Few data are available on the circadian rhythmicity in cancer patients. Since monitoring the disease usually implies the follow-up of blood concentrations of a number of biological variables, it would be of value to examine the profile of the circadian variations of serum cortisol and tumour marker antigens. This we did in 33 cancer patients (13 breast cancer patients and 20 ovarian cancer patients). The profiles of serum cortisol were documented, since this hormone is considered as a strong marker of circadian rhythms. This study shows that 8 out of 13 breast cancer patients and 15 out of 20 ovarian cancer patients had deeply altered cortisol circadian patterns. The modifications were either high levels along the 24 h scale and/or erratic peaks and troughs and/or flattened profiles. Within 24 h, variations of tumour marker antigens as large as 70% were observed but no typical individual circadian patterns could be found. No relationship between cortisol subgroups and concentration of tumour marker antigens at 8 h could be observed (Kolmogorov-Smirnov's test). The question thus arises as to the origin of these alterations, and whether they are related to a cause or a consequence of the disease, and their possible incidence upon therapeutic designs.