The soluble fraction chemical release rates from contaminated stream bed sources to overlying water is more significant than once thought. This fact is supported by field measurements on the transport coefficients in rivers during time-periods when particle resuspension is absent. The numerical values of soluble release coefficients, kf (cm/d), for PCBs in the Hudson, Grasse and Fox Rivers display similar magnitudes and an annual cyclic pattern. The lower values, 3 to 10 cm/d, occur in the winter while much higher values, 20 to 40 cm/d, occur in the summer. Candidate theoretical transport processes with characteristics capable of quantifying the observed magnitude and behavioral pattern of kf were selected from the literature and reviewed. The bed-side process of bioturbation driven biodiffusion coupled to a water-side benthic boundary layer resistance was selected as the theoretical transport algorithm. With flow, temperature and other data from the Fox River numerical coefficient calculations were made using a PCB congener. Based on these results and other evidence it was shown that the proposed theoretical algorithm predicted the correct magnitudes of kf and was capable of mimicking its cyclic behavior.