Respiration Physiology (1976) 27, 115-129; North-Holland Publishing Company, Amsterdam

AQUATIC LIFEI AT HIGH ALTITUDE: RESPIRATORY ADAPTATIONS IN

THE LAKE TITICACA FROG, TELMATOBIUS CULEUS

VICTOR H. HUTCHISON, HOWARD B. HAINES and

GUSTAV ENGBRETSON
Department of Zoology, University of Oklahoma, Norman, Okla. 73069, U.S.A

Abstract. Tebnatobius culeus has a combination of behavioral, morphological and physiological adapta-
tions which allows an aquatic life in cool (10 C) O,-saturated (at 100 mm Hg) waters at high altitude
(3812 m). The skin surface area is increased by pronounced folds and the cutaneous capillaries penetrate
to the outer layers of the skin. The erythrocyte volume (394 r3) is the smallest reported for amphibians.
The P,, (15.6 at pH 7.65 and 10 C) is the lowest, and the erythrocyte count (729 . IO/mm) the highest
for an anuran. The 0, capacity (11.7 vol%), hemoglobin (8.1 g/100 ml), hemoglobin concentration
(0.281 pg/r) and hematocrit (27.9 %) measured at 18 C dnd 3800 m are all elevated in comparison with
most amphibians. The 0, dissociation curve is sigmoid (n = 2), the Bohr factor is small @log P,,/dpH =
- 0.30) and the buffering capacity ( - 8.9 mM HCO; . 1- . dpH - ) is typical for an aquatic amphibian.
The metabolic rate (14.1 ~1 . g-r . h-t) is th e Iowest reported for a frog and among amphibians only the
giant salamanders (Amphiuma, Necturus and Siren) have lower values. If prevented from surfacing in
hypoxic waters, the frogs ventilate the skin by bobbing behavior; if allowed to surface, they will ventilate
the small lungs and the metabolic rate increases to 23 pl *g-r . h-r.

Adaptation to high altitude Hemoglobin

Amphibians life P
Bohr effect RZi cell count
Buffering capacity Skin respiration

Due to their adjustment to high altitudes and cold waters the aquatic frogs (Telmato-
bius) of the high Andean lakes have interested biologists for many years (Allen,
1922; Parker, 1940; Schmidt, 1954; Vellard, 1952, 1954, 1956; Macedo, 1960;
Diaz y Gomez, 1965). The most studied species is T. cdeus of Lake Titicaca which

Accepted for publication 19 March 1976.

1 Our work in Bolivia was made possible through the aid of Srs. Juan Hatjes E. and Wagner Terrazas U.
Field work in Bolivia was supported by a grant from the National Geographic Society; laboratory work
in the United States was supported in part by the National Science Foundation (Grant No. GB 36701 to
V.H.H.) and by the University of Oklahoma Research Council (to H.B.H.).

115
116 V. H. HUTCHISON et al.

reaches a snout-vent length of 140 mm and a weight of over 250 g. The lungs of this
species are greatly reduced (to less than one-third that of a ranid frog of equivalent
size), flattened and poorly vascularized with only primary alveoli and often a narrow
caudal end (Macedo, 1960). The sluggish, bottom-dwelling animals have never been
seen to surface for air, yet upon dissection the lungs contain small amounts of gas
(Allen, 1922). The loose, fibrous, and glandular skin, which serves as a gill, is
expanded by numerous large folds which hang from the dorsum, sides and hind legs.
These folds are highly vascularized by subepidermal plexi with capillaries which
penetrate the outermost layers of the stratum corneum. The buccal cavity is also
richly vascularized (Macedo, 1960).
Lake Titicaca is at an elevation of 3812 m with a maximum depth of 281 m. The
surface waters average 10 C with less than a 4 C annual range; deeper waters are a
nearly constant 10 C. The water is saturated with 0, (at about 100 mm Hg Po,)
due to strong wind and wave action. The frogs reportedly occur from near the surface
to the maximum depth and are often taken in fishing nets pulled along the lake
bottom. We observed them on rock ledges within two feet of the lake surface.
We have examined the respiratory properties of the blood and the aerial-aquatic
gas exchange to determine some of the physiological adaptations that allow an
aquatic existence at such high altitude in this unique salientian.

Materials and methods

ANIMALS

Adult frogs were collected by surface diving to a depth of 2 m at Copacabana and

the Tequina Straits, Bolivia in December 1973 and January 1974. The animals were
maintained in wet cloth bags and, plastic aquaria in LaPaz (elevation, 3759 m)
and exposed to the natural photoperiod (approximately LD 12 : 12) and 16-20 C.
Frogs not used for studies in LaPaz were shipped by air to the University of Okla-
homa (335 m) where they were maintained in plastic aquaria with aerated water.
The frogs were fed live earthworms, neonatal mice and aquatic insects. Aquaria
were housed in temperature controlled chambers at 10 + 1.O C and a photoperiod
of LD 12 : 12. Frogs were acclimatized at least three weeks in Oklahoma before
analyses were started.

HEMATOCRIT, HEMOGLOBIN, ERYTHROCYTE COUNTS

Erythrocytes were counted in LaPaz with hemacytometers; in Oklahoma erythrocyte

counts and size distributions were determined with a Coulter Model B particle
counter and Model J plotter. Hematocrits were measured with standard hematocrit
tubes and centrifuge and hemoglobin by the cyanomethoglobin method.
 RESPIRATION IN LAKE TITICACA FROG 117

OXYGEN CAPACITY, pH, Po,, PcoZ, DISOCIATION CURVES, BUFFERING CAPACITY

The 0, and CO, contents of blood were determined in LaPaz with Natelson micro-
gasometers. The mixture of 95 % 0, and 5 % CO, was repeatedly flushed into and
out of heparinized syringes containing the blood until an equilibrium was reached.
Dissolved 0, was subtracted according to the solubility coefficients of Sendroy
et al. (1934). In LaPaz blood pH was determined by the insertion of a semimicro
combination electrode (Bradley-James 9100) directly into a freshly incised ventricle
with no exposure to air; in 10 animals pH, Pco2 and PoZ were measured with an
Instrumentation Laboratory Model 127 analyzer at 18 C; pH and Pco2 were cor-
rected to 10 C according to Severinghaus (1965).
In Oklahoma the 0, dissociation curves were determined by measuring the 0,
content of whole blood at various PO2 tensions. The procedure was as follows:
(1) Blood (2-3 ml) was drawn into a 5 ml syringe and deoxygenated by repeated
introduction of a N&O, gas mixture. The syringe was shaken for five minutes to
facilitate equilibration. (2) After deoxygenation, approximately 1 ml of gas was
expelled from the syringe :and replaced with 1 ml of a gas mixture containing 21 %
0, and the same content of CO, as the deoxygenation mixture. Again blood and gas
equilibration was facilitated by shaking the syringe. (3) The 0, content of 20 ~1
of the equilibrated blood was measured coulometrically with a Lex-O&on oxygen
analyzer (Lexington Instrument Co.). The sensing element in the instrument is a
fuel cell which generates current from the combination with oxygen. The instrument
resolves 0.1 ml 0, per 100 ml sample in samples of gas or blood and is unaffected by
PH, P coZ, phosphates and other characteristics of the sample. (4) The Po, of the
equilibrated blood was measured with a polarographic electrode (Radiometer
E5046 with gas monitor PHA 927). The electrode was standarized with distilled
water equilibrated with room air at 10 C and an O,-free solution of 0.01 M borax
containing 20 mg/ml Na,SO,. (5) Steps 2,3 and 4 were repeated in several steps until
the blood was saturated with 0,. Blood pH was altered with gas mixtures of different
CO, contents. The oxygen-free gas mixtures were prepared in gas burettes from the
combination of instrument grade commercial mixtures of 9.43 y, CO,, balance N,
or 5.44 % CO,, balance N,. Lower CO, values were obtained by adding pure N,
to these mixtures. The mixtures which contained 0, were prepared similarly from
the addition of air to commercial mixtures of 20.9 % O,, 9.62 y0 CO,, balance N,
or 20.9 o/o0,, 5.25 y0 CO,, balance N,. Blood pH was measured with a micro glass
electrode (Radiometer G297/G2). The blood-gas equilibrations and Po, determina-
tions were made at 10 C. Twenty dissociation curves were measured with Pco,
values which ranged from almost 0 (air) to 69 mm Hg, with an average of 23 mm Hg.
Two curves were determined at 20 C. This method was evaluated by construction
of dissociation curves of normal human blood and comparison of the results to the
standard human curve (Altman and Dittmer, 1971). Nine such tests resulted in a
PsO of 26.0+ 1 mm Hg (37 C, pH 7.4, Pco2 40 mm Hg, base excess 0) compared to
the standard value of 26.6 mm Hg.
118 V. H. HUTCH&ON et ai.

Carbon dioxide dissociation curves were determined as follows : (1) Blood samples
were divided into two equal portions; one aliquot was equilibrated with a gas mixture
of 20.9 % O,, 5.44 % CO,, balance N,. (2) Blood CO, content was determined with
a Natelson microgasometer and PcoI, with a Severinghaus electrode (Radiometer,
E5036). The two gas mixtures were then mixed 1 : 1 with either air or 100 % N, and
equilibrated with blood samples. (3) Another 1 : 1 dilution with air or N, was made,
the blood equilibrated and CO, content and PcoZ determined. (4) This serial dilution
was repeated and the blood finally equilibrated with air or 100 % N, and values of
CO, content and Pco2 again recorded. The curves were determined at 27 C and
the values corrected to 10 C according to Severinghaus (1965).
Buffer capacity was computed with the Henderson-Hasselbalch equations from
the pH values of oxygenated blood in equilibrium with gases of known PcoZ. Ap-
propriate values for the solubility of CO, in plasma and pK, were from Severinghaus
(1965).

AQUATIC-AERIAL OXYGEN EXCHANGE

Cutaneous and pulmonary 0, uptake was measured at 10 C in aerial-aquatic

respirometers (Guimond and Hutchison, 1972, fig. 1) for periods of 8 to 15 hours.
The water below the mineral oil layer, which separated the aerial from the aquatic
portions of the respirometers, was at saturation with 0, at the beginning of each
experiment. The 0, was not replaced during the measurements. Experiments were
terminated when the water P,, fell below 30 mm Hg and sufficient measurements
had been made to calculate aerial V,,.
Observations on behavior of frogs under respiratory stress due to low dissolved
PO, were made in the respirometers or in aquaria in which dissolved 0, had been
removed with N, or by the addition of small amounts of Na,SO,. No differences in
behavior were noted when the two methods of 0, removal were used, except for an
increased mucous flow from animals in the water treated with Na,SO,. Animals
treated with Na,SO, were used for behavioral observation and not for studies on
blood or gas exchange.
Unless otherwise noted all values are given as the mean plus or minus the standard
error of the mean (SE).

RtSUltS
HEMATOLOGY

The mean cell volume of the nucleated erythrocytes was 394 pm3 (table 1, fig. 1).
The erythrocyte count observed in LaPaz was 729 x 103/mm3, while that in frogs
examined after acclimatization of l-6 months at lower altitude in Oklahoma had
decreased to 557 x 103,/mm3 (0.10 > P > 0.05). Hemoglobin concentration, 0,
capacity and hematocrit were also lower but not significantly (table l), in the animals
 RESPIRATION IN LAKE TITlCACA FROG 119

CELL VOLUME (CL))

Fig. 1. Erythrocyte size distribution for Telmutobiur The distribution is skewed such that the most
culeus.
cells are between 300 and 400 pm and the average cell is between 400 and 500 pm.

maintained at the lower elevation. Mean hemoglobin (and mean cell hemoglobin)
concentration increased slightly in animals acclimatized in Oklahoma (table 1).

BLOOD RESPIRATORY PROPERTIES

The pH of ventricular blood from 20 small animals (R = 42.9 f 1.6 g) measured

with the semimicro electrode at 18 C was 7.529+0.002; the blood pH of 10 large
animals (X = 161.8 f 12.4 g) determined with the IL 127 analyzer was 7.565 f0.063
(table 1) and was not significantly different from that of the small animals. The mean
of all observations is pH 7.541 and at 10 C, the average water temperature in Lake
Titicaca, this pH value wouldbe 7.63.
The blood of 20 small frogs in Bolivia had an oxygen capacity of 1.44+0.16 ml/g
hemoglobin, which was not significantly different from the 1.17 + 0.26 ml/g of 12
larger animals acclimatized in Oklahoma; these values indicate that virtually all
the hemoglobin was functional. Ventricular blood of 21 small frogs contained 7.5 ml
0, and 38 ml CO,/lOOml. The PO, (13.8k1.5 tot-r) and Pco, (11.3k2.6) of blood
from large frogs was measured in samples drawn anaerobically from the ventricles
(table 1). The 0, content of samples from the large frogs averaged 4.7 ml/100 ml,
notably less than in the small frogs, but is probably accurate because this 0, content
and the in uiuoPo, align reasonably well with the oxygen dissociation data.
The oxygen dissociation curve for whole blood is slightly sigmoidal with an average
coefficient of interaction (n) of 2.0 LO.15 (table 1, fig. 2). Although there was no
 TABLE 1
Snmmary of blood characteristics of Tehafobiw c&us at elevations of 3749 m (La Paz, Bolivia) and
335 m (Norman, Okla., USA)

Bolivia (18 C) Oklahoma (10 C)

N Z&SE N t&SE t P

Body weight (g) 48* -42.9 & 1.6 13 167.3 kl4.1 - -

lo** 161.8 f12.4 - - -

Hematocrit (/.) 47 27.9 1.1 8 23.6 f 2.2 1.51 ns

Hemo~obin {g/l00 ml) 47* 8.1 & 0.4 12 7.3 & 0.5 0.87 ns

0, capacity/hemoglobin
(ml/g) 201 1.44 k 0.16 12 1.17 + 0.26 1.27 ns

0, capacity (ml/100 ml) 11.7 - 8.5 - -

0, content, ventricular
blood, (ml/l~ ml) 21* 7.5 f 0.6 - - -

CO, content, ventricular

blood, (ml/ 100 ml) 21* 38 *1.4 -

Erythrocytes (103/mm3) 48* 129 rt32 8 551 +49 1.81 0.1 > Pp 0.05

Mean cell volume &m3) 45* 394 f18 8 423 +25 0.65 ns

Mean cell hemo~obin

(pgi=tl) 46* 110 f5 8 137 t-12 1.96 0.1 > P > 0.05

Mean cell hemoglobin

concentration (pg/$m) 44* 0.281 k 0.009 8 0.320& 0.016 1.68 0.1 < P<O.O5

@ventricular blood 20+ 7.5295 0.002 -

10** 7.565i 0.063 - -
corrected to IO C 7.63 -

PO,, (torr), ventricular

blood 10** 13.8 * 1.5 -

pCOr, ventricular blood

(mm Hg) IO** 11.3 & 2.6 -
corrected to 10 C 8.0 -

Hill coefficient - - 20 2.00 + 0.15 -

P,,, at pH 7.63 (mm Hg) - - 20 15.6 + 0.4*** -

Bohr factor
@lo8 Ps,/dpH) - - 20 -0.30 f 0.09, - -

Buffering capacity
(mM HCO; . l- +(pH
unit)- - - 14 - 8.9 -

* Observations made on smaller frogs Q wt = 42.9 g).

** Observations made on targer frogs (i wt = 161.8 g).
*+* Standard error of the estimate.
+ Standard error of the slope.
++ Computed from the slope of the least squares tit of log pC0, versus pH (log P,, = 10.44- 1.18 pH;
r = -0.95) for Pco, of 13-69 mmHg and pH of 7.29-7.80. CO, solubility coeffkicni was 0.0629 mM
CO, . I- * - (mmHg)- (Severinghaus, 1965).
 RESPIRATION IN LAKE TITICACA FROG 121

PO, (mm Hg)

Fig . 2. Representative oxygen dissociation curves for whole blood.

significant difference between n measured at pH extremes, a trend appeared such

that n between pH 7.30 and 7.63 averaged 2:3, and between pH 7.7 and 8.2 averaged
1.3. The Bohr factor was -0.30+0.09, dlog.P,,/dpH (fig. 3). The P,, at pH 7.63
was 15.6 mm with a lower 95 % confidence limit of 13.4 mm. Two dissociation curves
determined at 20 C had a temperature coeffkient of 0.02 dlog P,,/dT. If this co-
efficient is used to correct for the temperature effect, we calculate a P,, of approx-
imately 25 mm Hg at a blood pH of 7.54 for the frdgs studied in UPaz at 18%
(table 1).
In the physiological ranges of less than 6 mm Hg Pco, the total blood CO, content

Telmotobius

I 0%

n :- 0
7-
6- 1 Bohr Wctor = -0.30
5 1 I
7.2 7.4 76 7.0 8.0 8.2
QH
Fig. 3. The Bohr effect in the Lake Titicaca frog. The line is a least square tit for the data points shown.
122 V. H. HUTCHISON et cd.

is approximately 35 ml/100 ml..The Haldane effect was not detected on the three
pairs (oxygenated and unoxygenated) of curves constructed for fig. 4. The variable
hemoglobin content of the blood resulted in appreciable variability in the CO,
dissociation curves (fig. 4). Although these curves were determined in Oklahoma,
they agree well with the CO, data obtained in Bolivia; using the Henderson-Hassel-
balch equation with the Pco, observed in Bolivia at 18 C (table 1) with the appro-
priate solubility factor and pK, (Severinghaus, 1965),we estimated a total of 32-39
ml CO~/l~ml blood at pH 7.54 The observed value was 38~l.Oml/l~ml
(table 1). The observed Pco, at 18 C (table 1) corrects to 8 mm at 10 C (Severing-
haus, 1965) and this partial pressure corresponds to 33-37 ml CO,/100 mf on the
dissociation curve (fig. 4).
Buffering capacity was estimated from the pH of oxygenated blood which had been
equilibrated with gases of known Pea,. The slope of the line which relates log Pco2 to
pH converts to a buffer capacity of -8.9 mM HCO; * l- 1 (unit pH)- .

AQUATIC-AERIAL OXYGEN EXCHANGE

As in their natural habitat frogs in the laboratory do not surface to breath air when
the water is farily well oxygenated. However, when Po, in the res~romete~ is
decreased to 35-89 mm Hg (jz = 61.6 If:5.9, table 2) the frogs move to the surface,
place the nostrils, which open upward as in Xenopus (Macedo, 1960), into the air
and ventilate the small lungs. They maintain this position as long as the dissolved
Po, remains below normal. The metabolic rate during air-breathing is about 39 %
greater than the normal rate in well oxygenated water where all of the respiratory
gas exchange is across cutaneous surface (table 2). Results obtained from one frog

4V
r

Teimatcbios
w
IQC

M ZOg lib / 100 ml blood

I I I I * I, I I, I L ,I,
5 IO 15
Pco, (mm tig)

Fig. 4. Three CO, dissociation curves for the Lake Titicaca frog. The Pa1 was measured at 27 C and
corrected to 10 C (Severinghaus, 1965). The CO, concentration was measured at 10 C.
 BESPIBATION IN LAKE TITICACA PBOG 123

200 , , , ( , , , , , , , I 1 I 1 I

T~mtobh a!&!?
- IO@%, 1116

160 - 160

- 140

I20 3

- 100 I
E
E
-60 0

- 60

- 40

- 20

L L L I
2 4 6 6 IO I2 I4
HOURS

Fig. 5. Aerial aquatic oxygen uptake by a Lake Titicaca frog in a respirometer. The slope shown in the
regression equations is equal to 3o, in PI g- 1 h- . l PO, of water shown on right ordinate; l total 0,;
0 cutaneous 0,; 0, pulmonary 0,.

TABLE 2
Summary of data on aerial-aquatic oxygen consumption in Telmatobiw c&us at IO C

N f SE SD Range

Body weight (g) 22 122.3 5.4 26.1 7g204

PO, of water at start of air breathing

(mm Hg) 22 61.6 5.9 27.0 35-89

Po, of water at point cutaneous exchange

starts to decline (mm Hg) 21 37.6 3.5 15.8 14-75

Po, of water at end of experiments (mm Hg) 21 9.3 2.2 9.7 O-28

vo, cutaneaus only @I g- . h- ) 21 14.1 1.3 5.9 5.3-26.7

voz after start of air breathing @lg- .h-) 21 23.1 2.3 10.5 10.M.6
124 v. H. HUTCHISON et al.

are shown in fig. 5; in this experiment air-breathing began when the Po, of the water
dropped to 80 mm Hg. The slopes of the regression equations in fig. 5 are equal to the
vo2 in pl .g- .h -I. In most cases the cutaneous uptake remained constant for
several hours before declining after the initiation of lung ventilation. In the example
(fig. 5) the rate of cu~neous 0, uptake did not decline until the PO, of the water
had decreased below 25 ; the mean for all experiments was 37.6 f 3.5 mm Hg (table 2).
In all cases the total ?oi after the initiation of lung ventilation did not decline, even
when the water PO, fell to zero. Thus, the p~mona~ gas exchange alone is adequate
to maintain the metabolic rate for at least several hours.
If the frogs are prevented from reaching tne water surface by a wire screen when
the lowered Po, would otherwise stimulate them to surface for air, they do not struggle
to reach the surface, but stand on the bottom with legs and toes extended with a
maximum skin surface exposed to the water. About once every six seconds they bob
vigorously by quickly pushing up with their hind legs; this action usually lifts the
entire body into the water. As they slowly sink back towards the bottom the large
skin Raps were lifted passively by the force of the water and appear to wave back
and forth, thus breaking up the boundary layer of water and ventilating the skin
surfaces. If the animal is kept submerged and the dissolved 0, not replaced, the Po,
continues to decrease to near zero before the frogs will die. Thus, they have an
efficient mechanism .for utilization of 0, in low concentration and an apparent
ability to sustain a high degree of hypoxia for several hours.

The oxygen transport properties of the blood show several distinct adaptations for
an aquatic life at high altitude. The frogs studied in LaPaz had erythrocyte counts
(0.729 x 106/mm3) greater than that reported for any frog and among all amphibians
is exceeded (1 .I36 to 1.948) only in the tiger salamander, Ambysfoma regrind,
at elevations ofabout 2100 m in the Rocky Mountains (Roofe, 1961). Theerythrocyte
volume is the smallest (394 m3) known among amphibians; B. taitams has ery-
throcytes of 588 pm3 while that of ranid frogs ranges from 659 to 768 pm3 (Gahien-
beck amd Bartels, 1968 ; Rouf, 1969 ; Wintrobe, 1933 ; a summary table of amphibian
blood characteristics is provided in Wood et al., 1975). The hematocrit (4) of 27.9
is within the average range of most anurans and is similar to that of Rana catesbeiana
(23.5),, R. esculenta (27.3), and R.pipiens (24.6); slightly greater than most plethodon-
tid ~lamanders (21.9-25.7; Reynolds and Pickard, 1973); less than Bu~paracnemis
(34; Johansen and Ditadi, 1966), Heurodeies Walt/ii (38 %; Deparis et al., 1975),
Cryptobranchus (40.1 to 43.3, Jerrett and Mays, 1973), B. taitanus (40; Wood et al.,
1975), and high altitude populations of Ambystoma tigrin~ (48 ; Roofe, 1961).
The oxygen capacity (ml/100 ml) of 11.7 is fairly high among amphibians; it is
exceeded only by the 14.8 of the giant toads B&o marine (Hall, 1966) and B. parac-
nem& (Johansen and Ditadi, 1966) and the apodan B. taitanus (Wood et al., 1975).
 RFBPIRATION IN L.AKF! TITICACA FROG 125

The h~o~ob~ content (g/100 ml) falls within the upper range of arnp~~~~ ;
the highest recorded values (10.0-12.9) are from the fossorial spadefoot toads
Scaphiopus (Seymour, 1973) and the high altitude (3200 m) Bufo boccourti of Central
America (Stuart, 1951); the lowest (4.4) is from the salamander Dicampfodon ensatus
(Wood, 1971) and the gilled salamander Necturus macdosus (4.5, Lenfant and Jo-
hansen, 1967). The.mean cell hemo~obin ~n~ntmtion ~/~rn3) of 0.281 is in the
upper range of those previously observed in amphibians and is exceeded only by
R. esculenta (0.289; Gahlenbeck and Bartels, 1968).
After acclimatization at a lower altitude in Oklahoma the erythrocyte counts,
hemoglobin concentration and hematocrit decreased (table 1). Although the declines
were not statistically significant, we observed a steady decline over the lo-week
period of acclimati~tion in Oklahoma which indicates that these cha~cte~sti~
are subject to environmental influence. Oxygen capacity per gram of hemoglobin
was not different from the theoretical value of 1.34 ml/g in frogs examined in LaPaz
(1.44) or in Oklahoma (1.17), nevertheless the decrease observed in Oklahoma
parallels the alterations in erythrocyte counts, hemoglobin concentration, and
hematoc~t.
This animal has the lowest P,, of any frog at comparable temperature (10 C)
and pH (7.65) and among amphibians is comparable only to the aquatic salamander
Necturus maculosus with a P,, of 14.5 mm Hg at 20-22 C and a P&o, of 4.4 mm Hg
(Lenfant and Johansen, 1967).
We detected no particular adaptations of the blood divergent from those of other
amphibians for CO, transport or pH regulation. The quantity of CO, combined at
physiological partial pressures (table 1) is slightly greater than reported for other
amphibians, but this is due to the greater solubility of CO, at 10 C versus that at
20-22 C (Lenfant and Johansen, 1967). The low buffering capacity (8.9 mM
HCO; . 1-l I (pH unit)- in Tei~tobi~ is similar to that of other zaquatic am-
phibians including the gill and skin breathing Necturus (8.0) and axolotl(9.0) and the
lung and skin gas exchanging Amphiuma (9.0) (Erasmus er af., 1970/71 ;.Lenfant and
Johansen, 1967). A low buffering capacity is correlated with aquatic gas exchange
and the associated low level of Pa,,, and with lower metabolic rates and the associated
lower production rate of acid metaboli& (Lenfant and Johansen, 1967).
The 39 % increase in qo, during aerial ventilation over than in aquatic respiration
may be due to the : (1) added cost of breathing, especially when the small lung volume
is considered; (2) removal of an 0, debt accrued as the Po, of the water decreased
prior to the initiation of lung ventilation; (3) elevation of the metabolic rate due to an
increase in the availability of 0, ; (4) combination of all three factors. The metabolic
rate in amphibians may be determined in part by the availa~ity of 0, ; the lungless
plethodontid salamanders have lower metabolic rates than lunged forms of equi-
valent body size at the same temperatures (Whitford and Hutchison, 1967). Siren
lacertina, tin aquatic salamander with small gills and efficient well-developed lungs
(Guimondand Hutchison, 1974), has a higher metabolic rate when breathing bi-
modally than when submerged and exchanging gases solely in the aquatic mode
126 v. H. l3LWHBON et d.

(Ultsch, 1974,1976). Thus, the availability of O2 and the 0, transport capacity may
partially govern the \fo, rather than all such regulation being controlled at the cellular
level.
The behavior of the frogs when kept submerged in water at low Po, is remini~nt
of that of the aquatic hellbender Cryptobranchus alleganiensis, which periodically
rocks or sways from side to side to break up the boundary layer between the water
and skin to ventilate the cutaneous surfaces. The skin of this salamander is similar
to that in Telmutobius: richly vascularized with the capiharies penetrating into the
epide~is and with an increased surface area to volume ratio provided by a dorso-
ventrally flattened body and reticulated folds and flaps of skin (Guimond and
Hutchison, 1973).
The critical 0, tension (P,) in Telmatobius is apparently low compared to other
aquatic amphibians. Although we did not measure the P, directly, metabolic rate
does not decline during aquatic exchange (prior to start of air breathing), even when
the water Po, drops as low as 35 mm Hg. (fig. 5, table 2). The P, in S. lacer&a
varied between 70 and 170 mm Hg for a size range of 0.36-1310 g (Ultsch, 1973).
In the sirenid Pseudobranchusstriatusthe P, was 80 mm Hg and in the newt,,Diemic-
tylus uiridescens, 130 mm Hg (Ultsch, 1976).A low P, would be necessary for main-
tenance of metabolic rate in Lake Titicaca where the PO, at saturation would only be
approximately 100 mm Hg.
The metabolic rate of the Titicaca frog measured at low altitude is less than that
of any frog previously measured (table 1, x = 14.1 ~1 0, *g- *h- at 10 C).
Among all amphibians only the giant salamanders (Amphiuma, Cryptobrunchusand
Necturus) have lower metabolic rates. Even when breathing air, the metabolic rate
(K = 23.1 ~1 * g- ah-) is less than that reported for other anurans (Hutchi~n,
1971). At the lower Po, of the natural habitat one might expect that the metabolic
rate would be even lower than the values we obtained at lower altitude.
The continued fall in PO2after the start of air breathing, although at a slower rate,
shows that cutaneous 0, uptake continues even when the Po, reaches very low values
(fig. 5). The cutaneous uptake did not begin to decline until the Po, dropped to a
mean of 37.6 mm Hg, although air breathing began at a mean of 61.6 mm Hg
(table 2). These results cannot be due solely to changes in ventilation and perfusion
rates and suggest that the conductances (Piiper and Scheid, 1975) for 0, between
the respiratory medium and the blood are not equal in both directions at low en-
vironmental Po,.
The ability of Telmatobi~ to remove 0, efficiently from the water could result in
part from circulatory changes. A functional separation between the oxygen-rich
and oxygen-depleted blood in their passage through the univentricular heart occurs
in Bufo paracnemis (Johansen and Ditadi, 1966) and Rana pipiens (DeLong, 1962).
Oxygen tensions in systemic (80 mm Hg) and pulmocutaneous (60 mm Hg) vessels
provide evidence for a separation of blood flows in the ventricle of Xenopus laevis
when ventilating air; during a dive the tensions fell in all measured parts of the
circulatory system (Emilio and Shelton, 1974). A marked increase in pulmonary
 REsPIRATIOti IN LAKE TITICACA FROG 127

blood flow and decrease in pulmocutaneous pressure as a part of the normal diving
emergence behavior also occurs in X. laeuis(Emilio and Shelton, 1972).
In more terrestrial frogs such as Runaridibundadiving is accompanied by a marked
fall of P,, in the arterial blood a few minutes after submergence, showing that cu-
taneous gas exchange is inadequate to compensate for the lack of pulmonary ventila-
tion (Emilio, 1974). In the highly aquatic X. lueuisthe rate of 0, uptake from the
lungs was high at the surface, but fell rapidly during the first few minutes of a dive;
both lungs and blood serve as limited 0, stores and the contribution of cutaneous
gas exchange to aquatic respiration is fairly minor (Emilio and Shelton, 1972, 1974).
Xenopus is tropical, lives in often poorly oxygenated waters and is thus quite dif-
ferent from Telmatobius,which is unique in both its habitat and combination of
respiratory adaptations.
The Lake Titicaca frog is well adapted for existence as an aquatic animal at high
elevation where the Po, is only 100 mm Hg at saturation. The specilized skin serves
as a gill and allows the animal to obtain 0, to meet its metabolic requirements, the
lowest recorded for anurans and one of the lowest among all amphibians. The very
small erythrocytes and high hematocrit, erythrocyte count, and oxygen affinity of
the blood enhance the uptake of 0,. The waters of the lake serve as an infinite
pool (Piiper and Scheid, 1975) from which 0, is removed and an infinite sink into
which CO, is released. Behavioral adaptations allow an increased ventilation of the
cutaneous surfaces and some pulmonary gas exchange should the normal respiratory
medium become hypoxic. No other amphibian is known to possess this combination
of adaptations.

Acknowledgements

We are greatly appreciate the assistance of Donald Howard in both field and la-
boratory. The following persons and organizations contributed greatly to our efforts :
Mildred Fung and Mary E. Kanak; Drs. Alfonso Criales Q., Enrico Linares, Jean
Couderts, Jorge Rioja Rota, Hugo Villegas, Lydia Ruiz, Luis Hartmann ; Srs. Mario
Diaz M., Juan Nogales 0.; Director General de Agricultura Gover Barja B., Vice-
almirante Xavier Pinto T., Lt. Kelly Hughes, U.S.N.; Universidad Mayor de San
Andres, Ministerio de Agricultura y Ganaderia, Instituto Boliviano de Biologia de
Altura, Fuerza Naval Boliviana..

References

Allen, W. R. (1922). Notes on the Andean frog Tehatobius c&us (Garman). Copeia 1922: 52-54.
Altman, P. L. and D. S. Dittmer (editors) (1971). Respiration and &ulation, Biological Handbooks.
Bethesda, Md., Fed. Am. Sot. Exp. Biol., 930 p.
DeLong, K. T. (1962). Quantitative analysis of blood circulation through the frog heart. Science 138:
69S694.
128 v. H. HUTCHISON f?t Ul.

Deparis, P., J. C. Beetschen and A. Jaylet (1975). Red blood cells and hemoglobin concentration in normal
diploid and several types of polypoid salamanders. Camp. B&hem. P&o/. 50A: 263-266.
Diaz y Gomez, E. (1975). La respiration en un anuro delaltiplano andino (Tebnatobius culi escomeli
Angel). Bol. R. Sot. Esp. Hist. Nat. Sect. Biol. 63: 265-269.
Emilio, M. G. and G. Shelton (1972). Factors affecting blood flow to the lungs in the amphibian. Xenopus
laevis. .I. Exp. Biol. 56: 67-17.
Emilio, M. G. (1974). Gas exchange and blood gas concentrations in the frog Rana ridibunda. J. Exp.
Biol. 60: 901-908.
Emilio, M. G. and G. Shelton (1974). Gas exchange and its effect on blood acid concentrations in the
amphibian Xenopus laevis. J. Exp. Biol. 60: 567-579.
Erasmus, B., B. J. Howell and H. Rahn (1970/71). Ontogeny of Acid-base balance in the bullfrog and
chicken. Respir. Physiol. 11: 4653.
Gahlenbeck, H. and H. Bartels (1968). Temperaturadaptation der Sauerstoffafftnit8t des Blutes von
Rana esculenta. Z. vergl. Physiol. 59: 232-240.
Guimond, R. W. and V. H. Hutchison (1972). Pulmonary, branchial and cutaneous gas exchange in the
mudpuppy, Necturus maculosus maculosus (Rafinesque). Comp. Biochem. Physiol. 42A: 367-392.
Guimond, R. W. and V. H. Hutchison (1973). Aquatic respiration: An unusual strategy in the hellbender
Cryptobranchus alleganiensis alleganiensis (Daudin). Science 182: 12631265.
Guimond, R. W. and V. H. Hutchison (1974). Aerial and aquatic respiration in the Congo eel Amphiuma
means means (Garden). Respir. Physiol. 20: 147-159.
Hall, F. G. (1966). Hemoglobin functions in the blood of Bufo marinus. J. Cell. Comp. Physiol. 68: 67-74.
Hutchison, V. H. (1971). Oxygen Consumption. Part IV. Amphibians. In: Respiration and Circulation,
Biological Handbooks, edited by P. L. Altman and D. S. Dittmer. Bethesda, M& Fed. Am. Sot.
Exp. Biol., pp. 481485.
Jerrett, D. P. and C. E. Mays (1973). Comparative hematology of the hellbender Cryptobranchus allega-
niensis in Missouri. Copeia 1973: 331;337.
Johansen, K. and S. S. F. Ditadi (1966). Double circulation in the giant toad Bufo paracnemis. Physiol.
Zool. 39: 140-150.
Lenfant, C. and K. Johansen (1967). Respiratory adaptations in selected amphibians. Respir. Physiol.
2: 247-260.
Macedo, H. de (1960). Vergleichende Untersuchungen an Arten der Gattung Telmatobius (Amphibia,
Anura). Z. wiss. Zoo/. 163: 355-396.
Parker, H. W. (1940). The Percy Sladen Trust Expedition to Lake Titicaca in 1937. XII. Amphibia.
Trans. Linn. Sot. London., Ser. 3, 1: 203-216.
Piiper, J. and P. Scheid (1975). Gas transporf of gills, lungs and skin: theory and application. ,Respir.
Physiol. 23 : 209221.
Reynolds, A. E. and B. L. Pickard (1973). Normal blood values in some plethodontid salamanders.
Herpetologica 29 : 184-188.
Roofe, P. G. (1961). Blood constituents ofAmblystoma tigrinum. Anat. Rec. 140: 337-340.
Rouf, M. A. (1969). Hematology of the leopard frog Ranapipiens. Copeia 1969: 682687.
Schmidt, K. P. (1954). Notes on frogs of the genus .Telmatobius with descriptions of two new Peruvian
species. Fieldiana 34: 277-287.
Sendroy, J., Jr., R. Dillon and D. Van Slyke (1934). Studies of gas and electrolyte equilibria in blood.
XIX. The solubility and physical state of uncombined oxygen in blood. J. Biol. Chem. 105: 597632.
Severinghaus J. W. (1965). Bloodgas concentrations. In: Handbook ofPhysiology, Section 3: Respiration
Vol. VI., edited by W. 0. Fenn and H. Rahn. Washington, D. C., American Physiological Society,
pp. 1475-1487.
Seymour, R. S. (1973). Gas exchange in spadefoot toads beneath the ground. Copeia 1973: 4524.
Stuart, L. C. (1951). The distributional implications of temperature tolerances and hemoglobin values in
the toads Bufo marks (Linnaeus) and Bufo bocourti Brocchi. Copeia 1951: 22&229.
 RESPIRATION IN LAKE TITICACA FROG 129

Ultsch, G. R. (1973). A theoretical and experimental investigation of the relationships between metabolic
rate, body size, and oxygen exchange capacity. Respir. Physiof. 18: 143-160.
Ultsch, G. R. (1974). Gas exchange and metabolism in the Sirenidae (Amphibia: Caudata). I. Oxygen
consumption of submerged sirenids as a function of body size and respiratory surface area. Comp.
Biochem. Physiol. 47A: 485-498.
Ultsch, G. R. (1976). Respiratory surface area as a factor ~ntroiling the metabolic rate of aquatic sala-
manders. Respir. Physiol. 26: (In press).
Vellard, J. (1952). Adaptation des batraciens a la vie a grande hauteur darts les Andes. Bull. Sot. Zoof.
Franc. 77: 169-187.
Vellard, J. (1954). Etudes sur le Lac Titicaca. V. Les ~e~to~i~ du ham-plateau Interandin. 2%~. Zn,rt.
Franc. Etudes Andines 4 : l-57.
Vellard, J. (1956). Repartition des batraciens dans les Andes au sud de lbquateur. Trav. Inst. Franc.
Etudes Andines 5: 141-162.
Whitford, W. G. and V. H. Hutchison (1967). Body size and metabolic rate in salamanders. Physiof. Zooi.
40: 127-133.
Wintrobe, M. M. (1933). Variations in the sire and hemoglobin content of erythrocytes in the blood of
various vertebrates. Folia Haematol. Leipzig 51: 32-36.
Wood, S. C. (1971). Effects of metamorphosis on blood respiratory properties and erythrocyte adenosine
t~phosphate level of tbe salamander ~ic~pto~n ensatus (~h~hol~) Respir. Phys~I. 12: 53-65.
Wood, S. C., R. E. Weber, G. M. L. Maloiy and K. Johansen (1975). Oxygen uptake and blood respiratory
properties of the caecilian Boulengerula taitanus. Respir. Physiol. 24: 355-363.