Clinical studies to assess the benefits of blood transfusion or haemodilution in critical illness should take account of measured Circulating Blood Volume (CBV) before, during and after intervention. Surrogate measures of CBV are inadequate and studies based on these must be considered Incomplete, because they cannot distinguish between effects of changes in haemoglobin concentration and changes in blood volume.The choice of a suitable technique for measuring CBV depends on the facilities available locally. In general, methods based on labelled red cells are more reliable but are technically demanding and time consuming. Those based on albumin are likely to yield false high values and this is particularly true in all patients with impaired capillary integrity. The most promising plasma marker is hydroxyethyl starch which may be particularly useful when the polysaccharide is labelled with a fluorescent dye. Attaching fluorescein to hydroxyethyl starch is not difficult and, should demand be sufficient, it may well become available from manufacturers who are already capable of providing other fluorescent polysaccharides. The clinical benefits of such a development would include more rational schedules of i.v. fluid and blood transfusion management in surgical and intensive care patients.