1. A study has been made of kidney function in the crab-eating frog, Rana cancrivora, of south-east Asia.

  2. This frog can live in full-strength sea water; in such concentrated media its blood is slightly hypertonic to the medium, and a considerable part of the osmotic concentration is due to urea.

  3. In concentrated media the excretion of urea is greatly diminished. This is not due to active tubular reabsorption of urea, but primarily to a low urine flow caused by increased tubular reabsorption of water and reduced glomerular filtration.

  4. In concentrated media, as compared with dilute media, only a few percent of the filtered urea appears in the urine.

  5. Osmoregulation of the crab-eating frog in sea water resembles that of elasmobranchs except in that there is no evidence of active tubular reabsorption of urea in the frog.

Until recently it has been generally held that no amphibians could survive in media with a salinity much in excess of 1 %, and in particular that no amphibian could adjust to a life in sea water. This belief was based on attempts at acclimatizing various amphibians to increasing salt concentrations, attempts that on the whole were unsuccessful. On the other hand, it has been reported that some amphibians, mostly frogs and toads, occur in brackish or sea water in various parts of the world. Numerous such reports have been compiled and reviewed by Schmidt (1957) and by Neill (1958).

More recently, one reportedly euryhaline frog, the crab-eating frog of south-east Asia (Rana cancrivora) has been the subject of physiological investigations (Gordon, Schmidt-Nielsen & Kelly, 1961a, b). This frog is found in mangrove swamps along the coast of the Gulf of Thailand, where it swims and seeks its food in full-strength sea water. In the laboratory these frogs survived continued immersion in 4/5 of fullstrength sea water. At high salinities the frogs are slightly hypertonic to the medium. Part of the high osmotic concentration in their body fluids is due to a salt concentration about twice the level found in ordinary freshwater amphibians, but in addition the concentration of urea is exceptionally high (about 0·3 M). Thus the main osmoregulatory devices of the crab-eating frog seem similar to those of elasmobranchs.

Since urea has a major function in osmoregulation, it is of interest to study the kidney function of the frogs in order to establish how this substance is withheld from excretion and conserved as a physiologically desirable agent. In ordinary freshwater frogs urea is excreted by active tubular secretion (Marshall, 1932; Walker & Hudson, 1937; Forster, 1954). In elasmobranchs, on the other hand, urea is retained because of its osmotic function, and the excretion in the urine is minimized by active tubular reabsorption (Smith, 1936). Is the renal function of the crab-eating frog similar to that of sharks, or to their closer relatives, the freshwater frogs? This paper describes experiments designed to answer this question. We were able to obtain a small number of frogs from Thailand and studied the adjustment of blood concentrations to the external medium, the relative roles of urea and sodium in osmoregulation, and the renal function in various dilutions of sea water.

The animals used in these experiments were crab-eating frogs (R. cancrivora) which had been caught in Thailand and shipped by air to the United States. The mortality in transit was high, apparently because the container had been placed in an unheated freight compartment, and only about 25 % of the frogs survived the transport. On arrival the survivors were promptly placed in small aquaria with sea water diluted to 1/3 or 2/3 of full strength. The frogs had access to resting platforms just above the surface of the water, and they actually spent much time out of water. All experiments reported here were made after the frogs had adjusted to a given salinity for 2 days or more. Previous experience indicates that these frogs attain a relatively stable steady state in relation to the external medium in about this time (Gordon et al. 1961 b).

A total of twenty-eight frogs was used in the studies. Their body weights ranged from 10 to 55 g. The frogs were kept, and all experiments carried out, at room temperature which varied from 27 to 31° C.

The chemical determinations were made as follows : urea and ammonia by the microdiffusion method of Conway & O’Malley (1942), sodium by flame photometry with lithium as internal standard, and creatinine photometrically by the alkaline picrate method (Clark & Thompson, 1949 ; Sanz, 1957). Osmotic concentration was determined by a freezing point (actually melting-point) method similar to that described by Gross (1954). Warming rate of the bath was about 2° C./hr., and the unknown samples were observed, together with a set of four sodium chloride standards with melting points in the same range. The precision of the method was about ± 20 m-osm./l.

Glomerular filtration rate was measured as creatinine clearance. In order to increase the creatinine concentration in the blood (minimize the influence of chromogens other than creatinine) a solution of 5 % creatinine in 0·9 % sodium chloride was injected into the dorsal lymph sac 12–18 hr. prior to the experiment. The dose of creatinine solution was 1 ml./100 g. body weight. Urine was collected by means of a polyethylene catheter held in place in the cloaca with a purse-string ligature. Urine samples contaminated with faecal material were discarded. Blood samples were withdrawn directly from the heart in a heparinized syringe immediately after the last urine collection. Blood sampling before urine collection was usually omitted in order to avoid an effect on kidney function (Schmidt-Nielsen & Forster, 1954). The slope of the plasma creatinine curve was established by repeated blood samples on a few frogs, and the decrease rate was used to calculate the mid-point plasma concentration in those frogs where only a final blood sample was obtained. Since the frogs are small heart puncture is rather traumatic and no frog was used for more than one clearance experiment.

Urine and plasma samples were immediately analysed for urea and ammonia. Even a few hours’ storage of urine samples gives an appreciable increase in ammonia at the expense of urea. However, since the samples had been accumulating for a period of time in the cloacal catheter before collection, the observed urea concentrations must be considered as minimum values, and an unknown part of the observed ammonia may have been originally present as urea. The true urea value in the urine is therefore somewhere between the observed level and the sum of observed urea and ammonia. Samples for creatinine analysis were immediately measured and diluted in 0·1 N-HCI. The analyses were carried out the same day. Frozen samples were used for sodium analysis. Samples for freezing-point determination were sealed in glass capillaries, frozen, and analysed later.

Osmotic relation to the medium

The osmotic concentration of the plasma of Rana cancrivora in relation to the concentration of the environment is presented in Fig. 1. The frogs are hypertonic to the medium at all concentrations, as was reported by Gordon et al. (1961 a, b). Even in the most concentrated solutions tested, 4/5 of coastal Atlantic sea water, the frogs remained slightly hypertonic. At the lower concentrations the hypertonicity was greater, and in fresh water the osmotic concentration of the plasma was on the average close to equivalent to 370 m-osm./l. This is somewhat higher than reported by Gordon et al. which may be due to the different treatment of frogs examined in Thailand and those shipped to the United States by air, or to seasonal or individual differences. In any case, R. cancrivora is definitely hypertonic to normal frogs (see curve for R. esculenta in Fig. 1) at all external concentrations.

Fig. 1.

Plasma (▴) and urine (•) osmotic concentration of R. cancrivora in various external concentrations. The six groups of frogs were in fresh water, 1/8, 1/3, 2/3, 3/4 and 4/5 coastal sea water. The average value for each group is denoted by (○). Data for R. escalenta (×) are from Adolph (1933). The 45° line denotes equal plasma and medium concentrations.

Fig. 1.

Plasma (▴) and urine (•) osmotic concentration of R. cancrivora in various external concentrations. The six groups of frogs were in fresh water, 1/8, 1/3, 2/3, 3/4 and 4/5 coastal sea water. The average value for each group is denoted by (○). Data for R. escalenta (×) are from Adolph (1933). The 45° line denotes equal plasma and medium concentrations.

In these experiments no attempt was made to find the upper limit for the osmotic tolerance for the frogs. Since only a limited number was avilable, all our experiments were done well below the concentration of 4/5 sea water (670 m-osm/1.) where Gordon et al. found that many frogs survived. None of the frogs that we placed in 2/3 or 3/4 of coastal sea water (560 and 630 m-osm./l., respectively) seemed adversely influenced and they did not die sooner than other frogs placed in more dilute solutions or in fresh water.

Urine concentrations

The urine of R. cancrivora was hypotonic to the blood at all concentrations of the external medium (Fig. 1). In a dilute medium and in fresh water this is as expected and corresponds to what is known for amphibians in general. Since the body fluids are appreciably more concentrated than the medium, there is a steady osmotic inflow of water. The excess water is therefore excreted as a relatively dilute urine.

At higher concentrations the urine remained osmotically hypotonic to the medium and to the plasma. This is different from what is found in other amphibians where the urine becomes isotonic with the plasma if the animals are placed in relatively concentrated solutions, such as 1 % sodium chloride. The reason for the production of a hypotonic urine in R. cancrivora, even in the most concentrated media, seems to be that the frogs remain hypertonic to the medium. There is, therefore, a continued osmotic inflow of water which is eliminated as a relatively dilute urine. In this way the frogs remain hypertonic, and water is available for urine formation without the frogs having to drink the medium (as marine fish do).

Plasma concentrations

The plasma concentrations of sodium and of urea at various concentrations of the external medium are plotted in Fig. 2. The plasma sodium increases with the concentration of the medium and may be as high as 250 m-equiv./l., about twice as high as the maximum tolerated by most other amphibians.

Fig. 2.

A. Plasma sodium concentration of R. cancrivora in various external concentrations. B. Plasma urea concentration of R. cancrivora in various external concentrations.

Fig. 2.

A. Plasma sodium concentration of R. cancrivora in various external concentrations. B. Plasma urea concentration of R. cancrivora in various external concentrations.

As previously reported by Gordon et al., urea is a major contributor to the osmotic concentration of the plasma. The concentration in our frogs was higher in the more concentrated media than in dilute media. However, we did not observe any concentrations as high as the highest reported by Gordon et al. (480 mM./l.). This is perhaps due to the fact that our frogs were not in as good condition as the freshly caught animals used by Gordon et al.

Urea concentration in the urine

Since urea presumably is used as a major agent for osmoregulation in R. cancrivora, the substance becomes a physiological asset that must be retained to a suitable extent. The skin must be relatively impermeable to urea, and, furthermore, the renal excretion of urea must be reduced to the necessary minimum. Since the frogs are nearly isotonic with the concentrated medium, and must remain so because of the water-permeable skin, the retention of adequate amounts of urea permits the electrolyte concentration to be correspondingly low. This is probably the main physiological advantage of the urea retention.

Our observations on urea concentrations in the urine are related to the simultaneous plasma concentrations in Fig. 3. In virtually every case the urea concentration is higher in the urine than in the plasma. In one sample both the urea and the sum of urea and ammonia were clearly below the plasma concentration. The sample was too small for a repetition of the analysis and it is therefore impossible to say whether there was an error. A difficulty in interpreting the borderline data is that an unknown fraction of the free ammonia may have been present as urea when the urine was formed. The true urea concentration should therefore be at some point between observed urea concentration and the sum of urea plus ammonia. (The plasma urea is not subject to this error because ammonia is virtually absent from freshly drawn blood.)

Fig. 3.

Relationship between urine and plasma urea concentrations in R. cancrivora. (•) = urea concentration (total ammonia less free ammonia), (× ) = total ammonia (urea-N plus free ammonia). The pair of values obtained for each urine sample is joined by a vertical dashed line. The 45° line represents equal urine and plasma urea concentration.

The samples were analysed immediately after collection, but part of the urea may have been already split to NH, by bacterial action during the collection period. The concentration of urea in the urine when it was formed is therefore between the value for urea (•) and urea plus ammonia (×).

Fig. 3.

Relationship between urine and plasma urea concentrations in R. cancrivora. (•) = urea concentration (total ammonia less free ammonia), (× ) = total ammonia (urea-N plus free ammonia). The pair of values obtained for each urine sample is joined by a vertical dashed line. The 45° line represents equal urine and plasma urea concentration.

The samples were analysed immediately after collection, but part of the urea may have been already split to NH, by bacterial action during the collection period. The concentration of urea in the urine when it was formed is therefore between the value for urea (•) and urea plus ammonia (×).

The data in Fig. 3 indicate that urine urea is close to diffusion equilibrium with the plasma. If the urine concentration were lower than the simultaneous plasma concentration, this would mean reabsorption by an active process. With the one exception, there was no definite indication of an active reabsorption of urea from the urine. This is contrary to the statement by Gordon et al. (1961b), who reported urine urea concentration below the plasma concentration. These earlier observations were made on a small number of frozen samples that were shipped to the United States for analysis. Since they were pooled from several frogs they are not directly comparable to the plasma samples. This could account for the differences between the present observations and the earlier results. (The difference could also be due to another circumstance; our samples were collected in an indwelling catheter while the earlier samples were of bladder urine which could have been modified during the relatively long stay in the bladder.)

The amount of urea excreted at various external concentrations

The amount of urea excreted in the urine by R. cancrivora decreases with increasing external concentration (Fig. 4). In some individuals in concentrated media the amount of urea excreted per unit time was only a few per cent of the amount excreted by frogs in fresh water. This points out a conspicuous difference between those frogs that should have a definite advantage in retaining urea, and those that have little or no use for urea retention because they are in a dilute medium.

Fig. 4.

Excretion of urea by R. cancrivora in various external concentrations. Note the decrease in amount of urea excreted with increasing external concentrations.

Fig. 4.

Excretion of urea by R. cancrivora in various external concentrations. Note the decrease in amount of urea excreted with increasing external concentrations.

Since there is a definite difference in urea excretion between frogs in dilute and in concentrated media, and since the difference is not due to a demonstrable tubular reabsorption of urea (see Fig. 3), other aspects of the renal function must be responsible unless there are changes in the permeability of the skin.

Urine flow

The most obvious difference in urinary excretion between frogs in dilute and in concentrated media is that the urine flow is greatly diminished in those in concentrated media (Fig. 5 A). The urine flow in frogs in sea water with an osmotic concentration of 670 m-osm./l. was as low as 0·4 ml./kg.. hr, only about 1 % of the flow in some of the frogs in fresh water (as high as 31 ml./kg.. hr.). This reduction in urine flow in a concentrated medium is apparently the major factor in the retention of urea, for the plasma urea concentration varied less than two-fold as between the frogs in fresh water and those in salt water.

Fig. 5.

A. Urine flow of R. cancrivora in various external concentrations. (○) = average value for each group. B. Glomerular filtration rate (GFR) of R. cancrivora in various external concentrations. (○) = average value for each group. C. Creatinine U/P ratio (i.e. concentration index of urine) vs. external concentration in R. cancrivora. (○) = average for each group. Figures in parentheses indicate the number of observations. Note that water reabsorption increases with increasing external concentration.

Fig. 5.

A. Urine flow of R. cancrivora in various external concentrations. (○) = average value for each group. B. Glomerular filtration rate (GFR) of R. cancrivora in various external concentrations. (○) = average value for each group. C. Creatinine U/P ratio (i.e. concentration index of urine) vs. external concentration in R. cancrivora. (○) = average for each group. Figures in parentheses indicate the number of observations. Note that water reabsorption increases with increasing external concentration.

Glomerular filtration rate

The glomerular filtration rate, as indicated by the exogenous creatinine clearance, was appreciably higher in frogs in fresh water and dilute media than in those in concentrated media (Fig. 5B). Unfortunately only a few determinations were successfully carried out in the concentrated media (one difficulty was the low urine flow at these concentrations), but the data indicate that the filtration rate is in the order of magnitude of 1/3 of the filtration rate in dilute media.

As far as any conclusions can be drawn from the few observations in Fig. 5B, one can say that there is little or no reduction in filtration rate up to about 1/3 of fullstrength sea water. This means that the conspicuous decrease in urine flow indicated by Fig. 5A is not primarily due to a reduction in filtration rate, but is caused by an increased tubular reabsorption of water as well.

Tubular water reabsorption

Fig. 5 C shows the reabsorption of water which takes place between the formation of the glomerular filtrate and the discharge of the final urine. This is given as the creatinine U/P ratio, which indicates how many times the final urine is concentrated above the original filtrate. The water reabsorption increases with increasing medium concentration, becoming on the average more than ten times as great in the most concentrated media.

The conclusion is therefore that the main cause of the diminished urea excretion in concentrated external media is due to the reduced urine flow, caused by an increased tubular reabsorption of water and to a minor extent by a reduction in glomerular filtration rate.

Fraction of filtered urea reabsorbed

Since the urea diffuses with relative ease from the tubular fluid to the plasma, it is of interest to see what influence the combined effect of water reabsorption and reduction in glomerular filtration rate has on the urea excretion. Fig. 6 indicates what fraction of the filtered urea is excreted in the urine as the concentration ratio of the urine (water reabsorption) increases. A urea/creatinine clearance ratio above 1 indicates an active tubular secretion of urea. Such active secretion of urea has frequently been observed in many freshwater frogs (Marshall, 1932; Walker & Hudson, 1937; Forster, 1954), but only one individual among our frogs indicated that such secretion might occur in R. cancrivora. A few frogs had urea/creatinine clearance ratios approaching 1, but most had a ratio below 0·5, which means that more than half of the filtered urea was reabsorbed in the renal tubules. With an increasing creatinine U/P ratio (tubular reabsorption of water) the fraction of the filtered urea which appeared in the urine diminished rapidly. In one extreme case the amount of urea in the urine was only 1·4% of the amount filtered, but the reabsorption of the remaining 98-6% could, as explained above, be ascribed to a passive rather than to an active reabsorption.

Fig. 6.

Reabsorption of urea and sodium in relation to reabsorption of water. Urea/creatinine clearance ratio (•) indicates the fraction of filtered urea which is excreted in the urine. Sodium/ creatinine clearance ratio (× ) indicates the fraction of filtered sodium which is excreted in the urine. Clearance ratios less than 1 indicate reabsorption. Abscissa (creatinine U/P ratio) indicates the concentration index of the urine.

Fig. 6.

Reabsorption of urea and sodium in relation to reabsorption of water. Urea/creatinine clearance ratio (•) indicates the fraction of filtered urea which is excreted in the urine. Sodium/ creatinine clearance ratio (× ) indicates the fraction of filtered sodium which is excreted in the urine. Clearance ratios less than 1 indicate reabsorption. Abscissa (creatinine U/P ratio) indicates the concentration index of the urine.

The crab-eating frog remains somewhat hypertonic to the external medium, even in the most concentrated solutions that have been tested. This situation is maintained by the formation of a urine which is more dilute than the plasma. Since the skin is permeable to water, the inevitable result is an osmotic inflow of water from the medium, and this inflow permits the formation of urine without any necessity for the frog to drink the external medium. There may be several advantages in this situation. One is that the gastro-intestinal tract is not loaded with an excessive intake of magnesium and sulphate, which constitute roughly one-tenth of the salts in sea water.

Another advantage in the formation of a dilute urine is that this urine, if retained in the bladder, can serve as a water reservoir. It has frequently been pointed out that some desert frogs (Cyclorana, Chiroleptes), when water is available after heavy rains, form large amounts of dilute urine which are stored in the bladder. During prolonged dry periods the frogs aestivate deep in the ground and slowly reabsorb the bladder urine. The water reservoir is so large that Australian aborigines used the distended frogs as a supply of drinking water (Buxton, 1923). Since the information available about the ecology of the crab-eating frogs is inadequate, it is difficult to estimate the value of a dilute urine to this animal. During high tide this frog swims and seeks its food in full-strength sea water, and during low tide it remains in the tidal area where it sits under mangrove roots or in small holes. It is quite possible that it could have some advantage from keeping a supply of relatively dilute urine in its bladder. This question deserves further attention.

A parallel situation has been observed by Ruibal (1962), who investigated a number of anuran amphibians found in inland saline waters in Argentina. None of these amphibians was particularly tolerant to concentrated salt solutions, but one of the species remained slightly hypertonic to the medium, even in the most concentrated solutions that could be tolerated. This species (Pleurodema nebulosa) also formed a dilute urine. Interestingly this particular species occurred in a more saline habitat than any of the other species, although in the laboratory it did not display any higher tolerance to salinity of the medium. It therefore seems that the ability to form a dilute urine is of advantage in an environment where salinity is the major ecological problem.

In the crab-eating frog the retention of urea is the major mechanism for remaining hypertonic, aside from a considerable tolerance to increased plasma electrolyte concentrations. This makes the retention of urea physiologically desirable. In elasmobranchs, which also retain urea for osmoregulatory purposes, the retention of urea is based on a relative impermeability of the gills to urea, as well as on a tubular reabsorption of urea in the kidney (Smith, 1936). In the crab-eating frog the skin presumably is relatively impermeable to urea, but the data presented in this paper indicate that there is no active tubular reabsorption of urea. However, there is a considerable decrease in urea excretion in concentrated media, caused primarily by a reduction in urine flow. This reduction is due mainly to an increased tubular reabsorption of water, and to a lesser extent to a reduction in glomerular filtration rate. Since the volume of urine is determined primarily by the osmotic inflow of water through the skin, and since this again depends on the extent to which the frog is hypertonic to the medium, there is to some degree an automatic control of the urea loss. With an increased urine flow more urea is lost and the degree of the hypertonicity is reduced. The osmotic inflow of water is thereby decreased, urine volume goes down, and more urea is conserved.

It seems probable that the permeability of the skin to water would change with the amount of antidiuretic hormone, as it does in other frogs (Ussing et al. 1960). A study of the effect of antidiuretic hormone on the skin of the crab-eating frog as well as a study of the interaction between endocrine function and osmoregulation and renal function would be most interesting.

We would like to acknowledge our gratitude to Dr M. S. Gordon who transmitted to us a number of live frogs in his possession, as well as the valuable help of Dr Albert H. Banner and his sons who caught the frogs in Thailand and arranged for their shipment by air. Supported by Grant no. H 2228 from the National Institutes of Health.

Adolph
,
E. F.
(
1933
).
Exchanges of water in the frog
.
Biol. Rev
.
8
,
224
40
.
Buxton
,
P. A.
(
1923
).
Animal life in Deserts
,
176
pp.
London
:
Arnold
.
Clark
,
L. C.
, Jr
. &
Thompson
,
H. L.
(
1949
).
Determination of creatine and creatinine in urine
.
Anal. Chem
.
21
,
1218
21
.
Conway
,
E. J.
&
O’malley
,
E.
(
1942
).
Microdiffusion methods. Ammonia and urea using buffered absorbents
.
Biochem. J
.
36
,
655
61
.
Forster
,
R. P.
(
1954
).
Active cellular transport of urea by frog renal tubules
.
Amer. J. Physiol
.
179
,
372
7
.
Gordon
,
M. S.
,
Schmidt-Nielsen
,
K.
&
Kelly
,
H. M.
(
1961a
).
Osmotic regulation in the euryhaline crab-eating frog (Rana cancrivora) of southeast Asia
.
Fed. Proc
.
20
,
208
.
Gordon
,
M. S.
,
Schmidt-Nielsen
,
K.
&
Kelly
,
H. M.
(
1961b
).
Osmotic regulation in the crab-eating frog (Rana cancrivora)
.
J. Exp. Biol
.
38
,
659
78
.
Gross
,
W. J.
(
1954
).
Osmotic responses in the sipunculid Dendrostomum zostericolum
.
J. Exp. Biol
.
31
,
402
23
.
Marshall
,
E. K.
, Jr
. (
1932
).
Secretion of urea in the frog
.
J. Cell. Camp. Physiol
,
2
,
349
53
.
Neill
,
W. T.
(
1958
).
The occurrence of amphibians and reptiles in saltwater areas, and a bibliography
.
Bull. Mar. Sci. Gulf. Caribbean
,
8
,
1
97
.
Ruibal
,
R.
(
1962
).
Osmoregulation in amphibians from heterosaline habitats
.
Physiol. Zool. (in the Press)
.
Sanz
,
M. C.
(
1957
).
Ultramicro methods and standardization of equipment
.
Clin. Chem
.
3
,
406
19
.
Schmidt
,
K. P.
(
1957
).
Amphibians
.
Mem. Geol. Soc. Amer
.
67
,
1
,
1211
12
.
Schmidt-Nielsen
,
B.
&
Forster
,
R. P.
(
1954
).
The effect of dehydration and low temperature on renal function in the bullfrog
.
J. Cell. Comp. Physiol
.
44
.
233
46
.
Smith
,
H. W.
(
1936
).
The retention and physiological role of urea in the elasmobranchii
.
Biol. Rev
.
11
,
49
82
.
Ussinc
,
H. H.
et al. 
. (
1960
).
The Alkali Metal Ions in Biology
.
Berlin
:
Springer-Verlag
.
Walker
,
A. M.
&
Hudson
,
C. L.
(
1937
).
The role of the tubule in the excretion of urea by the amphibian kidney
.
Amer. J. Physiol
.
118
,
153
66
.