"Reconstructed" gill epithelia on filter supports were grown in primary culture from dispersed gill cells of freshwater rainbow trout (Oncorhynchus mykiss). This preparation contains both pavement cells and chloride cells, and after 7-9 days in culture, permits exposure of the apical surface to true freshwater while maintaining blood-like culture media on the basolateral surface, and exhibits a stable transepithelial resistance (TER) and transepithelial potential (TEP) under these conditions. These epithelia were used to develop a possible in vitro version of the biotic ligand model (BLM) for silver; the in vivo BLM uses short-term gill binding of the metal to predict acute silver toxicity as a function of freshwater chemistry. Radio-labeled silver ((110m)Ag as AgNO(3)) was placed on the apical side (freshwater), and the appearance of (110m)Ag in the epithelia (binding) and in the basolateral media (flux) over 3 h were monitored. Silver binding (greater than the approximate range 0-100 mug l(-1)) and silver flux were concentration-dependent with a 50% saturation point (apparent K(d)) value of about 10 mug l(-1) or 10(-7) M, very close to the 96-h LC50 in vivo in the same water chemistry. There were no adverse effects of silver on TER, TEP, or Na(+), K(+)-ATPase activity, though the latter declined over longer exposures, as in vivo. Silver flux over 3 h was small (<20%) relative to binding, and was insensitive to water chemistry. However, silver binding was decreased by elevations in freshwater Na(+) and dissolved organic carbon (humic acid) concentrations, increased by elevations in freshwater Cl(-) and reductions in pH, and insensitive to elevations in Ca(2+). With the exception of the pH response, these effects were qualitatively and quantitatively similar to in vivo BLM responses. The results suggest that an in vitro BLM approach may provide a simple and cost-effective way for evaluating the protective effects of site-specific waters.