1. Fourth deep lumbrical muscles were dissected out, with their nerve supply, from juvenile rats aged 8-15 days (a period corresponding to maximal rate of decline of polyneuronal innervation), and aged 28, 29 and 30 days (when developmental synapse elimination is complete). Preparations were superfused with rat Ringer solution at 25 degrees C. 2. Isometric twitches and tetani were recorded from the whole muscles and from a single motor unit in the muscle. Unit isolation was by partial section of the sural or lateral plantar nerves. The axon of a single unit occurred naturally in the sural nerve in some cases. 3. Fibres in single units were depleted of glycogen by repetitive stimulation, and studied histologically in frozen midbelly sections of the muscle, stained for glycogen with periodic acid-Schiff's reagent (PAX). Most fibre counts were based on transmittance measurements made with an image analysis system. Contralateral muscles were unstimulated and acted as controls. 4. Motor unit sizes were estimated from tetanic tensions and from muscle fibre cross-sectional area measurements. Comparison of the two methods indicated that in most units glycogen depletion was not complete. This effect mas maximal at 8 and 10 days postnatally. It is suggested that this is due to weak neuromuscular transmission at synapses in the process of natural elimination during development. 5. Other sections (serial and semi-serial) were immunostained with a polyclonal antibody raised against slow myosin. Fibres staining for the antibody (slow; S-fibres) contribute about 12-9% of muscle fibres depending on age. In some muscles, fibre types were determined by myosin ATPase staining following alkali pre-incubation. Fast fibres (F-fibres) contained no slow myosin. 6. Some units had no S-fibres (i.e. they were homogeneous), and many units had a small proportion of X-fibres, though less than in the whole muscle (i.e. they were heterogeneous but composition was biased in favour of F-fibres). 7. One unit from a 10-day-old rat contained more S- than F-fibres. Many of the F-fibres were small. It is proposed that this was a developing II C/II A unit, a type known to occur in adults. 8. It is concluded that mismatched connections in developing motor units possibly become weak early (by 8 days postnatally) in the process of synapse elimination (which is complete by 20 days postnatally); but that the time course of actual withdrawal cannot be followed by the technique of glJrcogen depletion.
