Occasionally interpretation guidelines from validation studies are difficult to apply to real forensic casework, especially in the case of mixed samples. Exogenous contamination, an unknown number of contributors or unbalanced proportion of each one in the sample and a varied degree of degradation of the biological materials, contribute to the difficulties in the interpretation of sample profiles. In this paper we have reviewed all the mixed genetic STR profiles encountered in our laboratory over 4 years (1997-2000) and evaluated the problems in the interpretation of the results. From 1547 criminal cases with 2424 samples typed, 163 showed a mixed profile (6.7%). We have observed that occasionally, a mixture appeared in the same sample with one multiplex amplification kit (e.g. Blue) and not with another (e.g. Green). From our results, it can be suggested that technical characteristics of the different fluorochrome groups in the multiplexes override the molecular characteristics of each STR in their capacity to detect mixtures.