Procedures for graphic and numerical parameter fitting from Scatchard plots for the case of two different independent sets of equivalent binding sites are described. These methods are compared with other procedures and their advantages and disadvantages are discussed. Estimates of how much one parameter can be changed without changing the Scatchard curve significantly are made. This variability of the parameters is discussed. It is shown that low saturation points are of considerable importance in evaluating the binding parameters of the high affinity binding site. An example of the parameter fitting procedures is given and the utility and limitations of the procedures are discussed. It has been stated [6] that a logarithmic plot should be used to exclude other forms of data representation. The logarithmic scale allows one t o plot all theoretically obtainable data. The doublereciprocaI plot [7] and the Scott or half-reciprocal plot [8] have open upper limits on the abscissa as has the absorption isotherm. I n the Scatchard plot, all theoretically obtainable data can be plotted provided that the degree of binding is used on the abscissa and the nonspecific binding sites with large binding capacity are absent. An objection to the Sctachard plot was that linearisation of binding data, which is only possible in the rare case of only one single set of binding sites, should result in a loss of information [4]. Nevertheless, we studied the representation of binding data with the Scatchard plot, mainly because of its wide-spead use, but being all the time aware of its drawbacks in certain cases.A similar study has been undertaken by Feldman 191, where a least-squares parameter fitting has been used for the numerical method. Least-squares parameter fitting was also used by Nagy [lo]. A graphic parameter fitting procedure has been described by Berson and Yalow [ll], and Rosenthal [12], while others [13,14] used a semiautomated parameter fitting on a small computer. The aim of the present investigation was to study the determination of the parameters of binding systems with different means of expenditure. The graphical and numerical curve fitting have been studied using a programmable calculator and a computer. The numerical parameter fitting of the computer program does not use least-squares regression, therefore some difficulties could be overcome. To test the applicability of the fitting procedure a practical example will be given as an illustration.
BASIC RELATIONSHIPSThe outlined procedures are based on the following assumptions : all ligands are identical, the activity (chemical, biological) will not be altered except by binding, two independent sets of binding sites are present, all sites of each set are equivalent (identical), there is no cooperativity, each set of binding sites obeys the mass law action.At equilibrium for each system containing one single set of binding sites, the Scatchard plot [3] gives a straight line following the equation Eur. J. Biochem. 42 (1974)