This study reviews the literature on the premenstrual syndrome specifically prevalence, definitional issues, etiological assumptions, and methodological issues. The conclusion is reached that the area requires the input of further well-controlled research. The technique of self-report is also reviewed with particular emphasis on applications, methodological advantages and limitations. The particular relevance of self-report to the study of the premenstrual syndrome is addressed as is the validity of self-report in the area. The first aim was to collect sufficient daily mood data from target subjects to determine correlations with their menstrual cycle. Secondly, data was collected from significant other observers and cyclicity was measured. Lastly, the relationship between target-rated and observer-rated mood was analysed for commonly occurring cyclicity. The statistical technique involved was Spectral Analysis. Thirteen subjects and observers were involved and results showed clear evidence of PMS in three subjects confirmed by observer recordings. Lack of coherence in cyclicity between target mood and observer rated mood was shown. It was concluded that 1) some of these women may not have been suffering from PMS and that 2) lack of coherence between target and observer on mood ratings was due to the 'private' nature of mood or to poor observation.