The range in the sizes of the synaptic areas between the various axons seemed to be a continuous

distribution with no obvious steps between those with large areas and those with small areas (Figure 4D). Previous work showed that over time, as the dominant axon comes to occupy most of the neuromuscular junction site, it comes to have a larger axon caliber than the axons that are in the process of being eliminated (Keller-Peck et al., 2001 and Walsh and Lichtman, 2003). Interestingly, we find here that even at birth, the axons with the most synaptic contact have the largest axonal caliber at the entrance site Protein Tyrosine Kinase inhibitor of the junctions (Figure 4E). Therefore, the axon’s caliber at the neuromuscular junction entrance site in newborns is an excellent measure

of the area of overlap with AChRs and strongly correlates with the number of contact Trichostatin A concentration sites. The small area of contact of virtually all motor axon inputs (area of contact ranged from 10%–30% of the AChR plaque) suggests that many are too weak to bring the muscle fiber to threshold, consistent with physiological evidence of low-quantal-content neuromuscular axons in the perinatal period (Colman et al., 1997 and Kuno et al., 1971). Subthreshold axonal inputs would be invisible to postsynaptic activity-based assays such as glycogen depletion or muscle tension, explaining the disparity between these results with physiological measures of motor unit size (see Discussion). The large number of converging axons raised the possibility that at birth, muscle fibers were innervated by a substantial fraction or perhaps even all of the axons that innervated the region of muscle they resided in. As already described (Figure 3), Phosphoprotein phosphatase in some muscles, axons project to a limited region of the endplate band at birth just as they do in later life. From axonal

reconstructions at postnatal day 8 from a previous study (Keller-Peck et al., 2001), we analyzed the area of the endplate band occupied by single motor units and found that, on average, axons in the sternomastoid muscle occupied ∼18% (0.42 ± 0.12 μm2, n = 6) of the endplate band area in the muscle as a whole. Because there are in the range of 50–60 primary motor axons innervating the sternomastoid muscle (Nguyen et al., 1998), we anticipate that 18% of these or 9–11 motor axons should project to any one region. This number roughly matches the number of innervating axons per junction at birth, suggesting that, at least in some cases, all the motor axons within the vicinity of a muscle fiber innervate it at birth. Hence, we found no evidence for any synaptic selectivity in the initial innervation pattern as might have been expected if axons preferentially innervated muscle fibers of a particular type.