Summary (1) Using the tarachograph developed by Corti et al., a study was made of postural tonus during normal waking and sleeping in 10 adult cats of both sexes. In order to distinguish the awake from the sleeping condition, the study was combined with EEG and electro-oculogram recordings, assessment of the awakening threshold, and inspection of posture. (2) On the base of those criteria distinguishment was made between dozing, orthodox and paradoxical sleep. During the whole sleeping period of about 2 h the paradoxical phase was interrupted up to 15 times by orthodox phases. The paradoxical sleep of anxious cats alternated in the same way with spells of dozing. In all instances, paradoxical sleep preceded spontaneous awakening. If the animal was roused from sleep, the ‘awakening threshold’ was seen to rise during the first 2–3 sleep cycles, and was higher during the paradoxical phase than during the orthodox or the dozing phase. The higher the ‘awakening threshold’, the more quickly did the animal fall asleep again. (3) The seismographic recording of minor tremor in the postural muscles showed that minor tremor rate ranged between 8–13 c/sec, with predominance of 10 c/sec, during all stages of wakefulness or sleep. It never changed its frequency. In contrast to this the amplitude of the minor tremor was greatest during quiet wakefulness and smallest during the paradoxical-orthodox sleep cycles, respectively, during the paradoxical sleep and dozing cycles of the anxious animals. Between paradoxical and orthodox phase, respectively, between paradoxical phase and cyclic dozing the amplitude showed no difference. But there was a change depending on respiration. The amplitude was increasing with inspiration and diminishing with expiration. (4) A correlation could be established between the awaking threshold and postural tonus during orthodox sleep and dozing. In contrast, no such relationship could be established with regard to paradoxical sleep (associated with high awakening threshold). (5) It is suggested that the relative constancy of the minor tremor rate, amounting to about 10 c/sec, is due to the contraction time (approx. 100 msec) of the various tonic motor units subserving skeletal muscle tone. 10 c/sec would represent the net effect of asynchronously occurring contractions of 100 msec duration. This would imply no direct relation between minor tremor rate and the rate of discharge of the spinal motoneurons. Minor tremor amplitude, however, appears to depend upon the number of motor units activated.