BACKGROUND New sleep disturbance model in rats is useful for estimating the characteristics of some hypnotics. OBJECTIVE The present study was undertaken to investigate the utility of a sleep disturbance model by placing rats on a grid suspended over water using three kinds of hypnotics, that is, short-acting benzodiazepine (triazolam), intermediate-acting benzodiazepine (flunitrazepam) and long-acting barbiturate (phenobarbital). METHODS Electrodes for measurement of EEG and EMG were implanted into the frontal cortex and the dorsal neck muscle of rats. EEG and EMG were recorded with an electroencephalogram. SleepSign ver.2.0 was used for EEG and EMG analysis. Total times of wakefulness, non-REM and REM sleep were measured from 0900 to 1500 hours. RESULTS In rats placed on the grid suspended over water up to 1 cm under the grid surface, not only triazolam but also flunitrazepam and phenobarbital caused a shortening of sleep latency. Both flunitrazepam and phenobarbital were effective in increasing of total non-REM sleep time in rats placed on sawdust or the grid, and the effects of both drugs in rats placed on the grid were larger than those in rats placed on sawdust. Measurement of the hourly non-REM sleep time was useful for investigating the peak time and duration of effect of the three hypnotics. Phenobarbital showed a decrease in total REM sleep time in rats placed on the grid, although both triazolam and flunitrazepam were without effect. CONCLUSIONS The present insomnia model can be used as a sleep disturbance model for testing not only the sleep-inducing effects but also the sleep-maintaining effects including non-REM sleep and REM sleep of hypnotics.