1. It has been long suggested that central 5HT-mediated systems may be involved in modulation of anxiety and in the anxiolytic effect of benzodiazepines. However, recent evidence has questioned this hypothesis, particularly with respect to the mode of action of benzodiazepines. 2. Development of 5HT agonists and antagonists selective for different 5HT receptor sub-types (5HT1A, 5HT1B, 5HT2, 5HT3) has opened a new avenue for investigation of the potential role of 5HT in anxiety. 3. Buspirone is clinically active in the treatment of anxiety and it, and other anxiolytic candidates, gepirone and isapirone, may act as agonists (or perhaps partial agonists) on 5HT1A receptors. 4. The prototype 5HT1A agonist 8OH-DPAT may also have potential anxiolytic effects. 5HT1A agonists may act to suppress the activity of 5HT neurones as a major part of their action. 5. Although there is some supporting evidence, there is no clear indication of anxiolytic activity with agonists with some selectivity for 5HT1B sites (RU24969, mCPP, TMPP). 6. A selective 5HT2 antagonist, ritanserin, has anxiolytic effects in clinical studies but, like the 5HT1A agonists, does not show a similar profile to benzodiazepines in models of anxiety. 7. This raises the question of clinical predictivity of the various models used. 8. A recently developed 5HT3 antagonist, GR38032F, has been claimed to possess potential anxiolytic activity but its mode of action in this respect requires further elucidation.