1. Isometric twitch and tetanic tensions were recorded from whole muscles and single motor units in isolated fourth deep lumbrical muscles from neonatal rats (most at 3-5 days old) and from older rats of various ages. 2. Whole-muscle time to peak contraction reduced from about 120 ms at birth to about 20-25 ms at 20 days and older. 3. The number of motor units in the muscle was constant with age (eleven on average) and there was no significant branching of motor axons below the common peroneal nerve branching point in the thigh. 4. In the 3-5 days age range, mean twitch:tetanus ratio for whole muscles was 0.299 and for single units was 0.177. As a consequence, mean motor unit size (as a percentage of whole-muscle tension) was greater for tetani (29.7%) than for twitches (19.9%). This was not the case in muscles from animals 22 days or older. Evidence is given that the cause of this is low junctional efficacy in some neuromuscular junctions in neonatal muscle. Intracellular recordings supported this view. 5. The relationships of motor-unit size to the contraction time, to the ratio of contraction time:half-relaxation time, and to fatigue index are given. There was no indication of clear segregation of motor units into more than one population, but it is concluded that small motor units are more likely to contain a higher proportion of slowly contracting, fatigue-resistant fibres than large units. 6. The level of overlap by axons in the lateral plantar nerve onto muscle fibres in a single sural nerve motor unit was greater in tetani than in twitches. The results indicate that the distribution of weak and strong inputs was not random, but that there was a tendency for one strong input to accompany a number of weak inputs (on average about two) on each muscle fibre. 7. Intracellular recording indicates that about 12% of fibres at 3-5 days may be electrically coupled.