ABSTRACT Fluxes of potassium ions across biological membranes can be measured using 42K as a tracer. This isotope, however, has a half-life of only 12-4 h and it is expensive. Rubidium belongs to the alkali metal series of elements and rubidium ions behave in many respects, physicochemically and biologically, like K+. Furthermore, 86Rb has a half-life of 18-7 days and is less expensive than 42K. Thus 86Rb has been used quite widely as a tracer for estimating K+ transport. In mammals this usage has included erythrocytes (Beauge & Adragna, 1971), the crystalline lens (Becker, 1962), renal tubules (Ellison, Velazquez & Wright, 1986) and large intestine (Tannen, Marino & Dawson, 1986; Freel, 1987). In these instances, the ratios of the measured fluxes, JK/JRb were not found to differ significantly from unity. In non-mammals, 86Rb has been shown to be a useful substitute for K+ in transport studies using flounder intestine (Frizzell et al. 1984), salmon erythrocytes (Bourne & Cossins, 1984) and frog skin (Zerahn, 1983). In fish gills, the ratio JK/JRB was consistently found to be about 1·3 (Sanders & Kirschner, 1983), thus allowing for correction of estimates of K+ flux based on the use of 86Rb as a tracer. In the American silkworm (Hyalophora cecropia) gut, this ratio was 1·1 in a flat-sheet preparation (Wood & Harvery, 1979), but in a spherical preparation it was rather unpredictable with a mean of 1 -7 (Zerahn, 1980). None of these authors considered 86Rb to be a reliable substitute for 42K. Eddy (1985) has drawn attention to such possible disparities of K+ metabolism estimated by the use of 86Rb in fish. We have also found differences in estimates of K+ transport using 42K and 86Rb as tracers in the avian intestine.