To determine the suitability and estimate the sensitivity of an immunohistochemical (IHC) test for disease-associated prion protein (PrP(Sc)) in biopsy specimens of rectoanal mucosa-associated lymphoid tissue (RAMALT) for diagnosis of scrapie in sheep. 762 sheep at high risk for having scrapie and indemnified by the National Scrapie Eradication Program. The IHC test for PrP(Sc) was applied to 2 RAMALT and 2 third-eyelid biopsy specimens and a postmortem RAMALT specimen from each sheep. Results were compared with those of a reference test in which results for tissues from obex and retropharyngeal lymph nodes, tonsil, or both were considered in parallel. The reference test identified 139 sheep as having scrapie. Biopsy-related complications occurred in 3 sheep. Sensitivity of the IHC test in RAMALT ranged from 85.3% to 89.4%, depending on the anatomic location from which RAMALT was obtained. Results for the test applied to 1 RAMALT specimen were similar to results interpreted in parallel for 2 third-eyelid specimens (sensitivity, 87.0%). The proportion of inconclusive test results attributable to insufficient lymphoid follicles in biopsy specimens was lower when considering results for 2 RAMALT specimens in parallel (10.1%) than when considering results for 2 third-eyelid specimens in parallel (23.7%). Specimens of RAMALT that were inappropriately collected from an area caudal to the rectoanal interface yielded a high proportion of inconclusive results (33.3% to 50.0%). The IHC test for PrP(Sc) in RAMALT was an effective means of detecting subclinical scrapie in live, high-risk sheep.