Summary— This review focuses on the history of investigations into the behavioural reaction resulting from excess stimulation of postsynaptic 5-hydroxytryptamine receptors and the relative risk of this occurring with different combinations of drugs. Other aspects, particularly treatment with 5-hydroxytryptamine receptor antagonists, are reviewed in a recent separate paper . The first human case was in 1955 and animal work had defined the characteristic features by 1958, and established they were lessened by chlorpromazine. Substantial evidence of a ‘dose-effect’ relationship existed by 1984. The relative risk with different drug combinations is assessed from available evidence and argued to be strongly associated with the degree of elevation of 5-hydroxytryptamine, which is greatest following combinations of irreversible inhibitors of monoamine oxidase A and B with potent serotonin reuptake inhibitors. The various serotonergic drugs that may be implicated in serotonin syndrome are tabulated and discussed in relation to the relative risk. It is suggested that the proposed ‘diagnostic criteria’ for serotonin syndrome are inappropriate since there is a continuous spectrum from side effects to toxicity. The term ‘serotonin syndrome’ may encourage the presumption that it is an idiosyncratic response, as neuroleptic malignant syndrome is usually considered to be. The terms ‘toxic serotomimetic reaction’ or ‘toxic serotonin syndrome’ may be preferable alternatives. The differences between serotonin syndrome and neuroleptic malignant syndrome are highlighted with examples from difficult or questionable cases in the recent literature. It is proposed that more systematic national collection of toxicity data is essential in order to quantify the relative risk of serotonin syndrome with various combinations of serotonergic drugs.