Research on water exchange in frogs has historically assumed that blood osmotic potential drives water exchange between a frog and its environment, but here we show that the “seat patch” (the primary site of water exchange in many anurans), or other sites of cutaneous water uptake, act as an anatomic “compartment” with a water potential controlled separately from water potential of the blood, and the water potential of that compartment can be the driver of water exchange between the animal and its environment. 

We studied six frog species (Xenopus laevis, Rana pipiens, R. catesbeiana, Bufo boreas, Pseudacris cadaverina and P. regilla) differing in ecological relationships to environmental water. We inferred the water potentials of seat patchesfrom water exchanges by frogs in sucrose solutions ranging in water potential from 0 to 1000 -kPa. 

Terrestrial and arboreal species had seat patch water potentials that were more negative than the water potentials of more aquatic species, and their seat patch water potentials were similar to the water potential of their blood, but the water potentials of venters of the more aquatic species were different from (and less negative than) the water potentials of their blood. 

These findings indicate that there are physiological mechanisms among frog species that can be used to control water potential at the sites of cutaneous water uptake, and that some frogs may be able to adjust the hydric conductance of their skin when they are absorbing water from very dilute solutions. 

Largely unexplored mechanisms involving aquaporins are likely responsible for adjustments in hydric conductance, which in turn, allow control of water potential at sites of cutaneous water uptake among species differing in ecological habit and the observed disequilibrium between sites of cutaneous water uptake and blood water potential in more aquatic species.

The seat patch water potential was inferred from experiments using different environmental water potentials to find the conditions where frogs do not exchange water with the environment – i.e., when the water potential of the environment and seat patch are equal (Tracy and Rubink 1978). Prior to each experiment, frogs were placed in pure water for two hours to allow them to become fully hydrated. Then, frogs were catheterized to remove any water in the bladder. The body mass of the frog with an empty bladder was then recorded as the “standard body mass” (Ruibal 1962; McClanahan 1972; Tracy 1976), and then the frogs were dehydrated in a wind tunnel to 90% of their standard body mass which took from one to several hours depending on species and activity of the frogs. Once they reached this target level of dehydration, the frogs were placed individually into an apparatus similar to the one used in Tracy and Rubink (1978) consisting of a plastic container containing a sucrose solution so that only the ventral surface of the frog’s body was in contact with the solution, and the rest of its body was exposed to the air. For the following hour, the frogs’ body masses were measured every 10 min after the frog was blotted dry with a paper towel. Each frog was tested in the various sucrose solutions, with at least 7 days between experiments. The order of testing began with pure water and progressed towards more negative water potentials.

Cutaneous evaporative water loss during the water uptake experiment was estimated from experiments using frog models made from 3% agar molded into the shape of the frog, which evaporates as a free water surface of the given size and shape (Spotila and Berman 1976). Negative molds of each species of frogs were made by pouring dental alginate on live frogs that had been given MS222 (Tricaine methanesulfate) as an anesthetic to minimize the frog’s movements while the mold was being made. Alginate sets in less than one minute, so no harm is caused to the frog. Plaster of Paris was then poured into the alginate mold to create a positive mold of the frog. Latex was then painted onto the plaster of Paris mold in several layers to make a thick and durable negative mold that could be reused many times. A 3% agar solution was then poured into the latex mold and allowed to set. By this approach, the agar frog models were made to be the same size, shape, and posture as the living frogs from which molds were made. Water absorption by the resulting agar model was prevented by coating the venter of the model with fingernail polish approximately where the frog would be in contact with the sucrose solution. The agar frog model was placed in the experimental apparatus for measuring water uptake, but the liquid solution was not allowed to come into contact with agar frog model. Thus, the relative humidity in the apparatus was the same as that in experiments for measuring seat patch water potential. The body mass of the agar models was measured every ten minutes.

The mean change in frog body mass in the apparatus reflects the sum of water uptake and evaporative water loss, so the water influx into the frog was obtained by adding the evaporative water loss (estimated from the agar models as described above) to the water exchange (measured in the apparatus described above). This calculated water influx was then plotted against water potential of the environment (sucrose solutions) to obtain the x-intercept of this relationship, which is the point at which no liquid water was exchanged because the water potential of the seat patch was equal to the water potential of the sucrose solution. Data were discarded in cases when excessive frog activity resulted in questionable results. The bladders of the frogs were voided prior to being dehydrated to 90% standard body mass to decrease the likelihood of urination, but the data were discarded when either urination or defecation occurred during the experiments. Inspection of plots of water uptake as a function of sucrose water potential showed two patterns for each species except P. cadaverina:  a sloping linear relationship at lower water potentials, and a horizontal relationship at high water potentials. That is, for dilute solutions (near 0 kPa), water uptake rates were similar despite the frogs being subjected to different water potentials, but at more negative water potentials there was a sloping linear relationship between water potential and water uptake. Thus, because it was clear that different mechanisms governed the points at different water potentials for these species, only the points along the slope were used in regression analyses to determine the seat patch water potentials (intercept of the X-axis). In P. cadaverina, the X-intercept was not different if the point for the highest water potential was used in the regression or not. We used an ANOVA followed by a Tukey HSD to test for differences of seat patch water potentials among the six species. 

Osmotic Potential of Blood 

Several months after the experiments described above were completed, the osmotic potential of the blood was measured for each frog at 100% and 90% of their standard (fully hydrated) body mass, with a one week interval between the two measurements. Although most species were allowed to dehydrate in air, X. laevis had to be dehydrated osmotically by placing them in a sucrose solution of -600 kPa because this species showed signs of distress in air (perhaps they cannot supply mucus to the skin as can more terrestrial anurans). Blood samples (approximately 25 mL) were collected from all frogs via the abdominal midline vein using a 23-gauge hypodermic needle, and whole blood osmolality was measured immediately with a freezing point osmometer (Advanced Instruments 3MO). 

            Water Potential is expressed in the SI units of kPa, but the osmometer output was in units of mOsm. Blood osmotic potentials were converted to units of kPa using the van't Hoff Law (Salisbury and Ross 1969).

The osmotic potentials of blood for each species at 100% and 90% hydration were compared using a paired t–test. For each species, the water potential of the seat patch and the osmotic potential of blood at 90% hydration were also compared using a paired t-test.