The diagnostic accuracy of two different methods of sectioning rectal biopsies was evaluated prospectively. The frequency of focal and diffuse findings in more than 80 serial sections from a limited portion of the well-oriented areas of 74 biopsies was compared with that found in three or more step sections distributed throughout the well-oriented areas of the same rectal biopsies from patients with known or suspected inflammatory bowel disease (IBD). Significantly more focal abnormalities (granulomas, giant cells, histiocytic collections) were detected by serial sectioning than by step sectioning. For example, serial sectioning increased the ability to detect granulomas by 50% compared to step sectioning. Serial sectioning provided no significant advantage in the detection of diffuse abnormalities (abnormal architecture, decreased mucus, nonspecific inflammation, crypt abscess). Serially sectioning part of a biopsy is superior to step sectioning more of the biopsy in detecting focal abnormalities.