In order to understand the clinical and biological implications of prostate cancer multifocality and heterogeneity, we investigated their occurrence in relation to variables such as tumour volume, local invasion, and biopsy findings. In a series of 61 completely sectioned whole-mount radical prostatectomy specimens with clinical stage T2 prostate cancer, we mapped histological grade heterogeneity and tumour multifocality. We also evaluated 55 prostate biopsy cases to assess the accuracy of pre-operative grading. Among all of the prostates, only 28 per cent had a single tumour and in 16 per cent one histological grade of cancer was evident. Extracapsular invasion was not restricted to the largest tumour in each case, but also occurred in tumours of relatively small volume and low histological grade. Variability of histological grade was directly proportional to tumour volume. Both grade heterogeneity and tumour multifocality of the prostatectomy specimen showed no significant relationship to the grade accuracy of biopsies. Biopsy grading error proved greatest among small, well-differentiated, tumours. Whole-mount sectioning of prostatectomy specimens in patients with clinically localized adenocarcinoma demonstrates that grade heterogeneity is most closely related to tumour volume; that the largest (index) tumour lesion may not be representative of the pathological stage; and that grading error in prostate needle biopsies can be only partly explained by grade heterogeneity or tumour multifocality.