A number of equations for the correlation of retention data for a series of solutes on a given stationary phase (or solvent) have been investigated with the aim of characterising stationary phases. The two most successful equations are, SP =c+dδ₂+sπ₂*+aα₂^H+bβ₂^H+I log L¹⁶(a), SP =c+rR₂+sπ₂*+aα₂^H+bβ₂^H+I log L¹⁶(b) In the present case the dependent variable SP is log L– log L^Decane and the explanatory variables are solute parameters as follows: δ₂ is an empirical polarisability correction term, R₂ is a polarisability parameter that reflects the ability of a solute to interact with a solvent through π and n electron pairs, α₂^H is the solute hydrogen–bond acidity, β₂^H is the solute hydrogen–bond basicity, π₂* is the solute dipolarity/polarisability, and L¹⁶ is the Ostwald solubility coefficient of the solute on n-hexadecane at 298 K. The constants c, r, s, a, b, and l in the more useful equation (b) are found by the method of multiple linear regression analysis, and serve to characterise a solvent phase in terms of specific solute/solvent interactions. Application of equation (b) to the five stationary phases examined by Laffort et al. shows that the magnitude of these constants is in accord with general chemical principles, and that the present procedure constitutes a new, general method for the characterisation of gas chromatographic stationary phases.