REVIEW ARTICLE

Physiology of Hibernation Under

the Ice by Turtles and Frogs
DONALD C. JACKSON1 AND GORDON R. ULTSCH2y
1
Department of Molecular Pharmacology, Physiology, and Biotechnology, Brown University,
Providence, Rhode Island
2
Department of Biology, University of Florida, Gainesville, Florida

ABSTRACT Successful overwintering under ice by an air-breathing vertebrate requires either effective aquatic
respiration if dissolved O2 is available or the capacity for prolonged anaerobic metabolism if O2
supplies are limiting. Frogs can remain aerobic for many weeks when submerged at low
temperature, even at water PO2 as low as 30 mmHg, but are unable to survive even 1 week in
anoxic water. Fuel reserves of hibernating frogs limit aerobic submergence, whereas acidosis may
limit anoxic submergence. Freshwater turtles can also satisfy all or most of their O2 needs in well-
aerated water at low temperature by aquatic respiration, but certain species, in particular painted
and snapping turtles, can also survive for up to 4–5 months without O2. Key adaptations of the
painted turtles, and presumably snapping turtles, include metabolic depression and the
exploitation of the shell and other bones to buffer lactic acid. As in frogs, glycogen and glucose
are the only fuel sources during anoxia, and stores do not seem to be limiting in the painted turtle.
Significant differences in anoxia tolerance exist among chelonian species that can be attributed, at
least in part, to the magnitude of metabolic depression, the effectiveness of lactic acid buffering,
and the size of glycogen stores. J. Exp. Zool. 313A:311–327, 2010. & 2010 Wiley-Liss, Inc.
J. Exp. Zool.
313A:311–327,
How to cite this article: Jackson DC, Ultsch GRy. 2010. Physiology of hibernation under the ice
2010 by turtles and frogs. J. Exp. Zool. 313A:311–327.

Frogs and freshwater turtles are obligate air breathers at summer
temperatures, but species overwintering at high latitudes in the
northern hemisphere may be unable to breathe air for many
weeks while their habitats are ice-covered. Indeed, as much as
half of their lifetimes may be spent holding their breath during
under-ice hibernation, accompanied by low levels of activity and
greatly reduced metabolic rates. During this time, they do not
feed and must survive by mobilizing stored deposits of metabolic
substrates. Important events, such as mating and reproduction,
will not occur until after emergence. In effect, they are in a
holding pattern, conserving their resources until normal activities
can resume in the spring.
Of central importance for the submerged animal is the
availability of dissolved oxygen and how well any available
oxygen can be delivered to metabolizing cells. Photosynthesis or
flowing water may keep the water PO2 (PwO2) near ambient air
levels, but heavy snow cover or eutrophication may produce
severely hypoxic or even anoxic conditions. Animals burying
farther north than that of turtles, so that the length of time frogs
may be submerged at low temperature can exceed that of turtles.
Maintenance of adequate fuel reserves is also a key requirement
for successful overwintering, particularly if reproductive activity
in the spring occurs before feeding and replenishment of reserves.
If sufficient oxygen can be obtained from the water while
hibernating, fat stores can be exploited, but fat is not utilizable
under anaerobic conditions. Turtles are clearly less well adapted
for cutaneous exchange than are frogs; nevertheless, all species
of turtles tested in normoxic water during simulated hibernation
can support significant aerobic metabolism using aquatic
respiration, and some species can remain fully aerobic.
The remarkable feature of turtles, in contrast to frogs, is their
capacity to survive long periods without oxygen. The physiolo-
gical traits that permit this extraordinary behavior have been
Correspondence to: Donald C. Jackson, Department of Molecular

into the bottom mud may also have little or no available oxygen,
regardless of the PwO2 of the overlying water. Although both
frogs and turtles can benefit from available aquatic oxygen, the
more effective cutaneous gas exchange of frogs enables them to
remain aerobic even in water that is well below normoxic (i.e.,
atmospheric) PO2 levels. The geographical range of frogs extends
Pharmacology, Physiology, and Biotechnology, Brown University, Providence,
RI 02912. E-mail: donald_jackson@brown.edu
y
Courtesy Professor.
Received 20 October 2009; Revised 18 December 2009; Accepted 17
February 2010
Published online 26 March 2010 in Wiley InterScience (www.
interscience.wiley.com). DOI: 10.1002/jez.603

& 2010 WILEY-LISS, INC.

312
the subject of numerous studies since the pioneering work of
Belkin (’63) and Robin et al. (’64) awakened interest in this
subject. Of particular note is the capacity of several species of
freshwater turtles (e.g., Chrysemys picta and Chelydra serpentina)
to survive for several months submerged at winter temperatures
with reduced or no dissolved oxygen. An anoxic state requires a
shift to anaerobic glycolysis, an inefficient pathway energetically
because of the low yield of ATP per glucose molecule, and a
pathway that disturbs acid-base balance owing to the production
of lactic acid. Only free glucose and glycogen can serve as
metabolic substrates during anoxia, so adequate availability of
these compounds is essential. Coping with these potentially
limiting stresses requires unusual adaptive traits in these animals
that include, in particular, metabolic depression and effective
acid buffering. In addition, considerable research has been
directed toward understanding brain and heart function of turtles
under anaerobic conditions, because these organs are usually
quite vulnerable to low oxygen in other vertebrates. Discovering
the mechanisms that permit their continued function and
survival in anoxic turtles is of fundamental importance.
The purpose of this review is two-fold: first, to discuss the
physiological traits that permit long-term winter submergence in
frogs and turtles, including a consideration of environments both
with and without normoxic levels of dissolved oxygen, and
second, to point out the incompleteness of our understanding in
many of these areas. Because the focus of our work has been at
the level of whole-animal physiology, we emphasize this aspect
in our review with less discussion on cellular and molecular
mechanisms. Readers interested in the latter approaches can
profit from recent reviews by Bickler and Buck (2007), Boutilier
(2001), and by a series of articles in a recent issue of Comparative
Biochemistry and Physiology (Part A, Volume 147, pp 261–343,
2007) that were published to honor the memory of the late Peter
Lutz, who was a major contributor to our understanding of
anoxia tolerance in turtles. Additional recent reviews on the
general subject include Boutilier et al. (’97), Jackson (2000, 2002),
Ultsch (2006), Tattersall and Ultsch (2008), and Milton (2008).
SUBMERGENCE IN WATER WITH DISSOLVED OXYGEN
Aquatic Gas Exchange
Frogs and other amphibians are well known to rely on their skin
for a variety of exchange processes with their aquatic environ-
ment, including gas exchange. Their moist skin is heavily
vascularized and the diffusion distance from the surface of the
epidermis to the dermal capillaries is relatively short compared
with other vertebrates, on the order of 20–50 m, and in extreme
cases as little as 12 m (Czopek, ’65). This distance, though, is long
compared with lung or gill diffusion paths, so skin exchange,
even in amphibians, has been considered to be diffusion limited
(Gatz et al., ’75). Even at summer temperatures, the skin can
contribute significantly to gas exchange, particularly for CO2 loss
J. Exp. Zool.
JACKSON AND ULTSCH
(Hutchison et al., ’68) and the relative contribution of the skin to
exchange of both O2 and CO2 increases as temperature decreases
(Feder and Burggren, ’85). In contrast, the integument of
freshwater turtles is designed more to resist exchange with the
environment than to facilitate it, and at summer temperatures
extrapulmonary exchange is trivial for most, though not all,
turtles. Nevertheless, aquatic gas exchange is critically important
to turtles as well as to frogs during overwintering.
Extrapulmonary gas exchange can occur by at least three
different pathways in submerged turtles: via the skin, bucco-
pharyngeal surface, or across the surface of the cloaca or cloacal
bursae. The relative contributions of these pathways differ among
species (King and Heatwole, ’94). In certain subtropical Australian
turtles, aquatic ventilation of the cloacal bursae can supply most
of the oxygen requirements even at warm temperatures (Gordos
and Franklin, 2002). In North American species, the cloacal
bursae are poorly vascularized or absent and probably play less
of a role in gas exchange (Peterson and Greenfields, 2001). The
softshell turtle (Apalone spinifera) and the musk turtle (Ster-
notherus odoratus), both North American temperate species, have
significant but still not sufficient aquatic gas exchange at
summer temperatures (Dunson ’60; Belkin, ’68), owing to gas-
permeable skin and additionally, in the case of the softshell, to
buccopharyngeal respiration (Wang et al., ’89). The painted turtle,
C. picta, a species representative of more typical turtle
morphology, relies principally upon its skin for aquatic O2
uptake (Jackson et al., 2004) but, like its close relative Trachemys
scripta (Belkin, ’68), cutaneous O2 uptake at summer tempera-
tures is a minor component of total O2 uptake (Wasser et al., ’91).
At winter temperatures, however, all the turtles tested, including
Chrysemys and Trachemys, can obtain enough O2 from aerated
water to supply most or all of their metabolic requirements. The
species S. odoratus (Ultsch and Cochran, ’94), A. spinifera (Reese
et al., 2003), and Graptemys geographica (Reese et al., 2001) can
remain fully aerobic while submerged at 31C in aerated water,
and C. picta (Reese et al., 2004) and C serpentina (Reese et al.,
2002) rely only to a relatively modest extent on anaerobic
metabolism under similar experimental circumstances. In addi-
tion, extrapulmonary loss of CO2 prevents or minimizes
respiratory acidosis during winter submergence (e.g., Ultsch and
Jackson, ’82).
The greater relative importance of cutaneous gas exchange at
low temperature is because low temperature depresses cellular
metabolism much more than it depresses diffusional gas
exchange. Therefore, the magnitude of aquatic O2 uptake by a
painted turtle that is inconsequential at 201C can meet most of
the cellular demands for O2 at 31C when metabolism is extremely
low (Jackson et al., 2001a). For frogs, and for turtles such as
Apalone and Sternotherus with effective aquatic exchange, the
temperature at which aquatic O2 uptake can fully satisfy resting
aerobic demands is higher than 31C; so, at hibernating
temperatures these species have a reserve O2 uptake capacity
HIBERNATION UNDER THE ICE BY TURTLES AND FROGS
that can increase aerobic scope or permit aerobic function in
hypoxic environments. Those species with more effective aquatic
gas exchange can be said to have a higher cutaneous
conductance for respiratory gases. Conductance (G) is defined
_ 2 and the gas partial pressure
as the ratio between gas flux VO
difference between water and arterial blood. Thus, for O2:
_ 2 =ðPw O2  Pa O2 Þ:
GO2 ¼ VO
Based on this relationship, it is obvious that under limiting
ambient PwO2 conditions, the aerobic state can potentially be
maintained by either increasing GO2 or by metabolic depression
(i.e., decreasing O2 demand). Metabolic depression beyond that
attributable to low temperature alone has been observed in frogs
(R. temporaria) under severe hypoxic conditions (Donohoe and
Boutilier, ’98) at 31C. In other experiments in the same study,
metabolic rates of frogs were observed to fall in response to
decreasing PwO2 until at PwO2 of 60 mmHg, O2 consumption was
reduced to only 25% of the normoxic value. Indirect evidence
also suggests that turtles become hypometabolic when sub-
merged at low temperature in normoxic water, discussed in detail
in the next section.
Supply can also be matched to demand by increasing skin
conductance for respiratory gases. Conductance has been
reported to increase during cold submergence in frogs to
facilitate gas exchange. Pinder (’87) submerged paralyzed
R. catesbeiana at 51C and observed constant diffusional O2
uptake as PwO2 was reduced from 140 to 80 mmHg, a response he
interpreted as increased GO2. Below 80 mmHg, O2 uptake
decreased sharply, although GO2 continued to increase. Later
studies on smaller species, which were also not treated with any
drug (those used by Pinder were curarized), suggest that the
critical O2 tension of cold submerged frogs is much lower (Fig. 1),
in the neighborhood of 30–40 mmHg (Bradford, ’83; Tattersall
and Boutilier, ’97, ’99a; Ultsch et al., 2004), and may be even
lower during chronic hypoxia (see review of Tattersall and Ultsch,
2008). Based on other studies (e.g., Burggren and Moalli, ’84), the
Figure 1. Oxygen consumption (V̇O2) as a function of ambient PO2
in Rana pipiens acclimated and tested at 31C during progressive
hypoxia (from Ultsch et al., 2004). Pc is the critical oxygen tension.
313
likely mechanism for increased skin conductance is recruitment
of earlier closed skin capillaries. More recently, Tattersall and
Boutilier (’99b) reported an increase in skin CO2 conductance
between 0.2 and 71C in R. temporaria submerged in aerated
water. Blood PCO2 and pH remained constant over this range,
indicating that cutaneous gas exchange capacity was matched to
metabolic rate. The combined effects of metabolic depression and
increased skin gas exchange conductance that have been
described by Boutilier and his co-workers are important
adaptations that can enable frogs to survive periods of severe
hypoxia while minimizing reliance on anaerobic metabolism
(see Tattersall and Ultsch, 2008, for a review). Some evidence
suggests similar mechanisms in turtles, but the necessary studies
have not been performed to verify their presence.
Metabolic Rate and Energetics
The metabolic rates of turtles are greatly diminished during
hibernation, by as much as 95% when breathing air and water
and 99% when submerged in anoxic water, relative to that while
breathing air at 201C (Herbert and Jackson, ’85b). Whether frogs
are able to attain such a profound reduction in metabolic rate
while hibernating is an open question, but if they cannot, energy
limitations may be one reason why the overwintering survival of
frogs seems to be considerably less than that of turtles (Wilbur,
’75; Galbraith and Brooks, ’87; Elmberg, ’90; Congdon et al., ’94,
2000). The aerobic metabolic rate of ranid frogs in normoxic
water at 3–51C is about 6 mL O2 g1h1 (Bradford, ’83; Pinder,
’87; Tattersall and Boutilier, ’97; Ultsch et al., 2004). Herbert and
Jackson (’85b) reported the metabolic rate of the painted turtle (C.
picta bellii) in aerated water with air access at 31C to be 1.14 mL
O2 g1h1. For comparison with a 60 g frog, and assuming that
the slope of the log–log regression of body size against metabolic
rate in turtles is 0.75, it follows that a 60 g turtle (allowing for
most shell mass as relatively inactive tissue) would have a
metabolic rate of 2.27 mL O2 g1h1, which is only 38% that of a
frog of the same size. This is a conservative estimate, as a turtle’s
metabolic rate is probably lower than this because of denial to air
access. In fact, a series of studies of four species of turtles
submerged in respirometers at their winter hibernation sites
(2–91C: Graham and Forsberg, ’91; Graham and Graham, ’92;
Graham and Butler, ’93; Graham and Guimond, ’95) gave
metabolic rates (adjusted here to 60 g) of 0.68–1.83 mL O2 g1h1,
which are 11–30% of those measured for frogs. Even considering
all the assumptions in these calculations, it seems likely that the
metabolic rates of overwintering frogs are at least twice those of
turtles, and frogs, therefore, should be able to survive only half as
long if both taxa entered hibernation with similar relative energy
stores. Lower metabolic rates should also permit turtles to survive
a more severe hypoxia while remaining aerobic than frogs, by
virtue of a lower critical O2 tension (Ultsch et al., 2004).
As discussed above, continuous submergence in cold water
elicits a hypometabolism in frogs, and probably also in turtles.
J. Exp. Zool.
314
Rana temporaria submerged in normoxic water at 3oC accumu-
late no lactate initially, whereas those submerged in hypoxic
(PO2 5 60 mmHg) water have an initial increase in plasma lactate
to E9 mmol L1 that returns to control levels of E1.5 mmol L1
after 8 weeks of submergence (Donohoe and Boutilier, ’98).
A progressive hypometabolism, triggered by both submergence
and hypoxia, lowers energy demands enough that extrapulmon-
ary respiration is not only sufficient for the frog to become
completely aerobic, but also to metabolize the excess lactate.
Among turtles, the anoxia-tolerant C. picta and C. serpentina
accumulate relatively small amounts of plasma lactate (E20–25
mmol L1) after more than 100–150 days of submergence at 31C
in normoxic water, which they are not able to metabolize;
however, the rates of accumulation fall substantially after 10 days
of submergence, also suggesting a progressive hypometa-
bolism (Reese et al., 2002, 2004). Anoxia-intolerant species, such
as A. spinifera, G. geographica, and S. odoratus, when submerged
in normoxic water at 31C, accumulate either no or a
physiologically insignificant amount of lactate during long-term
submergence, and remain essentially aerobic (Reese et al., 2003).
Frogs enter hibernation with empty guts and do not feed until
emergence (Koskela and Pasanen, ’74; Bradford, ’83). Some
species feed after emerging and before breeding, such as bullfrogs
(R. catesbeiana) and green frogs (R. clamitans), which are
relatively late breeders (Martof, ’56; Willis et al., ’56; Hulse
et al., 2001). In contrast, some species do not feed until after
breeding in spring, such as northern European R. temporaria
(Pasanen and Koskela, ’74). In either case, the frogs must enter
hibernation with sufficient energy reserves for overwintering,
and in the latter case, enough additional stored energy must be
available for reproduction.
The major energy stores of frogs are glycogen and lipids. The
major storage depot for glycogen is the liver, followed by muscle
glycogen. Fat bodies have typically been assumed to be the major
depot for lipids, and their reduction or depletion during
hibernation, along with decreases in glycogen stores, has been
noted in numerous studies (e.g., Smith, ’50; Pasanen and Koskela,
’74; Byrne and White, ’75; Pinder et al., ’92). However, fat bodies
constitute o25% of the total body fat in R. pipiens and R.
clamitans before hibernation (Brenner, ’69; Brenner and Brenner,
’69), so calculations of the potential energetic contributions of fat
to the overwintering energy balance that utilize only lipids in fat
bodies are likely to be somewhat inaccurate, depending upon the
ability to mobilize lipids from other locations. Measured whole-
body lipid reserves were calculated to be sufficient for at least
4 months at overwintering temperatures in red-spotted newts
from Ohio (Notophthalmus viridescens; Jiang and Claussen, ’92),
which is the expected time of overwintering; the addition of
glycogen stores would extend the calculated survival time.
Similarly, Bradford (’83) calculated that R. muscosa could easily
survive its 8 months of hibernation using no more than half of its
whole-body fat stores, with the remaining being available for
J. Exp. Zool.
JACKSON AND ULTSCH
gametogenesis and breeding, which occurs soon after ice melt
(Lannoo et al., 2005) and, therefore, presumably before feeding.
Generally, the higher the latitude at which a frog occurs, the
longer the hibernation period, and presumably the more energy
reserves the frog must have for overwintering; conversely, the
shorter will be the period of activity available to accrue these
reserves. Thus, one could theorize that energetics limits the
northern distribution of frogs. From the viewpoint of the
hibernation period, data are conflicting. A key factor is the rate
of energy depletion during overwintering (e.g., metabolic rate);
however, this rate is quite labile, being dependent at least upon
time submerged, temperature, and dissolved O2 (Boutilier et al.,
’97; Boutilier and St-Pierre, 2002; Ultsch et al., 2004). Boutilier
et al. (’97) calculated that liver glycogen and fat body energy
stores of R. temporaria would last 157 days for a frog submerged
in normoxic water at 31C, assuming a metabolic rate depressed to
the point they observed relative to that of an air-breathing frog,
but would last 308 days at a water PO2 of 50 mmHg, owing to a
further metabolic depression associated with hypoxia. Because
their study considered neither muscle glycogen nor fats other
than those in the fat bodies, both estimates are probably low.
Bradford (’84) submerged R. muscosa at 41C and they were still
alive a year later. These studies suggest that the energetics of
overwintering does not limit the northern range of frogs.
However, some species, such as R. temporaria, breed before
feeding, so the energetic requirements of overwintering are not
the only consideration; one must consider the total energy
required from last feeding during autumn until the next feeding
after breeding in spring.
An appropriate model to address energetic limitations on
northern range limits is R. temporaria in North (E641N) Finland,
where hibernation lasts 7–8 months (Pasanen and Koskela, ’74)
and emergent frogs breed before feeding. For males (complete
data not given for females), the percentage of the weight of the
liver expressed as a percentage of body weight was 5.07%
entering hibernation, 2.83% emerging from hibernation, and
2.04% after breeding. The glycogen content of the liver at the
corresponding times were 16.4, 14.0, and 3.1%, which at first
would seem to indicate that much more of the glycogen energy
stores were used during the post-hibernation breeding period
than during the much longer hibernation period. However, the
size of the liver was decreasing during overwintering. If one
considers the total glycogen stores, using data supplied on liver
mass and the percentage of liver mass present as glycogen, it can
be calculated that male frogs entered hibernation with 326.5 mg
of liver glycogen, emerged with 159.0 mg, and had only 24.9 mg
remaining after spawning. Thus, as an approximation, 51% of the
liver glycogen stores were used during hibernation and an
additional 41% were used during spawning. Only about 8% of the
original liver glycogen stores were left from the time of entering
hibernation until the time the frogs were ready to start feeding
again, which is tantamount to saying that liver glycogen stores
HIBERNATION UNDER THE ICE BY TURTLES AND FROGS
are close to depleted during the period from last feeding to first
feeding. The lack of after-spawning data for females prevents a
complete comparison to males, but for those entering hiberna-
tion, glycogen stores, calculated as above, are 253.8 mg
compared with 111.6 mg when emerging, or 56% utilization,
about the same as males over the same period. Similar
calculations for fat bodies (sexes combined) with a mean mass
of 258.0 mg at the start of hibernation yield 206.4 mg lipid. After
hibernation, there were only 14.0 mg of lipid left in the fat
bodies; this 93% utilization of lipids in the fat bodies indicates
that fats are the preferred energy source during overwintering.
Similar calculations for data supplied on liver lipids give a 69.6%
utilization of an original 133 mg of lipid. Combining liver lipids
with those from fat bodies and using a conversion of 9.4 cal g1
for lipids and 4.2 cal g1 for glycogen, indicates that 80.4% of the
energy used during hibernation from these sources came from
lipids. This favoring of lipid utilization is to be expected, so long
as there is sufficient oxygen available, because of their high
energy content relative to glycogen. Qualitatively, these data lead
to the conclusion that liver glycogen stores are about half
depleted during hibernation with most of the rest used for
breeding, and that lipids in fat bodies and liver lipid stores are
more reduced than liver glycogen stores after emergence from
hibernation, although the energy content of lipids remaining in
the liver and fat bodies is roughly equivalent to that of the
remaining liver glycogen. These data support the theory that
energetics may limit the northern range of this frog species. If it
occurred any farther north, there might not be enough time to lay
down sufficient energy stores during the shortened active season.
Studies with other species also suggest that frogs may be
energy-limited in their duration of hibernation. Smith (’50) found
fat bodies and liver glycogen to profoundly decrease during
overwintering in R. temporaria. Brenner and Brenner (’69) found
rates of decreases in whole-body lipids to be highly variable,
dependent upon sex and extent of darkness (continuous vs.
8 hr d1 of light), in R. pipiens at 51C. Initial body fat, expressed
as percentages of dry weight, ranged from approximately 11 to
17%, with only 10–24% of the total being found in fat bodies.
Extrapolating their data for total body fat to 0% gives times of
depletion from about 120 days for males exposed to darkness for
16 hr d1 to 350 days for females in continuous darkness.
However, it is unlikely that the frogs would survive if all lipids
were entirely consumed; so, for this species, which in the
northern part of its range may hibernate for 4200 days,
energetics may be a limiting factor. However, Bradford’s (’84)
finding that R. muscosa could survive hibernation for at least a
year argues that hibernation duration is not limited by energetics
in a natural setting. Perhaps the simplest test of the effects of
energetics in limiting overwintering survival would be to
submerge several species of frogs that hibernate aquatically in
cold normoxic water until death, and then compare their total
glycogen and lipid contents with those of frogs before
315
submergence. If a species can survive and recover from simulated
hibernation longer than it would ever have to hibernate in the
field, then energetics is clearly not a limiting factor; but if its
energy supplies are exhausted in a time near that of its natural
hibernation period, this would be evidence for the primacy of
energetics in determining hibernation survival time, and hence
the latitudinal range limits of frogs. Such data are currently
unavailable; therefore, whether energetics is the limiting factor in
determining the tolerable duration of hibernation in frogs in their
natural environments remains an open question.
Frogs store energy for overwintering as lipids in fat bodies,
liver, and elsewhere, in addition to storing glycogen in the liver
and other locations. As described above, this anticipatory energy
storage may be a requirement for successfully sustaining a
lengthy period of aphagia, including hibernation. Whether turtles
have such a requirement, considering their low rates of
metabolism while hibernating, is an open question. All northern
species tested can survive more than 200 days of submergence in
normoxic water at 31C without being supplied any extra food
before submergence (Ultsch GR, unpublished observations),
which suggests that energetics is not a factor limiting survival
during simulated overwintering periods in excess of what would
be expected naturally. Furthermore, several species can survive
for more than 100 days of submergence in anoxic water at 31C,
which means that they can persist almost entirely on glycogen
stores for more than 3 months without tapping lipid reserves.
Because lipid reserves supply more than twice the calories of
glycogen, total energy reserves seem to be more than sufficient
for typical periods of aphagia. Crawford (’94) found that, in C.
picta in Michigan, total carcass energy stores remaining after
hibernation were 62.2% of neutral lipids, 89.0% proteins, and
29.8% glycogen, percentages higher than that found for frogs
(see above). Crawford converted these data to energy equivalents
(kJ) for 2 animals and found that lipids contributed 56.5%,
glycogen 16.5%, and protein 27%, noting that the substantial
role of proteins is often overlooked; he also noted that substantial
energy reserves were available upon emergence. These data
suggest that the energetic costs of overwintering are not a
limiting factor in the life histories of turtles, as also suggested by
Congdon et al. (’82), who calculated that only 4% of the annual
energy budget of Michigan C. picta is used during overwintering.
Nevertheless, such data do not preclude the accumulation of
energy stores before hibernation. Certainly, some energy stores
are used during the period of aphagia that starts with decreasing
water temperatures in autumn, continues through hibernation,
and a period of water warming following emergence in spring
(painted turtles typically do not feed at water temperatures
o151C). This energy must be restored during the activity
season, but there are few data tracking such replacement. In G.
pseudogeographica from South Dakota, there was a progressive
3–4-fold increase in liver glycogen from June 10 to September 3,
but with a 34% decrease in liver lipids (Emerson, ’67). Because
J. Exp. Zool.
316
map turtles are aerobic during hibernation (Crocker et al., 2000a, b;
Reese et al., 2001), fat accumulation should be energetically
favored over glycogen accumulation. As Emerson’s study did not
measure total carcass lipids, which can be as high as 1.24 g of lipid
for every gram of lean body mass (Congdon and Tinkle, ’82), this
possibility cannot be ruled out. An informative study would be to
follow total body lipid, glycogen, and protein levels throughout an
annual cycle in a natural population of a northern turtle, such as
the painted turtle.
Acid-Base and Ion Balance
Frogs show little or no change in acid-base variables during
submergence in aerated water at low temperature. In
R. temporaria, Donohoe et al. (’98) did observe a transient
respiratory acidosis after 1 week at 31C, but blood pH returned to
control values for the remaining 15 weeks. Likewise, blood pH
remained unchanged during 125 days submergence in R. pipiens
and 150 days submergence in R. catesbeiana, both at 31C (Stewart
et al., 2004), although Pinder (’85, cited in Pinder et al., ’92) did
observe significant increases in blood pH owing to respiratory
alkalosis in R. pipiens submerged in aerated and in ‘‘mildly
hypoxic’’ water (PO2 5 80 mmHg) at 51C for 14 days. Except for
the short-term increase seen in R. temporaria (Donohoe et al.,
’98), blood PCO2 generally remains at control levels or falls
throughout submergence (Stewart et al., 2004). In hypoxic water,
blood PCO2 fell and pH rose (Boutilier et al., ’97), which these
authors suggested increased hemoglobin affinity for O2 and
facilitated O2 uptake from the water. Plasma lactate levels
remained at control levels throughout in all these studies,
consistent with an aerobic metabolic state.
Ion balance is only modestly affected in frogs submerged in
normoxic water at low temperatures, although a significant fall
in plasma [Na1] was observed in both R. pipiens and R.
catesbeiana (Stewart et al., 2004); plasma osmolality fell in R.
pipiens but not in R. catesbeiana, and neither [Cl] nor [K1] fell
in either species, so the osmoregulatory significance of the Na1
change is not clear. Donohoe et al. (2000) observed a significant
decrease in plasma [Na1] and also in skeletal muscle [K1], but no
change in either plasma or muscle [Cl] nor in ventricular muscle
[K1]. Despite these ionic changes, both types of muscles retained
control resting membrane potentials throughout submergence.
These authors postulated that Na1 may be selectively lowered to
reduce influx into cells and, thereby, lower the cellular work of
ion pumping. They also observed a 30% reduction in skeletal
muscle Na1/K1 pump activity as measured by ouabain–inhibi-
table Na1 and Rb1 flux.
The effect of submergence in cold aerated water on acid-base
and ionic statuses of turtles depends on the effectiveness of their
aquatic O2 uptake. Those species with effective O2 exchange that
remain fully aerobic during submergence had either a small
increase (softshell turtle; Reese et al., 2003) or no change in blood
pH (musk turtle; Ultsch, ’88; map turtle; Reese et al., 2001).
J. Exp. Zool.
JACKSON AND ULTSCH
In both the latter species, blood PCO2 fell during submergence,
suggesting either decreased metabolism or increased skin CO2
conductance; plasma [HCO 3 ] also fell, although plasma [lactate]
remained close to control levels. Inorganic ions tended to fall, an
effect owing most likely to an increase in body water. In the map
turtles, body weight increased by 6.8% after more than 150 days
of submergence. In the turtles with less effective aquatic gas
exchange (painted turtle; Jackson et al., 2000a; snapping turtle;
Reese et al., 2002), lactate increased significantly early in the
submergence period, but then tended to level off. In the painted
turtle, C. picta bellii, blood pH fell early by about 0.2 U and
remained at that value for the duration of the 125 day
submergence (Jackson et al., 2000a); elevated plasma [lactate]
and unchanged blood PCO2 indicated that this was a relatively
mild metabolic (or nonrespiratory) acidosis. In the snapping turtle
(Reese et al., 2002), blood pH remained at the control level
throughout the 150 days of submergence, although plasma
[lactate] was elevated throughout. Proportional decreases in
blood PCO2 and plasma [HCO 3 ] accounted for the constant pH. In
the painted and snapping turtles, a pattern of ionic changes was
observed that were qualitatively similar to those that have been
observed in submerged anoxic turtles (see below), including a
decrease in plasma [Cl], and increases in plasma [K1], [Ca21],
and [Mg21], as depicted in Figure 2. The changes were small,
however, compared with the anoxic state. Here and elsewhere in
this article the concentrations, [Ca21] and [Mg21], refer to total
Figure 2. Ion balance diagrams (Gamblegrams) for the painted
turtle, Chrysemys picta bellii, at 31C, either predive with access to
air, submerged in normoxic water for 125 days, or submerged in
anoxic water for 90 days. Ion identities for the Predive and
Normoxic bars are in the same order as in the labelled Anoxic bars.
Missing anions are presumably largely net negative charges on
plasma protein. Although not depicted in this diagram, the majority
of Ca21 and Mg21 in the anoxic state is combined with lactate.
Data from Jackson and Heisler (’82) and Jackson et al. (2000a).
HIBERNATION UNDER THE ICE BY TURTLES AND FROGS
calcium and magnesium as measured by atomic absorption
spectrophotometry, i.e., existing both as free ions and combined
with other solutes, such as protein or lactate.
RESPONSES TO SUBMERGENCE IN ANOXIC OR SEVERELY
HYPOXIC WATER
Overall Comparison of Frogs and Turtles
Frogs are distinctly less tolerant of anoxia than are freshwater
turtles. Even the softshell, A. spinifera, the least tolerant of
anoxia among turtles (Reese et al., 2003), survives experimental
anoxic submergence at 31C for about 3–4 times as long as either
R. pipiens or R. catesbeiana (Stewart et al., 2004; Tattersall and
Ultsch, 2008), neither of which survived more than 4 days. When
the comparison is with the most anoxia-tolerant turtle, C. picta
bellii, a 50-fold or more difference is observed (Jackson et al.,
2000a; Tattersall and Ultsch, 2008). Differences in tolerance
duration are associated with very different rates of development
of metabolic acidosis in the three species (Fig. 3). Because of its
more dramatic capacity to withstand anoxia, the following
discussion of mechanisms underlying anoxic survival will focus
on the painted turtle and its special adaptations, and the
comparatively limited corresponding traits of frogs will be
treated more briefly. It is important to emphasize at the outset,
however, that the resistance of frogs to anoxia, although modest
in comparison with turtles, could nevertheless have significant
ecological importance by permitting them to survive brief
Figure 3. Henderson–Hasselbalch diagram showing the change in
acid-base status during submergence anoxia at 31C in the painted
turtle, Chrysemys picta bellii (Reese et al., 2004), the softshell
turtle, Apalone spinifera (Reese et al., 2003), and the leopard frog,
Rana pipiens (Stewart et al., 2004). Note the similarity among the
species in the slopes of pH decrease but the very different time
courses involved.
317
episodes (a few days) of severe hypoxia in their overwintering
environments (Stewart et al., 2004). But long periods without
oxygen, on the order of a week or more, would likely prove fatal
for frogs, as is suggested by frequent reports of frog winterkills
(Manion and Cory, ’52; Merrell, ’77; Bradford, ’83; Licht, ’91).
Metabolic Rate
Two key metabolic responses are necessary for turtles to survive
anoxia at low temperature for extended periods. First, metabo-
lism must be greatly depressed in order to slow the rates of
substrate depletion and metabolic waste production. Second, it is
necessary to maintain adequate cellular energy balance by
matching the rate at which ATP is being produced to the rate at
which it is being consumed. From a whole-animal descriptive
perspective, several factors contribute to achieving metabolic
reduction in the cold, anoxic turtle. First, turtles are ectotherms,
so their standard metabolic rate at any temperature is low
compared with that of an endotherm. Second, reducing
temperature further reduces metabolic rate owing to the well-
known effect of temperature on chemical reactions. In an
animal’s temperature range when active, this effect typically
lowers rates by 2–3-fold for each 10oC fall in temperature
(Q10 5 2–3). However, at temperatures below the normal activity
range, such as winter temperatures near 01C, Q10 values can be
much higher (Hochachka and Somero, 2002). Measurements of
aerobic metabolic rates in turtles (C. picta bellii) revealed that Q10
values were as high as 8.5 between 10 and 31C (Herbert and
Jackson, ’85b). Finally, prolonged submergence and anoxia
further reduced metabolic rate by some 10-fold compared with
the aerobic rate at the same temperature. Anaerobic metabolic
rate has been measured in the turtles, T. scripta, at 241C by direct
calorimetry (Jackson, ’68) and C. picta at 31C by lactate
accumulation (Herbert and Jackson, ’85b); individual organ
anaerobic rates have been measured in brain (Lutz et al., ’85) and
liver (Buck et al., ’93). Together, these studies indicate a metabolic
rate in a cold anoxic turtle that is only about 0.6% of the rate of
an aerobic turtle at 201C (Herbert and Jackson, ’85b), and less
than 0.01% of a similarly sized resting mammal at its euthermic
body temperature (Jackson, 2000).
Profound metabolic depression greatly slows the rate at which
stored substrates are depleted. For an anoxic turtle that relies
exclusively on anaerobic glycolysis, its principal substrate is
glycogen, the bulk of which is in liver and skeletal muscle.
Glycogen breakdown leading to lactate as the end-product is
quite inefficient energetically. Only 3 molecules of ATP are
derived from each glucosyl unit (2 from glucose) in contrast to
some 29 molecules for full aerobic oxidation (Brand, 2003). The
less the demand for ATP, the longer the glycogen reserves will
last. In addition, low metabolism slows the rate at which lactic
acid is produced. Two molecules of lactate are produced from
each glycosyl unit or glucose molecule along with two protons, if
the formation and hydrolysis of glycolytically derived ATP
J. Exp. Zool.
318
remain in balance (Hochachka and Mommsen, ’83). Under
anaerobic conditions, these protons impose a significant acid
load on the organism. Either exhaustion of glycogen reserves or
accumulation of acid could be limiting factors in surviving
anoxic submergence, and reduced metabolism will serve to slow
the development of each condition. In addition, as discussed
below, large glycogen reserves and effective acid buffering are
critical for extending anoxic survival in freshwater turtles.
Much of what we know about the responses of turtles to
prolonged submergence, including their metabolic rate, has been
learned from experiments in which the animals were artificially
denied access to the surface, i.e., using a forced-submergence
protocol. It is important to justify this approach and to critically
consider how well it simulates normal behavior. It is well known
(Belkin, ’64; Ackerman and White, ’79; Burggren and Shelton,
’79; Gatten, ’81) that at temperatures in the vicinity of 201C,
turtles in a free-diving situation emerge to breathe before blood
oxygen falls to critical levels. In other words, they do not
voluntarily become severely hypoxic in a laboratory setting.
Forced submergence at room temperature, therefore, imposes a
stress on the animals, and probably leads to an increased
metabolism and an acceleration of oxygen depletion and the
switch to anaerobic metabolism. Turtles typically struggle early
in the submergence period as they attempt to reach the surface to
breathe (e.g., Jackson, ’68). In a forced-diving experiment at
201C, turtles (C. picta) were anoxic (PaO2 o1 mmHg) after 1 hr
(Wasser et al., ’91). In contrast, in a study by Burggren and
Shelton (’79) on free-diving turtles (T. scripta) at 18–201C, a
turtle remained submerged voluntarily for about 3 hr with lung
and arterial PO2 falling to only about 20 mmHg. Results obtained
from forced dives at room temperature, thus, probably under-
estimated the turtle’s capacity for surviving long periods of
submergence.
This confounding response may be less serious at winter
temperatures (E31C). At low temperature, turtles are typically
quiescent and only rarely surface to breathe (Ultsch and Jackson,
’82). Covering the water surface to prevent breathing is not
obviously different from the situation in nature when a turtle
finds its access to the surface blocked by ice. Because the bulk of
the discussion in this review concerns experimental submer-
gences at winter temperatures, we believe that these data can
reasonably approximate what occurs in nature, a supposition
supported by field studies (Crocker et al., 2000a, b), at least for
hibernation in water with a high PO2. However, conditions in the
laboratory are carefully controlled and a fully anoxic state may
be imposed experimentally for many weeks, whereas it is not
certain whether or how often this occurs naturally. Laboratory
studies, therefore, reveal what turtles are capable of and not
necessarily what they commonly experience in their normal
environment.
Although whole-body anaerobic metabolic rate has been
measured at 241C in T. scripta by direct calorimetry (Jackson, ’68),
J. Exp. Zool.
JACKSON AND ULTSCH
rates at low temperature have only been estimated indirectly,
using the accumulation of lactate in the various body compart-
ments (Herbert and Jackson, ’85b; Jackson, ’97). The latter
method requires accurate knowledge of the size and lactate
concentrations of the various compartments. The initial calcula-
tions (Herbert and Jackson, ’85b) were made before it was
realized that a sizable fraction of the turtle’s postanoxic lactate
was situated in the shell and skeleton, but later estimates taking
this into account with more direct measurements of extracellular
volume gave similar values (Jackson, ’97, 2002). The exceedingly
low rate of heat production by the cold anoxic turtle makes direct
calorimetry under these conditions technically difficult if not
impossible, but whole-body homogenization (Gatten, ’81) can
provide a more direct and accurate measure of final whole-body
lactate. The rate of lactate accumulation in plasma and, thus,
presumably anaerobic metabolic rate itself, slows considerably
during the course of anoxic submergence in turtles; so, final total
lactate must take this into account in determining the minimal
rate.
Metabolic depression is now known to be a common response
of many organisms to a variety of stresses (Guppy and Withers,
’99). In the anoxic turtle as in other organisms, a coordinated
reduction in both ATP production and ATP utilization occurs that
minimizes the decrease in cellular ATP. Indeed, ATP levels in a
variety of tissues of turtles are maintained or even increased after
5 hr of anoxia at 181C in T. scripta (Kelly and Storey, ’88). On the
production side, decreased activity of key glycolytic enzymes
slows flux through the glycolytic pathway and prevents or
minimizes the Pasteur Effect (Storey, ’96). On the utilization side,
reduced protein synthesis and transmembrane ion pumping
secondary to reduced ion channel leak pathways contribute to
slowing the rate of ATP hydrolysis (Buck et al., ’93; Bickler and
Buck, 2007). One open question is how this downregulation is
initiated and controlled, although some oxygen or energy-
sensing mechanism has been postulated (Hochachka et al., ’96).
A related question is how these two processes, reduction of ATP
production and reduction of ATP use, are matched to maintain
the cell’s energy balance, although the biochemical mechanisms
involved are beginning to be explored (Staples and Buck, 2009).
Comparative studies of submergence in turtles have revealed
that considerable differences in anoxia tolerance occur among
species that naturally encounter near-freezing winter conditions.
The basis for these differences is probably complex, but variable
metabolic rate depression is likely a significant contributor. As
already noted, the rate of plasma lactate increase during anoxic
submergence can be used as an estimate of anaerobic metabolic
rate, and this variable was used to compare four species of North
American freshwater turtles after 11 days of anoxic submergence
at 31C. The more anoxia-tolerant species, the painted turtle,
C. picta bellii (Herbert and Jackson, ’85a; Jackson et al., 2000a), and
the snapping turtle, C. serpentina (Reese et al., 2003), accumulated
lactate at an average rate of 3.2 and 3.6 mmol L1 d1,
HIBERNATION UNDER THE ICE BY TURTLES AND FROGS 319

respectively, whereas among less anoxia-tolerant species, the map

turtle, G. geographica (Reese et al., 2001), the softshell turtle,
A. spinifera (Reese et al., 2003), and the musk turtle, S. odoratus
(Ultsch and Cochran, ’94), lactate increased by 4.7, 5.6, and
6.0 mmol L1 d1, respectively. These higher rates in the latter
species accelerate the development of lactic acidosis (Fig. 4) and the
depletion of glycogen stores, and can thereby shorten the time the
animal can remain anoxic.
As discussed earlier, the aerobic metabolic rate of frogs (e.g.,
R. temporaria), decreases during submergence at low temperature
in aerated (Donohoe et al., ’98) and hypoxic water (Boutilier et al.,
’97). Anurans subjected to anoxia at room temperature also
depress their energy metabolism by 75–80%, as measured by
direct calorimetry in the toad Bufo bufo (Leivestad, ’60) and in the
frog R. temporaria (Schulz et al., ’91), which is comparable to the
85% decrease observed using direct calorimetry in the turtle,
T. scripta (Jackson, ’68). Similarly, heat production of isolated
sartorius muscle of R. temporaria decreased by about 80%
when subjected to 1 hr of sustained anoxia following 2 hr of
progressive hypoxia at 201C (West and Boutilier, ’98). In frogs
exposed to anoxia at 201C, in contrast to what occurs in turtles,
brain ATP levels progressively fall (Wegener and Krause, ’93), a
condition described as ‘‘slow death’’ by Knickerbocker and Lutz
(2001). Frog muscle ATP, on the other hand, remains unchanged
during anoxia (Wegener and Krause, ’93).
At low temperature, anoxic metabolic rate can only be
estimated by the rate of accumulation of lactate in body tissues.
Figure 4. Time courses of changes in plasma [lactate] (top panel)
At 51C, blood and lymph lactate concentrations increased from
and blood pH (bottom panel) in Chrysemys picta bellii, Apalone
1 and 0.7 mmol L1 to 32.6 and 39.2 mmol L1, respectively, after
spinifera, and Rana pipiens submerged in anoxic water at 31C (see
4–6 days of anoxic submergence in R. pipiens (Christiansen and
legend of Fig. 3 for references). Standard error bars are shown
Penney, ’73). More recently, Stewart et al. (2004) observed plasma
although in some cases they are hidden within the symbol.
lactate levels averaging 38.5 mmol L1 and 37.3 mmol L1 after
4d or 3d anoxic submergence at 31C in R. pipiens and R.
catesbeiana, respectively. These latter rates of accumulation, 9.6 conservation is particularly important to ensure continued
and 12.4 mmol L1 d1, are 2–4 times higher than observed in delivery of glucose to the brain and heart.
various species of turtles as given above, suggesting significantly Brain tissue has limited glycogen reserves that may support
higher rates of anaerobic metabolism in the frogs. The higher some anaerobic metabolism (Partata and Marques, ’94), but
rates of lactate production contribute to a more rapid fall in blood becomes rapidly depleted during anoxia at 221C (Penney, ’74).
pH in frogs (Fig. 4). Heart muscle has substantial glycogen content in control
animals, but its depletion is typically faster than in liver and
Glycogen Stores and Utilization muscle (Daw et al., ’67; Penney, ’74; Wasser et al., ’91) and, like
Under anaerobic conditions, glycogen and free glucose are the the brain, the heart must then become more dependent on
metabolic substrates available to the turtle. Survival depends circulating glucose to support continued anaerobiosis. Anoxic
upon making supplies of these substrates last through the winter turtle heart studied in vitro preferentially consumed endogenous
months when hypoxic or anoxic conditions might occur. Turtles, glycogen but then could shift to exogenous glucose with no
compared with most other vertebrates, have relatively large change in function (Reeves, ’63).
reserves of glycogen primarily located in liver and skeletal Based on estimated metabolic depression, Hochachka (’82)
muscle (Daw et al., ’67). Skeletal muscle glycogen may only be calculated that sufficient liver glycogen alone was available to
accessible to muscle cells where it resides because lack of sustain anoxic turtles for up to 6 months at 31C, probably longer
glucose-6-phosphatase may prevent the production of exportable than any turtle would be subjected to anoxia in the wild. A recent
glucose from glucose-6-phosphate. Glucose derived from liver study (Jackson DC, unpublished), in which lactate levels
glycogen is available to all tissues of the body; however, its and glycogen contents were measured on control painted turtles

J. Exp. Zool.
320
Table 1. Glycogen concentrations in selected tissues of turtles
(Chrysemys picta) at 31C either air-breathing or after 3 months
submergence in severely hypoxic (PO2 o10 mmHg) water (Jackson
DC, unpublished data).
Air-breathing glycogen Anoxic glycogen
(mg gww1) (mg gww1)
Heart ventricle 68.773.7 17.271.3
Liver 141.174.7 95.5710.7
Pectoralis muscle 37.671.1 24.473.6
Retrahens capitis 25.471.6 14.272.0
muscle
(C. picta bellii) and on turtles after 3 months of submergence in
severely hypoxic water (PwO2 o10 mmHg) at 3oC, supported
Hochachka’s calculation (Table 1), although it is important to
note that considerable glycogen is also present in the large
muscle mass of the turtle. Less than half the available liver
glycogen was utilized by the anoxic turtles even though their
plasma (ECF) lactate concentrations had reached nearly 100 mmol
L1. The highest plasma lactates that have been reported,
E200 mmol L1 (Ultsch and Jackson, ’82), are explainable based
on these glycogen levels. The stoichiometry between lactate
accumulation and glycogen depletion was also found to be
reasonable in the recent study. Estimated total lactate production,
calculated from measured changes in glycogen concentrations
and estimated organ masses (Stecyk et al., 2004; Jackson DC,
unpublished measurements), was 58 mmol and total accumulated
lactate, based on measured values in plasma and shell using an
earlier published analysis (Jackson, ’97), was 66 mmol. The
somewhat lower value, based on glycogen depletion, may be
because the calculations assumed that glycogen was present only
in liver and skeletal muscle and neglected other possible sites of
glycogen storage. Exhaustion of glycogen reserves does not,
therefore, seem to be an important limiting factor in anoxic
survival for the painted turtle (C. picta bellii) for submergences up
to about 3 months when plasma lactate levels are 100–150 mmol
L1. A recent study on the red-eared slider (T. scripta), however,
suggests that depletion of glycogen reserves may be a critical
factor in anoxic survival of this species (Warren et al., 2006).
Normoxic glycogen levels in liver and skeletal muscle were
significantly less than published values for the painted turtle, and
stores after less than 6 weeks of anoxic submergence at 31C were
more than 80% exhausted. In two separate studies (Ultsch, ’85;
Warren et al., 2006), survival duration of T. scripta has been
found to be significantly less than in C. picta bellii, and smaller
glycogen reserves in Trachemys have been suggested as a
possible contributing factor to this difference. It is worth noting
that anoxic survival in the painted turtle could be limiting if on-
board glycogen reserves at the start of hibernation were low and
insufficient oxygen was available to utilize lipid reserves.
J. Exp. Zool.
JACKSON AND ULTSCH
Limited information is available on the importance of
glycogen reserves in anoxic survival at low temperature in
submerged frogs. However, the study by Christiansen and Penney
(’73) suggests that substrate availability may not be the limiting
factor in R. pipiens. Whole-body glycogen levels fell by only
25% after 4–6 d at 51C, indicating that adequate reserves were
still available. Like the turtle, however, cardiac glycogen was
nearly depleted during this time, although an elevated circulating
glucose level presumably was able to support cardiac function.
As discussed below, acidosis is more likely the crucial factor in
anoxic tolerance in cold, anoxic frogs.
Lactic Acid Buffering and Acid-Base Balance
Anaerobic metabolism of a glucose molecule or a glycosyl unit of
glycogen produces two molecules of lactate and two protons.
Acid-base homeostasis requires that the generated protons are
buffered so that body fluid pH does not fall to an incapacitating
or lethal level. Freshwater turtles, as revealed particularly from
studies of the painted turtle, C. picta bellii, have extraordinary
buffering capacity compared with other vertebrates, and this trait
is crucial for their capacity to survive long periods of anoxia and
anaerobic metabolism. The bulk of this animal’s buffering under
these conditions is contributed by its mineralized tissues, shell,
and skeleton, which together account for 35–40% of its fresh
body weight; however, its first line of defense against an acid
load, as in other organisms, is its extra- and intracellular fluid
buffering.
Lactic acid is produced within the cells of an anaerobic animal
and initial buffering occurs there. Intracellular buffering is not
exceptional in turtle tissues (Shi et al., ’97), but extracellular
buffering capacity is high compared with other vertebrates
(Smith, ’29). Plasma [HCO 3 ] in painted turtles is typically about
40–45 mmol L1, compared with 24 mmol L1 in humans and
other mammals (e.g., Herbert and Jackson, ’85a; Wasser and
Jackson, ’88). In addition, as also reported by Smith (’29), the
pericardial and peritoneal fluids of some species of freshwater
turtles have comparatively large volumes and [HCO 3 ] in the
range of 120 and 80 mmol L1, respectively. Subsequent
measurements confirmed these observations in T. scripta
(Jackson and Silverblatt, ’74) and in C. picta bellii (Jackson and
Heisler, ’84). Even these high levels of extracellular HCO-3,
however, are inadequate to cope with the acid load observed after
anoxic submergence lasting 10 days at 101C (Herbert and
Jackson, ’85a) or 3 months or more at 31C (Jackson and Heisler,
’82; Ultsch and Jackson, ’82), when lactate levels can range from
100 to 200 mmol L1. At these levels of lactate, supplemental
buffering from the shell and skeleton is necessary to keep pH in a
viable range.
Indirect evidence for supplemental buffering was provided in
a study by Robin et al. (’81), who observed a ‘‘cation’’ gap in
turtles (Pseudemys ( 5 Trachemys) scripta elegans) of about
24 meq L1 after 6 hr of forced diving at 241C. These authors
HIBERNATION UNDER THE ICE BY TURTLES AND FROGS
postulated that titration of negative sites on plasma proteins
likely accounted for the gap, but acknowledged that ‘‘theoreti-
cally, there might be an increase in various cationic species.’’
Soon thereafter, we reported that a far larger cation gap in the
painted turtle (C. picta bellii), submerged in anoxic water for up
to 5 months at 31C, could be largely explained by increases in
plasma [Ca21] and [Mg21] (Jackson and Ultsch, ’82). These
changes, plus significantly increased plasma [K1] and decreased
[Cl] and a small presumed decrease in protein anionic sites,
could fully account for ionic charge balance in the extracellular
fluid (Fig. 2). Subsequent work has shown that a similar pattern
of plasma ionic change occurs in C. picta bellii submerged at
different temperatures (Herbert and Jackson, ’85a) and in other
species of freshwater turtles at 3oC (Ultsch and Cochran, ’94;
Reese et al., 2001, 2003). Interestingly, little change in plasma
[Na1] was observed in these studies. These changes in plasma ion
concentrations represent severe disturbances of normal ionic
homeostasis, but can be considered as essential for preventing a
fatal metabolic acidosis by maintaining a positive strong ion
difference (Stewart, ’83; Herbert and Jackson, ’85a). The net
increase in strong cations was associated with the addition of
buffer anions that kept the decrease in plasma [HCO 3 ] much
smaller than the increase in plasma [lactate] (Fig. 2).
Both indirect and direct evidence point to the shell and other
skeletal structures as the source of the divalent cations, Ca21 and
Mg21. First, most of the body stores of these cations reside in the
mineralized tissues, making them the most likely source. In
addition, measurements of their concentrations in skeletal muscle
cells of anoxic turtles revealed modest increases in concentration
of both Ca21 and Mg21 (Jackson and Heisler, ’83), ruling out the
possibility that these cells, or presumably other tissue cells, could
be the source of their elevated ECF levels. Finally, direct
measurements of shell–mineral composition revealed that
[Mg21] was significantly reduced in turtles following 3 months
of anoxic submergence at 31C compared with control turtles
(Warburton and Jackson, ’95). Because Ca21 is 50 times more
abundant in bone than Mg21, but only increases in plasma by
about 3 times as much (Jackson and Ultsch, ’82; Jackson and
Heisler, ’82), statistically significant changes in bone Ca21 were
undetectable. Based on abundance in bone, therefore, Mg21 is
actually released preferentially to Ca21, an observation that may
relate to the site of origin within bone for the released cations.
Studies of mammalian bone have indicated that buffer is
released in response to acidosis primarily from the hydration
shell of bone crystal in the form of Ca(HCO3)2 (Bushinsky, 2001;
Arnett, 2003). If this is also the case in turtles and if their release
probability is the same, then the ratio of Ca21 to Mg21 in this
accessible site may be about 3:1. Present evidence suggests that
the apatite mineral of turtle bone is not broken down in response
to anoxic acidosis in turtles because phosphate levels have not
been observed to increase appreciably either in the plasma of
anoxic turtles (Jackson et al., 2000a) or in acidic solution bathing
321
in vitro incubated turtle bone powder (Jackson et al., ’99). The
bulk of the bone calcium is located in the apatite and may,
therefore, be inaccessible as a buffering source under these
conditions. Based on the mammalian data, this suggests that
bicarbonate may be the buffer anion that is released from turtle
bone. However, both in vitro and in vivo observations indicate
that carbonate is released from turtle bone. When bone powder
was incubated under pH-stat conditions, using 1 M HCl as the
titrant, CO2 was evolved from the incubation flask in approxi-
mately a 1:2 molar relationship to HCl, strongly suggesting that
carbonate was released (Jackson et al., ’99). In vivo comparisons
of molar decreases in bone and extracellular CO2 with total
lactate accumulated also are consistent with carbonate as the
ionic form released from bone (Warburton and Jackson, ’95;
Jackson et al., 2000a; Jackson DC, unpublished observations).
The presence of bicarbonate in mammalian bone is inferred from
comparing bone CO2 before and after drying. Labile bicarbonate
is lost and total bone CO2 is reduced (Poyart et al., ’75). Results
from our laboratory are contradictory in this regard; earlier work
(Warburton, ’92) found no loss of CO2 from bone after heat
drying, whereas more recent tests using a different method for
measuring CO2 did reveal a modest but consistent loss of CO2
(Taylor SE and Jackson DC, unpublished observations). Our
tentative conclusion is that accessible carbonate, associated with
both Ca21 and Mg21, is the effective buffer in turtle bone,
possibly located outside the crystalline structure of bone.
Additional study is required to verify this hypothesis, however.
As noted above, plasma [Na1] does not usually change
during prolonged anoxic submergence in turtles; however, both
shell and skeletal [Na1] decreased significantly more than 15%
after 3 months of anoxia in C. picta bellii (Jackson et al., 2000a).
It is not clear where this Na1 went, because a large increase in
ECF [Na1] that would be expected from the magnitude of the
bone Na1 loss was not observed. We hypothesize that Na1 is lost
to the surrounding water together with Cl-. Decreased plasma
[Cl] has been observed consistently in anoxic turtles (e.g.,
Jackson and Ultsch, ’82; Jackson et al., 2000a), but the fate of the
lost Cl- has never been determined. Coupled excretion of Na1
and Cl- could resolve both these uncertainties. Any loss of salt
from the body may not be associated with water loss because
body weight of anoxic turtles increases (Ultsch and Jackson, ’82),
presumably owing to uptake of water from the environment.
Direct measurements of possible Na1 and Cl- loss from anoxic
turtles are necessary to confirm or refute this hypothesis.
An uncertainty regarding buffer release from the turtle’s shell
is whether it is passive or active. In mammalian bone, buffer
release is thought to be mainly cell mediated, involving an
upregulation of osteoclast activity and downregulation of
osteoblast activity (Arnett, 2003). Evidence for osteoclast dissolu-
tion of bone mineral is the presence of so-called resorption pits on
the surface of the bone. In preliminary experiments, shell samples
from postanoxic turtles were examined but resorption pits were
J. Exp. Zool.
322 JACKSON AND ULTSCH

not found, consistent with the premise that buffer release is not
cell mediated (Arnett TR and Jackson DC, unpublished observa-
tions). In addition, the pattern of buffer release in an in vitro
preparation in which powdered bone was incubated at constant
and low pH was similar to the pattern observed in vivo during
experimental submergence anoxia (Jackson et al., ’99). These are
not conclusive experiments, however, and further study is
required to resolve the question of how buffer is released.
The shell and skeleton of the turtle contribute in another
important way to lactic acid buffering during anoxic submer-
gence. Lactic acid is taken up into bone, where it is buffered and
sequestered (Jackson et al., ’96, ’99; Davis and Jackson, 2007).
Figure 5. Summary of reactions and exchanges believed to be
After long anoxic exposures at 3oC, more than 40% of the total
involved in buffering lactic acid in the anoxic turtle, as discussed in
body lactate in C. picta bellii resides in the bone (Jackson, ’97).
the text. All lactic acid is produced in anaerobic cells, but only 20%
Similar uptake has been observed in mineralized tissues of other
remains and is buffered in that compartment. The balance moves
species (Jackson et al., 2001b, 2003; Warren and Jackson, 2005)
into the ECF where buffering is accomplished principally by
suggesting a generalized phenomenon. It is of greater quantita- 2
endogenous HCO 3 and by CO3 coming from bone and shell
tive significance in turtles because of their large mass of
accompanied by Ca (and, not shown, by Mg21 and Na1). About
21
mineralized tissue and the prolonged high levels of circulating
36% of the total lactate remains in the ECF. The balance of the
lactate that enable full equilibration with bone. The observed
lactic acid (44%) enters bone and shell where it is presumably
concentrations of lactate in turtle bone can reach levels that far
buffered by CO2 3 . The CO2 generated by this reaction diffuses into
exceed what can be explained by simple equilibration of the fluid
the ECF and from thence, together with CO2 generated by ECF
phase of bone with ECF lactate. This disequilibrium strongly
buffering, out of the animal into the surrounding water. A large
suggests that lactate exists in some combined form, perhaps
fraction of the ECF Ca21 complexes with lactate to form CaLa1.
complexed with bone calcium. It is known that a calcium lactate
The fate of the bone and shell lactate is uncertain, thus the
complex forms in the ECF of anoxic turtles (Jackson and Heisler,
question mark, but we hypothesize that much of it complexes with
’82); however, most of the ECF lactate is still in the ionized,
calcium. To simplify the diagram, some of the reactions are
dissolved state. Additional study is necessary to understand how
abbreviated and their stoichiometries are not exact. The large open
the bone of anoxic turtles can contain such high levels of lactate.
arrows depict exchanges between compartments and the small
The importance of this mechanism is clear, though. The
solid arrows depict chemical reactions or dissociations.
combined effect of the release of buffer into the ECF and the
buffering of lactic acid within shell and skeleton of anoxic turtles
accounts for as much as 75% of the total lactic acid buffering mineralized shell with only 25% as much Ca21 and buffering
(Jackson, ’97). The sequestration of lactate within bone relieves the capacity as the painted turtle shell (Jackson et al., 2000b). Its
other body fluids of a substantial fraction of the buffering that poorer bone buffering results in less buffer release and a greater
would otherwise be required. At the end of anoxia, when the turtle fall in blood pH for a given rise in blood lactate, compared with
once again becomes aerobic and ECF lactate begins to fall, lactate more anoxia-tolerant species (Fig. 4). Correlations in anoxia
is gradually released from bone as it continues to equilibrate with tolerance and bone buffering characteristics also exist among
the circulating blood. The lactate can then be utilized as an aerobic turtles with well-mineralized shells, although the relative shell
substrate or converted back into glycogen (Warren and Jackson, masses differ among these turtles as well (Iverson, ’84). Based on
2008). An additional uncertainty is how the lactate enters bone, comparisons of lactate uptake into shell disks, buffering by
whether by simple diffusion alone or via a transport mechanism. powdered shell samples, and shell total CO2 concentrations, the
Preliminary evidence indicates a small contribution by a lactate more anoxia-tolerant C. picta and C. serpentina exceed the less
carrier mechanism (Warren DE, unpublished observations), but anoxia-tolerant G. geographica and S. odoratus in all these
further investigations will be required to confirm this. The reactions indicators of shell buffering effectiveness (Jackson et al., 2007).
and exchanges that we hypothesize to occur in order to buffer the Field observations indicate that the less anoxia-tolerant species
lactic acid of an anoxic painted turtle are summarized in Figure 5. tend to select hibernation sites that have a dependably high PO2,
Differences in shell buffering also contribute to the variability such as streams or rivers, in contrast with the more anoxia-
in anoxia tolerance among turtle species described above. The tolerant species that can also be found in ponds, marshes, swamps,
most obvious example is the softshell turtle, A. spinifera, which is and other habitats that may become hypoxic (Ultsch, 2006).
the least anoxia-tolerant North American freshwater species As discussed above, cold anoxic frogs accumulate lactate in
studied to date (Reese et al., 2003) and which has a poorly their blood at a faster rate than do freshwater turtles, suggesting

J. Exp. Zool.
HIBERNATION UNDER THE ICE BY TURTLES AND FROGS
that metabolic depression is not as deep in the anurans. In
addition, and perhaps of even greater significance, is the poorer
whole-body buffering capacity of frogs. Extracellular [HCO 3 ] in
frogs is in the order of 25 mmol L1 (Stewart et al., 2004) in
comparison with values of 40 mmol L1 or higher in painted
turtles (e.g., Herbert and Jackson, ’85a). Frogs also have
considerably less mineralized tissue than turtles, although they
do possess subcutaneous and endolymphatic deposits of CaCO3
(Simkiss, ’68) that can contribute, as can their skeleton, to
buffering lactic acid during anoxic submergence (Warren and
Jackson, 2005). Nonetheless, acidosis develops very rapidly in
anoxic frogs compared with turtles, even the relatively anoxia-
intolerant softshell turtle (Fig. 4).
CONCLUDING COMMENTS
The physiological mechanisms of frogs and freshwater turtles
that permit them to survive weeks or months without breathing
in ice-covered bodies of water are now understood, at least in
broad terms. In normoxic water, both groups of animals can
satisfy all or most of their tissue O2 demands by uptake of O2
from the water. The high cutaneous gas conductance of frogs
supports aerobic metabolism even in hypoxic water, although
frogs are less able than turtles to survive in water when the PwO2
is too low to support aerobic metabolism. Metabolic depression in
both turtles and frogs in response to submergence and hypoxia/
anoxia is well documented, although the cellular signaling
pathways mediating these metabolic changes are poorly under-
stood.
Surviving under anoxic conditions is a more complex
response that involves a shift of the metabolic system from
aerobic to anaerobic production of ATP, with less efficient use of
substrate and the production of the strong acid end-product,
lactic acid. What limits anoxia tolerance is uncertain, but likely
factors include adequacy of glycogen reserves, effectiveness of
buffering of lactic acid, and ultimately maintenance of cellular
ATP levels. Frogs, in contrast with turtles, have less effective
buffering and develop severe metabolic acidosis far more rapidly
than do turtles. In addition, their metabolic rate is apparently
higher than turtles, accelerating the decline in function; frogs are
less able than turtles to sustain brain ATP levels. It is clearly
important for frogs to select hibernating sites that are reliably
oxygenated or that only become anoxic or severely hypoxic for
periods no longer than a few days.
Some species of freshwater turtles, in particular painted and
snapping turtles, can survive several months at 31C while anoxic,
although their normal body fluid homeostasis is severely
perturbed in the process. Some of these changes, such as elevated
ECF Ca21 and Mg21, are associated with buffer release from shell
and skeleton, which is a key mechanism for slowing the
development of metabolic acidosis. The contribution of the shell
and skeleton is central to the turtle’s anaerobic tolerance, but
important questions still remain unanswered. Is the release of
323
buffer in the form of bicarbonate or carbonate? Is there any
breakdown of the apatite crystal structure of bone? Is the release
of buffer a passive physicochemical event or is it cell mediated?
When lactate enters bone, what is its chemical state? Is it
combined with Ca21? Does the mechanical strength of the shell
and skeleton suffer as a result of the buffer release? The extreme
nature of the turtle shell’s contribution to acid-base regulation,
accounting for 75% of the total lactic acid buffering, may make it
an ideal model structure for studying the role of vertebrate bone
in acid buffering.
Because our understanding of the responses of turtles to
anoxic submergence is based almost entirely on simulated
experimental hibernation in the laboratory, questions remain as
to how accurately these laboratory studies describe the
physiological status of animals in the field. Laboratory studies
show us what these animals can do, but more field work is clearly
needed to test these conclusions under natural conditions. How
likely is it that a painted turtle will experience 3 months of
anoxic water in nature, and if it did, could it spontaneously
recover? It is hoped that with further developments in remote
sensing and in noninvasive measurement of physiological
variables, these and other important ecophysiological questions
will be addressed and answered.
LITERATURE CITED
Ackerman RA, White FN. 1979. Cyclic carbon dioxide exchange in the
turtle Pseudemys scripta. Physiol Zool 52:378–389.
Arnett T. 2003. Regulation of bone cell function by acid-base balance.
Proc Nutr Soc 62:511–520.
Belkin DA. 1963. Anoxia: tolerance in reptiles. Science 139:492–493.
Belkin DA. 1964. Variation in heart rate during voluntary diving in the
turtle Pseudemys concinna. Copeia 1964:321–330.
Belkin DA. 1968. Aquatic respiration and underwater survival of two
freshwater turtle species. Respir Physiol 4:1–14.
Bickler PE, Buck LT. 2007. Hypoxia tolerance in reptiles, amphibians,
and fishes: life with variable oxygen availability. Annu Rev Physiol
69:145–170.
Boutilier RG. 2001. Mechanisms of cell survival in hypoxia and
hypothermia. J Exp Biol 204:3171–3181.
Boutilier RG, St Pierre J. 2002. Adaptive plasticity of skeletal muscle
energetics in hibernating frogs: mitochondrial proton leak during
metabolic depression. J Exp Biol 205:2287–2296.
Boutilier R, Donohoe PH, Tattersall GJ, West TG. 1997. Hypometabolic
homeostasis in overwintering aquatic amphibians. J Exp Biol
200:387–400.
Bradford DF. 1983. Winterkill, oxygen relations and energy metabo-
lism of a submerged dormant amphibian, Rana muscosa. Ecology
64:1171–1183.
Bradford DF. 1984. Water and osmotic balance in over-
wintering tadpoles and frogs, Rana muscosa. Physiol Zool 57:
474–480.
J. Exp. Zool.
324
Brand M. 2003. Approximate yield of ATP from glucose, designed by
Donald Nicholson. Biochem Mol Biol Educ 31:2–4.
Brenner FJ. 1969. The role of temperature and fat deposition in
hibernation and reproduction in two species of frogs. Herpetologica
25:105–113.
Brenner FJ, Brenner PE. 1969. The influence of light and temperature
on body fat and reproductive conditions of Rana pipiens. Ohio J Sci
69:305–312.
Buck LT, Land SC, Hochachka PW. 1993. Anoxia-tolerant hepatocytes:
model system for study of reversible metabolic suppression. Am J
Physiol 265:R49–R56.
Burggren WW, Moalli R. 1984. ‘‘Active’’ regulation of cutaneous gas
exchange by capillary recruitment in amphibians: experimental
evidence and a revised model for skin respiration. Respir Physiol
55:379–392.
Burggren WW, Shelton G. 1979. Gas exchange and transport
during intermittent breathing in chelonian reptiles. J Exp Biol
82:75–92.
Bushinsky DA. 2001. Acid-base imbalance and the skeleton. Eur J Nutr
40:238–244.
Byrne JJ, White RJ. 1975. Cyclic changes in liver and muscle glycogen,
tissue lipid, and blood glucose in a naturally occurring population of
Rana catesbeiana. Comp Biochem Physiol A 50:709–715.
Christiansen J, Penney D. 1973. Anaerobic glycolysis and lactic acid
accumulation in cold submerged Rana pipiens. J Comp Physiol
87:237–245.
Congdon JD, Tinkle DW. 1982. Reproductive energetics of the painted
turtle (Chrysemys picta). Herpetologica 38:228–237.
Congdon JD, Dunham AE, Tinkle DW. 1982. Energy budgets and life
histories of reptiles. In: Gans C, Pough FH, editors. Biology of the
reptilia, Vol 13. New York: Academic Press. p 233–271.
Congdon JD, Dunham AE, van Loben Sels RD. 1994. Demographics of
common snapping turtle (Chelydra serpentina): implications for
conservation and management of long-lived organisms. Am Zool
34:397–408.
Congdon JD, Nagle RD, Kinney OM, Osentoski M, Avery HW, van
Loben Sels RC, Tinkle DW. 2000. Nesting ecology and embryo
mortality: implications for hatching success and demography of
Blanding’s turtles (Emydoidea blandingii). Chelonian Conserv Biol
3:569–579.
Crawford KM. 1994. Patterns of energy substrate utilization in
overwintering painted turtles, Chrysemys picta. Comp Biochem
Physiol A 109:495–502.
Crocker CE, Feldman RA, Ultsch GR, Jackson DC. 2000a. Overwintering
behavior and physiology of eastern painted turtles (Chrysemys picta
picta) in Rhode Island. Can J Zool 78:936–942.
Crocker CE, Graham TE, Ultsch GR, Jackson DC. 2000b. Physiology of
common map turtles (Graptemys geographica) hibernating in
the Lamoille River, Vermont. J Exp Zool (Mol Dev Evol)
286:143–148.
Czopek J. 1965. Quantitative studies on the morphology of respiratory
surfaces in amphibians. Acta Anat 62:296–323.
J. Exp. Zool.
JACKSON AND ULTSCH
Davis EC, Jackson DC. 2007. Lactate uptake by skeletal bone in anoxic
turtles, Trachemys scripta. Comp Biochem Physiol A 146:299–304.
Daw JC, Wenger DP, Berne RM. 1967. Relationship between cardiac
glycogen and tolerance to anoxia in the western painted turtle,
Chrysemys picta bellii. Comp Biochem Physiol 22:69–73.
Donohoe PH, Boutilier RG. 1998. The protective effects of metabolic
rate depression in hypoxic cold submerged frogs. Respir Physiol
111:325–336.
Donohoe PH, West TG, Boutilier RG. 1998. Respiratory, metabolic, and
acid-base correlates of aerobic metabolic rate reduction in
overwintering frogs. Am J Physiol 274:R704–R710.
Donohoe PH, West TG, Boutilier RG. 2000. Factors affecting
membrane permeability and ionic homeostasis in the cold-
submerged frog. J Exp Biol 203:405–414.
Dunson WA. 1960. Aquatic respiration in Trionyx spinifer asper.
Herpetologica 16:277–283.
Elmberg J. 1990. Long-term survival, length of breeding season, and
operational sex ratios in a boreal population of common frogs,
Rana temporaria L. J Zool 68:121–127.
Emerson DN. 1967. Preliminary study on seasonal liver lipids and
glycogen, and blood sugar levels in the turtle Graptemys
pseudogeographica (Gray) from South Dakota. Herpetologica
23:68–70.
Feder ME, Burggren WW. 1985. Cutaneous gas exchange in vertebrates:
design, patterns, control and implications. Biol Rev 60:1–45.
Galbraith DA, Brooks RJ. 1987. Survivorship of adult females in a
northern population of common snapping turtles, Chelydra
serpentina. Can J Zool 65:1581–1586.
Gatten RE. 1981. Anaerobic metabolism in freely diving painted
turtles (Chrysemys picta). J Exp Zool (Mol Dev Evol) 216:377–385.
Gatz RN, Crawford EC, Piiper J. 1975. Kinetics of inert gas
equilibration in an exclusively skin-breathing salamander, Desmog-
nathus fuscus. Respir Physiol 24:15–29.
Gordos MA, Franklin CE. 2002. Diving behavior of two bimodally
respiring turtles, Rheodytes leukops and Emydura macquarii, in a
natural setting. J Zool 258:335–342.
Graham TE, Butler BO. 1993. Metabolic rates of wintering Blanding’s
turtles, Emydoidea blandingii. Comp Biochem Physiol A 106:
663–665.
Graham TE, Forsberg JE. 1991. Aquatic oxygen uptake by naturally
hibernating wood turtles Clemmys insculpta. Copeia 1991:
836–838.
Graham TE, Graham AA. 1992. Metabolism and behavior of wintering
common map turtles, Graptemys geographica, in Vermont. Can
Field Nat 106:517–519.
Graham TE, Guimond RW. 1995. Aquatic oxygen consumption by
wintering red-bellied turtles. J Herpetol 29:471–474.
Guppy M, Withers PC. 1999. Metabolic depression in animals:
physiological perspectives and biochemical generalizations. Biol
Rev 74:1–40.
Herbert CV, Jackson DC. 1985a. Temperature effects on the responses
to prolonged submergence in the turtle Chrysemys picta bellii. I.
HIBERNATION UNDER THE ICE BY TURTLES AND FROGS
Blood acid-base and ionic changes during and following anoxic
submergence. Physiol Zool 58:655–669.
Herbert CV, Jackson DC. 1985b. Temperature effects on the responses
to prolonged submergence in the turtle Chrysemys picta bellii. II.
Metabolic rate, blood acid-base and ionic changes, and cardiovas-
cular function in aerated and anoxic water. Physiol Zool 58:
670–681.
Hochachka PW. 1982. Anaerobic metabolism: living without oxygen.
In: Taylor CR, Johansen K, Bolis L, editors. A companion to
animal physiology. Cambridge, UK: Cambridge University Press.
p 138–150.
Hochachka PW, Mommsen TP. 1983. Protons and anaerobiosis.
Science 219:1391–1397.
Hochachka PW, Somero GN. 2002. Biochemical adaptation: mechan-
isms and process in physiological evolution. Oxford: Oxford
University Press
Hochachka PW, Buck LT, Doll CJ, Land SC. 1996. Unifying theory of
hypoxic tolerance: molecular/metabolic defense and rescue
mechanisms for surviving oxygen lack. Proc Natl Acad Sci USA
93:9493–9498.
Hulse AC, McCoy CJ, Censky EJ. 2001. Amphibians and reptiles of
Pennsylvania and the Northeast. Ithaca, New York: Cornell
University Press.
Hutchison VH, Whitford WG, Kohl M. 1968. Relation of body size
and surface area to gas exchange in anurans. Physiol Zool 41:
65–85.
Iverson JB. 1984. Proportional skeletal mass in turtles. Florida Sci 47:
1–11.
Jackson DC. 1968. Metabolic depression and oxygen depletion in the
diving turtle. J Appl Physiol 24: 503–509.
Jackson DC. 1997. Lactate accumulation in the shell of the turtle,
Chrysemys picta bellii, during anoxia at 3 and 101C. J Exp Biol
200:2295–2300.
Jackson DC. 2000. Living without oxygen: lessons from the freshwater
turtle. Comp Biochem Physiol A. 125:299–315.
Jackson DC. 2002. Hibernating without oxygen: physiological
adaptations of the painted turtle. J Physiol 543:731–737.
Jackson DC, Heisler N. 1982. Plasma ion balance of submerged anoxic
turtles at 3oC: the role of calcium lactate formation. Respir Physiol
49:159–174.
Jackson DC, Heisler N. 1983. Intracellular and extracellular acid-base
and electrolyte status of submerged anoxic turtles at 31C. Respir
Physiol 53: 187–201.
Jackson DC, Heisler N. 1984. The contribution of the alkaline
pericardial fluid of freshwater turtles to acid buffering during
prolonged anoxia. J Exp Biol 109:55–62.
Jackson DC, Silverblatt H. 1974. Respiration and acid-base status of
turtles following experimental dives. Am J Physiol 226:903–909.
Jackson DC, Ultsch GR. 1982. Long-term submergence at 31C of the
turtle, Chrysemys picta bellii, in normoxic and severely hypoxic
water. II. extracellular ionic responses to extreme lactic acidosis.
J Exp Biol 96:29–43.
325
Jackson DC, Toney VI, Okamoto S. 1996. Lactate distribution and
metabolism during and after anoxia in the turtle, Chrysemys picta
bellii. Am J Physiol 271:R409–R416.
Jackson DC, Goldberger Z, Visuri S, Armstrong RN. 1999. Ionic
exchanges of turtle shell in vitro and their relevance to shell
function in the anoxic turtle. J Exp Biol 202:513–520.
Jackson DC, Crocker CE, Ultsch GR. 2000a. Bone and shell
contribution to lactic acid buffering of submerged turtles
Chrysemys picta bellii at 31C. Am J Physiol 278:R1564–R1571.
Jackson DC, Ramsey AL, Paulson JM, Crocker CE, Ultsch GR. 2000b.
Lactic acid buffering by bone and shell in anoxic softshell and
painted turtles. Physiol Biochem Zool 73:290–297.
Jackson DC, Crocker CE, Ultsch GR. 2001a. Mechanisms of home-
ostasis during long term diving and anoxia in turtles. Zoology-Anal
Complex Sy 103:150–156.
Jackson DC, Wang T, Koldkjaer P, Taylor EW. 2001b. Lactate
sequestration in the carapace of the crayfish Austropotamobius
pallipes during exposure in air. J Exp Biol 204:941–946.
Jackson DC, Andrade DV, Abe AS. 2003. Lactate sequestration by
osteoderms of the broad-nose caiman, Caiman latirostris, following
capture and forced submergence. J Exp Biol 206:3601–3606.
Jackson DC, Rauer EM, Feldman RA, Reese SA. 2004. Avenues of
extrapulmonary oxygen uptake in western painted turtles (Chrys-
emys picta bellii) at 101C. Comp Biochem Physiol A 139:221–227.
Jackson DC, Taylor SE, Asare VS, Villarnovo D, Gall JM, Reese SA. 2007.
Comparative shell buffering properties correlate with anoxia
tolerance in freshwater turtles. Am J Physiol 292:R1008–R1015.
Jiang S, Claussen DL. 1992. A bioenergetic budget for overwintering
newts (Notophthalmus viridescens) from southern Ohio: their fat
reserves and aerobic metabolic rates in water. Comp Biochem
Physiol A 101:743–750.
Kelly DA, Storey KB. 1988. Organ-specific control of glycolysis in
anoxic turtles. Am J Physiol 255:R774–R779.
King P, Heatwole H. 1994. Partitioning of aquatic oxygen uptake
among different respiratory surfaces in a freely diving pleurodiran
turtle, Elseya latisternum. Copeia 1994:802–806.
Knickerbocker DL, Lutz PL. 2001. Slow ATP loss and the defense of ion
homeostasis in the anoxic frog brain. J Exp Biol 204:3547–3551.
Koskela P, Pasanen S. 1974. The wintering of the common frog, Rana
temporaria L. in northern Finland. Aquilo Ser Zool 15:1–7.
Lannoo M, Gallant AL, Nanjappa P, Blackburn L, Hendricks R. 2005.
Species accounts. In: Lannoo M, editor. Amphibian declines: the
conservation status of United States species. Berkeley, CA:
University of California Press. p. 349–914.
Leivestad H. 1960. The effect of prolonged submersion on the
metabolism and the heart rate in the toad (Bufo bufo). Univ Bergen
Arbok Naturvitenskop Rekke 5:1–15.
Licht LE. 1991. Habitat selection of Rana pipiens and Rana sylvatica
during exposure to warm and cold temperatures. Am Midl Nat
125:259–268.
Lutz PL, Rosenthal M, Sick TJ. 1985. Living without oxygen: turtle
brain as a model of anaerobic metabolism. Mol Physiol 8:411–425.
J. Exp. Zool.
326
Manion JJ, Cory L. 1952. Winter kill of Rana pipiens in shallow ponds.
Herpetologica 8:32.
Martof B. 1956. Growth and development of the green frog, Rana
clamitans, under natural conditions. Am Midl Nat 55:101–117.
Merrell DJ. 1977. Life history of the leopard frog, Rana pipiens, in
Minnesota. Occas Pap Bell Mus Nat Hist Univ Minne 15:1–23.
Milton SL. 2008. The physiology and anatomy of anoxia tolerance in
the freshwater turtle brain. In: Wyneken J, Godfrey MW, Bels V,
editors. Biology of turtles. Boca Raton FL: CRC Press. p 301–344.
Partata WA, Marques M. 1994. Effects of fasting and seasonal
variations in brain glycogen disposition in the turtle (Chrysemys
dorbigni). Comp Biochem Physiol A 107:727–730.
Pasanen S, Koskela P. 1974. Seasonal and age variation in the
metabolism of the common frog, Rana temporaria L., in Northern
Finland. Comp Biochem Physiol A 47:635–654.
Penney DG. 1974. Effects of prolonged diving anoxia on the
turtle Pseudemys scripta elegans. Comp Biochem Physiol A 47:
933–941.
Peterson CC, Greenfields D. 2001. Negative test for cloacal drinking in
a semi-aquatic turtle (Trachemys scripta) with comments on
the functions of cloacal bursae. J Exp Zool (Mol Dev Evol) 290:
247–254.
Pinder AW. 1987. Cutaneous diffusing capacity increases during
hypoxia in cold submerged bullfrogs (Rana catesbeiana). Respir
Physiol 70:85–95.
Pinder AW, Storey KB, Ultsch GR. 1992. Estivation and hibernation. In:
Feder ME, Burggren WW, editors. Environmental physiology of the
amphibians. Chicago: The University of Chicago Press. p 250–274.
Poyart CF, Bursaux E, Fréminet A. 1975. The bone CO2 compartment:
evidence for a bicarbonate pool. Respir Physiol 25:89–99.
Reese SA, Crocker CE, Carwile ME, Jackson DC, Ultsch GR. 2001. The
physiology of hibernation in common map turtles (Graptemys
geographica). Comp Biochem Physiol A 130:331–340.
Reese SA, Jackson DC, Ultsch GR. 2002. The physiology of
overwintering in a turtle that occupies multiple habitats, the
common snapping turtle (Chelydra serpentina). Physiol Biochem
Zool 75:432–438.
Reese SA, Jackson DC, Ultsch GR. 2003. Hibernation in freshwater
turtles: softshell turtles (Apalone spinifera) are the most intolerant
of anoxia among northern North American species. J Comp Physiol B
173:263–268.
Reese SA, Stewart ER, Crocker CE, Jackson DC, Ultsch GR. 2004.
Geographic variation of the physiological response to overwintering
in the painted turtle (Chrysemys picta). Physiol Biochem Zool
77:619–630.
Reeves RB. 1963. Control of glycogen utilization and glucose uptake
in the anaerobic turtle heart. Am J Physiol 205:23–29.
Robin ED, Vester JW, Murdaugh HV, Millen JE. 1964. Prolonged
anaerobiosis in a vertebrate: anaerobic metabolism in the fresh-
water turtle. J Cell Comp Physiol 63:287–297.
Robin ED, Robin DA, Ackerman R, Lewiston N, Hance AJ, Caligiuri M,
Theodore J. 1981. Prolonged diving and recovery in the freshwater
J. Exp. Zool.
JACKSON AND ULTSCH
turtle, Pseudemys scripta—I. lung and blood gases, pH, lactate
concentrations and ‘‘cation’’ gap. Comp Biochem Physiol
70A:359–364.
Schulz C, Thuy M, Wegener G. 1991. Heat production of frogs under
normoxic and hypoxic conditions: a microcalorimetric study using a
gas flow system. Thermochim Acta 187:71–78.
Shi H, Hamm PH, Meyers RS, Lawler RG, Jackson DC. 1997.
Mechanisms of pHi recovery from NH4Cl-induced acidosis in anoxic
isolated turtle heart: a 31P-NMR study. Am J Physiol 272:
R6–R15.
Simkiss K. 1968. Calcium and carbonate metabolism in the frog
(Rana temporaria) during respiratory acidosis. Am J Physiol 214:
627–634.
Smith CL. 1950. Seasonal changes in blood sugar, fat body, liver
glycogen, and gonads in the common frog, Rana temporaria. J Exp
Biol 26:412–429.
Smith HW. 1929. The inorganic composition of the body fluids of the
Chelonia. J Biol Chem 82:651–661.
Staples JF, Buck LT. 2009. Matching cellular metabolic supply and
demand in energy-stressed animals. Comp Biochem Physiol A
153:95–105.
Stecyk JAW, Overgaard J, Farrell AP, Wang T. 2004. a-adrenergic
regulation of systemic peripheral resistance and blood flow
distribution in the turtle (Trachemys scripta). J Exp Biol 207:
269–283.
Stewart ER, Reese SA, Ultsch GR. 2004. The physiology of hibernation
in Canadian leopard frogs (Rana pipiens) and bullfrogs (Rana
catesbeiana). Physiol Biochem Zool 77:65–73.
Stewart PA. 1983. Modern quantitative acid-base chemistry. Can J
Physiol Pharmacol 61:1442–1461.
Storey, KB. 1996. Metabolic adaptations supporting anoxia
tolerance in reptiles: recent advances. Comp Physiol Biochem B
113:23–35.
Tattersall GJ, Boutilier RG. 1997. Balancing hypoxia and hypothermia
in cold-submerged frogs. J Exp Biol 200:1031–1038.
Tattersall GJ, Boutilier RG. 1999a. Behavioural oxy-regulation by cold-
submerged frogs in heterogeneous oxygen environments. J Exp Biol
200:1031–1038.
Tattersall GJ, Boutilier RG. 1999b. Constant set points for pH and
PCO2 in cold-submerged skin-breathing frogs. Respir Physiol
118:49–59.
Tattersall GJ, Ultsch GR. 2008. Physiological ecology of aquatic
overwintering in ranid frogs. Biol Rev 83:119–140.
Ultsch GR. 1985. The viability of nearctic freshwater turtles
submerged in anoxia and normoxia at 3 and 101C. Comp Biochem
Physiol A 81:607–611.
Ultsch GR. 1988. Blood gases, hematocrit, plasma ion concentrations,
and acid-base status of musk turtles (Sternotherus odoratus) during
simulated hibernation. Physiol Zool 61:78–94.
Ultsch GR. 2006. The ecology of overwintering among turtles:
where turtles overwinter and its consequences. Biol Rev 81:
339–367.
HIBERNATION UNDER THE ICE BY TURTLES AND FROGS
Ultsch GR, Cochran BM. 1994. Physiology of northern and southern
musk turtles (Sternotherus odoratus) during simulated hibernation.
Physiol Zool 67:263–281.
Ultsch GR, Jackson DC. 1982. Long-term submergence at 31C of
the turtle, Chrysemys picta bellii, in normoxic and severely hypoxic
water. I. survival, gas exchange and acid-base status. J Exp Biol
96:11–28.
Ultsch GR, Reese SA, Stewart ER. 2004. Physiology of hibernation in
Rana pipiens: Metabolic rate, critical oxygen tension, and the
effects of hypoxia on several plasma variables. J Exp Zool (Mol Dev
Evol) 301A:169–176.
Wang ZX, Sun NZ, Sheng WF. 1989. Aquatic respiration in
softshelled turtles, Trionyx sinensis. Comp Biochem Physiol A
92:593–598.
Warburton SJ. 1992. Calcium metabolism in anoxic poikilotherms:
painted turtles (Chrysemys picta bellii) and bullfrogs (Rana
catesbeiana). Ph.D. dissertation, Brown University.
Warburton SJ, Jackson DC. 1995. Turtle (Chrysemys picta bellii) shell
mineral content is altered by exposure to prolonged anoxia. Physiol
Zool 68:783–798.
Warren DE, Jackson DC. 2005. The role of mineralized tissue in the
buffering of lactic acid during anoxia and exercise in the leopard
frog, Rana pipiens. J Exp Biol 208:1117–1124.
327
Warren DE, Jackson DC. 2008. Lactate metabolism in anoxic turtles:
an integrative review. J Comp Physiol B 178:133–148.
Warren DE, Reese SA, Jackson DC. 2006. The factors that limit
survival of red-eared slider turtles, Trachemys scripta, during long-
term anoxic submergence at 31C. Physiol Biochem Zool 79:
736–744.
Wasser JS, Jackson DC. 1988. Acid-base balance and the control of
respiration during anoxic and anoxic-hypercapnic gas breathing in
turtles. Respir Physiol 71:213–226.
Wasser JS, Warburton SJ, Jackson DC. 1991. Extracellular and
intracellular acid-base effects of submergence anoxia and nitrogen
breathing in turtles. Respir Physiol 83:239–252.
Wegener G, Krause U. 1993. Environmental and exercise anaerobiosis
in frogs. In: Hochachka PW, Lutz PL, Sick T, Rosenthal M, van den
Thillart G, editors. Surviving hypoxia: mechanisms of control and
adaptation. Boca Raton, FL: CRC Press. p 217–236.
West TG, Boutilier RG. 1998. Metabolic suppression in anoxic frog
muscle. J Comp Physiol B 168:273–280.
Wilbur HM. 1975. Evolutionary and mathematical demography of
turtle Chrysemys picta. Ecology 56:64–77.
Willis YL, Moyle DL, Baskett TS. 1956. Emergence, breeding,
hibernation, movements and transformation of the bullfrog, Rana
catesbeiana, in Missouri. Copeia 1956:30–41.
J. Exp. Zool.