In this study, colonic biopsy specimens from 87 subjects (37 men and 50 women; mean age 51 +/- 14 years) were investigated by a new method of quantitation morphometric analysis. Subjects were classified in 3 groups: I, control group (25 cases); II, constipation and/or abdominal pain (36 cases); III, diarrhea (26 cases). All biopsy specimens were considered normal by both colonoscopic and microscopic observations. Morphometric analysis was performed with a specialized automated image processor. Complete analysis of 1 biopsy specimen required about 10 minutes. Glandular area showed no difference in the ascending, transverse, and descending colonic sites in the three groups. By contrast, the rectal glandular area was significantly larger than those from the 3 colonic sites in all groups (p less than 0.001). The mean cellular density in the control group was 172.7 +/- 24 nuclei/unit area of lamina propria at all biopsy sites. The cellular densities of groups II and III were significantly increased in the ascending colon as compared with the group I subjects (195.1 +/- 22.8, p less than 0.01 and 198.2 +/- 24.7 p less than 0.001, respectively). The cellular density in group III was significantly increased in the other sites as compared with group I (p less than 0.01). Morphometric analysis does provide a definition of the normal range of the lamina propria cellularity and can identify mild abnormalities in patients in whose both colonoscopy and conventional histologic interpretation have been considered as normal by experienced endoscopists and pathologists. A statistical increase in cellular density in groups of patients with clinical symptoms should be regarded as consistent with the concept of microscopic colitis and detectable by morphometry. Such a method reduces interobserver bias and permits objective comparison of groups.