Use of the Rotterdam Symptom Checklist (RSCL) to measure psychological and physical distress as experienced by cancer patients, is discussed in this paper. The stability of the structure of the RSCL was assessed in principal component analyses in three studies: one concerning cancer patients during either chemotherapy or follow-up (n = 86), one done in patients undergoing chemotherapy for advanced ovarian cancer (n = 56), and the third dealing with cancer patients under treatment, disease-free 'patients', and 'normal' controls (n = 611). The psychological dimension proved to be stable across populations. A scale based on this factor was highly reliable (Cronbach's alpha 0.88-0.94). The physical distress is reflected by several dimensions in a homogeneous population (pain, fatigue, gastrointestinal complaints) and undimensionally in a heterogeneous population. Reliability of the physical distress scales is good (0.71-0.88). The current components of the RSCL and the use of individual and disease specific symptoms are discussed.