1. Using large ramp and triangular stretches a survey has been made of the effect of stimulating single gamma fusimotor fibres on primary endings of muscle spindles in the peroneus brevis to see whether 'intermediate' types of fusimotor action could be recognized, falling between the well known static and dynamic types. 2. Responses were classified into six groups, as detailed on pp. 844-846, ranging from apparently 'pure' dynamic action (category I) to apparently 'pure' static action (category IV). Models for a putative mixed action were produced by combining the stimulation of a static and of a dynamic fibre to the same spindle. The clearest sign of static action was firing on the releasing phase of the stretch. The essential sign of dynamic action, which survived combination with the more dominant static action, was the low adaptive decay of firing with a time constant of about 0-5 sec that occurs on the plateau of the ramp stretch. 3. Out of 153 responses, each elicited from a primary ending on stimulation of a single fusimotor fibre, 67% were apparently 'pure' examples of dynamic and static action. The remaining 33% of responses were to some degree suggestive of an admixture, in various proportions, of static and dynamic actions. For only 18% of them was there firm indication of such admixture. 4. When a given fibre was tested on more than one ending then, with one exception out of thirty-six instances, its action always proved to be either predominantly static or predominantly dynamic. There was no special tendency for an axon with a mixed action on one spindle to have a similarly mixed action on other endings so that individual fusimotor fibres were best classified as static or dynamic without intermediate grades. 5. Simultaneous stimulation of two fusimotor fibres eliciting apparently 'pure static and dynamic actions, could mimic all the intermediate types of action. 6. The results are discussed in relation to recent studies, especially those based on glycogen depletion. It was concluded that dynamic action arises from activation of the bag1 intrafusal muscle fibre, and that static action arises from the bag2 and chain fibres, whether acting individually or collaboratively. The intermediate actions are suggested to arise from an overlap of motor innervation to contrasting types of intrafusal muscle fibre. 7. On the basis of effects on the regularity of the afferent discharge the findings support the view that a given static action axon can innervate bag2 and chain fibres in various proportions in different spindles, so that they do not provide separable effector pathways. 8. Responses to large amplitude sinusoidal stretching were also studied in relation to our classification.