Two important qualities of controlled clinical trials are that they reduce dependence on historical standards for evaluating therapy and separate the effect of treatment from the confounding influence of time. Whatever the theory of the clinical trial, however, time has not easily been banished from the analysis of medical experiments in practice, and many doctors still prefer to evaluate treatments by comparing now with then rather than treatment with control. This is especially so in trials of bronchodilators in the treatment of asthma where, it is shown, baselines are used in ways which can increase the variability of measurement and even bias the results.