Aim: Abnormally functioning immunocompetent cells in the endometrium are thought to be responsible for at least some cases of recurrent reproductive failure [recurrent miscarriage or recurrent in vitro fertilisation (IVF) failure], but their detailed investigation has been hampered by a lack of a standardised protocol of counting such cells in study or control patients. The purpose of this study is to use a standardised protocol for the assessment of immune cells in the endometrial biopsies of a large cohort ofwomen with recurrent reproductive failure and establish relevant reference ranges. Method: In a recent study, we reported the presence and distribution of selected immune cells and macrophages in the endometria from 222 women who had a routine endometrial biopsy for investigation of recurrent miscarriage or IVF failure. Since the completion of that study, a further 1767 cases have been examined, using the same assessment parameters of the earlier study. Results: This updated analysis of 1989 endometrial biopsies provides reference ranges for CD8 +, CD163+, CD56+ and CD57+ cells for individual 'days' of a normalised menstrual cycle. CD8+ T-cells displayed a modest (50%) increase in numbers in the luteal phase and periglandular aggregation was a useful indicator of a subtle focal endometritis, possibly of infective origin, and generally not identified in H&E sections. A rapid accumulation of CD163+ macrophages occurs in the superficial stroma after day 22 of the cycle, while a significant number of cases displayed single or clustered macrophages within glandular lumens of the superficial endometrium in luteal phase, especially after day 20 of the cycle. The significance of this change is unclear but may relate to a macrophage response to abnormal glandular secretion or to bleeding occurring at the time of ovulation. CD56+ uterine natural killer (uNK) cells show such a dramatic rise in both absolute numbers and percentage of stromal cells from day 22 of the standardised 28 day cycle that this needs to be taken into account in all clinical studies or individual assessments of endometrial biopsies. CD57+ NK cells are seen in small numbers in most cases and cell counts of greater than 10 per mm2 are regarded as abnormal. Conclusions: This large database provides a daily range which is the most accurate survey yet of uNK cell numbers. Co-location of CD8+ T-cells and CD56+ uNK cells in perviascular aggregates has been demonstrated.