Endometrial biopsy specimens (n = 62) were evaluated by five pathologists to assess the effect of interobserver variation on histologic dating of the endometrium. The potential effect of this variation on the diagnosis of luteal phase defects (LPDs) and resulting clinical management was also determined. Mean (+/- standard error) interobserver variation was 0.96 +/- 0.08 days, comparable to results reported by other investigators. The magnitude of the variation was not affected by whether the biopsy specimen was obtained in the mid or late luteal phase, the degree of lag between the dating and subsequent menses, or the presence of an LPD. Redating of a specimen by another pathologist would have resulted in a change in the determination of "in" or "out" of phase in 22% of cases. The subsequent probability of changing patient management altered ranged from 22% to 39% depending on the clinical setting.