Infertile women exhibiting a short luteal phase in 3 consecutive menstrual cycles as judged from basal body temperature recordings underwent a prolactin (PRL) stimulation test with metoclopramide (MTCL) on day 22 of the 4th cycle. While serum progesterone (P) concentrations were decreased (less than 35 nmol/l), PRL was increased only in about two thirds of the women. When compared to normally ovulating women, 3 groups could be segregated: in group I basal (less 500 mU/l) as well as MTCL stimulated PRL (less 8000 mU/l) was normal; in group II-patients only basal PRL levels were elevated; in group III baseline levels of serum PRL were at the borderline, but PRL release after MTCL stimulation was exaggerated. Correlation of sleep induced PRL peaks, observed during the night preceding the MTCL test, with maximal MTCL stimulated PRL showed a positive relation (r = 0.7, p less than 0.05) in group I- and III-patients as well as in normal volunteers. Suppression of PRL release by bromocriptin (2.5 mg bid) from the day following the MTCL test resulted in significantly reduced basal as well stimulated serum PRL on day 22 of the treatment cycle. Luteal P secretion was normalized in group II- and III-patients only. The data presented indicate that some of the patients with luteal phase defects react to various stimuli with an exaggerated PRL release, while others exhibit permanently increased serum PRL levels. Inappropriate luteal P secretions is not only caused by PRL elevations, but also by other pathophysiological mechanisms. This is underlined by the fact that bromocriptin treatment resulted in normalization of luteal P output only, when PRL was elevated.