Plots were made of multiunit activity versus ankle joint position for receptors in each of the 12 muscles crossing the cat ankle joint, except peroneus tertius, by recording from populations of afferent fibers in muscle nerves. The discharge was measured 15 or 30 sec after terminating the movements that altered the position of the joint. These recordings were dominated by large-spike activity that would be expected to originate mainly from primary spindle endings. Seven of the 12 muscles also cross other joints. Their responses at a given ankle joint position were so altered by changes in the position of the knee or toe joints that they could not reliably signal the position of the ankle joint. As judged from multiunit recording, receptors in each of the five muscles specific to the ankle joint were influenced by more than one axis of ankle joint displacement. Single-unit recording from dorsal root filaments was used to determine whether primary or secondary spindle receptors in soleus and tibialis anterior could selectively signal one axis of ankle joint rotation. Individual soleus receptors were tested both on the flexion-extension axis and with a combined adduction-eversion movement. For 38 of the 70 soleus receptors examined (54%), firm adduction-eversion produced a level of activity greater than that caused by 10 degrees of flexion, and for 77% the level of activity was greater than that caused by 5 degrees of flexion. For 168 of the 184 tibialis anterior receptors studied (91%), firm abduction-inversion produced a level of activity greater than that caused by 10 degrees of extension. Thus few receptors were found that responded exclusively to one axis of rotation. One way in which the position of the ankle joint could be specified in the face of multiaxial receptor activity is by examining the receptor discharge from more than one muscle. A suggestion for how the nervous system might do this is given in the discussion.