In a study of 1529 patients with primary operable breast carcinoma we have assessed the effect of applying both histological grade and tumour type to determine their comparative value as prognostic factors in human breast cancer. The prognostic group the patient was placed in, based on histological type alone, was less accurate than using grade and type together for many tumours. The importance of performing histological grading of ductal/no special type carcinoma (50% of the women in this series) is confirmed in this series. The 10-year-survival varied from 76% for women with grade 1 carcinoma to 39% for those with grade 3 tumours. Some of the 'special types' of breast carcinoma including tubular, tubulo-lobular, invasive cribriform and grade 1 mucinous carcinomas behaved as would be predicted, with a greater than 80% 10-year-survival in this series. Others, including grade 2 mucinous carcinomas, however, behaved less well with a 60% to 80% 10-year-survival. Indeed, many of the histological tumour types including tubular mixed, ductal/no special type, mixed ductal with special type and lobular carcinomas of classical, solid or mixed types showed a variation in behaviour that could not be predicted by typing alone. Histological grade and tumour type, when used together, more accurately predicted prognosis. In multivariate analysis of a larger group of 2658 cases of primary breast carcinomas (including the 1529 study cases) when histological grade, lymph node status and tumour size were entered, grade was the most important factor in predicting for survival.(ABSTRACT TRUNCATED AT 250 WORDS)