EDITORIALBiochemical hypotheses of premenstrual tension syndrome 1 As information has evolved concerning the incidence of cyclical emotional changes preceding the onset of menses, the premenstrual tension syndrome (PMT) has gained increased attention in both the lay and professional literature. Although there has been some disagreement as to the exact nature, the time of occurrence, or even the existence of PMT (Koeske, 1981), there is considerable consensus that the PMT syndrome indeed exists (Janowsky et al. 1966(Janowsky et al. , 1967(Janowsky et al. , 1969(Janowsky et al. , 1973. Steiner et al. (1980) and Halbreich & Endicott (1982) have now operationalized psychodiagnostic criteria for PMT which can allow separate studies to attain diagnostic consensus more accurately.Several methodological problems in PMT research have limited exploration of how premenstrual physiology might influence mood. A major problem has been the strong reliance on the use of correlative techniques. Many hormones,, electrolytes, neurotransmitters and somatic parameters change within the menstrual cycle, and a tradition has existed for assuming the aetiological importance of a favourite correlate which changes premenstrually. Although such correlations are a reasonable place to start, it may be naive to believe that temporal correlations necessarily imply aetiology.Another problem in PMT research is the general assumption that circulating ovarian-linked hormones as such are of fundamental significance in triggering and sustaining premenstrual emotional symptoms. It is equally possible that central neurotransmitters, neuromodulators, and neuroendocrine mechanisms control the onset and intensity of PMT, and that peripheral ovarian and other hormonal changes are merely simultaneous epiphenomena, and/or are driven by the above central influences, but do not actually exert behavioural effects.The use of retrospective reports in PMT research has caused additional problems, with exaggerated incidences being reported, probably due to societal expectations (Koeske, 1981; Parlee, 1973;Sommer, 1978). In contrast, the incidence of PMT symptoms reported usually is relatively low in prospective PMT studies. Thus, many subjects who report PMT do not show symptoms in actual experiments. Also, since the intensity of PMT may vary from month to month (Green, 1982), such variation may introduce serious inconsistencies and variance problems into research studies. One recent step forward in this area has been the administration of repeated mood and behavioural ratings to prospective subjects over several cycles to screen for and document cyclic emotional changes before a given subject is used in a specific PMT study.Also, there is some evidence that the stress of the experimental method itself may influence and confound studies exploring menstrual cycle-behavioural relationships (Koeske, 1981;Parlee, 1982). Conversely, a major impediment to understanding and treating PMT has been its remarkable sensitivity to treatment by placebo. Many promising aetiol...