One hundred and twenty unpremedicated patients undergoing gynaecological surgery were randomly allocated to one of three equal treatment groups to assess the effectiveness of ethyl chloride in producing instant skin anaesthesia to prevent the pain of venepuncture from a 20 G cannula. They received either no anaesthetic, 0.2 ml one per cent lidocaine plain intradermally or a ten-second spray of ethyl chloride at the cannulation site. Ethyl chloride produced skin anaesthesia that significantly reduced the pain of venepuncture. However, it was not as effective as intradermal lidocaine. It had no effect on vein visualisation or ease of cannulation. Ethyl chloride can be recommended as a method of producing instant skin anaesthesia.