1. A method was developed for locating the rostral part of the medial lemniscus in anaesthetized cats and then exploring it with a micro-electrode selective for single axons. Records were made from 165 axons, all shown histologically to lie in the lemniscus.2. Almost all lemniscal axons responded at short latency to a shock through surface electrodes over the dorsal columns at C2. The great majority probably belonged to the dorsal column-lemniscal system, though some may have belonged to other (e.g. spinocervicothalamic) systems.3. Resting discharge was seen in almost all axons in the absence of any stimulation, and must have been generated almost entirely in the relevant relay nuclei, particularly since in many axons it was easily depressed or totally inhibited by appropriately placed mechanical stimulation of skin or a shock to the dorsal columns.4. For each fibre held for an adequate length of time, the receptive field, if accessible, was classified as accurately as possible. Fifty-two axons were precisely categorized in this way: many more were studied for long enough to yield useful information.5. One half (twenty-six) of the best categorized axons had receptive fields suggesting excitation by only one type of receptor: fifteen by tylotrich hairs, four by rapidly adapting tactile foot pad receptors, two by claw movement, two by cutaneous touch corpuscles and three by Type II cutaneous receptors. Rigidly held probes driven by electromechanical transducers were used to establish stimulus/response relations. Adjacent or surround inhibition was seen in nearly all these fields, except for the Type II category.6. The other half (twenty-six) of the best categorized axons showed various degrees of inter-receptive excitatory convergence. Five responded to all types of hair, twelve to hair movement and foot-pad displacement, and nine to hair movement combined with inputs from a variety of slowly adapting receptors in skin or deep tissues, thresholds for the latter ranging from light contact to noxious pressure.7. Forty axons responded with a slowly adapting discharge to joint movement, some with properties suggesting that their receptors did not lie in the joint capsule itself. The high threshold of most of these axons to dorsal column stimulation suggested that the relevant primary axons lay either deep in the dorsal column or in some other tract.8. Of axons whose receptive fields were accurately located, 88% lay in forelimb or upper trunk - the remainder in lower trunk, hind limb or tail. The forepaw accounted for 41% of the former group. Axons with receptively ;pure' properties tended to lie in central or deep parts of the main lemniscal mass at the level studied. Axons responding to joint movement tended to lie deep in the main mass and in the ventromedial lemniscal bundle. There was some clustering of axons with identical receptive properties.