Case material and retrospective studies support the use of both lorazepam and ECT in treating catatonia, but few prospective investigations exist and none employ quantitative monitoring of response. In this study we test their efficacy in an open, prospective protocol, and define a "lorazepam test' with predictive value for treatment. Twenty-eight patients with catatonia were treated systematically with parenteral and/or oral lorazepam for up to 5 days, and with ECT if lorazepam failed. Outcome was monitored quantitatively during the treatment phase with the Bush-Francis Catatonia Rating Scale (BFCRS). In 16 of 21 patients (76%) who received a complete trial of lorazepam (11 with initial intravenous challenge), catatonic signs resolved. A positive response to an initial parenteral challenge predicted final lorazepam response, as did length of catatonic symptoms prior to treatment. Neither demographic variables nor severity of catatonia predicted response to lorazepam. Four patients failing lorazepam responded promptly to ECT. It is concluded that lorazepam and ECT are effective treatments for catatonia. The rating scale has predictive value and displays sensitivity to change in clinical status.