The plasma volume and total red cell mass are controlled by different mechanisms and pregnancy provides the most dramatic example of the way in which that can happen. A healthy woman bearing a normal sized fetus, with an average birth weight of about 3.3 kg, will increase her plasma volume by an average of about 1250 ml, a little under 50% of the average non-pregnant volume for white European women of about 2600 ml. There is little increase during the first trimester, followed by a progressive rise to a maximum at about 34-36 weeks, after which little or no further increase occurs. It seems certain that the frequently observed fall in plasma volume in the last six weeks of pregnancy is an artefact of measurement due to poor mixing of tracer when the woman lies supine and obstructs the circulation to her lower limbs. The maximum increase depends largely on the size of the conceptus. It is somewhat increased, perhaps to a mean of 1300 ml, in association with the bigger baby of multiparae and increases still more with twins, triplets and quadruplets. Red cell mass increases by relatively much less, a rise of about 250 ml (some 18% of the non-pregnant volume) in women who take no supplemental iron, and between 400 and 450 ml when iron supplements are taken. The rise is probably linear from the end of the first trimester to term, and there is some evidence of a preliminary fall in red cell mass during the first trimester. As a result of the relatively much greater increase in plasma volume, red cells in the blood are 'diluted' and the venous haematocrit drops from a non-pregnant average of about 40 to about 33 during the last trimester. The differential changes are biologically plausible: red cell mass rises proportionately to the need to carry the extra oxygen taken up in pregnancy; the greater plasma volume increment is needed to cope with the very large increases in blood flow to organs which require little extra oxygen, the skin and the kidneys.