Staging of rectal carcinoma before surgical treatment was performed in a prospective blind study, comparing digital rectal exploration and transrectal linear ultrasonography (TRUS) with the resulting pathological examination. TRUS underestimated the depth of penetration in three of 33 patients and overestimated it in nine of 74. The figures for digital examination were five of 18 and 20 of 76, respectively. Penetration of the rectal wall was correctly identified in 56 of 61 patients by digital rectal examination and in 59 of 61 by TRUS. Specimens without penetration of the rectal wall were identified in 26 of 33 patients by TRUS, but in no more than 13 of 33 by digital examination. Regional lymph node metastases were present in 19 patients; none were diagnosed by digital examination, but TRUS identified 11 of the 19. It is concluded that TRUS will result in more patients being given the possibility of curative local surgery.