The presence of a trigger point is essential to the myofascial pain syndrome. This study centres on identifying clearer criteria for the presence of trigger points in the quadratus lumborum and gluteus medius muscle by investigating the occurrence and inter-rater reliability of trigger point symptoms. Using the symptoms and signs as described by Simons' 1990 definition and two other former sets of criteria, 61 non-specific low back pain patients and 63 controls were examined in general practice by 5 observers, working in pairs. From the two major criteria of Simons' 1990 definition only 'localized tenderness' has good discriminative ability and inter-rater reliability (kappa > 0.5). This study does not find proof for the clinical usefulness of 'referred pain', which has neither of these two abilities. The criteria 'jump sign' and 'recognition', on the condition that localized tenderness is present, also have good discriminative ability and inter-rater reliability. Trigger points defined by the criteria found eligible in this study allow significant distinction between non-specific low back pain patients and controls. This is not the case with trigger points defined by Simons' 1990 criteria. Concerning reliability there is also a significant difference between the two different criteria sets. This study suggests that the clinical usefulness of trigger points is increased when localized tenderness and the presence of either jump sign or patient's recognition of his pain complaint are used as criteria for the presence of trigger points in the M. quadratus lumborum and the M. gluteus medius.