This paper reports on the phenomenon of heterogeneity of DNA distribution in colorectal carcinoma. The aim of this study was to find a new strategy in sampling methodology as a solution to the heterogeneity problem by studying ploidy topography. The study was carried out by using image analysis for densitometric measurements of tissue imprints of 19 colorectal specimens after Feulgen staining. Heterogeneity of ploidy in colorectal carcinoma is well known; therefore, we wished to determine whether this heterogeneity is random within different parts of a colorectal tumour. For each tumour, five systematic and reproducible samples were taken from the peripheral, intermediate and central tumoural areas. In addition two samples were taken from adjacent non-tumoural areas: one from the proximal and the other from the distal site with respect to the neoplastic lesion. By using image analysis three parameters were obtained. Mean DNA content was computed for each sample and expressed in arbitrary units (DNA-a.u.) from the measurement of integrated optical density according to the Beer Lambert law. Secondly mean DNA content expressed in relative units (DNA-r.u.) was computed according to an internal euploid control. Finally entropy was computed from each histogram of DNA content. Experimental design was based on a repeated measures analysis of variance with a priori orthogonal comparisons. We found that DNA content and particularly entropy are not randomly distributed. Furthermore, there was a significant difference between the two non-tumoural epithelia.