This in vitro study was undertaken to examine whether water proton spin-lattice relaxation in biological tissues is adequately described by a single time constant T1, to define under what circumstances a multi-exponential approach is indicated, and to study the implications of multi- exponentiality for quantitative NMR imaging. Water proton relaxation curves were measured with the 180-tau-90 method at 60 MHz. Uni- and bi-exponential curves were fitted to the empirical curves using chi 2 as a criterion for the goodness of fit. An F-test was applied to test the validity of each exponential term as it was added to the fitting function. Taking into account experimental accuracy, the uni-exponential model appeared to be an adequate description of the relaxation data for necrotic tissue. Eyelens and fat showed distinct bi- exponentiality , while liver, spleen, salivary gland, tumour, and muscle presented intermediate cases. The bi-exponential analysis generally yields a minor component with a fast relaxation time, T11 less than 20 ms, and a slow relaxation major component with T12 greater than 300 ms. A simplified bi-exponential model is proposed for implementation in quantitative NMR imaging. The results seem to be consistent with current views about water proton spin-lattice relaxation in biological tissues.