Motor-unit size was measured by tension recording in neonatal (3-5 day) rat skeletal muscle (fourth deep lumbrical muscle). Each unit was then depleted of glycogen and its fibres studied in mid-belly frozen sections, by staining for glycogen (periodic acid-Schiff reagent and antibody labelling for slow myosin. The contralateral muscle acted as control, and further controls for the method are described. 2. All the motor units contained both slow-myosin-containing (S; antibody-positive) and slow-myosin-free (F; antibody-negative) fibres. 3. The proportion of each unit that was made up of S fibres was compared with the whole muscle. Of the twelve units studied seven were not selectively innervated, four may have been selectively innervated in favour of F fibres, and one was selectively innervated in favour of S fibres. The last unit was much smaller than the others. 4. Fibre cross-sectional areas were measured in units and in the whole muscles. Mean cross-sectional areas for individual F fibres in all the motor units were smaller than in the corresponding whole muscles (ratio 0.71), implying that small fibres have higher levels of polyneuronal innervation than larger ones (each small fibre occurring in more overlapping units than each larger fibre). There was no such difference in S fibres (ratio 0.96). 5. Motor-unit sizes (as a percentage of whole muscle) were smaller when obtained from summed fibre cross-sectional areas than from fibre counts (this follows from 4, above). Comparisons with unit sizes from tension recording are discussed. 6. Controls show that there is little, if any, non-specific fatigue of muscle fibres that are not part of the unit subjected to glycogen depletion. 7. Evidence is given that muscle fibre conduction block occurs during the depletion regime, leading to less glycogen depletion towards the ends of the muscle fibres than in the end-plate zone.
