A method is described for the evaluation of the speed of onset of i.v. anaesthesia in mice. The technique involves (a) determination of the median hypnotic dose (HD50) by plotting the probit value of the percentage of mice sleeping, against dose on a logarithmic scale, (b) plotting mean induction time over a range of doses against the logarithm of the dose and (c) comparison of induction times at 1.25 HD50. All doses were given over 1 s or 10 s. A 1–s injection was thought to be of most value in the evaluation of structure activity effects whereas the 10–s injection produced results similar to those which have been reported in man. With this technique, and using 1–s injection times, induction times were found to be similar with thiopentone and Althesin. Those following methohexitone, etomidate and propanidid were marginally longer whereas ketamine and pentobarbitone were obviously slower in onset.