Summary The regional distribution of blood flow to the LBDS, fibrosarcoma, transplanted into the subcutaneous site in rats, was investigated using the readily diffusible compound '4C-iodo-antipyrine ('4C-IAP). Quantitative autoradiography was used to establish absolute values of specific blood flow F for 100 x 100 x 20 1sm adjacent tissue volumes of the unperturbed tumour. Mean blood flow to whole tumours was found to decrease with increase in tumour size. This relationship was abolished if blood flow was only measured in sections cut from the periphery of the tumours. Detailed analysis of a sub-group of tumours showed that blood flow to individual tumours was heterogeneous. The range of blood flow was large, indicating that mean blood flow to a whole tumour is a poor reflection of the blood perfusion pattern of that tumour. Necrotic tumour regions were usually very poorly perfused. With the exception of the smallest tumours studied, blood flow was lower in the centre of tumours than in the periphery. Necrosis also tended to develop centrally. However, the peripheral to central gradient of blood flow was apparent even when densely cellular, viable tumour regions and necrotic regions were analysed separately. The decrease in blood flow with tumour size was also apparent in densely cellular, viable tumour regions when analysed separately. Qualitative comparison of tumour histology and regional blood flow showed that there were areas of very low blood flow associated with viable tumour regions. Less common were areas of rather high blood flow associated with necrotic tumour regions. A complicated relationship exists between tumour histology and blood flow. The quantitative autoradiography technique is suitable for investigating the most poorly perfused and the most well perfused viable fractions of animal tumours which may limit the efficacy of different types of therapy.The blood flow to tumours plays an important role in their treatment. In radiotherapy, the efficacy of treatment depends upon cellular oxygen concentration which is governed by respiration rate and local blood flow. In chemotherapy, blood flow determines the efficiency of drug delivery. In hyperthermia, the temperature elevation achieved and the sensitivity of cells to heat is influenced by blood flow. Blood flow is therefore a critical parameter to measure both experimentally and clinically.It is known that blood flow to both transplanted animal tumours and to human tumours is heterogeneous (Chaplin et al., 1987;Ito et al., 1982;Vaupel & Frinak, 1980). Therefore global measurements may not be the most pertinent means of describing tumour blood flow. For instance, global measurements will include blood flow to necrotic tumour regions which are not relevant for therapy. For hyperthermia, a knowledge of high blood flow in discrete tumour regions would be important because these regions are the ones most likely to limit efficacy of treatment.Blasberg and colleagues used the inert, readily diffusible compound iodo-antipyrine to measure blood flow to tr...