Pre-eclampsia is usually defined on the basis of new onset hypertension and albuminuria developing after 20 weeks of pregnancy. There are difficulties with measurement of these variables. Conventional sphygmomanometry remains the gold standard for blood-pressure measurement. The value of ambulatory blood-pressure measurement has yet to be established. Oedema is now omitted from all definitions of preeclampsia, although the finding of widespread severe oedema of sudden onset should not be ignored for clinical purposes. Definitions of pre-eclampsia based solely on hypertension and proteinuria ignore the wide clinical variability in this syndrome. Women with no proteinuria but who do have hypertension and other features such as severe headache or other symptoms, thrombocytopenia, hyperuricaemia, disordered liver function, and fetal compromise are likely to have pre-eclampsia. This notion is accepted in the new Australasian definition of pre-eclampsia and more than hinted at in the new American College of Obstetricians and Gynecologists' definition. Definitions used for clinical purposes should be as safe as practical; they are likely to include a considerable number of false positives. Most research studies are weakened if patients without the disease are included. Therefore, a separate stringent research definition of pre-eclampsia we also suggest.