Two forms of hypoxia are thought to exist in tumours: (1) hypoxia caused by limitations of its diffusion (chronic hypoxia); and (2) hypoxia caused by changes in perfusion (acute hypoxia). Indirect information suggests the existence of perfusion-limited hypoxia, but there is no direct proof that fluctuations in blood flow can lead to hypoxia, nor is there any information regarding potential causes of fluctuant flow. In this study, we have begun to explore these questions using R3230AC tumours transplanted into rat dorsal-flap window chambers. Two types of fluctuant flow have been observed. The first type, usually confined to single vessels, is typified by instability of flow magnitude and direction, and total vascular stasis occurs, but only for a few seconds at a time (4% incidence). The second type of fluctuation occurs in groups of vessels and is cyclic, with cycle times ranging from 20-60 min. Total vascular stasis does not necessarily occur, but the fluctuations in red cell flux are accompanied by changes in vascular oxygen content, as measured by microelectrodes. Another source of chronic hypoxia has also been identified in these experiments. Nine per cent (9%) of vessels examined had plasma flow, but very low or absent red cell flux over periods of many minutes.