Rather high values of standard deviation could be ascribed to the heterogeneity between matrices used for validation procedure. In fact complex matrices, such as food (in particular meat and fishery products), can bind colorants in different sites. Solvent extraction may be not effective enough to remove all dyes from the matrix at optimum levels. Furthermore,the extracts prior to the clean up step are various mixtures containing different substances coming from animal and vegetable tissues. Polyamide resin used in SPE cartridges adsorbs polar compounds containing groups that can be protonated. Since the sample extract is at pH 2 before the SPE step, the dyes are adsorbed to the polyamide stationary phase by Van der Waals’ interactions. Other hydrophilic substances can mask SPE interaction sites by reducing their binding power for the dyes and consequently reducing the capacity of the cartridges. Furthermore, some substances, as amaranth, are strongly restrained by SPE cartridges and the ammonia solution used for elution could be insufficient for its release (low recoveries). For solid food matrices, LOQ values calculated as 10 the standard deviations of blank noise were 10 mg kg1 for groups AeC and 5 mg kg1 for group B.