AIM--Proliferative activity of tumours reflects their malignant potential. In colorectal adenomas, a subjective impression of the number of mitoses is a criterion often used to assess the degree of dysplasia. Since these subjective impressions of mitotic activity may lack reproducibility, the aim of this study was to perform an objective analysis. METHODS--Mitotic counts were conducted in tissue sections of 59 colorectal adenomas. Of these, 20 showed mild, 20 moderate, and 19 severe dysplasia, according to blind duplicate assessments by two pathologists. Forty three were classified as tubular adenomas and 16 as "villous" adenomas (tubulo-villous and villous). The number of mitoses, both per unit area of epithelium (area weighted mitotic counts, AWMC) and per colonic crypt (mitotic counts per colonic crypt MCCC), was scored in the most dysplastic area within the adenoma. Mitotic figures were counted using a light microscope (ocular x 10, objective x 40, NA 0.75), and the area of the glandular epithelium was measured using an interactive video overlay measurement system. Twenty glands per specimen were assessed. In the intra-observer reproducibility tests, the coefficients of error for the AWMC and MCCC were 4.5% and 7.4% respectively. RESULTS--For the AWMC a significant difference was found between mild and moderate as well as between mild and severe dysplasia, but not between moderate and severe dysplasia. The results of the MCCC showed the same trend, but the differences did not reach a significant level. Furthermore, cases classified as mild dysplasia were found that showed numerous mitoses, while cases classified as severe dysplasia were found with only very few mitoses. No significant difference in AWMC was found between tubular and villous adenomas. Thus the different malignant potential of tubular and villous adenomas was not reflected by a difference in AWMC. A seemingly strong difference for MCCC between tubular and "villous" adenomas appeared to depend completely on the difference in crypt size between these two groups. CONCLUSIONS--The area weighted mitotic count, rather than the mitotic count per colonic crypt, may be useful for assessing the proliferation rate in colorectal adenomas.