The resolving power of multicomponent spectral analysis and the computation reliability of the stability constants and molar absorptivities determined for five variously protonated anions of physostigmine salicylate by the SQUAD(84) and SPECFIT/32 programs has been examined with the use of simulated and experimental spectra containing overlapping spectral bands. The reliability of the dissociation constants of drug was proven with goodness-of-fit tests and by examining the influence of pre-selected noise level s inst( A) in synthetic spectra regarding the precision s(p K) and also accuracy of the estimated dissociation constants. Precision was examined as the linear regression model s(p K) = β 0 + β 1 s inst( A). In all cases the intercept β 0 was statistically insignificant. When an instrumental error s inst( A) is small and less than 0.5 mAU, the parameters’ estimates are nearly the same as the bias Δp K = p K a,calc − p K a,true is quite negligible. In all four dissociation constants the bias seems to be quite small even though for p K a4 it is a little bit higher, i.e., +0.05 for s inst( A) about 1.0 mAU. In the interval of s inst( A) from 0.1 to 1.0 mAU all four dissociation constants p K i are accurate enough. Of the various regression diagnostics considered, the goodness-of-fit is the most efficient criterion of whether the parameters found adequately represent the data. The magnitude of instrumental error s inst( A) only slightly affects the shape of a Cattel's scree graph s k ( A) = f( k) to determine the true number of light-absorbing species in the equilibrium mixture.