freeze tolerance (Storey, as yolk materials are mobilized to other tissues 2004a). In freeze-tolerant organisms, changes in (Hemmings and Storey, 2000). Costanzo et al. protein and gene expression occur preparatory to (2006) found no association between residual winter or are triggered directly by the freezing yolk mass and freeze-tolerance capacity among event (Storey, ’99, 2004b). Most of the relevant nine taxa of temperate North American turtles,

J. Exp. Zool. 356 J.P. COSTANZO ET AL. including both freeze-tolerant and freeze-intoler- exposures (Churchill and Storey, ’92b; Packard ant ones. However, this result may not be et al., ’93, ’99b; Costanzo et al., ’95b, 2006; Baker definitive because little yolk remained when the et al., 2003; Dinkelacker et al., 2005b); yet this turtles were tested, in late winter. limited capacity is distinct from that of other Maintaining a large glycogen reserve could northern species (e.g., S. odoratus, K. flavescens, contribute to freezing survival in hatchling turtles. A. spinifera), which succumb to even mild freeze/ Freeze-tolerant species have ample supplies of this thaw stress (Costanzo et al., ’95b, 2006). compound, which, for example, may account for Freeze tolerance has been examined in hatchl- over one-half of the liver’s mass (Hemmings and ings from all four families (Chelydridae, Emydi- Storey, 2000). Glycogen, by virtue of its water dae, Kinosternidae, Trionychidae) of freshwater binding, helps reduce ice formation in tissues and turtles found in North America. Such studies have also provides the freezing-mobilized solutes, glu- disproportionately focused on the Emydidae, a cose and lactate, putative cryoprotectants (Church- family of mostly North American pond turtles ill and Storey, ’91; Hemmings and Storey, 2000). (Ultsch, ’89), although we suspect that freeze Dinkelacker et al. (2005b) found that liver mass tolerance is not common in other family groups; (normalized to body size) varied among hatchlings for example, only C. serpentina (Chelydridae) of seven turtle taxa, perhaps reflecting similar tolerates even a modicum of freezing (Packard variation in glycogen reserves; however, these et al., ’93; Costanzo et al., ’95b, 2006; Dinkelacker authors found no support for the expected relation- et al., 2005b). No studies of freeze tolerance are ship between liver mass and capacities for freeze available for any testudinid (tortoise), geoemydid tolerance or anoxia tolerance. That glycogen supply (Asian pond turtle), sea turtle (cheloniid or does not limit anoxia tolerance in hatchling turtles dermochelyid), or side-necked turtle (chelid, pelo- is further supported by the finding that 33–50% of medusid, or podocnemid), although all of these are the initial amount remains at the limit of survival subtropical or tropical in nesting distribution. (Reese et al., 2004b). Among emydids, freeze tolerance is found in both clades, occurring in at least two genera of the (Eurasian and North American terres- Taxonomic and ecological associations trial forms) and in at least two genera of the Distinguishing degrees of freeze tolerance among (North American aquatic forms), turtle taxa is not a simple matter, partly because suggesting that it is widespread in this family. the thermal range for survival extends only to À41C Nevertheless, at least one northern emydid in even the most profoundly freeze-tolerant species. (G. geographica) is decidedly intolerant of freezing In addition, comparing literature reports from (Baker et al., 2003) and further research may different laboratories is confounded by inconsis- identify others. tency in methodologies and criteria for assessing Various questions about phylogenetic aspects of freeze tolerance. A few studies, in which investiga- the evolution of chelonian freeze tolerance remain tors tested hatchlings using identical methods, unanswered because too few species and too few attest that freeze tolerance varies considerably major groupings have been investigated. The among species (Costanzo et al., ’95b, 2006; Packard northern distribution and life histories of certain et al., ’99b; Willard et al., 2000; Dinkelacker et al., turtles (e.g., E. orbicularis, T. horsfieldii) make 2005b). In a study of nine taxa of temperate North them interesting candidates for freeze-tolerance American turtles, viability of hatchlings subjected testing. Unfortunately, many species may never be to experimental freezing (72 h at À31C) ranged investigated owing to concerns regarding their 0–100% and recovery rates varied among those conservation status. Clues to their freeze-toler- surviving (Costanzo et al., 2006). Furthermore, the ance status could be gathered through careful field consensus of reports on individual species is that observation. For example, Chase et al. (’89) noted freeze tolerance is more fully developed in some that bog turtles (C. muhlenbergii) in Maryland taxa than in others. Hatchling C. picta, T. ornata, overwintered successfully in close proximity to E. blandingii,andM. terrapin can recover from frozen mud and ice, questioning whether these freezing to relatively low temperatures and for long turtles had a special physiological mechanism periods (Storey et al., ’88; Dinkelacker et al., 2004; allowing them to survive in the frost zone. In Baker et al., 2006; Costanzo et al., 2006). In Quebec, hatchling wood turtles (C. insculpta), contrast, hatchling T. scripta, G. geographica,and monitored as they dispersed from a natal nest in C. serpentina tolerate only moderate freezing late September, apparently became frozen during

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 357

TABLE 5. Summary of results of experimental tests of freeze tolerance in hatchlings of various taxa of North American turtles

Equilibrium Hibernation Freeze-tolerance temperature Duration tested Taxon habitat capacity tested (1C) (h) References

Apalone spinifera Aquatic Low À3.0 33 1 Chelydra Aquatic Low –2.0 to –4.0 24–168 1–4 serpentina Chrysemys picta Natal nest High –2.0 to –4.0 24–226 1–10 bellii C. p. marginata Natal nest High –2.5 to –4.0 16–264 4, 9, 11 Emydoidea Unknown; High –2.0 to –3.6 24–168 2–4, 12, 13 blandingii possibly terrestrial Graptemys Natal nest Low –2.5 to –3.0 24–72 3, 4, 14 geographica Kinosternon Below natal Low –2.6 26 1 flavescens nest Malaclemys Natal nest, High –2.5 to –3.5 26–288 4, 15 terrapin other terrestrial, or aquatic Sternotherus Aquatic Low –3.0 72 4 odoratus Terrapene ornata Below natal High –2.5 to –3.0 36–168 1, 3, 4 nest Trachemys Natal nest Equivocal –2.0 to –3.0 24–168 2–4, 7 scripta

Freeze-tolerance capacity was considered high if most experiments resulted in survival after freezing at À2.51C for 48 h. Survey included experiments in which turtles began freezing at Z–2.01C, were exposed to equilibrium temperatures between –2.0 and –4.01C, and remained frozen Z24 h. References: (1) Costanzo et al. (’95b); (2) Packard et al. (’99b); (3) Dinkelacker et al. (2005b); (4) Costanzo et al. (2006); (5) Attaway et al. (’98); (6) Packard and Packard (2004); (7) Willard et al. (2000); (8) Packard and Packard (2003b); (9) Churchill and Storey (’92a); (10) Costanzo et al. (2004); (11) Storey et al. (’88); (12) Packard et al. (2000); (13) Dinkelacker et al. (2004); (14) Baker et al. (2003); (15) Baker et al. (2006); (16) Churchill and Storey (’92b).

an overnight frost, but before noon of the follow- cold-adapted species (Baker et al., 2006). Although ing day had thawed and revived, most having hatchling C. picta reportedly survive freezing at already resumed migration (A. Walde, personal À2.51C for at least 11 d (Churchill and Storey, communication). ’92a), M. terrapin, which does not range as far Freeze tolerance generally is well developed in north, apparently cannot. species known (or suspected) to hibernate in There are some notable exceptions to the microenvironments where freezing is likely and general association between hibernation habitat poorly developed in species that overwinter in and ability to withstand freezing/thawing. For frost-free sites (Table 5). Terrestrial hibernators instance, although G. geographica sometimes are not necessarily freeze tolerant, however. For encounters subzero temperatures while overwin- example, hatchling K. flavescens do not tolerate tering within the natal nest, they do not tolerate freezing, but they commonly hibernate in the soil even mild freezing; instead, they are adept at column beneath the nest, well below the reach of freeze avoidance (Baker et al., 2003). Hatchling frost (Costanzo et al., ’95b). Hatchling T. ornata C. serpentina, which usually hibernate under also overwinter beneath the natal nest, although water, can tolerate freezing under relatively mild they do not always evade frost (Doroff and Keith, conditions (Packard and Packard, ’90; Packard ’90); this species tolerates somatic freezing et al., ’93; Costanzo et al., ’95b, 2006; Dinkelacker (Costanzo et al., ’95b, 2006). Freeze tolerance in et al., 2005b). Terrestrial overwintering of these hatchling M. terrapin, which can successfully hatchlings occurs at least occasionally (Bleakney, overwinter on cold, exposed estuarine beaches, is ’63; Obbard and Brooks, ’81a; Costanzo et al., robust, but nevertheless inferior to that of more ’95b), but whether they benefit from such limited

J. Exp. Zool. 358 J.P. COSTANZO ET AL. tolerance is equivocal. Possibly, this ability (Sinclair et al., 2003a). Physical characteristics of enables them to survive brief, shallow freezing the hibernaculum microenvironment, such as episodes that occur in autumn at high latitudes, factors affecting freezing risk, are particularly before nest emergence. Freeze tolerance is weakly important (Ansart and Vernon, 2003), as are developed in hatchling T. scripta, which usually certain life-history traits. For instance, body size, overwinter inside the natal nest (Churchill and which influences capacities for supercooling and Storey, ’92b; Packard et al., ’99b; Dinkelacker freeze tolerance, may strongly influence develop- et al., 2005b; Costanzo et al., 2006). However, this ment of cold-hardiness strategies (Ansart and species is primarily southern in its distribution Vernon, 2003). and has not successfully colonized far northern Some authors have posited various phylogenetic, habitats (Cagle, ’50; Holman, ’94). ontogenetic, and ecological arguments that freeze Additional study is needed to clarify phylogenetic, avoidance is the basal modality, and that freeze geographic, and ecological associations in chelonian tolerance, which has independently evolved multi- freeze tolerance. The available evidence argues ple times, is a relatively recent adaptation (Vernon against the hypothesis that the ability to withstand and Vannier, 2002; Sinclair et al., 2003a). There is freezing/thawing is widespread among the hatchl- general agreement that freeze tolerance derives ings of North American turtles and is a shared from preexisting, fundamental responses to stres- ancestral trait (Packard et al., ’99b). Rather, it ses, such as desiccation and anoxia (Holmstrup, suggests that freeze tolerance has been improved 2003; Costanzo et al., 2006; Storey, 2006), and that by natural selection for its current role in cold- its development through natural selection is hardiness strategies in some northern species. facilitated in species whose ancestral traits include tolerance to cold and limited freezing (Convey, ’97; Sinclair, ’99; Voituron et al., 2002a; Sinclair et al., Evolutionary and ecological implications 2004). Research on the winter biology of hatchling Freeze tolerance and freeze avoidance via super- turtles has helped elucidate principle mechanisms cooling are important mechanisms for coping with underpinning the evolution of cold hardiness in subzero temperatures in a diverse array of ec- ectothermic animals. tothermic animals (see recent reviews by Ramlov, 2000; Zachariassen and Kristiansen, 2000; Packard Freeze tolerance or supercooling? and Packard, 2001a; Wharton, 2002). Cryoprotec- tive dehydration is an important strategy of certain The painted turtle presents an unusual and small, soil-dwelling invertebrates (Holmstrup, interesting example of cold-hardiness evolution in 2003), but has not yet been reported in vertebrates. that it can survive chilling episodes through either Although it is widely recognized that freeze freeze tolerance or supercooling. This bimodal tolerance and supercooling modalities are usually strategy, which affords plasticity in response to mutually exclusive, it remains unclear why a environmental variability, has probably allowed particular taxon adopts one strategy over the other. C. picta to broadly extend its geographical range. Some attention has been paid to the costs and Although sound evidence indicates this animal benefits associated with these strategies, particu- belongs to the small group of ectotherms whose larly with respect to energetics, water conservation, cold-hardiness repertoire includes both strategies and potential for significant cryoinjury (Zacharias- (Stover, ’73; Ring, ’82; Horwath and Duman, ’84; sen, ’85; Block, ’91; Duman et al., ’95). Voituron Shimada and Riihimaa, ’88; Costanzo et al., ’95a; et al. (2002a) advanced the theoretical argument Bale et al., 2001; Voituron et al., 2002a), the that the modality adopted by a particular organism literature concerning C. picta is marked by must provide the greatest likelihood of survival competing views of the putative roles of freeze whilst also minimizing energy use. tolerance and supercooling in its winter survival. Myriad biotic and abiotic factors influence evolu- The laboratory of K. B. Storey, which offered the tionary development of cold-hardiness strategies in seminal report of freezing survival in hatchling ectothermic animals. Environmental variation on turtles, advocates freeze tolerance as the major spatial and temporal scales probably has significant cold-hardiness mechanism in hatchling C. picta consequences for the genesis of cold-hardiness (Storey et al., ’88). Working primarily with north- adaptations (Sinclair et al., 2003b) as, for example, ern populations, this research group demonstrated freeze tolerance is a particularly useful strategy in both limited supercooling capacity and profound thermally mild and unpredictable environments freeze tolerance, including survival of hatchlings

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 359 exposed to À41C with over 50% of their body water esis. According to this view, supercooling is the frozen and of hatchlings kept frozen for at least predominant strategy when environmental and 11 d (Storey et al., ’88; Churchill and Storey, ’92a). physiological conditions are conducive to freeze According to these authors, freeze tolerance in avoidance, but a tolerance for freezing becomes hatchling C. picta is a finely tuned adaptation that critical in other instances. Subsequent work by has evolved from a complex suite of genetic, these investigators has revealed the complexity of molecular, and physiological responses to stress these relationships. Although some authors have (Storey and Storey, ’92). considered this hypothesis plausible (Ultsch, ’89; In numerous editorials, G. C. Packard and Nagle et al., 2000), it has not enjoyed universal coauthors (Packard and Packard, ’95b, 2001a, acceptance (Packard and Packard, ’95a; Weisrock 2003c; Packard, 2004a,b) have zealously main- and Janzen, ’99). tained that supercooling, not freeze tolerance, is Difficulty in resolving questions about the the primary survival mechanism of hatchling putative roles of freeze tolerance and supercooling C. picta. These authors initially advocated that stems from a paucity of information about the hatchlings inside nests are incapable of resisting actual physical states of hatchling turtles during inoculation and therefore rely on freeze tolerance natural hibernation. Packard and Packard (2003c) for winter survival (Packard et al., ’89; Packard addressed this limitation, stating that it is and Packard, ’90). This position was later sup- virtually impossible to obtain such evidence, citing planted by their claim that supercooling, not their own failed attempts to examine hatchlings freeze tolerance, is key to survival, and that inside frozen nests and to confirm frozen status by turtles must resist inoculation to survive at distinguishing individual animal exotherms from subzero temperatures. They further asserted that variation in soil temperatures (Packard, ’97). reported instances of freezing survival in hatchl- Excavating nests in winter is a challenging under- ing turtles were ‘‘nothing more than artifacts of taking, but can be done, especially if nests are laboratory research’’ (Packard and Packard, ’93b) constructed in friable (sandy) soils. Costanzo et al. and that somatic freezing is ‘‘fatal to hatchlings, (2004) extricated C. picta from several nests in regardless of the minimum temperature to which early February, following an extended period of they are exposed’’ (Packard and Packard, ’95a). freezing weather, finding that many were inani- More recently, however, these authors acknowl- mate, ice-covered, and frozen; yet, several revived edged freezing survival of hatchling turtles, even on warming. Direct determinations of the state of publishing their own accounts (Attaway et al., ’98; naturally hibernating hatchlings that do not Packard et al., ’99b; Willard et al., 2000; Packard disturb the nest or its occupants would involve and Packard, 2003b), but have continued to argue methods that have not yet been applied. One against its adaptive significance. They maintain possible technique is acoustic sensing, which is that the ability to survive somatic freezing is based on the premise that sound waves will travel common among chelonians, but is merely an through an unfrozen turtle at a different rate than incidental consequence of some other aspect of through a frozen one. Also, indirect methods not their biology (Packard et al., ’99b; Packard and previously used to address this question might be Packard, 2001a). However, the implication that brought to bear. For example, the emerging field freeze tolerance arose by happenstance is unlikely of metabonomics, a process of characterizing considering the unique challenges presented by multiparametric responses associated with meta- freezing and thawing stresses and the strong bolic dysfunction, may be helpful. Preliminary selection pressure needed for evolution of such experiments suggest that the NMR-based metabo- specific coping mechanisms. Moreover, evidence nomic approach can be used to distinguish tissues from several laboratories not only points to that have been frozen from those that have been definite, species-specific differences in this trait supercooled (Fig. 15). (Costanzo et al., ’95b, 2006; Packard et al., ’99b; Lacking direct observations on free-living ani- Willard et al., 2000; Dinkelacker et al., 2005b) but mals, investigators have relied on laboratory also indicates that such variation tends to associ- simulations and measurements of the hibernacu- ate with freezing risk in nature (Dinkelacker lum microenvironment to explain the remarkable et al., 2005b; Costanzo et al., 2006). cold hardiness exhibited by hatchling C. picta. Is there any middle ground? Predicated on their Proponents of the freeze-tolerance hypothesis 5-yr field and laboratory study, Costanzo et al. advocate that, although hatchlings are prone to (’95b) have promulgated a dual-modality hypoth- freezing, their capacity to withstand this stress

J. Exp. Zool. 360 J.P. COSTANZO ET AL.

from notions about the thermal relations of overwintering hatchlings and their capacities for supercooling, inoculation resistance, and freeze tolerance. However, as we discuss in the following sections, some of this information may be inaccu- rate or interpreted improperly. One of the major shortcomings is that direct measurements of body temperature in hibernating turtles are lacking, probably owing to the technical difficulties of collecting such data. In practice, investigators infer body temperature from the out- put of data loggers placed inside or near the hibernaculum. Although some workers deem these data an acceptable proxy (e.g., Breitenbach et al., ’84; Packard, ’97), it is noteworthy that hibernacula are three-dimensional structures with complex heating/cooling dynamics and, therefore, actual body temperatures may be higher or lower than those recorded by such devices. In addition, the data output does not necessarily reflect temperatures of all individuals using the same hibernaculum. Strong thermal gradients can occur inside turtle nests (Wilhoft et al., ’83; Thompson, ’88; Georges, ’92; Tucker, ’99) and probably cause siblings to experi- ence different thermal conditions (Packard et al., ’97a). Additionally, the heat of fusion liberated during freezing of some turtles could, in principle, alter the thermal profiles of nearby individuals. Similarly, cold-hardy slugs are known to ‘‘huddle’’ during chilling, possibly as a means to improve Fig. 15. Using a metabonomics approach to distinguish supercooled and frozen states in hatchling painted turtles survival (Cook, 2004), and various plants derive (Chrysemys picta bellii). Results of preliminary experiments, protectionfromtheheatoffusionwhencertainof showing (A) representative metabolite profile of brain tissue their parts freeze (Zachariassen and Kristiansen, of a cold-acclimated hatchling, produced by 1H-NMR spectro- 2000). Note that the typical 7-g hatchling, having a scopy in a Bruker 500 MHz NMR spectrometer with an hr- water concentration of 80% (fresh mass) and an MAS probe and (B) principle components analysis (PCA) assumed bound water fraction of 20%, liberates suggesting differences in metabolite levels among cold- acclimated (control), supercooled, and frozen/thawed hatchl- 1.0 kJ whilst two-thirds of its body water freezes. ings. Eggs were collected from gravid females at a nesting area Taking values for the specific heat of liquid in Garden Co., NE, and incubated in the laboratory at 291C. water (4.18 J gÀ1 1CÀ1), ice (2.09 J gÀ1 1CÀ1), and After hatching in August, turtles were kept in darkness and dry hatchling tissue (1.56 J gÀ1 1CÀ1; Claussen and progressively acclimated to 41C. In December, the control turtle was sampled directly from its cage; other turtles were Zani, ’91), we calculate that only 19–25 J, or r2.5% sampled without rewarming after being held supercooled or of the energy liberated by a single freezing turtle, frozen for 48 h at À31C. (G. Lorigan, T. Muir, and P. Baker, must be absorbed by an adjacent turtle, either unpublished data.) supercooled or frozen, to raise its body temperature 11C under steady-state thermal conditions. Of (with respect to both thermal and temporal limits) course, the nest hibernaculum is incompletely closed is sufficient to allow recovery from most natural to heat exchange and tends to be cooling whilst chilling episodes. Proponents of the freeze-avoid- turtles are freezing. However, considering the ance hypothesis argue that, relative to their insulative properties of frozen soil and the close meager freeze tolerance, frigorific conditions physical association of the entombed siblings, this inside hibernacula are too severe and turtles scenario seems worthy of study. When considering therefore survive winter by virtue of their well- the cold-hardiness mechanisms of hatchling turtles, developed capacities to supercool and resist workers should recognize the limitations of their inoculative freezing. Both views draw support methodology and make allowance for imprecision in

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 361 knowing the true thermal relations of hibernating significance of freezing as a winter mortality animals. factor; rather, we suggest that authors not Claims about the importance of freeze tolerance discount the biological significance of freeze in the winter life histories of hatchling turtles tolerance solely on the basis of generalized claims have been criticized largely on grounds that freeze about the thermal environment inside terrestrial tolerance is so limited that there could be few hibernacula. instances in which turtles would survive freezing. Capacities for freeze tolerance, supercooling, For example, some authors contend that tempera- and inoculation resistance in hatchling turtles tures in C. picta nests commonly fall below À41C, are known only from laboratory studies; those of the apparent lower limit of freeze tolerance, and free-living animals have not as yet been deter- that subzero temperatures persist long beyond the mined. Although many of these studies are temporal limit to survival in the frozen state instructive, results of experiments that were (Packard and Packard, 2003c). These concerns are improperly designed or executed have confounded not unfounded, but the argument appears over- the literature. In particular, failing to control ice stated for two reasons. First, in assessing the nucleation, freezing specimens too rapidly, testing disparity between environmental demands and unacclimatized animals, and denying hatchlings physiological tolerances, Packard (’97) focused on adequate recovery time have led authors to minimum temperatures and maximum durations underestimate the capacity for freeze tolerance. of subzero chilling episodes, placing special em- Furthermore, recent study has shown that ther- phasis on the coldest periods of the coldest winters mal history can strongly influence freeze tolerance rather than more routine temperatures and in some species (Lee et al., 2006b). Supercooling typical durations. Secondly, the principal context capacity has been overestimated (sometimes for this argument is a C. picta population in grossly) through the practice of hatching and northcentral Nebraska, where severe winters can rearing experimental subjects on an artificial, occur owing to a continental climate that is not INA-free medium. In tests of inoculation resis- only very cold but also lacking in insulating snow tance, chilling hatchlings in mock nests without cover, and yet populations thrive. A more balanced initiating freezing of the soil matrix, as well as interpretation of the available data is that, even in constructing nests of exotic clayey soil (rather severe winters, most subzero chilling episodes are than indigenous sand), has biased results toward relatively benign and, in moderate winters, sub- freeze avoidance (Packard and Packard, ’93a,b, zero temperatures occur infrequently, if at all, in ’95a; Packard and Janzen, ’96; Packard et al., many nests (Packard et al., ’89; Paukstis et al., ’97c). On the other hand, the convention of ’89; Costanzo et al., ’95b, 2004; Packard, ’97). It is completely enveloping individuals in a soil matrix, noteworthy that extreme cold rarely occurs in rather than testing groups of hatchlings in open- C. picta nests in the Midwest, Northeast, and chambered nests containing eggshell fragments, other parts of the species’ range, in part owing to probably accentuates freezing risk. Investigators the moderating effect of snow cover (Table 2). For should realize that even the most carefully example, in multi-year studies of C. picta over- conducted experiments are crude simulations of wintering in Michigan, Breitenbach et al. (’84) the natural environment and that cold-hardiness reported that nest temperatures rarely fell below characteristics of free-living animals may differ À2.01C (although one nest reached À3.31C), and considerably from those determined for hatchlings Nagle et al. (2000) reported that nests remained reared and/or tested under laboratory conditions. frost-free. In one winter in Illinois, only half of the monitored C. picta nests cooled below À41C Resolving the debate (Weisrock and Janzen, ’99); mortality was slight and occurred only in nests where temperatures Proponents of the freeze-avoidance hypothesis fell below À81C. In northern Indiana, minimum argue that freeze tolerance is an ineffectual temperatures inside C. picta nests during two strategy because hibernaculum temperatures can consecutive winters, including a severe one, fall below À41C, the approximate lower lethal ranged from À1.2 to À4.71C (Costanzo et al., temperature for frozen turtles. However, by the 2004). Hatchlings in some populations probably same token, supercooling capacity is also limited; can survive most, if not all, chilling episodes that is, hibernaculum temperatures can fall below wholly by virtue of their freeze tolerance, limited À7toÀ81C, the critical minimum temperature for as it is. Our point here is not to minimize the turtles hatched in natural nesting soil and cooled

J. Exp. Zool. 362 J.P. COSTANZO ET AL. in perfect isolation from environmental ice and long periods, providing a more thorough under- INA (Packard et al., 2001; Costanzo et al., 2003; standing of their winter habits, habitat require- Packard and Packard, 2003a, 2006). In one field ments, and survival rates. study (Weisrock and Janzen, ’99), winter mortal- Field biologists need to conduct additional ity of hatchling C. picta occurred only in nests that studies of natural hibernacula and environmental fell below À81C, this threshold matching the lower conditions confronting hatchling turtles, particu- limit of supercooling for that population (Packard larly those overwintering under water. Results and Janzen, ’96). Otherwise, there is poor con- of laboratory studies indicate that hatchlings, currence between winter mortality rates and unlike the older individuals of some species, are results of laboratory studies delimiting cold unable to survive long-term submergence in hardiness in hatchling turtles (e.g., Packard, ’97; anoxic water. Submergence in cold, normoxic Costanzo et al., 2003, 2004). Neither adaptation, water may cause iono-osmoregulatory perturba- insofar as our current understanding allows, can tions in some species, particularly ones whose explain how hatchling C. picta successfully emerge hatchlings usually hibernate on land. in spring from nests that earlier had cooled below Overwintering in terrestrial habitats precludes À101C (Woolverton, ’63; DePari, ’96; Packard, ’97; winter feeding, but hatchlings seem not to be at risk Packard et al., ’97a; Costanzo et al., 2004), or even of starvation, probably because their metabolism is À151C (Costanzo et al., 2003). Clearly, more markedly reduced during hibernation. Winter mor- ecological and physiological information is needed tality is caused by flooding, dehydration, and to clarify the roles of these cold-hardiness re- exposure to severe cold; anoxia is a potential threat, sponses in winter survival. Caution must be used but has not been investigated. Terrestrial hiberna- not only in designing suitable experiments, but tors exhibit adaptations that promote survival in also in extrapolating laboratory findings to field these harsh conditions. For example, these species conditions. The salient relationships are highly better resist EWL than species whose hatchlings complex and investigators will need to exercise usually hibernate under water; this difference patience in elucidating them. apparently reflects variation in morphology and characteristics of the integument. CONCLUSIONS Considerable study has been devoted to the cold- hardiness adaptations exhibited by terrestrial Temperate species of turtles hatch from eggs in hibernators. Hatchlings of all temperate species late summer. Although the hatchlings of some tolerate a degree of chilling, but survival of species leave their natal nest to hibernate else- repeated, transient exposure to subzero tempera- where on land or under water, others remain tures depends on either freeze avoidance through inside the nest until spring. Most species have an supercooling or on freeze tolerance, each mechan- affinity for one behavior or the other, but several ism having its own advantages and disadvantages. exhibit plasticity in this trait, sometimes exhibit- Freeze tolerance permits revival after the tissues ing multiple behaviors at the same locale. It is not have frozen and thawed, but can protect turtles known why this occurs. Studies addressing phy- only at modest temperatures (ZÀ41C). Super- siological regulation of nest emergence behavior in cooling permits survival at much lower tempera- hatchling turtles are lacking. Investigating the tures (to perhaps À121C), but the supercooled endocrine status of emerging and nonemerging state is inherently precarious because turtles are hatchlings, and of populations exhibiting both fall susceptible to inoculation by endogenous and and spring emergence might prove rewarding. exogenous ice-nucleating agents (INA) and contact Post-hatching behavior strongly influences the with ice in the environment. Freeze tolerance and winter ecology and physiology of hatchling turtles. supercooling usually are dichotomous strategies, The costs and benefits associated with overwinter- but at least one species apparently can switch ing inside vs. outside the nest have long been between them, the one utilized in a particular contemplated, but the implications of hibernacu- instance of chilling depending on prevailing lum choice remain incompletely understood physiological and environmental conditions. because fundamental information about hatchling Freeze avoidance through supercooling is pro- behavior and microenvironmental conditions are moted by behavioral, anatomical, and physiologi- lacking. Improvements in radiotelemetry and cal features that minimize risk of inoculation by other tracking techniques would enable investiga- ice and INA. Neonates harbor ice-nucleating tors to monitor hatchling behavior for sufficiently substances of endogenous origin, but some of

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 363 these agents are shed when gut contents are Ackerman RA. 1977. The respiratory gas exchange of sea purged. The intrinsic supercooling limit is never turtle nests (Chelonia, Caretta caretta). Respir Physiol 31: reached by hatchlings in nature because freezing 19–38. Ackerman RA. 1997. The nest environment and the embryo- is initiated by contact with exogenous agents. Ice nic development of sea turtles. In: Lutz PL, Musick JA, is a particularly potent catalyst, but inoculation editors. The biology of sea turtles. Boca Raton, FL: CRC can also result from contact with various inorganic Press. p 83–106. and organic INA ubiquitously found in terrestrial Allard HA. 1935. The natural history of the box turtle. Sci hibernacula. Susceptibility to inoculative freezing Mon 41:325–338. is governed by temperature and various factors Allard HA. 1948. The eastern box turtle and its behavior. J Tenn Acad Sci 23:307–321. that mediate intimacy with these agents, and also Anderson PK. 1958. The photic responses and water-approach by hatchling behavior, morphology, and physiol- behavior of hatchling turtles. Copeia 1958:211–215. ogy. Generally, terrestrial hibernators exhibit Andersson S, Johansson L. 2001. Cold hardiness in the boreal better-developed capacities for supercooling and adder, Vipera berus. Cryo-Letters 22:151–156. inoculation resistance than species whose hatchl- Andjus RK. 1955. Suspended animation in cooled, supercooled and frozen rats. J Physiol (Lond) 128:547–556. ings usually hibernate under water. In principle, Andreas B, Paul R. 1998. Clutch size and structure of breeding smaller hatchlings are better able to avoid freezing chambers of Emys o. orbicularis in Brandenburg. In: through supercooling and inoculation resistance, Fritz U, Joger U, Podloucky R, Servan J, Buskirk JR, and the evidence suggests that hatchlings from editors. Proceedings of EMYS symposium, Vol. 10, Dresden, populations in cold climates are relatively small. Mertensiella, p 29–32. Ansart A, Vernon P. 2003. Cold hardiness in molluscs. Acta The possibility that the winter environment is a Oecologia 24:95–102. selective force shaping the life-history traits of Ansart A, Vernon P. 2004. Cold hardiness abilities vary with turtles merits further consideration. the size of the land snail Cornu aspersum. Comp Biochem Freeze tolerance is promoted by a complex suite Physiol A 139:205–211. of molecular, biochemical, and physiological Ansart A, Vernon P, Daguzan J. 2001. Freezing tolerance responses enabling hatchlings of certain species versus freezing susceptibility in the land snail Helix aspersa (Gastropoda: Helicidae). Cryo-Letters 22:183–190. to survive the freezing and thawing of extracel- Ansart A, Vernon P, Daguzan J. 2002. Effects of a freezing lular fluids. The adaptive strategies of freeze event during hibernation on further survival, reproduction tolerance include control of the ice-nucleation and growth in the partially freezing tolerant land snail Helix event, management of the redistribution of water aspersa Mu¨ller (Gastropoda, Helicidae). Cryo-Letters 23: and solutes, hypometabolism, upregulation of anti- 269–274. Aresco MJ. 2004. Reproductive ecology of Pseudemys floridana oxidant defense systems, and expression of genes and Trachemys scripta (Testudines: Emydidae) in north- involved in homeostasis and somatic repair. Such western Florida. J Herpetol 38:249–256. protective mechanisms probably originated from Attaway MB, Packard GC, Packard MJ. 1998. Hatchling fundamental responses to osmotic and hypoxic stress. painted turtles (Chrysemys picta) survive only brief freezing of their bodily fluids. Comp Biochem Physiol A 120: ACKNOWLEDGMENTS 405–408. Auger PJ, Giovannone P. 1979. On the fringe of exist- ence: diamondback terrapins at Sandy Neck. Cape Nat 8: Many colleagues have provided information and 44–58. insights concerning the various topics contained Averill-Murray RC, Martin BE, Bailey SJ, Wirt EB. 2002. herein. We thank them all for their considerable Activity and behavior of the Sonoran in input to this project. We thank G. Lorigan for Arizona. In: Van Devender TR, editor. The Sonoran Desert sharing unpublished data with us, and M. Elnitsky tortoise, natural history, biology, and conservation. Tucson, for reviewing an early draft of this article. The AZ: University of Arizona Press and the Arizona-Sonora Desert Museum, p 135–158. submitted paper was strengthened by the com- Baker PJ, Costanzo JP, Iverson JB, Lee RE. 2003. Adaptations mentaries and suggestions offered by J. Iverson to terrestrial overwintering of hatchling northern map turtles, and an anonymous referee. The writing, literature Graptemys geographica. J Comp Physiol B 173:643–651. review, and some of the research highlighted Baker PJ, Costanzo JP, Lee RE. 2004. Environmental herein were supported by National Science Foun- physiology of terrestrial hibernation in hatchling turtles. In: Barnes BM, Carey HV, editors. Life in the cold: dation Grants (IBN 9817087 and IAB 0416750 to evolution, mechanisms, adaptation, and application. Fair- J. P. C., and IBN 00765592 to G. R. U.). banks: Institute of Arctic Biology, University of Alaska Fairbanks. p 329–338. LITERATURE CITED Baker PJ, Costanzo JP, Herlands R, Wood RC, Lee RE. 2006. Inoculative freezing promotes winter survival in hatchling Aarset AV. 1982. Freezing tolerance in intertidal inverte- diamondback terrapin, Malaclemys terrapin. Can J Zool brates (a review). Comp Biochem Physiol A 73:571–580. 84:116–124.

J. Exp. Zool. 364 J.P. COSTANZO ET AL.

Baker PJ, Costanzo JP, Lee RE. 2007. Oxidative stress and Bogert CM, Cowles RB. 1947. Moisture loss in relation to antioxidant capacity of a terrestrially-hibernating hatchling habitat selection in some Floridian reptiles. Am Mus Nov turtle. J Comp Physiol B 177:875–883. 1358:1–34. Bale JS. 1993. Classes of insect cold hardiness. Funct Ecol Bohn M, Lu¨thje S, Sperling P, Heinz E, Do¨rffling K. 2007. 7:751–753. Plasma membrane lipid alterations induced by cold acclima- Bale JS. 1996. Insect cold hardiness: a matter of life and death. tion and abscisic acid treatment of winter wheat seedlings Eur J Entomol 93:369–382. differing in frost resistance. J Plant Physiol 164:146–156. Bale JS, Hansen TN, Baust JG. 1989. Nucleators and sites of Booth DT. 1998. Nest temperatures and respiratory gases nucleation in the freeze tolerant larvae of the gallfly Eurosta during natural incubation in the broad-shelled river turtle, solidaginis (Fitch). J Insect Physiol 35:291–298. expansa (Testudinata: ). Aust J Zool Bale JS, Worland MR, Block W. 2001. Effects of summer frost 46:183–191. exposures on the cold tolerance strategy of a sub-Antarctic Bouchard RW, Carillo MA, Kells SA, Ferrington LC. 2006. beetle. J Insect Physiol 47:1161–1167. Freeze tolerance in larvae of the winter-active Diamesa Barton AJ, Price JW. 1955. Our knowledge of the bog turtle, mendotae Muttkowski (Diptera: Chironomidae): a contrast Clemmys muhlenbergi, surveyed and augmented. Copeia to adult strategy for survival at low temperatures. Hydro- 1955:159–165. biologia 568:403–416. Baust JG. 1991. The freeze tolerance oxymoron. Cryo-Letters Bozhansky AT, Orlova VF. 1998. Conservation status of the 12:1–2. European pond turtle, Emys orbicularis (Linnaeus, 1758), in Baust JG, Rojas RR. 1985. Review—insect cold hardiness: European Russia. In: Fritz U, Joger U, Podloucky R, Servan facts and fancy. J Insect Physiol 31:755–759. J, Buskirk JR, editors. Proceedings of EMYS symposium, Belkin DA. 1963. Anoxia: tolerance in reptiles. Science Vol. 10, Dresden, Mertensiella, p 41–46. 139:492–493. Bra¨nna¨s E. 1995. First access to territorial space and exposure Belmore B. 1980. The basic ecology of the spotted turtle to strong predation pressure: a conflict in early emerging Clemmys guttata in Massachusetts. J Northern Ohio Assoc Atlantic salmon (Salmo salar, L.). Evol Ecol 9:411–420. Herpetol 6:5–13. Breckenridge WJ. 1960. A spiny soft-shelled turtle nest study. Bennett AF. 1997. Adaptation and the evolution of physiolo- Herpetologica 16:284–285. Breitenbach GL, Congdon JD, van Loben Sels RC. 1984. gical characters. In: Dantzler WH, editor. Handbook of Winter temperatures of Chrysemys picta nests in Michigan: physiology, Sect 13: comparative physiology. New York: effects on hatchling survival. Herpetologica 40:76–81. Oxford University Press. p 3–16. Britson CA, Gutzke WHN. 1993. Antipredator mechanisms of Bentley PJ, Schmidt-Nielsen K. 1966. Cutaneous water loss in hatchling freshwater turtles. Copeia 1993:435–440. reptiles. Science 151:1547–1549. Brooks RJ, Brown GP, Galbraith DA. 1991. Effects of a sudden Bernstein NP, Black RW. 2005. Thermal environment of increase in natural mortality of adults on a population of the overwintering ornate box turtles, Terrapene ornata ornata, (Chelydra serpentina). Can J Zool in Iowa. Am Midl Nat 153:370–377. 69:1314–1320. Bettelheim MP. 2004. The , Clemmys Buech RR, Nelson CE, Hanson LG, Ewert MA. 2004. Clemmys marmorata—a natural history of the species. Privately insculpta (Wood turtle). Hatching failure. Herpetol Rev printed, Walnut Creek, California. 35:54. Bickler PE, Buck LT. 2007. Hypoxia tolerance in reptiles, Buhlmann KA. 1998. Ecology, terrestrial habitat use, and amphibians, and fishes: life with variable oxygen availabil- conservation of a freshwater turtle assemblage inhabiting a ity. Annu Rev Physiol 69:145–170. seasonally fluctuating wetland with emphasis on the life Bigg EK. 1953. The supercooling of water. Proc Phys Soc B history of Deirochelys reticulara. PhD thesis, University of 66:688–694. Georgia, Athens, USA. Birchard GF, Packard GC. 1997. Cardiac activity in super- Buhlmann KA, Coffman G. 2001. Fire ant predation of turtle cooled hatchlings of the painted turtle (Chrysemys picta). nests and implications for the strategy of delayed emer- J Herpetol 31:166–169. gence. J Elisha Mitchell Sci Soc 117:94–100. Bischof JC, Wolkers WF, Tsvatkova NM, Oliver AE, Crowe Buhlmann KA, Gibbons JW. 2001. Terrestrial habitat use by JH. 2002. Lipid and protein changes due to freezing in aquatic turtles from a seasonally fluctuating wetland: dunning AT-1 cells. Cryobiology 45:22–32. implications for wetland conservation boundaries. Chel Bjurlin CD, Bissonette JA. 2004. Survival during early life Cons Biol 4:115–127. stages of the desert tortoise (Gopherus agassizii) in the Buhlmann KA, Vaughn MR. 1991. Ecology of the turtle south-central Mojave Desert. J Herpetol 38:527–535. Pseudemys concinna in the New River, West Virginia. Bleakney S. 1963. Notes on the distribution and life histories J Herpetol 25:72–78. of turtles in Nova Scotia. Can Field Nat 77:67–76. Burger J. 1976a. Behavior of hatchling diamondback terrapins Block W. 1991. To freeze or not to freeze? Inverte- (Malaclemys terrapin) in the field. Copeia 1976:742–748. brate survival of sub-zero temperatures. Funct Ecol 5: Burger J. 1976b. Temperature relations in nests of the 284–290. northern diamondback terrapin, Malaclemys terrapin. Block W. 1996. Cold or drought- the lesser of two evils for J Herpetol 32:412–418. terrestrial arthropods? Eur J Entomol 93:325–339. Burger J. 1977. Determinants of hatching success in diamond- Bloomer TJ, Bloomer DM. 1973. New Jersey’s bog turtle— back terrapin, Malaclemys terrapin. Am Midl Nat 97: destined to extinction? Bull N Y Herpetol Soc 9:8–12. 444–464. Blouin-Demers G, Prior KA, Weatherhead PJ. 2000. Patterns Burggren WW, Pinder AW. 1991. Ontogeny of cardiovascular of variation in spring emergence by black rat snakes (Elaphe and respiratory physiology in lower vertebrates. Annu Rev obsoleta obsoleta). Herpetologica 56:175–188. Physiol 53:107–135.

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 365

Burke RL, Hussain AA, Storey JM, Storey KB. 2002. Freeze Chessman BC. 1984. Evaporative water loss from three south- tolerance and supercooling ability in the Italian wall lizard, eastern Australia species of freshwater turtle. Aust J Zool Podarcis sicula, introduced to Long Island, New York. 32:649–655. Copeia 2002:836–842. Chilian WM. 1976. Physiological strategies of dormancy in Buskirk JR. 1991. An overview of the western pond turtle, Kinosternon flavescens. MS thesis, Texas Tech University, Clemmys marmorata. In: Beaman KR, Caporaso F, Tyler, USA. McKeown S, Graff MD, editors. Proceedings of the first Chippindale P. 1989. Courtship and nesting records for the international symposium on turtles and tortoises: conserva- spotted turtles, Clemmys guttata, in the Mer Bleue Bog, tion and captive husbandry. Orange, CA: Chapman Uni- southeastern Ontario. Can Field-Nat 103:289–291. versity. 16–23p. Christens E. 1990. Nest emergence lag in loggerhead sea Bustard HR. 1971. Temperature and water tolerances of turtles. J Herpetol 24:400–402. incubating sea turtle eggs. Br J Herpetol 4:196–198. Christens E, Bider JR. 1987. Nesting activity and hatching Butler JA. 2004. Nesting, nest predation and hatchling success of the painted turtle (Chrysemys picta marginata)in southwestern Quebec. Herpetologica 43:55–65. emergence of the Carolina diamondback terrapin, Malacl- Christiansen JL, Bickham JW. 1989. Possible historic effects emys terrapin centrata, in northeastern Florida. Am Midl of pond drying and winterkill on the behavior of Kinoster- Nat 152:145–155. non flavescens and Chrysemys picta. J Herpetol 23:91–94. Butler BO, Graham TE. 1993. Tracking hatchling Blanding’s Christiansen JL, Gallaway BJ. 1984. Raccoon removal, nesting turtles with fluorescent pigments. Herpetol Rev 24:21–22. success, and hatchling emergence in Iowa turtles with Butler BO, Graham TE. 1995. Early post-emergent behavior special reference to Kinosternon flavescens. Southwest Nat and habitat selection in hatchling Blanding’s turtles, 29:343–348. Emydoidea blandingii, in Massachusetts. Chel Cons Biol Christiansen JL, Vandewalle T. 2000. Effectiveness of three 1:183–196. trap types in drift fence surveys. Herpetol Rev 31:158–160. Butler JA, Hull TW. 1996. Reproduction of the tortoise, Christiansen JL, Cooper JA, Bickham JW, Gallaway BJ, Gopherus polyphemus, in northeastern Florida. J Herpetol Springer MD. 1985. Aspects of the natural history of the 30:14–18. yellow mud turtle Kinosternon flavescens (Kinosternidae) in Butler JA, Bowmann RD, Hull TW, Sowell S. 1995. Move- Iowa: a proposed endangered species. Southwest Nat ments and home range of hatchling and yearling gopher 30:413–425. tortoises, Gopherus polyphemus. Chel Cons Biol 1:173–180. Churchill TA, Storey KB. 1991. Metabolic responses to Cagle FR. 1937. Egg laying habits of the slider turtle freezing by organs of hatchling painted turtles Chrysemys (Pseudemys troostii), the painted turtle (Chrysemys picta), picta marginata and C. p. bellii. Can J Zool 69:2978–2984. and the musk turtle (Sternotherus odoratus). J Tenn Acad Churchill TA, Storey KB. 1992a. Natural freezing survival by Sci 12:87–95. painted turtles Chrysemys picta marginata and C. picta Cagle FR. 1942. Herpetological fauna of Jackson and Union bellii. Am J Physiol 262:R530–R537. Counties, Illinois. Am Midl Nat 28:164–200. Churchill TA, Storey KB. 1992b. Responses to freezing Cagle FR. 1944. Activity and winter changes of hatchling exposure of hatchling turtles Trachemys scripta elegans: Pseudemys. Copeia 1944:105–109. factors influencing the development of freeze tolerance by Cagle FR. 1950. The life history of the slider turtle, Pseudemys reptiles. J Exp Biol 167:221–233. scripta troostii (Holbrook). Ecol Monogr 20:31–54. Clark AB, Wilson DS. 1981. Avian breeding adaptations: Cagle FR. 1954. Observations on the life cycles of painted hatching asynchrony, brood reduction, and nest failure. turtles (Genus Chrysemys). Am Midl Nat 42:225–235. Q Rev Biol 56:253–277. Cagle KD, Packard GC, Miller K, Packard MJ. 1993. Effects of Clarke CJ, Buckley SL, Lindner N. 2002. Ice structuring the microclimate in natural nests on development of proteins—a new name for antifreeze proteins. Cryo-Letters embryonic painted turtles, Chrysemys picta. Funct Ecol 23:89–92. 7:653–660. Claussen DL, Costanzo JP. 1990. A simple model for estimating the ice content of freezing ectotherms. J Therm Cahn AR. 1933. Hibernation in the box turtle. Copeia Biol 15:223–231. 1933:13–14. Claussen DL, Kim Y. 1993. The effects of cooling, freezing, and Cahn AR. 1937. The turtles of Illinois. Ill Biol Monogr thawing on cardiac and skeletal muscle of the turtle, 16:9–218. Chrysemys picta. J Therm Biol 18:91–101. Cameron AT. 1930. Temperature and life and death. Trans R Claussen DL, Zani PA. 1991. Allometry of cooling, super- Soc Can 1930:53–92. cooling, and freezing in the freeze-tolerant turtle Chrysemys Carr AF. 1952. Handbook of turtles: the turtles of the United picta. Am J Physiol 261:R626–R632. States, Canada, and Baja California. Ithaca, NY: Cornell Claussen DL, Townsley MD, Bausch RG. 1990. Supercooling University Press. and freeze-tolerance in the European wall lizard, Podarcis Carr AF, Hirth H. 1961. Social facilitation in green turtle muralis, with a revisional history of the discovery of freeze- siblings. Anim Behav 9:68–70. tolerance in vertebrates. J Comp Physiol B 160:137–143. Carroll DM, Ultsch GR. 2007. Emergence season and survival Claussen DL, Daniel PM, Jiang S, Adams NA. 1991. Hiberna- in the nest of hatchling turtles in southcentral New tion in the eastern box turtle, Terrapene c. carolina. Hampshire. Northeast Nat 2:307–310. J Herpetol 25:334–341. Chase JD, Dixon KR, Gates JE, Jacobs D, Taylor GJ. 1989. Conant R. 1938. The reptiles of Ohio. Am Midl Nat 20:1–200. Habitat characteristics, population size, and home range of Congdon JD, Gibbons JW. 1985. Egg components and the bog turtle, Clemmys muhlenbergii, in Maryland. reproductive characteristics of turtles: relationships to body J Herpetol 23:356–362. size. Herpetologica 41:194–205.

J. Exp. Zool. 366 J.P. COSTANZO ET AL.

Congdon JD, Gibbons JW. 1990. Turtle eggs: their ecology and Costanzo JP, Lee RE, Wright MF. 1992. Cooling rate evolution. In: Gibbons JW, editor. Life history and ecology of influences cryoprotectant distribution and organ dehydra- the slider turtle. Washington, DC: Smithsonian Institution tion in freezing wood frogs. J Exp Zool 261:373–378. Press. p 109–123. Costanzo JP, Wright MF, Lee RE. 1993. Physiological Congdon JD, Gibbons JW, Greene JL. 1983a. Parental responses to freezing in the turtle Terrapene carolina. investment in the chicken turtle, Deirochelys reticularia. J Herpetol 27:117–120. Ecology 64:419–425. Costanzo JP, Grenot C, Lee RE. 1995a. Supercooling, ice Congdon JD, Tinkle DW, Breitenbach GL, van Loben Sels RC. inoculation, and freeze tolerance in the European common 1983b. Nesting ecology and hatching success in the turtle lizard, Lacerta vivipara. J Comp Physiol B 165:238–244. Emydoidea blandingii. Herpetologica 39:417–429. Costanzo JP, Iverson JB, Wright MF, Lee RE. 1995b. Cold Congdon JD, Tinkle DW, Rosen PC. 1983c. Egg components hardiness and overwintering strategies of hatchlings in an and utilization during development in aquatic turtles. assemblage of northern turtles. Ecology 76:1772–1785. Copeia 1983:264–268. Costanzo JP, Moore JB, Lee RE, Kaufman PE, Wyman JA. Congdon JD, Breitenbach GL, van Loben Sels RC, Tinkle DW. 1997. Influence of soil hydric parameters on the winter cold 1987. Reproduction and nesting ecology of snapping turtles hardiness of a burrowing beetle, Leptinotarsa decemlineata (Chelydra serpentina) in southeastern Michigan. Herpeto- (Say). J Comp Physiol 167:169–176. logica 43:39–54. Costanzo JP, Litzgus JD, Iverson JB, Lee RE. 1998. Soil Congdon JD, Gotte SW, McDiarmid RW. 1992. Ontogenetic hydric characteristics and environmental ice nuclei influ- changes in habitat use by juvenile turtles, Chelydra serpentina and Chrysemys picta. Can Field Nat 106:241–248. ence supercooling capacity of hatchling painted turtles, Congdon JD, Nagle RD, Kinney OM, Osentoski M, Avery HW, Chrysemys picta. J Exp Biol 201:3105–3112. van Loben Sels RC, Tinkle DW. 2000. Nesting ecology and Costanzo JP, Litzgus JD, Lee RE. 1999. Behavioral responses embryo mortality: implications for hatchling success and of hatchling painted turtles (Chrysemys picta) and snapping demography of Blanding’s turtles (Emydoidea blandingii). turtles (Chelydra serpentina) at subzero temperatures. Chel Cons Biol 3:569–579. J Therm Biol 24:161–166. Converse SJ, Iverson JB, Savidge JA. 2002. Activity, repro- Costanzo JP, Litzgus JD, Iverson JB, Lee RE. 2000a. Ice duction and overwintering behavior of ornate box turtles nuclei in soil compromise cold hardiness of hatchling (Terrapene ornata ornata) in the Nebraska Sandhills. Am painted turtles, Chrysemys picta. Ecology 81:346–360. Midl Nat 148:416–422. Costanzo JP, Litzgus JD, Iverson JB, Lee RE. 2000b. Seasonal Convey P. 1997. How are the life history strategies of changes in physiology and development of cold hardiness in Antarctic terrestrial invertebrates influenced by extreme the hatchling painted turtle, Chrysemys picta. J Exp Biol environmental conditions? J Therm Biol 22:429–440. 203:3459–3470. Cook RT. 2004. The tolerance of the field slug Deroceras Costanzo JP, Jones EE, Lee RE. 2001a. Physiological reticulatum to freezing temperatures. Cryo-Letters 25:187–194. responses to supercooling and hypoxia in the hatchling Cooke WW. 1910. Incubation period of box-turtle eggs. Proc painted turtle, Chrysemys picta. J Comp Physiol B Biol Soc Wash 23:124. 171:335–340. Costanzo JP. 1985. The bioenergetics of hibernation in the Costanzo JP, Litzgus JD, Iverson JB, Lee RE. 2001b. Cold- eastern garter snake Thamnophis sirtalis sirtalis. Physiol hardiness and evaporative water loss in hatchling turtles. Zool 58:682–692. Physiol Biochem Zool 74:510–519. Costanzo JP. 1986. Influences of hibernaculum microenviron- Costanzo JP, Litzgus JD, Larson JL, Iverson JB, Lee RE. ment on the winter life history of the garter snake, 2001c. Characteristics of nest soil, but not geographic origin, Thamnophis sirtalis. Ohio J Sci 86:199–204. influence cold hardiness of hatchling painted turtles. Costanzo JP. 1988. Ecophysiological adaptations to over- J Therm Biol 26:65–73. wintering in the eastern garter snake, Thamnophis sirtalis Costanzo JP, Baker PJ, Dinkelacker SA, Lee RE. 2003. sirtalis. PhD thesis, Miami University, Oxford, USA. Endogenous and exogenous ice-nucleating agents constrain Costanzo JP, Claussen DL. 1990. Natural freeze tolerance in supercooling in the hatchling painted turtle. J Exp Biol the terrestrial turtle, Terrapene carolina. J Exp Zool 206:477–485. 254:228–232. Costanzo JP, Dinkelacker SA, Iverson JB, Lee RE. 2004. Costanzo JP, Lee RE. 1995. Supercooling and ice nucleation in Physiological ecology of overwintering in the hatchling vertebrates. In: Lee RE, Warren GJ, Gusta LV, editors. painted turtle: multiple-scale variation in response to Biological ice nucleation and its applications. St. Paul, MN: American Phytopathological Society Press. p 221–237. environmental stress. Physiol Biochem Zool 77:74–99. Costanzo JP, Lee RE. 1996. Mini review: ice nucleation in Costanzo JP, Baker PJ, Lee RE. 2006. Physiological responses freeze-tolerant vertebrates. Cryo-Letters 17:111–118. to freezing in hatchlings of freeze-tolerant and -intolerant Costanzo JP, Lee RE. 2005. Cryoprotection by urea in a turtles. J Comp Physiol B 176:697–707. terrestrially-hibernating frog. J Exp Biol 208:4079–4089. Crawford KM. 1991a. The effect of temperature and seasonal Costanzo JP, Claussen DL, Lee RE. 1988. Natural freeze acclimatization on renal function of painted turtles, Chrys- tolerance in a . Cryo-Letters 9:380–385. emys picta. Comp Biochem Physiol A 99:375–380. Costanzo JP, Lee RE, Wright MF. 1991a. Effect of cooling rate Crawford KM. 1991b. The winter environment of painted on the survival of frozen wood frogs, Rana sylvatica. J Comp turtles, Chrysemys picta: temperature, dissolved oxygen, Physiol B 161:225–229. and potential cues for emergence. Can J Zool 69:2493–2498. Costanzo JP, Lee RE, Wright MF. 1991b. Glucose loading Crocker CA, Ultsch GR, Jackson DC. 1999. The physiology of prevents freezing injury in rapidly-cooled wood frogs. Am diving in a north-temperate and three tropical turtle J Physiol 261:R1549–R1553. species. J Comp Physiol B 169:249–255.

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 367

Croes SA, Thomas RE. 2000. Freeze tolerance and cryopro- Drobnis EZ, Crowe LM, Berger T, Anchordoguy TJ, Over- tectant synthesis of the Pacific tree frog Hyla regilla. Copeia street JW, Crowe JH. 1993. Cold shock damage is due to 2000:863–868. lipid phase transitions in cell membranes: a demonstration Danforth BN. 1999. Emergence dynamics and bet hedging in a using sperm as a model. J Exp Zool 265:432–437. desert bee, Perdita portalis. Proc R Soc Lond B 266: Du WG, Zheng RQ, Shu L. 2006. The influence of incubation 1985–1994. temperature on morphology, locomotor performance, and Danks HV. 2000. Dehydration in dormant insects. J Insect cold tolerance of hatchling Chinese three-keeled pond Physiol 46:837–852. turtles, Chinemys reevesii. Chel Cons Biol 5:294–299. Dantzler WH, Schmidt-Nielsen B. 1966. Excretion in fresh- Duman JG. 2001. Antifreeze and ice nucleator proteins in water turtle (Pseudemys scripta) and desert tortoise terrestrial arthropods. Annu Rev Physiol 63:327–357. (Gopherus agassizii). Am J Physiol 210:198–210. Duman J. 2002. The inhibition of ice nucleators by insect Deeming DC. 2004a. Reptilian incubation: environment, antifreeze proteins is enhanced by glycerol and citrate. evolution and behaviour. Nottingham: Nottingham Univer- J Comp Physiol B 172:163–168. sity Press. 349p. Duman JG, Xu L, Neven LG, Tursman D, Wu DW. 1991. Deeming DC. 2004b. Post-hatching phenotypic effects of Hemolymph proteins involved in insect low temperature incubation in reptiles. In: Deeming DC, editor. Reptilian tolerance: ice nucleators and antifreeze proteins. In: Lee incubation: environment, evolution and behaviour. Notting- RE, Denlinger DL, editors. Insects at low temperature. New ham: Nottingham University Press. p 229–251. York: Chapman & Hall. p 94–127. Degenhardt WG, Christiansen JL. 1974. Distribution and Duman JG, Olsen TM, Yeung KL, Jerva F. 1995. The roles of habitats of turtles in New Mexico. Southwest Nat 19:21–46. ice nucleators in cold tolerant invertebrates. In: Lee RE, DePari JA. 1988. Overwintering in the nest chamber by Warren GJ, Gusta LV, editors. Biological ice nucleation and hatchling painted turtles, Chrysemys picta, in northern New its applications. St. Paul, Minnesota: American Phytopatho- Jersey. PhD thesis, Rutgers University, Newark, USA. logical Society Press. p 201–219. DePari JA. 1996. Overwintering in the nest chamber by Duman J, Bennett V, Sformo T, Hochstrasser R, Barnes BM. hatchling painted turtles, Chrysemys picta, in northern New 2004. Antifreeze proteins in Alaskan insects and spiders. Jersey. Chel Cons Biol 2:5–12. J Insect Physiol 50:259–266. Derickson WK. 1976. Lipid storage and utilization in reptiles. Dundee HA, Rossman DA. 1989. The amphibians and reptiles Am Zool 16:711–723. of Louisiana. Baton Rouge: Louisiana State University DeVries AL. 1982. Biological antifreezes in coldwater fishes. Press. Comp Biochem Physiol 73A:627–640. Dunson WA. 1967. Sodium fluxes in fresh-water turtles. J Exp Dı´az-Paniagua C, Keller C, Andreu AC. 1997. Hatching Zool 165:171–182. success, delay of emergence and hatchling biometry of the Dunson WA, Seidel ME. 1986. Salinity tolerance of estuarine spur-thighed tortoise, Testudo graeca, in south-western and insular emydid turtles (Pseudemys nelsoni and Tra- Spain. J Zool (Lond) 243:543–553. chemys decussata). J Herpetol 20:237–245. Dinkelacker SA, Costanzo JP, Iverson JB, Lee RE. 2004. Cold Edgren RA. 1942. A nesting rendezvous of the musk turtle. hardiness and dehydration resistance of hatchling Bland- Chicago Nat 5:63. ing’s turtles (Emydoidea blandingii): implications for over- Einum S, Fleming IA. 2000. Selection against late emergence wintering in a terrestrial habitat. Can J Zool 82:594–600. Dinkelacker SA, Costanzo JP, Iverson JB, Lee RE. 2005a. and small offspring in Atlantic salmon (Salmo salar). Survival and physiological responses of hatchling Blanding’s Ecology 54:628–639. turtles (Emydoidea blandingii) to submergence in normoxic Epperson DM, Heise CD. 2003. Nesting and hatchling ecology and hypoxic water under simulated winter conditions. of gopher tortoises (Gopherus polyphemus) in southern Physiol Biochem Zool 78:356–363. Mississippi. J Herpetol 37:315–324. Dinkelacker SA, Costanzo JP, Lee RE. 2005b. Anoxia Ernst CH. 1966. Overwintering of hatchling Chelydra serpen- tolerance and freeze tolerance in hatchling turtles. J Comp tina in southeastern Pennsylvania. Phil Herpetol Soc Bull Physiol B 175:209–217. 14:8–9. Donohoe PH, Boutilier RG. 1998. The protective effects of Ernst CH. 1968. Evaporative water-loss relationships of metabolic rate depression in hypoxic cold submerged frogs. turtles. J Herpetol 2:159–161. Respir Physiol 111:325–336. Ernst CH. 1971. Population dynamics and activity cycles of Doody JS, Georges A. 2000. A novel technique for gathering Chrysemys picta in southeastern Pennsylvania. J Herpetol turtle nesting and emergence phenology data. Herpetol Rev 5:151–160. 31:220–222. Ernst CH. 1975. Growth of the spotted turtle, Clemmys Doroff AM, Keith LB. 1990. Demography and ecology of an guttata. J Herpetol 9:313–318. (Terrapene ornata) population in south- Ernst CH. 1976. Ecology of the spotted turtle, Clemmys central Wisconsin. Copeia 1990:387–399. guttata (Reptilia, Testudines, Testudinidae), in southwes- Draud M, Bossert M, Zimnavoda S. 2004. Predation on tern Pennsylvania. J Herpetol 10:25–33. hatchling and juvenile diamondback terrapins (Malaclemys Ernst CH. 1986. Ecology of the turtle, Sternotherus odoratus, terrapin) by the Norway rat (Rattus norvegicus). J Herpetol in southeastern Pennsylvania. J Herpetol 20:341–352. 38:467–470. Ernst CH. 2001. Some ecological parameters of the wood Drda WJ. 1968. A study of snakes wintering in a small cave. turtle, Clemmys insculpta, in southeastern Pennsylvania. J Herpetol 1:64–70. Chel Cons Biol 4:94–99. Drobenkov SM. 2000. Reproductive ecology of the pond turtle Ernst CH, Barbour RW, Lovich JE. 1994. Turtles of the (Emys orbicularis L.) in the northeastern part of the species United States and Canada. Washington, DC: Smithsonian range. Russian J Ecol 31:49–54. Institution Press. 578p.

J. Exp. Zool. 368 J.P. COSTANZO ET AL.

Etheridge K, Wit LC, Sellers JC. 1983. Hibernation in the (Emydidae: Terrapene carolina) in central Illinois. Herpetol lizard Cnemidophorus sexlineatus (Lacertilia: Teiidae). Rev 35:333–335. Copeia 1983:206–214. Friebele E, Swarth CW. 2005. Potential factors determining Ewert MA. 1985. Embryology of turtles. In: Gans C, Billett F, hatchling emergence patterns in red-bellied turtles (Pseud- Maderson PFA, editors. Biology of the Reptilia. New York: emys rubriventris). Abstracts of 2005 meeting of the John Wiley and Sons, p 75–267. American Society of ichthyologists and herpetologists, Ewert, MA. 1991. Cold torpor, diapause, delayed hatching, and Tampa, Fl. aestivation in reptiles and birds. In: Deeming DC, Ferguson Frisbie MP, Lee RE. 1997. Inoculative freezing and the MWJ, editors. Egg incubation: its effect on embryonic problem of winter survival for freshwater macroinverte- development in birds and reptiles. Cambridge, England: brates. J N Am Benthol Soc 16:635–650. Cambridge University Press, p 173–191. Frye FL. 1973. Husbandry, medicine, and surgery in captive Ewert MA, Wilson DS. 1996. Seasonal variation of embryonic reptiles. Bonner Springs, KS: V. M. Publishing. diapause in the striped mud turtle (Kinosternon baurii) and Fukuta N. 1966. Experimental studies of organic ice nuclei. general considerations for conservation planning. Chel Cons J Atmos Sci 23:191–196. Biol 2:43–54. Galbraith DA, Brooks RJ. 1987. Survivorship of adult females Ewert MA, Jackson DA, Nelson CE. 1994. Patterns of in a northern population of common snapping turtles, temperature-dependent sex determination in turtles. Chelydra serpentina. Can J Zool 65:581–586. J Exp Zool 270:3–15. Garrick LD. 1972. Temperature influences on hibernation in Ewing HE. 1933. Reproduction in the eastern box turtle Terrapene carolina carolina (Linne´). Copeia 1933:95–96. Sceloporus occidentalis. J Herpetol 6:195–198. Fahey KM. 1987. Aspects of the life history of the river cooter, Gehrken U. 1992. Inoculative freezing and thermal hysteresis Pseudemys concinna (Le Conte), in the Tallapoosa River, in the adult beetles Ips acuminatus and Rhagium inquisitor. Tallapoosa County, Alabama. PhD thesis, Auburn Univer- J Insect Physiol 38:519–524. sity, Auburn, USA. Gemmell DJ. 1970. Some observations on the nesting of the Farkas B. 2000. The European pond turtle Emys orbicularis western painted turtle, Chrysemys picta belli, in northern (L.) in Hungary. Stapfia 69, zugleich Kataloge des OO¨ Minnesota. Can Field-Nat 84:308–309. Landesmuseums, Neue Folge Nr 149, p 127–132. Georges A. 1992. Thermal characteristics and sex determina- Farrell RF, Graham TE. 1991. Ecological notes on the turtle tion in field nests of the pig-nosed turtle, Carettochelys Clemmys insculpta in northwestern New Jersey. J Herpetol insculpta (Chelonia: Carettochelydidae), from Northern 25:1–9. Australia. Aust J Zool 40:511–521. Fields PG, McNeil JN. 1986. Possible dual cold-hardiness Gibbons JW. 1969. Ecology and population dynamics of the strategies in Cisseps fulvicollis (Lepidoptera: Arctiidae). chicken turtle, Deirochelys reticularia. Copeia 1969: Can Entomol 118:1309–1311. 669–676. Filoramo N, Janzen FJ. 1999. Effects of hydric conditions Gibbons JW, Coker JW. 1977. Ecological and life history during incubation on overwintering hatchlings of the red- aspects of the cooter, Chrysemys floridana (Le Conte). eared slider turtle (Trachemys scripta elegans). J Herpetol Herpetologica 33:29–33. 33:29–35. Gibbons JW, Greene JL. 1978. Selected aspects of the ecology Finkler M. 1999. Influence of water availability during of the chicken turtle, Deirochelys reticularia (Latreille) incubation on hatchling size, body composition, desiccation (Reptilia, Testudines, Emydidae). J Herpetol 12:237–341. tolerance, and terrestrial locomotor performance in the Gibbons JW, Nelson DH. 1978. The evolutionary significance snapping turtle Chelydra serpentina. Physiol Biochem Zool of delayed emergence from the nest by hatchling turtles. 72:714–722. Evolution 32:297–303. Finkler MS, Knickerbocker DL, Claussen DL. 2000. Influence Glass ML, Hicks JW, Riedesel ML. 1979. Respiratory of hydric conditions during incubation and population on responses to long-term temperature exposure in the box overland movement of neonatal snapping turtles. J Herpetol turtle, Terrapene ornata. J Comp Physiol 131:353–359. 34:452–455. Glen F, Broderick AC, Godley BJ, Hays GC. 2005. Patterns in Finkler MS, Bowen JT, Christman TM, Renshaw AD. 2002. the emergence of green (Chelonia mydas) and loggerhead Effects of hydric conditions during incubation on body size (Caretta caretta) turtle hatchlings from their nests. Mar Biol and triglyceride reserves of overwintering hatchling snap- 146:1039–1049. ping turtles (Chelydra serpentina). Copeia 2002:504–510. Goode J, Russell J. 1968. Incubation of eggs of three species of Finkler MS, Steyermark AC, Jenks KE. 2004. Geographic variation in snapping turtle (Chelydra serpentina serpentina) Chelid tortoises, and notes on their embryological develop- egg components across a longitudinal transect. Can ment. Aust J Zool 16:749–761. J Zool 82:102–109. Graham TE, Doyle TS. 1978. Dimorphism, courtship, eggs, Finneran LC. 1948. Reptiles of Branford, Connecticut. and hatchlings of the Blanding’s turtle, Emydoidea blan- Herpetologica 4:123–132. dingii (Reptilia, Testudines, Emydidae) in Massachusetts. Fitch HS, Plummer MV. 1975. A preliminary ecological study J Herpetol 13:125–127. of the soft-shelled turtle muticus in the Kansas Greaves WF, Litzgus JD. 2007. Overwintering ecology of wood River. Isr J Zool 24:28–42. turtles (Glyptemys insculpta) at the species’’ northern range Forge TA, MacGuidwin AE. 1992. Effects of water potential limit. J Herpetol 41:32–40. and temperature on survival of the nematode Meloidogyne Greenway SC, Storey KB. 1999. Discordant responses of hapla. Can J Zool 70:1553–1560. mitogen-activated protein kinases to anoxia and freezing Forsythe P, Flitz B, Mullin SJ. 2004. Radio telemetry and exposures in hatchling turtles. J Comp Physiol B 169: post-emergent habitat selection on neonate box turtles 521–527.

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 369

Gregory PT. 1982. Reptilian hibernation. In: Gans C, Pough Herbert CV, Jackson DC. 1985. Temperature effects on the FH, editors. Biology of the Reptilia. New York: Academic responses to prolonged submergence in the turtle Chrys- Press. p 53–154. emys picta bellii: II. Metabolic rate, blood acid–base and Grobman AB. 1990. The effect of soil temperatures on ionic changes, and cardiovascular function in aerated and emergence from hibernation of Terrapene carolina and anoxic water. Physiol Zool 58:670–681. T. ornata. Am Midl Nat 124:366–371. Herman TB, Power TD, Eaton BR. 1995. Status of Blanding’s Gurian-Sherman D, Lindow SE. 1995. Differential effects of turtles, Emydoidea blandingii, in Nova Scotia, Canada. Can growth temperature on ice nuclei active at different Field-Nat 109:182–191. temperatures that are produced by cells of Pseudomonas Hermes-Lima M, Storey KB. 1993. Antioxidant defenses in the syringae. Cryobiology 32:129–138. tolerance of freezing and anoxia by garter snakes. Am Guschina IA, Harwood JL. 2006. Mechanisms of temperature J Physiol 265:R646–R652. adaptation in poikilotherms. FEBS Lett 580:5477–5483. Hermes-Lima M, Storey JM, Storey KB. 1998. Antioxidant Hall FG. 1922. The vital limit of exsiccation of certain animals. defenses and metabolic depression. The hypothesis of Biol Bull 42:31–51. preparation for oxidative stress in land snails. Comp Halpern EA. 1979. Supercooling in the iguanid lizard Biochem Physiol B 120:437–448. Sceloporus jarrovi. Physiol Zool 52:190–204. Hermes-Lima M, Zenteno-Savin T. 2002. Animal response to Halpern EA, Lowe CH. 1968. Metabolism of the iguanid lizard drastic changes in oxygen availability and physiological Uta stansburiana in the supercooled state. Physiol Zool oxidative stress. Comp Biochem Physiol C 133:537–556. 41:113–124. Hamilton WJ. 1940. Observations on the reproductive Hirano SS, Upper CD. 1995. Ecology of ice nucleation-active behavior of the snapping turtle. Copeia 1940:124–126. bacteria. In: Lee RE, Warren GJ, Gusta LV, editors. Hammer DA. 1969. Parameters of a marsh snapping turtle Biological ice nucleation and its applications. St. Paul, population. J Wildl Manage 33:995–1005. MN: American Phytopathological Society Press. p41–61. Hammer DA. 1972. Ecological relations of waterfowl and Hoffman NE, Bischof JC. 2005. Mechanisms of injury caused snapping turtle populations. PhD thesis, Utah State Uni- by in vivo freezing. In: Fuller B, Lane N, Benson EE, editors. versity, Logan, USA. Life in the frozen state. Boca Raton, FL: CRC Press. Harding JH. 1991. A twenty year wood turtle study in p 455–481. Michigan: implications for conservation. First international Holland DC. 1994. The western pond turtle: habitat and symposium of turtles and tortoises, conservation and captive history. Final Report to the US Department of Energy, husbandry, p 31–35. Bonneville Power Administration, Environment, Fish and Harding JH, Bloomer TJ. 1979. The wood turtle, Clemmys Wildlife Portland, OR. insculpta—a natural history. HERP—Bull N Y Herpetol Soc Holman JA. 1994. Status of the red-eared slider turtle 15:9–26. Trachemys scripta elegans (Wied) in Michigan: a prelimin- Harms HK, Paitz RT, Bowden RM, Janzen FJ. 2005. Age and ary report. Michigan Academian 26:471–477. season impact resource allocation to eggs and nesting Holman JA, Andrews KD. 1994. North American Quaternary behavior in the painted turtle. Physiol Biochem Zool cold-tolerant turtles: distributional adaptations and con- 78:996–1004. straints. Boreas 23:44–52. Hart DR. 1983. Dietary and habitat shift with size of red-eared Holmstrup M. 2003. Overwintering adaptations in earth- turtles (Pseudemys scripta) in a southern Louisiana popula- worms. Pedobiologia 47:504–510. tion. Herpetologica 39:285–290. Holmstrup M, Bayley M, Ramlov H. 2002. Supercool or Hartley LM, Packard MJ, Packard GC. 2000. Accumulation of dehydrate? An experimental analysis of overwintering lactate by supercooled hatchlings of the painted turtle strategies in small permeable Arctic invertebrates. Proc (Chrysemys picta): implications for overwinter survival. Natl Acad Sci USA 99:5716–5720. J Comp Physiol B 170:45–50. Holmstrup M, Overgaard J, Bayley M. 2007. Small Dendro- Hartweg N. 1944. Spring emergence of painted turtle baena earthworms survive freezing better than large worms. hatchlings. Copeia 1944:20–22. Cryobiology 54:298–300. Hawes T, C., Bale JS, Convey P, Worland MR. 2006. Holte DL. 1988. Nest site characteristics of the western pond Ecologically realistic modalities in arthropod supercooling turtle, Clemmys marmorata, at Fern Ridge Reservoir, in point distributions. Eur J Entomol 103:717–723. west central Oregon. MS thesis, Oregon State University, Hays GC, Speakman JR, Hayes JP. 1992. The pattern of Corvallis, USA. emergence by loggerhead turtle (Caretta caretta) hatchlings on Cephalonia, Greece. Herpetologica 48:396–401. Holub RJ, Bloomer TJ. 1977. The bog turtle, Clemmys Hays DW, McAllister KM, Richardson SA, Stinson DW. 1999. muhlenbergii...a natural history. HERP—Bulletin of the Washington State recovery plan for the western pond turtle. New York Herpetological Society 13:9–23. Washington Department of Fish and Wildlife, Olympia, Horne BD, Brauman MJ, Moore JC, Seigel RA. 2003. Washington. Reproductive and nesting ecology of the yellow-blotched Hayward SA, Rinehart JP, Sandro LH, Lee RE, Denlinger DL. map turtles, Graptemys flavimaculata: implications for 2007. Slow dehydration promotes desiccation and freeze conservation and management. Copeia 2003:729–738. tolerance in the Antarctic midge Belgica antarctica. J Exp Horwath KL, Duman JG. 1984. Yearly variations in the Biol 210:836–844. overwintering mechanisms of the cold-hardy beetle Den- Hemmings SJ, Storey KB. 2000. Hepatic changes in the droides canadensis. Physiol Zool 57:40–45. freeze-tolerant turtle Chrysemys picta marginata in Hotaling EC, Wilhoft DC, McDowell SB. 1985. Egg position response to freezing and thawing. Cell Biochem Funct and weight of hatchling snapping turtles, Chelydra serpen- 18:175–186. tina, in natural nests. J Herpetol 19:534–536.

J. Exp. Zool. 370 J.P. COSTANZO ET AL.

Houghton JDR, Hays GC. 2001. Asynchronous emergence by Jackson DC. 2002. Hibernating without oxygen: physiological loggerhead turtle (Caretta caretta) hatchlings. Naturwis- adaptations of the painted turtle. J Physiol (Lond) senschaften 88:133–136. 543.3:731–3:737. Hughes EJ, Brooks RJ. 2006. The good mother: does nest-site Jackson DC. 2004. Overwintering in submerged turtles. In: selection constitute parental investment in turtles? Can Barnes BM, Carey HV, editors. Life in the cold: evolution, J Zool 84:1545–1554. mechanisms, adaptation, and application. Fairbanks: In- Hulse AC, McCoy CJ, Censky E. 2001. Amphibians and stitute of Arctic Biology, University of Alaska Fairbanks. reptiles of Pennsylvania and the Northeast. Ithaca, NY: p 299–306. Comstock Publishing Associates. Jackson DR. 1988. Reproductive strategies of sympatric Ims RA. 1990. On the adaptive value of reproductive freshwater emydid turtles in northern peninsular Florida. synchrony as a predator-swamping strategy. Am Nat Bull Florida State Mus Biol Sci 33:113–158. 136:485–498. Jackson DR. 1994. Overwintering of hatchling turtles in Irwin JT, Lee RE. 2000. Mild winter temperatures reduce northern Florida. J Herpetol 28:401–402. survival and potential fecundity of the goldenrod gall fly, Jackson DC, Crocker CE, Ultsch GR. 2000. Bone and shell Eurosta solidaginis (Diptera: Tephritidae). J Insect Physiol contribution to lactic acid buffering of submerged turtles 46:655–661. Chrysemys picta bellii at 31C. Am J Physiol 278: Issartel J, Voituron Y, Odagescu V, Baudot A, Guillot G, R1564–R1571. Ruaud J-P, Vernon P, Hervant F. 2006. Freezing or Jackson DA, Taylor MJ, Asare VS, Villarnovo D, Gall JM, supercooling: how does an aquatic subterranean crustacean Reese SA. 2007. Comparative shell buffering properties survive exposures at subzero temperatures? J Exp Biol correlate with anoxia tolerance in freshwater turtles. Am 209:3469–3475. J Physiol 292:R1008–R1015. Iverson JB. 1977. Reproduction in freshwater and terrestrial Janzen FJ. 1993. An experimental analysis of natural turtles of north Florida. Herpetologica 33:205–212. selection on body size of hatchling turtles. Ecology Iverson JB. 1980. The reproductive biology of Gopherus 74:332–341. polyphemus (Chelonia: Testudinidae). Am Midl Nat 103: Janzen FJ. 1994. Vegetational cover predicts the sex ratio of 353–359. hatchling turtles in natural nests. Ecology 75:1593–1599. Iverson JB. 1990. Nesting and parental care in the mud turtle Janzen FJ, Morjan CL. 2002. Egg size, incubation tempera- Kinosternon flavescens. Can J Zool 69:385–391. ture, and posthatching growth in painted turtles (Chrys- Iverson JB. 1991a. Patterns of survivorship in turtles (Order emys picta). J Herpetol 36:308–311. Testudines). Can J Zool 68:230–233. Janzen FJ, Packard GC, Packard MJ, Boardman TJ, Zum- Iverson JB. 1991b. Life history and demography of the yellow brunnen JR. 1990. Mobilization of lipid and protein by mud turtle, Kinosternon flavescens. Herpetologica 47: embryonic snapping turtles in wet and dry environments. 373–395. J Exp Zool 255:155–162. Iverson JB. 1992. Correlates of reproductive output in turtles Joanisse DR, Storey KB. 1996. Oxidative damage and (Order Testudines). Herpetol Monogr 6:25–42. antioxidants in Rana sylvatica, the freeze-tolerant wood Iverson JB. 1999. Reproduction in the frog. Am J Physiol 271:R545–R553. Kinosternon integrum. J Herpetol 33:144–148. Johnson BB. 1990. Freeze tolerance in the adult painted Iverson JB, Smith GR. 1993. Reproductive ecology of the turtle, Chrysemys picta, and the effects of freezing on the painted turtle (Chrysemys picta) in the Nebraska Sandhills cardiac cycle. Oxford, OH: Miami University. and across its range. Copeia 1993:1–21. Joyal LA, editor. 1999. Spotted turtle—Clemmys guttata. In: Iverson JB, Balgooyen CP, Byrd KK, Lyddan KK. 1993. Hunter ML, Calhoun AJK, McCollough M, editors. Maine Latitudinal variation in egg and clutch size in turtles. Can amphibians and reptiles. Orono: University of Maine Press. J Zool 71:2448–2461. p 134–137. Iverson JB, Brown RM, Akre TS, Near TJ, Le M, Thomson Judd FW, McQueen JC. 1980. Incubation, hatching, and RC, Starkey DE. 2007. In search of the tree of life for turtles. growth of the tortoise, Gopherus berlandieri. J Herpetol In: Shaffer HB, FitzSimmons NN, Georges A, Rhodin AGJ, 14:377–380. editors. Defining turtle diversity: proceedings of a workshop Kalabukhov NI. 1958. The problem of freezing, undercooling on genetics, ethics, and taxonomy of tortoises and fresh- and vitrifying of animal organism. 2nd edition. Oxford: water turtles. Chelon Res Monogr 4:85–106. Blackwell Scientific Publications. p 101–111. Iverson JB, Meylan PA, Seidel ME. 2008. Testudines–turtles. Kam Y-C. 1994. Effects of simulated flooding on metabolism In: Crother BI, editor. Scientific and standard English and water balance of turtle eggs and embryos. J Herpetol names of amphibians and reptiles of North America North 28:173–178. of Mexico. Herpetol Circ 37:67–74. Kanwisher J. 1959. Histology and metabolism of frozen Izumi Y, Sonada S, Yoshida H, Danks HV, Tsumuki H. 2006. intertidal animals. Biol Bull 116:258–264. Role of membrane transport of water and glycerol in the Kinney JS, White FN. 1977. Oxidative cost of ventilation in a freeze tolerance of the rice stem borer, Chilo suppressalis turtle, Pseudemys floridana. Respir Physiol 31:327–332. Walker (Lepidoptera: Pyralidae). J Insect Physiol Klose MK, Robertson RM. 2004. Stress-induced thermopro- 52:215–220. tection of neuromuscular transmission. Integr Comp Biol Jackson DA. 2000a. Living without oxygen: lessons from the 44:14–20. freshwater turtle. Comp Biochem Physiol A 125:299–315. Knight CA, Duman JG. 1986. Inhibition of recrystallization of Jackson DC. 2000b. How a turtle’s shell helps it survive ice by insect thermal hysteresis proteins: a possible prolonged anoxic acidosis. News Physiol Sci 15:181–185. cryoprotective role. Cryobiology 23:256–262.

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 371

Kolbe JJ, Janzen FJ. 2001. The influence of propagule size and Layne JR, Costanzo JP, Lee RE. 1998. Freeze duration maternal nest-site selection on survival and behaviour of influences postfreeze survival in the frog Rana sylvatica. neonate turtles. Funct Ecol 15:772–781. J Exp Zool 280:197–201. Kolbe JJ, Janzen FJ. 2002. Experimental analysis of an early Lee RE. 1991. Principles of insect low temperature tolerance. life-history stage: water loss and migrating hatchling In: Lee RE, Denlinger DL, editors. Insects at low tempera- turtles. Copeia 2002:220–226. ture. New York: Chapman & Hall. p 17–46. Koonz WH. 1998. Western painted turtle hatchlings over- Lee RE, Costanzo JP. 1993. Integrated physiological responses wintering in Manitoba nests. Blue Jay 56:183–184. promoting anuran freeze tolerance. In: Carey C, Florant G, Kostal V, Berkova P, Simek P. 2003. Remodelling of Wunder BA, Horwitz B, editors. Life in the cold: ecological, membrane phospholipids during transition to diapause physiological, and molecular mechanisms. Boulder, CO: and cold-acclimation in the larvae of Chymomyza costata Westview Press. p 501–510. (Drosophilidae). Comp Biochem Physiol B 135:407–419. Lee RE, Costanzo JP. 1998. Biological ice nucleation and ice Kostal V, Vambera J, Bastl J. 2004. On the nature of pre- distribution in cold-hardy ectothermic animals. Annu Rev freeze mortality in insects: water balance, ion homeostasis Physiol 60:55–72. and energy charge in the adults of Pyrrhocoris apterus. Lee RE, Zachariassen KE, Baust JG. 1981. Effect of J Exp Biol 207:1509–1521. cryoprotectants on the activity of hemolymph nucleating Kotenko TI. 2000. The European pond turtle (Emys orbicu- agents in physical solutions. Cryobiology 18:511–514. laris) in the Steppe Zone of the Ukraine. Stapfia 69, zugleich Lee RE, Warren GJ, Gusta LV, editors. 1995. Biological ice Kataloge des OO¨ Landesmuseums, Neue Folge Nr 149, nucleation and its applications. St. Paul, MN: American p 87–106. Phytopathological Society Press. 370p. Kumai M. 1976. Identification of nuclei and concentrations of Lee RE, Costanzo JP, Mugnano JA. 1996. Regulation of chemical species in snow crystals sampled at the South Pole. supercooling and ice nucleation in insects. Eur J Entomol J Atmos Sci 33:833–841. 93:405–418. Kuzmin SL. 2002. The turtles of Russia and other Ex-Soviet Lee RE, Damodaran K, Yi S-X, Lorigan GA. 2006a. Rapid cold- Republics (former Soviet Union). Frankfurt am Main: hardening increases membrane fluidity and cold tolerance of Edition Chimaira. insect cells. Cryobiology 52:459–463. Lahanas PN. 1982. Aspects of the life history of the southern Lee RE, Elnitsky MA, Rinehart JP, Hayward SA, Sandro LH, black-knobbed sawback, Graptemys nigrinoda denticola Denlinger DL. 2006b. Rapid cold-hardening increases the Folkerts and Mount. MS thesis, Auburn University, freezing tolerance of the Antarctic midge Belgica antarctica. Auburn, USA. J Exp Biol 209:399–406. Landers JL, Garner JA, McRae WA. 1980. Reproduction of Legler JM. 1954. Nesting habits of the western painted turtle, gopher tortoises (Gopherus polyphemus) in southwestern Chrysemys picta bellii (Gray). Herpetologica 10:137–144. Georgia. Herpetologica 36:353–361. Lillywhite HB, Maderson PFA. 1988. The structure and Langham EJ, Mason BJ. 1958. The heterogenous and permeability of integument. Am Zool 28:945–962. homogeneous nucleation of supercooled water. Proc R Soc Lindeman PV. 1991. Survivorship of overwintering hatchling Lond 247A:493–504. painted turtles, Chrysemys picta, in northern Idaho. Can Lardie RL. 1973. Notes on courtship, eggs, and young of the Field-Nat 105:263–266. Florida red-bellied turtle, Chrysemys nelsoni. HISS News-J Long DR. 1986. Lipid content and delayed emergence of 1:183–184. hatchling yellow mud turtles. Southwest Nat 31:244–246. Lardie RL. 1975. Observations on reproduction in Kinoster- Loomis SH. 1991. Comparative invertebrate cold hardiness. non. J Herpetol 9:260–264. In: Lee RE, Denlinger DL, editors. Insects at low tempera- Larson JL. 2004. Aspects of the overwinter ecology and ture. New York: Chapman & Hall. p 301–317. physiology of painted turtle (Chrysemys picta) hatchlings. Loomis SH. 1995. Freezing tolerance of marine invertebrates. MS thesis, Miami University, Oxford, USA. Oceanogr Mar Biol 33:337–350. Layne JR. 1991. External ice triggers freezing in freeze- Loomis SH, Carpenter JF, Anchordoguy TJ, Crowe JH, tolerant frogs at temperatures above their supercooling Branchini BR. 1989. Cryoprotective capacity of end- point. J Herpetol 25:129–130. products of anaerobic metabolism. J Exp Zool 252:9–15. Layne JR. 2000. Postfreeze O2 consumption in the wood frog Lovich JE, Tucker AD, Kling DE, Gibbons JW, Zimmerman (Rana sylvatica). Copeia 2000:879–882. TD. 1991. Behavior of hatchling diamond terrapins (Malacl- Layne JR, Jones AL. 2001. Freeze tolerance in the gray emys terrapin) released in a South Carolina salt marsh. treefrog: cryoprotectant mobilization and organ dehydra- Herpetol Rev 22:81–83. tion. J Exp Zool 290:1–5. Lowe CH, Lardner PJ, Halpern EA. 1971. Supercooling in Layne JR, Kuharsky DK. 2001. Effects of prolonged freezing reptiles and other vertebrates. Comp Biochem Physiol and supercooling on body composition, pupariation, and 39A:125–135. adult emergence of Eurosta solidaginis (Diptera: Tephriti- Lundheim R, Zachariassen KE. 1993. Water balance of dae). Environ Entomol 30:12–16. overwintering beetles in relation to strategies for cold Layne JR, Peffer BJ. 2006. The influence of freeze duration on tolerance. J Comp Physiol 163:1–4. postfreeze recovery by caterpillars of Pyrrharctia isabella Lusena CV. 1955. Ice propagation in systems of biological (Lepidoptera: Arctiidae): when is survival enough to qualify interest. III. Effect of solutes on nucleation and growth of as recovery? J Exp Zool 305A:570–575. ice crystals. Arch Biochem Biophys 57:277–284. Layne JR, Lee RE, Huang JL. 1990. Inoculation triggers Lutterschmidt DI, LeMaster MP, Mason RT. 2006. Minimal freezing at high subzero temperatures in a freeze-tolerant overwintering temperatures of red-sided garter snakes frog (Rana sylvatica) and insect (Eurosta solidaginis). Can (Thamnophis sirtalis parietalis): a possible cue for emer- J Zool 68:506–510. gence? Can J Zool 84:771–777.

J. Exp. Zool. 372 J.P. COSTANZO ET AL.

Lutz PL, Storey KB. 1997. Adaptations to variations in oxygen Minton SA. 1972. Amphibians and reptiles of Indiana. In Acad tension by vertebrates and invertebrates. In: Dantzler WH, Sci Monogr 3:1–346. editor. Handbook of physiology section 13: comparative Mitchell JC. 1974. Statistics of Chrysemys rubriventris physiology. New York: Oxford University Press. p 1479–1522. hatchlings from Middlesex County, Virginia. Herpetol Rev MacCulloch RD, Secoy DM. 1983. Demography, growth, and 5:71. food of western painted turtles, Chrysemys picta bellii Mitchell JC. 1988. Population ecology and life histories of the (Gray), from southern Saskatchewan. Can J Zool 61: freshwater turtles Chrysemys picta and Sternotherus odor- 1499–1509. atus in an urban lake. Herpetol Monogr 2:40–61. Macdonald JA. 1981. Temperature compensation in the Mitchell JC. 1994. The reptiles of Virginia. Washington, DC: peripheral nervous system: Antarctic vs temperate poiki- Smithsonian Institution Press. p 352. lotherms. J Comp Physiol A 142:411–418. Mitchell JC. 2003. Pseudemys rubriventris (red-bellied cooter). Maginniss LA, Tapper SS, Miller LS. 1983. Effect of chronic Co-joined twins. Herpetol Rev 34:242. cold and submergence on blood oxygen transport in the Mitrus S. 2005. Headstarting in European pond turtles (Emys turtle, Chrysemys picta. Respir Physiol 53:15–29. orbicularis): does it work? Amphib-Rept 26:333–341. Maginniss LA, Ekelund SA, Ultsch GR. 2004. Blood oxygen Mitrus S, Zemanek M. 1998. Reproduction of Emys orbicularis transport in common map turtles during simulated hiber- (L.) in Central Poland. In: Fritz U, Joger U, Podloucky R, nation. Physiol Biochem Zool 77:232–241. Servan J, Buskirk JR, editors. Proceedings of EMYS Malard F, Hervant F. 1999. Oxygen supply and the adapta- symposium, Vol. 10, Dresden, Mertensiella, p 187–191. tions of animals in groundwater. Freshwat Biol 41:1–30. Marco A, Dı´az-Paniagua C. 2008. Aggregation protects Mitrus S, Zemanek M. 2000. Distribution and biology of Emys flexible-shelled reptile eggs from severe hydric stress. orbicularis (L.) in Poland. Stapfia 69, zugleich Kataloge des ¨ J Comp Physiol B 178:421–428. OO Landesmuseums, Neue Folge Nr 149, p 107–118. Martin RP. 1989. Notes on Louisiana Mitrus S, Zemanek M. 2003. European pond tortoise, Emys (Gopherus polyphemus) reproduction. Herpetol Rev 20: orbicularis, neonates overwintering in the nest. Herpetol 36–37. J 13:195–198. Mason BJ, Maybank J. 1958. Ice-nucleating properties of some Moll EO. 1973. Latitudinal and intersubspecific variation in natural mineral dusts. Q J R Meteorol Soc 84:235–241. reproduction of the painted turtle, Chrysemys picta. Herpe- Mason JC, Chapman DW. 1965. Significance of early emer- tologica 29:307–318. gence, environmental rearing capacity, and behavioral Moll EO, Legler JM. 1971. The life history of a neotropical ecology of juvenile coho salmon in stream channels. J Fish slider turtle, Pseudemys scripta (Schoepff), in Panama. Bull Res Bd Can 22:173–190. Los Angel Co Mus Nat Hist 11:1–102. Mazanaeva L, Orlova V. 2004. Distribution and ecology of Moran KL, Bjorndal KA, Bolten AB. 1999. Effects of the Emys orbicularis in Daghestan, Russia. Biol, Bratislava thermal environment on the temporal pattern of emergence 59:47–53. of hatchling loggerhead turtles Caretta caretta. Mar Ecol Mazur P. 2004. Principles of cryobiology. In: Fuller BJ, Lane Prog Ser 189:251–261. N, Benson EE, editors. Life in the frozen state. Washington, Morjan CL. 2003. Variation in nesting patterns affecting nest DC: CRC Press. p 3–65. temperatures in two populations of painted turtles (Chrys- McCallum ML. 2003. Graptemys geographica (northern map emys picta) with temperature-dependent sex determination. turtle). Nest overwintering. Herpetol Rev 34:241. Behav Ecol Sociobiol 53:254–261. McGowan A, Rowe LV, Broderick AC, Godley BJ. 2001. Nest Morjan CL, Stuart JN. 2001. Nesting record of a Big Bend factors predisposing loggerhead sea turtles (Caretta caretta) slider, Trachemys gaigeae, and overwintering of hatchlings clutches to infestation by dipteran larvae in northern in the nest. Southwest Nat 46:230–234. Cyprus. Copeia 2001:808–812. Muldrew K, Acker JP, Elliott JAW, McGann LE. 2004. The McNeil JA, Herman TB, Standing KL. 2000. Movement of water to ice transition: implications for living cells. In: hatchling Blanding’s turtles (Emydoidea blandingii)in Fuller B, Lane N, Benson EE, editors. Life in the frozen Nova Scotia in response to proximity to open water: a state. Boca Raton, FL: CRC Press. p 67–108. manipulative experiment. Chel Cons Biol 3:611–617. Muller JF. 1921. Notes on the habits of the soft-shell turtle— Miller K. 1993. The improved performance of snapping turtles mutica. Am Midl Nat 7:180–184. (Chelydra serpentina) hatched from eggs incubated on a wet Murphy DJ. 1983. Freezing resistance in intertidal inverte- substrate persists through the neonatal period. J Herpetol brates. Annu Rev Physiol 45:289–299. 27:228–233. Miller K, Packard GC, Packard MJ. 1987. Hydric conditions Mushinsky HR, Wilson DS. 1992. Seasonal occurrence of during incubation influence locomotor performance of Kinosternon baurii on a sandhill in central Florida. hatchling snapping turtles. J Exp Biol 127:401–412. J Herpetol 26:207–209. Miller K, Monteforte PB, Landis LF. 1993. Scaling of Nagle RD, Burke VJ, Congdon JD. 1998. Egg components and locomotor performance and enzyme activity in the leopard hatchling lipid reserves: parental investment in kinosternid frog, Rana pipiens. Herpetologica 49:383–392. turtles from the southeastern United States. Comp Biochem Milsom WK, Chan P. 1986. The relationship between lung Physiol B 120:145–152. volume, respiratory drive, and breathing pattern in the Nagle RD, Kinney OM, Congdon JD, Beck CW. 2000. Winter turtle, Chrysemys picta. J Exp Biol 120:233–247. survivorship of hatchling painted turtles (Chrysemys picta) Milton SL, Nayak G, Kesaraju S, Kara L, Prentice HM. 2007. in Michigan. Can J Zool 78:226–233. Suppression of reactive oxygen species production enhances Nagle RD, Lutz CL, Pyle AL. 2004. Overwintering in the nest neuronal survival in vitro and in vivo in the anoxia-tolerant by hatchling map turtles (Graptemys geographica). Can turtle Trachemys scripta. J Neurochem 101:993–1001. J Zool 82:1211–1218.

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 373

Nedved O. 2000. Snow White and the Seven Dwarfs: a Packard GC. 2004a. To freeze or not to freeze: adaptations for multivariate approach to classification of cold tolerance. overwintering by hatchlings of the North American painted Cryo-Letters 21:339–348. turtle. J Exp Biol 207:2897–2906. Nemuras KT. 1967. Genus Clemmys. Int Turtle Tortoise Soc Packard GC. 2004b. Natural freeze-tolerance in reptiles. Cryo- J 1967:38–40. Letters 25:235–236. Newman HH. 1906. The habits of certain tortoises. J Comp Packard GC, Janzen FJ. 1996. Interpopulational variation in Neurol Psychol 16:126–152. the cold-tolerance of hatchling painted turtles. J Therm Biol Nichols JT. 1933. Further notes on painted turtles. Copeia 21:183–190. 1933:41–42. Packard GC, Packard MJ. 1988. The physiological ecology of Nikolskii AM. 1915. Fauna of Russia and adjacent countries. reptilian eggs and embryos. In: Gans C, Huey RB, editors. Reptiles. Vol. 1. Chelonia and Sauria. 1963 translation. Nat Biology of the Reptilia. Washington, DC: Alan Liss. Sci Found Washington, DC. Israel Program For Scientific p 523–605. Translations, Jerusalem. Packard GC, Packard MJ. 1990. Patterns of survival at Noble GK, Breslau AM. 1938. The sense involved in the subzero temperatures by hatchling painted turtles and migration of young fresh-water turtles after hatching. snapping turtles. J Exp Zool 254:233–236. J Comp Psychol 25:175–193. Packard GC, Packard MJ. 1993a. Delayed inoculative freezing Norris-Elye LTS. 1949. The common snapping turtle (Chely- is fatal to hatchling painted turtles (Chrysemys picta). Cryo- dra serpentina) in Manitoba. Can Field-Nat 63:145–147. Letters 14:273–284. Novotny M, Danko S, Havas P. 2004. Activity cycle and reproductive characteristics of the European pond turtle Packard GC, Packard MJ. 1993b. Hatchling painted turtles (Emys orbicularis) in the Tajba National Nature Reserve, (Chrysemys picta) survive exposure to subzero temperatures Slovakia. Biol Bratislava 59:113–121. during hibernation by avoiding freezing. J Comp Physiol B Obbard ME, Brooks RJ. 1981a. Fate of overwintered clutches 163:147–152. of the common snapping turtle (Chelydra serpentina)in Packard GC, Packard MJ. 1995a. The basis for cold tolerance Algonquin Park, Ontario. Can Field-Nat 95:350–352. in hatchling painted turtles (Chrysemys picta). Physiol Zool Obbard ME, Brooks RJ. 1981b. A radio-telemetry and mark- 68:129–148. recapture study of activity in the common snapping turtles, Packard GC, Packard MJ. 1995b. A review of the adaptive Chelydra serpentina. Copeia 1981:630–637. strategy used by hatchling painted turtles to survive in the Olson RE. 1989. Notes on evaporative water loss in terrestrial cold. Herpetol Rev 26:8–12. chelonians. Bull Maryland Herpetol Soc 25:49–57. Packard GC, Packard MJ. 1997. Type of soil affects survival by Olsen TM, Sass SJ, Li N, Duman JG. 1998. Factors overwintering hatchlings of the painted turtle. J Therm Biol contributing to seasonal increases in inoculative freezing 22:53–58. resistance in overwintering fire-colored beetle larvae Packard GC, Packard MJ. 1999. Limits of supercooling and Dendroides canadensis (Pyrochroidae). J Exp Biol 201: cold-tolerance in hatchling painted turtles (Chrysemys 1585–1594. picta). Cryo-Letters 20:55–60. Orvar BL, Sangwan V, Omann F, Dhindsa RS. 2000. Early Packard GC, Packard MJ. 2001a. The overwintering strategy steps in cold sensing by plant cells: the role of actin of hatchling painted turtles, or how to survive in the cold cytoskeleton and membrane fluidity. Plant J 23:785–794. without freezing. Bioscience 51:199–207. O’Steen S. 1998. Embryonic temperature influences juvenile Packard GC, Packard MJ. 2001b. Environmentally induced temperature choice and growth rate in snapping turtles, variation in size, energy reserves and hydration of hatchling Chelydra serpentina. J Exp Biol 201:439–449. painted turtles, Chrysemys picta. Funct Ecol 15:481–489. O’Steen S, Janzen FJ. 1999. Embryonic temperature affects Packard GC, Packard MJ. 2003a. Influence of acclimation and metabolic compensation and thyroid hormones in hatchling incubation medium on supercooling by hatchling painted snapping turtles. Physiol Biochem Zool 72:520–533. turtles, Chrysemys picta. Funct Ecol 17:611–618. Oswood MW, Miller LK, Irons III JG. 1991. Overwintering of Packard GC, Packard MJ. 2003b. Cold acclimation enhances freshwater benthic macroinvertebrates. In: Lee RE, Den- cutaneous resistance to inoculative freezing in hatchling linger DL, editors. Insects at low temperature. New York: painted turtles, Chrysemys picta. Funct Ecol 17:94–100. Chapman & Hall. p 360–375. Packard GC, Packard MJ. 2003c. Natural freeze-tolerance in Overgaard J, Sørensen JG, Petersen SO, Loeschcke V, hatchling painted turtles? Comp Biochem Physiol 134: Holmstrup M. 2005. Changes in membrane lipid composi- 233–246. tion following rapid cold hardening in Drosophila melano- gaster. J Insect Physiol 51:1173–1182. Packard GC, Packard MJ. 2006. The relationship between Overgaard J, Sørensen JG, Petersen SO, Loeschcke V, gut contents and supercooling capacity in hatchling painted Holmstrup M. 2006. Reorganization of membrane lipids turtles (Chrysemys picta). Comp Biochem Physiol 144A: during fast and slow cold hardening in Drosophila melano- 98–104. gaster. Physiol Entomol 31:328–335. Packard MJ, Packard GC. 2001c. Why hatchling painted Packard GC. 1991. Physiological and ecological importance of turtles overwinter in their nest. Am Zool 40:1160–1161. water to embryos of oviparous reptiles. In: Deeming DC, Packard MJ, Packard GC. 2004. Accumulation of lactate by Ferguson MW, editors. Egg incubation: its effects on frozen painted turtles (Chrysemys picta) and its relationship embryonic development in birds and reptiles. New York: to freeze tolerance. Physiol Biochem Zool 77:433–439. Cambridge University Press. p 213–228. Packard GC, Paukstis GL, Boardman TJ, Gutzke WHN. 1985. Packard GC. 1997. Temperatures during winter in nests with Daily and seasonal variation in hydric conditions and hatchling painted turtles (Chrysemys picta). Herpetologica temperature inside nests of common snapping turtles 53:89–95. (Chelydra serpentina). Can J Zool 63:2422–2429.

J. Exp. Zool. 374 J.P. COSTANZO ET AL.

Packard GC, Packard MJ, Miller K, Boardman TJ. 1987. Petokas PJ, Alexander MM. 1960. The nesting of Chelydra Influence of moisture, temperature, and substrate on serpentina in northern New York. J Herpetol 14:239–244. snapping turtle eggs and embryos. Ecology 68:983–993. Phelps P, Giddings TH, Prochoda M, Fall R. 1986. Release of Packard GC, Packard MJ, McDaniel PL, McDaniel LL. 1989. cell-free ice nuclei by Erwinia herbicola. J Bacteriol 167: Tolerance of hatchling painted turtles to subzero tempera- 496–502. tures. Can J Zool 67:828–830. Pitler R. 1985. Malaclemys terrapin terrapin (Northern Packard GC, Ruble KA, Packard MJ. 1993. Hatchling Diamondback Terrapin). Behavior. Behavior. Herpetol Rev snapping turtles overwintering in natural nests are inocu- 16:82. lated by ice in frozen soil. J Therm Biol 18:185–188. Plummer MV. 1976. Some aspects of nesting success in the Packard GC, Fasano SL, Attaway MB, Lohmiller LD, Lynch turtle, Trionyx muticus. Herpetologica 32:353–359. TL. 1997a. Thermal environment for overwintering hatchl- Plummer MV. 1977. Reproduction and growth in the turtle ings of the painted turtle (Chrysemys picta). Can J Zool 75: Trionyx muticus. Copeia 1977:440–447. 401–406. Plummer MV, Williams BK, Skiver MM, Carlyle JC. 2003. Packard GC, Lang JW, Lohmiller LD, Packard MJ. 1997b. Effects of dehydration on the critical thermal maximum of Cold tolerance in hatchling painted turtles (Chrysemys the (Terrapene ornata luteola). J Herpetol picta): supercooling or freeze tolerance? Physiol Zool 70: 37:747–750. 670–678. Po¨rtner HO. 2002. Climate variations and the physiological Packard GC, Tucker JK, Nicholson D, Packard MJ. 1997c. basis of temperature dependent biogeography: systemic to Cold tolerance in hatchling slider turtles (Trachemys molecular hierarchy of thermal tolerance in animals. Comp scripta). Copeia 1997:339–345. Biochem Physiol A 132:739–761. Packard GC, Lang JW, Lohmiller LD, Packard MJ. 1999a. Power BA, Power RF. 1962. Some amino-acids as ice Resistance to freezing in hatchling painted turtles (Chrys- nucleators. Nature 194:1170–1171. emys picta). Can J Zool 77:795–801. Prange HD, Ackerman RA. 1974. Oxygen consumption and Packard GC, Packard MJ, Lang JW, Tucker JK. 1999b. mechanisms of gas exchange of green turtle (Chelonia Tolerance for freezing in hatchling turtles. J Herpetol 33: mydas) eggs and hatchlings. Copeia 1974:758–763. 536–543. Punzo F. 1975. Studies on the feeding behavior, diet, nesting habits and temperature relationships of Packard GC, Packard MJ, Lang JW. 2000. Why hatchling Chelydra serpentina osceola (Chelonia: Chelydridae). J Herpetol 9:207–210. Blanding’s turtles don’t overwinter inside their nest. Ralph CR, Reina RD, Wallace BP, Sotherland PR, Spotila JR, Herpetologica 56:367–374. Paladino FV. 2005. Effect of egg location and respiratory gas Packard GC, Packard MJ, McDaniel LL. 2001. Seasonal concentrations on developmental success in nests of the change in the capacity for supercooling by neonatal painted leatherback turtle, Dermochelys coriacea. Aust J Zool 53: turtles. J Exp Biol 204:1667–1672. 289–294. Packard GC, Packard MJ, Morjan CL, Janzen FJ. 2002. Cold- Ramlov H. 2000. Aspects of natural cold tolerance in tolerance of hatchling painted turtles (Chrysemys picta ectothermic animals. Hum Reprod 15:26–46. bellii) from the southern limit of distribution. J Herpetol Ramlov H, Wharton DA, Wilson PW. 1996. Recrystallization 36:300–304. in a freezing tolerant Antarctic nematode, Panagrolaimus Pappas MJ, Brecke BJ. 1992. Habitat selection of juvenile davidi, and an Alpine Weta, Hemideina maori (Orthoptera; Blanding’s turtles (Emydoidea blandingii). J Herpetol Stenopelmatidae). Cryobiology 33:607–613. 26:233–234. Rasmussen DH, MacKenzie AP. 1973. Clustering in super- Pappas MJ, Brecke BJ, Congdon JD. 2000. The Blanding’s cooled water. J Chem Phys 59:5003–5013. turtles (Emydoidea blandingii)ofWeaverDunes,Minnesota. Rathbun GB, Scott NJ, Murphy TG. 2002. Terrestrial habitat Chel Cons Biol 3:557–568. use by Pacific pond turtles in a Mediterranean climate. Parker WS, Brown WS. 1980. Comparative ecology of two Southwest Nat 47:225–235. colubrid snakes Masticophis t. taeniatus and Pituophis Ratterman RJ, Ackerman RA. 1989. The water exchange and melanoleucus deserticola, in northern Utah. Milwaukee hydric microclimate of painted turtle (Chrysemys picta) eggs Pub Mus Publ Biol Geol 7:1–104. incubating in field nests. Physiol Zool 62:1059–1079. Parren SG, Rice MA. 2004. Terrestrial overwintering of Reese DA, Welsh HH. 1997. Use of terrestrial habitat by hatchling turtles in Vermont nests. Northeast Nat western pond turtles, Clemmys marmorata: implications for 11:229–233. management. In: Proceedings: conservation, restoration, and Pasanen S, Karhapa¨a¨ M. 1997. Can boreal common frog (Rana management of turtles and tortoises. An international temporaria L.) survive in frost? Ann Zool Fenn 34:247–250. conference. New York Turtle and Tortoise Society. p352–357. Paukstis GL, Shuman RD, Janzen FJ. 1989. Supercooling and Reese SA, Crocker CA, Carwile ME, Jackson DC, Ultsch GR. freeze tolerance in hatchling painted turtles (Chrysemys 2001. The physiology of hibernation in common map turtles picta). Can J Zool 67:1082–1084. (Graptemys geographica). Comp Biochem Physiol A 130: Paukstis GL, Janzen FJ, Tucker JK. 1996. Response of 331–340. aerially-exposed zebra mussels (Dreissena polymorpha)to Reese SA, Jackson DC, Ultsch GR. 2002. The physiology of subfreezing temperatures. J Freshwat Ecol 11:513–519. overwintering in a turtle that occupies multiple habitats, Pell SM. 1941. Notes on the habits of the common snapping the common snapping turtle (Chelydra serpentina). Physiol turtle, Chelydra serpentina (Linn.) in central New York. MS Biochem Zool 75:432–438. thesis, Cornell University, Ithaca, USA. Reese SA, Jackson DC, Ultsch GR. 2003. Hibernation in Pe´rez-Pinzo´n MA, Rice ME. 1995. Seasonal- and temperature freshwater turtles: softshell turtles (Apalone spinifera) are dependent variation in CNS ascorbate and glutathione the most intolerant of anoxia among North American levels in anoxia-tolerant turtles. Brain Res 705:45–52. species. J Comp Physiol B 173:263–268.

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 375

Reese SA, Stewart ER, Crocker CA, Jackson DA, Ultsch GR. Mount Desert Island, Maine. MS thesis, Bard College, 2004a. Geographic variation of the physiological response to Annandale-On-Hudson, USA. overwintering in the painted turtle (Chrysemys picta). Rubinsky B, Hong J-S, Storey KB. 1994. Freeze tolerance in Physiol Biochem Zool 77:619–630. turtles: visual analysis by microscopy and magnetic reso- Reese SA, Ultsch GR, Jackson DC. 2004b. Lactate accumula- nance imaging. Am J Physiol 267:R1078–R1088. tion, glycogen depletion, and shell composition of hatchling Rutberg AT. 1987. Adaptive hypotheses of birth synchrony in turtles under simulated aquatic hibernation. J Exp Biol ruminants: an interspecific test. Am Nat 130:692–710. 207:2889–2895. Salt R. 1961. Principles of insect cold hardiness. Annu Rev Resetarits WJ. 1996. Oviposition site choice and life history Entomol 6:55–74. evolution. Am Zool 3:205–215. Salt RW. 1963. Delayed inoculative freezing of insects. Can Rhen T, Lang JW. 1999. Temperature during embryonic and Entomol 95:1190–1202. juvenile development influences growth in hatchling snap- Salt RW. 1966. Factors influencing nucleation in supercooled ping turtles, Chelydra serpentina. J Therm Biol 24:33–41. insects. Can J Zool 44:117–133. Richmond ND. 1945. Nesting habits of the mud turtle. Copeia Saunders DK, Patel KH. 1998. Comparison of blood viscosity 1945:217–219. in red-eared sliders (Trachemys scripta) adapted to cold and Rinehart JP, Hayward SAL, Elnitsky MA, Sandro LH, Lee RE, room temperature. J Exp Zool 281:157–163. Denlinger DL. 2006. Continuous up-regulation of heat shock Schmid WD. 1982. Survival of frogs in low temperature. proteins in larvae, but not adults, of a polar insect. Proc Natl Science 512:697–698. Acad Sci USA 103:14223–14227. Schmidt-Nielsen K, Bentley PJ. 1966. Desert tortoise Ring RA. 1982. Freezing-tolerant insects with low super- Gopherus agassizii: cutaneous water loss. Science 154:911. cooling points. Comp Biochem Physiol A 73:605–612. Schmidt-Nielsen K, Crawford EC, Bentley PJ. 1966. Discon- Ring RA, Danks HV. 1994. Desiccation and cryoprotection: tinuous respiration in the lizard Sauromalus obesus. Fed overlapping adaptations. Cryo-Letters 15:181–190. Proc 25:506. Ring RA, Tesar D. 1981. Adaptations to cold in Canadian Schneeweiss N, Andreas B, Jendretzke. 1998. Reproductive arctic insects. Cryobiology 18:199–211. ecology data of the European pond turtle (Emys Risley PL. 1933. Observations on the natural history of the o. orbicularis) in Brandenburg, northeast Germany. In: common musk turtle, Sternotherus odoratus (Latreille). Fritz U, Joger U, Podloucky R, Servan J, Buskirk JR, Papers Mich Acad Sci Art Lett 17:685–711. editors. Proceedings of EMYS symposium, Vol. 10, Dresden, Roberts JB, Lillywhite HB. 1980. Lipid barrier to water Mertensiella, p 227–234. exchange in reptile epidermis. Science 207:1077–1079. Schwarzkopf L, Brooks RJ. 1987. Nest-site selection and Roberts P, Hallett J. 1968. A laboratory study of the ice offspring sex ratio in painted turtles, Chrysemys picta. nucleating properties of some mineral particulates. Q J R Copeia 1987:53–61. Meteorol Soc 94:25–34. Seidel ME. 1978. Terrestrial dormancy in the turtle Kinos- Robertson RM. 2004. Modulation of neural circuit operation ternon flavescens: respiratory metabolism and dehydration. by prior environmental stress. Integr Comp Biol 44:21–27. Comp Biochem Physiol 61A:1–4. Rogers JS, Stall RE, Burke JM. 1987. Low-temperature Seidel ME, Reynolds SL. 1980. Aspects of evaporative water conditioning of the ice nucleation active bacterium, Erwinia loss in the mud turtles Kinosternon hirtipes and Kinoster- herbicola. Cryobiology 24:270–279. non flavescens. Comp Biochem Physiol 67A:595–598. Rojas RR, Charlet LD, Leopold RA. 1992. A differential Semlitsch RD, Gibbons JW. 1989. Lack of largemouth bass scanning colorimetric analysis of inoculative freezing in an predation on hatchling turtles (Trachemys scripta). Copeia insect. Cryo-Letters 13:355–362. 1989:1030–1031. Roosenberg WM. 1994. Nesting habitat requirements of the Semple RE, Sigsworth D, Stitt JT. 1970. Seasonal observa- diamondback terrapin: a geographic comparison. Wetland tions on the plasma, red cell and blood volumes of two turtle J 6:8–11. species native to Ontario. Can J Physiol Pharmacol 48: Roosenburg WM. 2003. Diamondback terrapin nesting on the 282–290. Poplar Island environmental restoration project. Proceed- Servan J. 1998. Ecological study of Emys orbicularis in Brenne ings of the 13th biennial coastal zone conference, Baltimore, (central France). Proceedings of EMYS symposium, Vol. 10, MD. Dresden, Mertensiella, p 227–234. Rose FL. 1969. Desiccation rates and temperature relation- Sexton OJ. 1957. Notes concerning turtle hatchlings. Copeia ships of Terrapene ornata following scute removal. South- 1957:229–230. west Nat 14:67–72. Sexton OJ. 1959. Spatial and temporal movements of a Rossler M. 2000. The ecology and reproduction of an Emys population of the painted turtle, Chrysemys picta marginata orbicularis population in Austria. Chelonii 2:69–72. (Agassiz). Ecol Monogr 29:113–140. Rostal DG, Lance VL, Grumbles JS, Alberts AC. 1994. Sexton OJ, Marion KR. 1981. Experimental analysis of Seasonal reproductive cycle of the desert tortoise (Gopherus movements by prairie rattlesnakes, Crotalus viridis, during agassizii) in the eastern Mohave Desert. Herpetol Monogr hibernation. Oecologia 51:37–41. 8:72–82. Shams I, Avivi A, Nevo E. 2005. Oxygen and carbon dioxide Rowe JW, Holy L, Ballinger RE, Stanley-Samuelson D. 1995. fluctuations in burrows of subterranean blind mole rats Lipid provisioning of turtle eggs and hatchlings: total lipid, indicate tolerance to hypoxic-hypercapnic stresses. Comp phosphoplipid, triacylglycerol and triacyl glycerol fatty Biochem Physiol 142A:376–383. acids. Comp Biochem Physiol B 112:323–330. Shealy RM. 1976. The natural history of the Alabama map Rozycki CB. 1998. Reproductive and nesting ecology of the turtle, Graptemys pulchra Baur, in Alabama. Bull Florida painted turtle (Chrysemys picta) in Acadia National Park, State Mus Biol Sci 21:47–111.

J. Exp. Zool. 376 J.P. COSTANZO ET AL.

Shelton G, Boutilier RG. 1982. Apnoea in amphibians and St. Clair RC, Gregory PT. 1990. Factors affecting the northern reptiles. J Exp Biol 100:245–273. range limit of painted turtles (Chrysemys picta): winter Shen JH, Klier K, Zettlemoyer AC. 1977. Ice nucleation by acidosis or freezing? Copeia 1990:1083–1089. micas. J Atmos Sci 34:957–960. Standing KL, Herman TB, Hurlburt DD, Morrison IP. 1997. Shimada K, Riihimaa A. 1988. Cold acclimation, inoculative Postemergence behavior of neonates in a northern periph- freezing and slow cooling: essential factors contributing to eral population of Blanding’s turtle, Emydoidea blandingii, the freezing tolerance in diapausing larvae of Chymomyza in Nova Scotia. Can J Zool 75:1387–1395. costata (Diptera: Drosophilidae). Cryo-Letters 9:5–10. Standing KL, Herman TB, Morrison IP. 1999. Nesting ecology Sims PA, Packard GC, Chapman PL. 2001. The adaptive of Blanding’s turtle (Emydoidea blandingii) in Nova Scotia, strategy for overwintering by hatchling snapping turtles the northeastern limit of the species’’ range. Can J Zool (Chelydra serpentina). J Herpetol 35:514–517. 77:1609–1614. Sinclair BJ. 1999. Insect cold tolerance: how many kinds of Starkey DE, Shaffer HB, Burke RL, Forstner RJ, Iverson JB, frozen? Eur J Entomol 96:157–164. Ultsch GR. 2003. Molecular systematics, phylogeography, Sinclair BJ, Addo-Bediako A, Chown SL. 2003a. Climatic and the effects of Pleistocene glaciations in the Painted variability and the evolution of insect freeze tolerance. Biol Turtle (Chrysemys picta) complex. Evolution 57:119–128. Rev 78:181–195. Steen DA, Sterrett SC, Miller SA, Smith LL. 2007. Terrestrial Sinclair BJ, Vernon P, Klok CJ, Chown SL. 2003b. Insects at movements and microhabitat selection of overwintering low temperatures: an ecological perspective. Trends Ecol subadult eastern mud turtles (Kinosternon subrubrum)in Evol 18:257–262. southwest Georgia. J Herpetol 41:532–535. Sinclair BJ, Marshall DJ, Singh S, Chown SL. 2004. Cold Steiner AA, Petenusci SO, Brentegani LG, Branco LGS. 2000. tolerance of Littorinidae from southern Africa: intertidal The importance of glucose for the freezing tolerance/ snails are not constrained to freeze tolerance. J Comp intolerance of the anuran amphibians Rana catesbeiana Physiol B 174:617–624. and Bufo paracnemis. Rev Bras Biol 60:321–328. Skorepa AC, Ozment JE. 1968. Habitat, habits, and variation Stephens PR, Weins JJ. 2003. Ecological diversification and of Kinosternon subrubrum in southern Illinois. Trans Ill phylogeny of emydid turtles. Biol J Linn Soc 79:1609–1614. Acad Sci 61:247–251. Steponkus PL. 1984. Role of the plasma membrane in freezing injury and cold acclimation. Annu Rev Plant Physiol Smith PW. 1961. The amphibians and reptiles of Illinois. Ill 35:543–584. Nat Hist Surv Bull 28:1–298. Steyermark AC, Spotila JR. 2000. Effects of maternal identity Smith HM, Chiszar D. 2006. Dilemma of name-recognition: and incubation temperature on snapping turtle (Chelydra why and when to use new combinations of scientific names. serpentina) metabolism. Physiol Biochem Zool 73:298–306. Herpetol Cons Biol 1:6–8. Steyermark AC, Spotila JR. 2001. Body temperature and Snieshkus E. 1998. Some observations on secondary sexual maternal identity affect snapping turtle (Chelydra serpenti- characteristics, sex ratio, and reproductive aspects of na) righting response. Copeia 2001:1050–1057. European pond turtle, Emys orbicularis (Linnaeus, 1758), Stitt JT, Semple RE. 1971. Sites of plasma sequestrat- in the former USSR. Proceedings of EMYS symposium, Vol. ion induced by body cooling in turtles. Am J Physiol 221: 10, Dresden, Mertensiella, p 253–258. 1189–1191. Somero GN, Yancey PH. 1997. Osmolytes and cell-volume Stone PA, Iverson JB. 1999. Cutaneous surface area in regulation: physiological and evolutionary principles. In: freshwater turtles. Chel Cons Biol 3:512–515. Hoffman JF, Jamieson JD, editors. Handbook of physiology: Stone PA, Dobie JL, Henry RP. 1992. Cutaneous surface area cell physiology. New York: Oxford University Press. and bimodal respiration in soft-shelled (Trionyx spiniferus), p 441–484. stinkpot (Sternotherus odoratus), and mud turtles (Kinos- Sømme L. 1982. Supercooling and winter survival in terres- ternon subrubrum). Physiol Zool 65:311–330. trial arthropods. Comp Biochem Physiol 73:519–543. Storey JM, Storey KB. 1985. Triggering of cryoprotectant Sømme L. 1996. The effect of prolonged exposures at low synthesis by the initiation of ice nucleation in the freeze temperatures in insects. Cryo-Letters 17:341–346. tolerant frog, Rana sylvatica. J Comp Physiol B Spaans EJA, Baker JM. 1996. The soil freezing characteristic: 156:191–195. its measurement and similarity to the soil moisture Storey KB. 1990a. Biochemistry of natural freeze tolerance in characteristic. Soil Sci Soc Am J 60:13–19. animals: molecular adaptations and applications to cryopre- Spellerberg IF. 1972. Temperature tolerances of southeast servation. Biochem Cell Biol 68:687–698. Australian reptiles examined in relation to reptile thermo- Storey KB. 1990b. Life in a frozen state: adaptive strategies regulatory behavior and distribution. Oecologia (Berlin) for natural freeze tolerance in amphibians and reptiles. Am 9:23–46. J Physiol 258:R559–R568. Spellerberg IF. 1976. Adaptations of reptiles to cold. In: d’A Storey KB. 1996. Oxidative stress: animal adaptations in Bellaires A, Cox C, editors. Morphology and biology of nature. Braz J Med Biol Res 29:1715–1733. reptiles: Linnaean Society symposium series, Vol. 3. Storey KB. 1997. Organic solutes in freezing tolerance. Comp p 261–285. Biochem Physiol A 117A:319–326. Spencer R-J, Thompson MB, Banks PB. 2001. Hatch or wait? Storey KB. 1999. Living in the cold: freeze-induced gene A dilemma in reptilian incubation. Oikos 93:401–406. responses in freeze-tolerant vertebrates. Clin Exp Pharma- Spotila JR, Berman EN. 1976. Determination of skin col Physiol 26:57–63. resistance and the role of the skin in controlling water loss Storey KB. 2004a. Strategies for exploration of freeze in amphibians and reptiles. Comp Biochem Physiol 55A: responsive gene expression: advances in vertebrate freeze 407–411. tolerance. Cryobiology 48:134–145.

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 377

Storey KB. 2004b. Vertebrate freeze tolerance: role of freeze- Thomas D. 1999. The influence of latitude and temperature on responsive gene expression. In: Barnes BM, Carey HV, the in vitro emergence of hatchling slider turtles, Trachemys editors. Life in the cold: evolution, mechanisms, adaptation, scripta. Abstracts of the 1999 meeting of the American and application. Fairbanks: Institute of Arctic Biology, Society of ichthyologists and herpetologists, Pittsburgh, PA. University of Alaska Fairbanks. p 299–306. Thompson MB. 1988. Nest temperatures in the Pleurodiran Storey KB. 2006. Reptile freeze tolerance: metabolism and turtle, macquarii. Copeia 1988:996–1000. gene expression. Cryobiology 52:1–16. Thumm K, Mahoney M. 2002. Hatching dynamics and Storey KB, Storey JM. 1988. Freeze tolerance in animals. bet-hedging in a temperate frog, Pseudophryne australis Physiol Rev 68:27–84. (Anura: Myobatrachidae). Amphib-Rept 23:433–444. Storey KB, Storey JM. 1990. Metabolic rate depression and Tinkle DW. 1961. Geographic variation in reproduction, size, biochemical adaptation in anerobiosis, hibernation, and sex ratio and maturity of Sternothaerus odoratus (Testudi- estivation. Q Rev Biol 65:145–174. nata: Chelydridae). Ecology 42:68–76. Storey KB, Storey JM. 1992. Natural freeze tolerance in Tinkle DW, Congdon JD, Rosen PC. 1981. Nesting frequency ectothermic vertebrates. Annu Rev Physiol 54:619–637. and success: implications for the demography of painted Storey KB, Storey JM. 1996. Natural freezing survival in turtles. Ecology 62:1426–1432. animals. Annu Rev Ecol Syst 27:365–386. Tomcala A, Tollarova M, Overgaard J, Simek P, Kostal V. Storey KB, Baust JG, Buescher P. 1981. Determination of 2006. Seasonal acquisition of chill tolerance and restructur- water ‘‘bound’’ by soluble subcellular components during ing of membrane glycerophospholipids in an overwintering low-temperature acclimation in the gall fly larva, Eurosta insect: triggering by low temperature, desiccation and solidaginis. Cryobiology 18:315–321. diapause progression. J Exp Biol 209:4102–4114. Storey KB, Storey JM, Brooks SPJ, Churchill TA, Brooks RJ. Toner GC. 1940. Delayed hatching in the snapping turtle. 1988. Hatchling turtles survive freezing during winter Copeia 1940:265. hibernation. Proc Natl Acad Sci USA 85:8350–8354. Trauth SE, Robison HW, Plummer MV. 2004. The amphibians Storey KB, McDonald DG, Duman JG, Storey JM. 1991. Blood and reptiles of Arkansas. Fayetteville: University of Arkan- chemistry and ice nucleating activity in hatchling painted sas Press. turtles. Cryo-Letters 12:351–358. Trobec TN, Stanley JG. 1971. Uptake of ions and water by the Storey KB, Layne JR, Cutwa MM, Churchill TA, Storey JM. painted turtle, Chrysemys picta. Copeia 1971:537–542. 1993. Freezing survival and metabolism of box turtles, Tucker JK. 1997. Natural history notes on nesting, nests, and Terrapene carolina. Copeia 1993:628–634. hatchling emergence in the red-eared slider turtle, Tra- Stover H. 1973. Cold resistance and freezing in Arianta chemys scripta elegans, in west-central Illinois. Ill Nat Hist arbustorum L. (Pulmonata). In: Wieser W, editor. Effects of Surv Biol Notes 140:1–13. temperature on ectothermic organisms. Berlin and New Tucker JK. 1999. Environmental correlates of hatchling York: Springer. p 281–290. emergence in the red-eared turtle, Trachemys scripta Surface HA. 1908. First report on the economy of Pennsylva- elegans, in Illinois. Chel Cons Biol 3:401–406. nia turtles. Zool Bull Div Zool Penn Dept Agric 6:105–196. Tucker JK. 2000a. Body size and migration of hatchling Swarth CW. 2004. Natural history and reproductive biology of turtles: inter- and intraspecific comparisons. J Herpetol 34: the red-bellied turtle Pseudemys rubriventris. In: Swarth 541–546. CW, Roosenberg WM, Kiviat E, editors. Conservation and Tucker JK. 2000b. Egg size in the red-eared slider (Trachemys ecology of turtles of the Mid-Atlantic Region: a symposium. scripta elegans). Herpetol Nat Hist 7:171–174. Salt Lake City, UT: Bibliomania. p 73–83. Tucker JK, Packard GC. 1998. Overwinter survival by Sweeney BW, Vannote RL. 1982. Population synchrony in hatchling sliders (Trachemys scripta) in west-central Illi- mayflies: a predator satiation hypothesis. Evolution 36: nois. J Herpetol 32:431–434. 810–821. Tucker JK, Paukstis GL. 1999. Post-hatching substrate Swingland IR, Coe MJ. 1978. The natural regulation of giant moisture and overwintering hatchling turtles. J Herpetol tortoise populations on Aldabra Atoll: recruitment. Phil 33:608–615. Trans R Soc Lond B 286:177–188. Tucker JK, Filoramo GL, Paukstis GL. 1998. Residual yolk in Szcerbak NN. 1998. The European pond turtle (Emys captive and wild-caught hatchlings of the red-eared slider orbicularis) in Ukraine. Proceedings of EMYS symposium, turtle (Trachemys scripta elegans). Copeia 1998:488–492. Vol. 10. Dresden, Mertensiella, p 259–266. Tucker JK, Paukstis GL, Janzen FJ. 2008. Does predator Takagi H, Sakai K, Morida K, Nakamori S. 2000. Proline swamping promote synchronous emergence of turtle hatchl- accumulation by mutation or disruption of the proline ings among nests? Behav Ecol 19:35–40. oxidase gene improves resistance to freezing and desiccation Tuma M. 1993. Ecology of the state endangered yellow mud stresses in Saccharomyces cerevisiae. FEMS Microbiol Lett turtle, Kinosternon flavescens in Henry Co., Illinois. MS 184:103–108. thesis, Eastern Illinois University, Charlestown, USA. Thacker TL. 1924. Notes on Bell’s painted turtle (Chrysemys Tuma M. 2006. Range, habitat use, and seasonal activity of the marginata bellii) in British Columbia. Can Field-Nat 38: yellow mud turtle (Kinosternon flavescens) in northwestern 164–167. Illinois: implications for site-specific conservation and The´ron A, Combes C. 1995. Asynchrony of infection timing, management. Chel Cons Biol 5:108–120. habitat preference, and sympatric speciation of schistosome Turner FB, Hayden P, Burge BL, Roberson JB. 1986. Egg parasites. Evolution 49:372–375. production by the desert tortoise (Gopherus agassizii)in Thomas K. 1972. The annual cycle of reproduction of the California. Herpetologica 42:93–104. emydine turtle, Pseudemys floridana floridana (Testudina- Tursman D, Duman JG, Knight CA. 1994. Freeze tolerance ta, Testudinidae) with observations on its ecology. MS adaptations in the centipede, Lithobius fortificatus. J Exp thesis, Auburn University, Auburn, USA. Zool 268:347–353.

J. Exp. Zool. 378 J.P. COSTANZO ET AL.

Turtle Taxonomy Working Group. 2007. An annotated list of Washington Department of Fish and Wildlife, Wildlife modern turtle terminal taxa with comments on areas of Management Program. instability and recent change. In: Shaffer HB, FitzSimmons Vernon P, Vannier G. 2002. Evolution of freezing suscept- NN, Georges A, Rhodin AGJ, editors. Defining turtle ibility and freezing tolerance in terrestrial arthropods. C R diversity: proceedings of a workshop on genetics, ethics, Biol 325:1185–1190. and taxonomy of tortoises and freshwater turtles. Chelon Vincent TK, Secoy DM. 1978. The effects of annual variation Res Monogr 4:173–199. in temperature on cold resistance in a northern population Tuttle SE, Carroll DM. 1997. Ecology and natural history of of the red-sided garter snake, Thamnophis sirtalis parieta- the wood turtle (Clemmys insculpta) in southern New lis. J Herpetol 12:291–294. Hampshire. Linnaeus fund research report. Chel Cons Biol Vogt RC. 1980. Natural history of the map turtles Graptemys 2:447–449. pseudogeographica and G. ouachitensis in Wisconsin. Tul Tuttle SE, Carroll DM. 2005. Movements and behavior of Stud Zool Bot 22:17–48. hatchling wood turtles (Glyptemys insculpta). Northeast Nat Vogt RC. 1981. Turtle egg (Graptemys: Emydidae) infestation 12:331–348. by fly larvae. Copeia 1981:457–459. Ultsch GR. 1989. Ecology and physiology of hibernation and Voituron Y, Mouquet N, de Mazancourt C, Clobert J. 2002a. overwintering among freshwater fishes, turtles, and snakes. To freeze or not to freeze? An evolutionary perspective on Biol Rev 64:435–516. the cold-hardiness strategies of overwintering ectotherms. Ultsch GR. 2006. The ecology of overwintering among turtles: Am Nat 160:255–270. where turtles overwinter and its consequences. Biol Rev 81: Voituron Y, Verdier B, Grenot C. 2002b. The respiratory 339–367. metabolism of a lizard (Lacerta vivipara) in supercooled and Ultsch GR, Anderson JF. 1986. The respiratory microenvir- frozen states. Am J Physiol 382:R181–R186. onment within the burrows of gopher tortoises (Gopherus Voituron Y, Eugene M, Barre H. 2003. Survival and metabolic polyphemus). Copeia 1986:787–795. responses to freezing by the water frog (Rana ridibunda). Ultsch GR, Jackson DC. 1982. Long-term submergence at 31C J Exp Zool 299A:118–126. of the turtle, Chrysemys picta bellii, in normoxic and Voituron Y, Joly P, Eugene M, H. B. 2005a. Freezing tolerance severely hypoxic water. I. Survival, gas exchange and of the European water frogs: the good, the bad, and the ugly. acid–base status. J Exp Biol 96:11–28. Am J Physiol 288:R1563–R1570. Ultsch GR, Reese SA. 2008. Ecology and physiology of Voituron Y, Servais S, Romestaing C, Douki T, Barre H. overwintering. In: Brooks RJ, Steyermark AC, Finkler M, 2005b. Oxidative DNA damage and antioxidant defenses in editors. The biology of the snapping turtle. Washington, DC: the European common lizard (Lacerta vivipara) in super- Smithsionian Institution Press. p 91–99. cooled and frozen states. Cryobiology 52:74–82. Ultsch GR, Reese SA, Stewart ER. 2004. The physiology of Wahlquist H, Folkerts GW. 1973. Eggs and hatchlings of hibernation in Rana pipiens: metabolic rate, critical oxygen Barbour’s map turtle, Graptemys barbouri Carr and Marc- tension, and the effects of hypoxia on several plasma hand. Herpetologica 29:236–237. variables. J Exp Zool 301A:169–176. Wallace BP, Sotherland PR, Spotila JR, Reina RD, Franks BF, Ultsch GR, Draud M, Wicklow BJ. 2008. Post-emergent Paladino FV. 2004. Biotic and abiotic factors affect the nests movements and overwintering of snapping turtle, Chelydra environment of embryonic leatherback turtles, Dermochelys serpentina, hatchlings in New York and New Hampshire. coriacea. Physiol Biochem Zool 77:423–432. Can Field-Nat (in press). Warren DE, Reese SA, Jackson DA. 2006. Tissue glycogen and Upper CD, Vali G. 1995. The discovery of bacterial ice extracellular buffering limit the survival of red-eared slider nucleation and its role in the injury of plants by frost. In: turtles during anoxic submergence at 31C. Physiol Biochem Lee RE, Warren GJ, Gusta LV, editors. Biological ice Zool 79:736–744. nucleation and its applications. St. Paul, MN: American Waye HL, Gillies C. 1999. Chrysemys picta bellii (Western Phytopathological Society Press. p 29–39. painted turtle). Early emergence. Herpetol Rev 30:94–95. Valerio PF, Kao MH, Fletcher GL. 1992. Fish skin: an effective Weatherhead PJ. 1989. Temporal and thermal aspects of barrier to ice crystal propagation. J Exp Biol 164:135–151. hibernation of black rat snakes (Elaphe obsoleta) in Ontario. Vali G. 1991. Nucleation of ice. In: Gotz G, Meszaros E, Vali G, Can J Zool 67:2332–2335. editors. Atmospheric particles and nuclei. Budapest: Aka- Weisrock DW, Janzen FJ. 1999. Thermal and fitness-related demiai Kiado. p 131–192. consequences of nest location in painted turtles (Chrysemys Vali G. 1995. Principles of ice nucleation. In: Lee RE, Warren picta). Funct Ecol 13:94–101. GJ, Gusta LV, editors. Biological ice nucleation and its Wharton DA. 2002. Life at the limits. Cambridge, UK: applications. St. Paul, MN: American Phytopathological Cambridge University Press. 307p. Society Press. p 1–28. Wilbur HM. 1975a. A growth model for the turtle Chrysemys van Waarde A. 1988. Biochemistry of non-protein nitrogenous picta. Copeia 1975:337–343. compounds in fish including the use of amino acids for Wilbur HM. 1975b. The evolutionary and mathematical anaerobic energy production. Comp Biochem Physiol B 91: demography of the turtle Chrysemys picta. Ecology 56: 207–228. 64–77. van der Lee GEM, de Winder B, Bouten W, Tietema A. 1999. Wilhoft DC. 1986. Eggs and hatchling components of the Anoxic microsites in Douglas fir litter. Soil Biol Biochem 31: snapping turtle (Chelydra serpentina). Comp Biochem 1295–1301. Physiol A 84:483–486. Van Leuven S, Allen H, Slavens K, Clark S, Anderson D. 2004. Wilhoft DC, Hotaling EC, Franks P. 1983. Effects of Western pond turtle head-starting and reintroduction Oct. temperature on sex determination in embryos of the 2003 thru Sept. 2004. Olympia and Vancouver, Washington: snapping turtle, Chelydra serpentina. J Herpetol 17:38–42.

J. Exp. Zool. WINTER BIOLOGY OF HATCHLING TURTLES 379

Willard R, Packard GC, Packard MJ, Tucker JK. 2000. The Wright JS. 1982. Distribution and population biology of the role of the integument as a barrier to penetration of ice into gopher tortoise, Gopherus polyphemus, in South Carolina. overwintering hatchlings of the painted turtle (Chrysemys MS thesis, Clemson University, Clemson, USA. picta). J Morphol 246:150–159. Wygoda ML, Chmura CM. 1990. Effects of shell closure on Willette DAS, Tucker JK, Janzen FJ. 2005. Linking climate water loss in the Sonoran mud turtle, Kinosternon sonor- and physiology at the population level for a key life-history iense. Southwest Nat 35:228–229. stage of turtles. Can J Zool 83:845–850. Yancey PH. 2005. Organic osmolytes as compatible, metabolic Williams J, Shorthouse JD, Lee RE. 2002. Extreme resistance and counteracting cytoprotectants in high osmolarity and to desiccation and microclimate-related differences in cold- other stresses. J Exp Biol 208:2819–2830. hardiness of gall wasps (Hymenoptera: Cynipidae) over- Yearsley JM, Kyriazakis I, Gordon IJ. 2004. Delayed costs of wintering on roses in southern Canada. J Exp Biol 205: growth and compensatory growth rates. Funct Ecol 18: 2115–2124. 563–570. Willmore WG, Storey KB. 1997. Antioxidant systems and Zachariassen KE. 1985. Physiology of cold tolerance in insects. Physiol Rev 65:799–831. anoxia tolerance in a freshwater turtle Trachemys scripta Zachariassen KE. 1991. The water relations of overwintering elegans. Mol Cell Biochem 170:177–185. insects. In: Lee RE, Denlinger DL, editors. Insects at low Wojakowski MM, Widrig AL, Burke RL. 2005. Emergence of temperature. New York: Chapman & Hall. p 47–63. the diamondback terrapin (Malaclemys terrapin) in Jamaica Zachariassen KE, Hammel T. 1976. Nucleating agents in the Bay Wildlife Refuge, New York City. Abstracts of the 2005 haemolymph of insects tolerant to freezing. Nature 262: meeting of the American Society of ichthyologists and 285–287. herpetologists, Tampa, FL. Zachariassen KE, Husby JA. 1982. Antifreeze effect of thermal Wolanczyk JP, Storey KB, Baust JG. 1990. Ice nucleating hysteresis agents protects highly supercooled insects. activity in the blood of the freeze-tolerant frog, Rana Nature 298:865–867. sylvatica. Cryobiology 27:328–335. Zachariassen KE, Kristiansen E. 2000. Ice nucleation and Wolfe J, Bryant G, Koster KL. 2002. What is ‘‘unfreezable antinucleation in nature. Cryobiology 41:257–279. water’’, how unfreezable is it and how much is there? Cryo- Zinenko O. 2004. Notes on egg-laying, clutch size and Letters 23:157–166. hatchling feeding of Emys orbicularis in the Kharkiv region, Woodbury AM, Hardy R. 1948. Studies of the desert tortoise, Ukraine. Biol Bratislava 59:149–151. Gopherus agassizii. Ecol Monogr 18:144–200. Zuffi MAL. 2000. Conservation biology of the European pond turtle Woolverton E. 1963. Winter survival of hatchling painted Emys orbicularis (L.) in Italy. Stapfia 69, zugleich Kataloge des turtles in northern Minnesota. Copeia 1963:569–570. OO¨ Landesmuseums, Neue Folge Nr 149, p 219–228.

J. Exp. Zool.