Previously, the algal absorption product (absorption coefflcicnt at wavelength X times cuvette depth) was estimated as D, -D750, the difference in optical densities of a suspension, measured against water in opal glass cuvettes. Such estimates are subject to "cuvette error" dependent on dimensions and materials of the cuvettes and to "A error" due to inadequacy of 750 nm as a reference wavelength. Monte Carlo calculations were performed to assess cuvette error; then limits of the combined cuvette and X errors were explored. With cuvettes and suspensions used previously, the combined error was usually + 10% (lo-20% at 440 nm), not 100% as supposed in a recent publication. Cuvette error is minimal in cubical cuvettes with black sidewalls, but substantial X error may still occur. Both errors arc avoided in a method using two pairs of opal cuvettes. Weidemann and Bannister (1986) and Weidemann et al. (1985) concentrated algae or lake particulates and measured absorption spectra of the suspensions against water in rectangular cuvettes with opal glass backwalls. The difference, DA -D750, between optical densities at a wavelength X in the visible and at 750 nm (where algal absorption was presumed negligible) was taken as an estimate of the product of cuvette depth (d) and the volume absorption coefficient (a) of the algae or particles at X.Values of ad determined in this way may over-or underestimate true values of the algal absorption product. There are two sources of error. One, a "cuvette error," arises from the fact that the walls of opal glass cuvettes contribute to the diffusion of light in the suspension, as a result of which average path length becomes larger than cuvette depth. The magnitude of cuvette error depends on the dimensions and materials of the cuvettes. Herein, cuvette error is defined as the factor by which DA -D,,, overestimates the absorption product, ad, when the particles and suspending medium are ascribed "ideal properties." Ideal behavior assumes that the suspending medium does not scatter, absorption by the medium is either insignificant or else independent of X, particles do not absorb at 750 nm, and particle scattering is independent of X. A second error, here called "X error," arises when properties of a suspension depart from ideal ones, D75,, then becoming an inadequate correction. Scattering by water is too weak (b < 0.0 1 m-l : Kirk 1983) to contribute to X error. The absorption coefficient of water is X-dependent, however, and rises to -2.5 m-l at 750 nm; in 25-mm cuvettes, a small but significant X error would occur. There is some evidence that the algal scattering coefficient is inversely proportional to wavelength (Kirk 1983), and some studies have indicated significant particle absorption at 740 nm (Morel and Bricaud 198 1). I shall show that a large X error can arise from particle absorption at 750 nm or X-dependent particle scattering.Weidemann and Bannister (1986) supposed that their suspensions scattered light mainly through small angles and that diffuse light scatte...