Hygroscopic equilibrium data have been obtained for a wide variety of materials, because the derived moisture sorption isotherms serve a number of useful functions. They may be employed : to define limits for the dehydration of fruits, vegetables, nuts, etc. (28); to estimate moisture content changes under any given condition of temperature and humidity (28); to evaluate processing variables and to distinguish differences between grades or varieties of agricultural commodities (23,27); to aid in the selection of packaging materials (23,28); and to define moisture or humidity conditions under which product deterioration (2, 6, 10,19) and microbial growth can be inhibited (5, 38, 39). In addition, water sorption isotherms of more homogeneous materials, such as natural and synthetic fibres, proteins and other purified materials, can be used in the determination of various physical constants (1, 7, 11, 12, 13,16, 18,24,29,36,40) and to provide information on the mechanisms of moisture binding (2 to 4, 9, 21, 29 to 35, 41, 46). Diversity in the shapes and amplitudes of the sigmoid moisture sorption isotherms is due to the relative amounts and individual moisture affinities of hygroscopic constituents in the whole material. I t would be advantageous to describe moisture sorption isotherms precisely by means of a general mathematical expression, since a minimum amount of experimental data would be required to construct an isotherm and it would permit precise and convenient interpolation of available data. In addition, it would permit the extrapolation to low and especially high humidities.Hygroscopic equilibrium data are difficult to obtain experimentally at extreme humidities because of the limitations of available hygrometers or because of the rapid deterioration and development of mold at high humidities. Several equations have been derived on empirical or theoretical bases for the description of sorption phenomena and tested against observed data (2, 3, 4, 7, 8, 12,16, 18,21,29,36). However, previous treatments were only partially successful in that they did not represent the observed isotherms over the complete relative humidity interval, or they exhibited major deviations from observed isotherms or did not adequately represent the sorption isotherms for particular hygroscopic materials. No attempt will be made to consider the extensive literature on this subject, since it has been reviewed in the work cited above and other publications referred to therein.