SUMMARY.-Within 48 hours of the institution of severe phenylhydrazineinduced anaemia in mice bearing ascites tumours or generalised leukaemia, a substantial proportion of the malignant cells disappeared respectively from the peritoneal cavity or infiltrated liver. The results of radiobiological experiments permitting determination of the proportion of viable leukaemia cells which were severely hypoxic and relatively radioresistant in the livers of leukaemic mice, showed that induction of anaemia was associated with a several hundredfold increase in the proportion of such cells. The proportion of hypoxic cells was greatly reduced when the anaemic leukaemic mice were transfused with packed erythrocytes or allowed to breathe oxygen under high pressure. Similar experiments with solid sarcomas indicated that a high proportion of the tumour cells were hypoxic in non-anaemic mice breathing air. The hypoxic fraction was not significantly reduced when tumour-bearing mice were made severely anaemic during growth of the tumour and were later transfused. Thus, the hypoxic cells in leukaemic livers and those in solid tumours are markedly different in their capacity for oxygenation following the induction of relative hyperoxaemia.IT is well known that anaemia is a common complication of human cancer and that a wide variety of factors are implicated in its pathogenesis. In cases in which obvious factors, such as haemorrhage or bone marrow replacement, can be excluded, it is probable that continuous extravasation of blood into the tumour itself is paramount. This process has been demonstrated in several transplanted animal tumours using radio-labelled erythrocytes (Greenfield, Godfrey and Price, 1958). Clinical awareness of the frequency of haematological deficiency, even in cases in which the tumour is quite small in relation to body weight, leads to rectification of anaemia before measures to control the tumour are undertaken.Contrasting with the clinical attentiveness to constitutional depradations associated with localised cancer is the frequent neglect of such considerations in the experimental study of rodent tumours. In our experience, rapidly increasing host anaemia is frequently present when transplanted mouse tumours attain a size at which significant observations of the tumour are likely to be made. For example, terminal depression of tumour growth rate (Laird, 1964) and the microanatomy of extremely large tumours (Inch and McCredie, 1968) have been interpreted without reference to the possible implication of progressive host anaemia in the changes described.