Synaptic competition and the elimination of polyneuronal innervation follwoing reinnervation of adult frog sartorius muscles

The elimination of polyneuronal innervation (synapse elimination) that occurs following reinnervation was studied in sartorius muscles of adult Rana pipiens. The percentage of neuromuscular junctions that were polyneuronally innervated declined from 47% at 40–80 days after nerve crush to 22% at greater than 250 days after nerve crush. We measured the size, synaptic strength, and position of competing nerve terminals at identified dually innervated neuromuscular junctions at these two different periods of synapse elimination. Our goal was to determine if any of these parameters play a role in the competition between nerve terminals that ultimately results in the elimination of polyneuronal innervation. Our data support the hypothesis that polyneuronal innervation will persist if competing nerve terminals are of similar synaptic efficacies but will be eliminated if the competing terminals are of different synaptic efficacies. We also tested, but failed to find any evidence, that the spatial proximity of competing nerve terminals at the same synaptic site influences the elimination of polyneuronal innervation.     

Synaptic competition and the elimination of polyneuronal innervation following reinnervation of adult frog sartorius muscles

The elimination of polyneuronal innervation (synapse elimination) that occurs following reinnervation was studied in sartorius muscles of adult Rana pipiens. The percentage of neuromuscular junctions that were polyneuronally innervated declined from 47% at 40-80 days after nerve crush to 22% at greater than 250 days after nerve crush. We measured the size, synaptic strength, and position of competing nerve terminals at identified dually innervated neuromuscular junctions at these two different periods of synapse elimination. Our goal was to determine if any of these parameters play a role in the competition between nerve terminals that ultimately results in the elimination of polyneuronal innervation. Our data support the hypothesis that polyneuronal innervation will persist if competing nerve terminals are of similar synaptic efficacies but will be eliminated if the competing terminals are of different synaptic efficacies. We also tested, but failed to find any evidence, that the spatial proximity of competing nerve terminals at the same synaptic site influences the elimination of polyneuronal innervation.     

Elevated levels of polyneuronal innervation persist for as long as two years in reinnervated frog neuromuscular junctions

When the nerve to an adult frog sartorius muscle is crushed, and axons are allowed to regenerate, the level of polyneuronal innervation at reinnervated neuromuscular junctions is higher than normal. With time, much of this polyneuronal innervation is reduced by the process of synapse elimination (Werle and Herrera, 1988). Using intracellular recording, we estimated the level of polyneuronal innervation in adult frog (Rana pipiens) sartorius muscles 2 years (range: 1.7-2.4 years) after crushing the sartorius nerve. We found that 27% (S.E. = 1.4%) of the junctions in muscles 2 years after reinnervation were polyneuronally innervated, whereas only 10% (S.E. = 1.2%) of the junctions in normal frog muscles were polyneuronally innervated. Thus, the synapse elimination that occurs following reinnervation does not restore the normal level of polyneuronal innervation. Histological comparisons of junctional structure between muscles 2 years after reinnervation and normal muscles revealed substantial differences. Reinnervated junctions had a greater length of synaptic gutter apposed by nerve terminal processes, more axonal inputs, more empty synaptic gutter, more instances of single synaptic gutters innervated by more than one axon, and longer lengths of nerve terminal processes that connect synaptic gutters within a junction. On the basis of this physiological and anatomical evidence, we conclude that nerve injury causes persistent changes in the pattern of muscle innervation.     

Synaptic competition and the persistence of polyneuronal innervation at frog neuromuscular junctions

Mechanisms governing the elimination of polyneuronal innervation were examined by correlating the morphology and physiology of competing nerve terminals at identified dually innervated neuromuscular junctions in sartorius muscles of adult frogs (Rana pipiens). Synaptic efficacy (endplate potential amplitude per unit nerve terminal length) was presumed to reflect the ability of a terminal to compete for synaptic space. The synaptic efficacies of two terminals at the same synaptic site were found to be surprisingly equal, with a median difference of 33%. Much more variation would be expected if dually innervated junctions were randomly innervated by pairs of terminals having the same range of synaptic efficacy as that found at singly innervated junctions in the same muscle. This finding supports the hypothesis that the weaker input is eliminated from dually innervated junctions when there is a large discrepancy in competitive efficacy, and that both inputs may persist if competitive efficacies are relatively equal. We also tested but failed to find support for the hypothesis that spatial proximity between competing terminals intensifies competition for synaptic space during synapse elimination.     

Mechanisms of elimination, remodeling, and competition at frog neuromuscular junctions

Mechanisms governing synapse elimination, synaptic remodeling, and polyneuronal innervation were examined in anatomical and electrophysiological studies of frog neuromuscular junctions. There was a substantial level of polyneuronal innervation in adult junctions and this varied seasonally. Nerve terminal retraction and synapse elimination occurred during normal growth and following reinnervation. Synapse elimination was not inevitable, however. Repeated in vivo observations of some identified junctions showed that polyneuronal innervation could persist for over a year, while at other junctions it arose de novo by terminal sprouting. We concluded that polyneuronal innervation in adult muscles was governed by an equilibrium between processes of retraction and elimination on one hand, and sprouting and synaptogenesis on the other. Other observations revealed that structural remodeling was a common feature of adult junctions. Most often, remodeling involved the simultaneous growth and retraction of different parts of the same junction. The net result was usually junctional growth that, in small frogs, appeared to provide a good match between synaptic size and the electrical demands of transmission. In larger animals, pre- and postsynaptic sizes were not as well matched, providing morphological evidence for a growth-associated decline in synaptic efficacy. Finally, electrophysiology was used to describe some of the functional correlates and consequences of competitive interactions between the terminals of different axons. These results are explained by a hypothetical mechanism that involves trophic support provided by the muscle to the motoneuron, the overall level of nerve-muscle activity, and the synchrony of pre- and postsynaptic activity.     

Activity and synapse elimination at the neuromuscular junction

The neuromuscular junction undergoes a loss of synaptic connections during early development. This loss converts the innervation of each muscle fiber from polyneuronal to single. During this change the number of motor neurons remains constant but the number of muscle fibers innervated by each motor neuron is reduced. Evidence indicates that a local competition among the inputs on each muscle fiber determines which inputs are eliminated. The role of synapse elimination in the development of neuromuscular circuits, other than ensuring a single innervation of each fiber, is unclear. Most evidence suggests that the elimination plays little or no role in correcting for errant connections. Rather, it seems that connections are initially highly specific, in terms of both which motor neurons connect to which muscles and which neurons connect to which particular fibers within these muscles. A number of attempts have been made to determine the importance of neuromuscular activity during early development for this rearrangement of synaptic connections. Experiments reducing neuromuscular activity by muscle tenotomy, deafferentation and spinal cord section, block of nerve impulse conduction with tetrodotoxin, and the use of postsynaptic and presynaptic blocking agents have all shown that normal activity is required for normal synapse elimination. Most experiments in which complete muscle paralysis has been achieved show that activity may be essential for the occurrence of synapse elimination. Furthermore, experiments in which neuromuscular activity has been augmented by external stimulation show that synapse elimination is accelerated. A plausible hypothesis to explain the activity dependence of neuromuscular synapse elimination is that a neuromuscular trophic agent is produced by the muscle fibers and that this production is controlled by muscle-fiber activity. The terminals on each fiber compete for the substance produced by that fiber. Inactive fibers produce large quantities of this substance; on the other hand, muscle activity suppresses the level of synthesis of this agent to the point where only a single synaptic terminal can be maintained. Inactive muscle fibers would be expected to be able to maintain more nerve terminals. The attractiveness of this scheme is that it provides a simple feedback mechanism to ensure that each fiber retains a single effective input.     

Synapse elimination occurs late in the hormone-sensitive levator ani muscle of the rat

Using tetranitroblue tetrazolium (TNBT) to stain neuromuscular synapses, we compared the development of the adult pattern of innervation in two fast-twitch muscles in the rat: the androgen-sensitive levator ani (LA) and the extensor digitorum longus (EDL), which is not thought to be androgen sensitive. We found that about 18% of adult LA muscle fibers, but only about 2% of adult EDL fibers, are multiply innervated. Moreover, synapse elimination occurs substantially later in the LA compared with the EDL. At 2 weeks after birth, the EDL is already predominantly singly innervated, whereas the LA is still predominantly multiply innervated. The apparent delay in the normal time course of synapse elimination in the LA corresponds to a similar delay in other aspects of neuromuscular development (the time course of appearance of axonal retraction bulbs, the growth of fibers, and the development of adult motor terminal morphology). Finally, motor terminals change during synapse elimination from morphologies resembling growth cones to the adult form of neuromuscular synapses. Because the period of synapse elimination is significantly different for muscles that differ in their androgen sensitivity, hormonal sensitivity may represent an important property of motoneurons or muscle fibers influencing the normal time course of neuromuscular synapse elimination in rats. Thus, androgen might regulate the normal ontogenetic process of synapse elimination.     

Effects of evoked and spontaneous motoneuronal firing on synapse competition and elimination in skeletal muscle

Synapse elimination is a general feature of the development of neural connections, including the connections of motoneurons to skeletal muscle fibers. Our work addressed two questions: (1) how the action potentials generated in the set of motoneurons innervating an individual muscle (i.e., in a motor pool) are correlated in time during development in vivo; (2) what influence different firing patterns exert on the processes of polyneuronal innervation and synapse elimination which characterize the establishment of muscle innervation. We recorded the spontaneous electromyographic activity of the tibialis anterior and soleus muscles of late embryonic and neonatal rats, identifying the firing of at least two single motor unit signals in each record. We found that a striking switch occurs a few days after birth from a highly synchronous type of firing to an asynchronous one, the first thus characterizing embryonic while the second one adult motoneurons. We also investigated the effects of an evoked synchronous type of discharge on neuromuscular synapse formation, measuring polyneuronal innervation and synapse elimination. This was done in an adult in vivo model of de novo synapse formation, while a chronic TTX nerve conduction block, placed centrally with respect to the stimulating electrodes, eliminated the natural activity of motoneurons. We found that the imposed synchronous activity greatly inhibits synapse elimination, causing polyneuronal innervation to persist. We conclude that the early synchronous firing, favors the establishment of polyneuronal innervation while the subsequent switch to an asynchronous one promotes synapse elimination.     

Activity-dependent switch from synapse formation to synapse elimination during development of neuromuscular junctions

The embryonic development of neuromuscular junctions consists of two successive epochs, an early period marked by exuberant synapse formation and a later period marked by synapse elimination. In the frog muscles we have studied, myogenesis is protracted and overlaps the periods of synapse formation and elimination. Thus, the formative and regressive events of synaptic development do not occur in synchrony across different fibers in the muscle. We propose that local activity orchestrates a shift from synaptogenesis to synapse elimination at the level of single muscle fibers. We also present evidence that perisynaptic Schwann cells and the expression of ion channels in the sarcolemma play important roles in the development of neuromuscular junctions. Questions for future study are outlined.     

Synapse elimination by fiber type and maturational state in rabbit soleus muscle

We have compared the development of fast and slow motor innervation in the neonatal rabbit soleus, a muscle which contains two distinct motor unit types during the early period of polyneuronal innervation. The innervation state of individual muscle fibers was ascertained using an intracellular electrode; a fluorescent dye was then injected into particular fibers to permit subsequent identification of histochemical type. We found no significant difference in the time course of synapse elimination for fast and slow motor units as judged by the percentage of fibers remaining polyneuronally innervated at two ages: 7-8 days, when most fibers are multiply innervated, and 10-11 days, when the level of polyinnervation is low. In a second experiment, we examined a phenomenon in which compound end-plate potentials were occasionally seen in muscle fibers at an age (17-23 days) well past the major episode of synapse elimination. We present evidence that this apparent polyinnervation in fact derives from an electrode-induced electrical coupling artifact and that genuinely polyinnervated fibers are very rare at this stage, if present at all.     

Muscles in a mouse model of spinal muscular atrophy show profound defects in neuromuscular development even in the absence of failure in neuromuscular transmission or loss of motor neurons

A mouse model of the devastating human disease "spinal muscular atrophy" (SMA) was used to investigate the severe muscle weakness and spasticity that precede the death of these animals near the end of the 2nd postnatal week. Counts of motor units to the soleus muscle as well as of axons in the soleus muscle nerve showed no loss of motor neurons. Similarly, neither immunostaining of neuromuscular junctions nor the measurement of the tension generated by nerve stimulation gave evidence of any significant impairment in neuromuscular transmission, even when animals were maintained up to 5days longer via a supplementary diet. However, the muscles were clearly weaker, generating less than half their normal tension. Weakness in 3 muscles examined in the study appears due to a severe but uniform reduction in muscle fiber size. The size reduction results from a failure of muscle fibers to grow during early postnatal development and, in soleus, to a reduction in number of fibers generated. Neuromuscular development is severely delayed in these mutant animals: expression of myosin heavy chain isoforms, the elimination of polyneuronal innervation, the maturation in the shape of the AChR plaque, the arrival of SCs at the junctions and their coverage of the nerve terminal, the development of junctional folds. Thus, if SMA in this particular mouse is a disease of motor neurons, it can act in a manner that does not result in their death or disconnection from their targets but nonetheless alters many aspects of neuromuscular development.     

Critical period for the androgenic block of neuromuscular synapse elimination

Juvenile androgen treatment during developmental synapse elimination changes the pattern of innervation in the adult levator ani (LA), an androgen-sensitive muscle (Jordan, Letinsky, and Arnold, 1989b). Most notably, such adult muscles contain an unusually high number of muscle fibers that are innervated by two or more axons indicating that these fibers are multiply innervated. Juvenile androgen treatment also increases the adult level of preterminal branching, the number of junctional sites per adult fiber, and the size of adult LA muscle fibers and motoneurons in the spinal nucleus of the bulbocavernosus (SNB). The present study was designed to determine when in development androgen treatment is most effective in maintaining multiple innervation in adulthood and whether there are different critical periods for the different effects of juvenile androgen treatment. Male rats were castrated on 7, 21, or 34 days after birth (roughly corresponding to the beginning, middle, and end of synapse elimination in the LA muscle) and treated daily with testosterone propionate for the next 2 weeks. All rats were sacrificed at 9 weeks and their spinal cords and LA muscles were stained and analyzed. Only during the first treatment period (7-20) did androgen treatment result in increased levels of multiple innervation at 9 weeks. During this period, androgen also increased the number of junctional sites per fiber and the size of SNB somata but did not influence the adult level of preterminal branching or the diameter of adult LA muscle fibers. Androgen treatment during the two later periods increased the level of preterminal branching and the size of LA muscle fibers without influencing the level of multiple innervation.(ABSTRACT TRUNCATED AT 250 WORDS)     

The relationship of neuromuscular synapse elimination to synaptic degeneration and pathology: Insights from WldS and other mutant mice

Neuromuscular synapse elimination, Wallerian degeneration and peripheral neuropathies are not normally considered as related phenomena. However, recent studies of mutant and transgenic mice, particularly the Wld ^S mutant—in which orthograde degeneration is delayed following axotomy—suggest that re-evaluation of possible links between natural, traumatic and pathogenic regression of synapses may be warranted. During developmental synapse elimination from polyneuronally innervated junctions, some motor nerve terminals progressively and asynchronously vacate motor endplates. A form of asynchronous synapse withdrawal, strongly resembling synapse elimination, also occurs from mononeuronally-innervated motor endplates following axotomy in young adult Wld ^S mutant mice. A similar pattern is observed in skeletal muscles of several neuropathic mutants, including mouse models of dying-back neuropathies, motor neuron disease and—remarkably—models of neurodegenerative diseases such as Huntington's and Alzheimer's diseases. Taken together with recent analysis of synaptic remodelling at neuromuscular junctions in Drosophila, a strong candidate for a common regulatory mechanism in these diverse conditions is one based on protein ubiquitination/deubiquitination. Axotomised neuromuscular junctions in Wld ^S mutant mice offer favourable experimental opportunities for examining developmental mechanisms of synaptic regression, that may also benefit our understanding of how degeneration in the synaptic compartment of a neuron is initiated, and its role in progressive, whole-cell neuronal degeneration.     

Axotomy of developing rat spinal motoneurons: Cell survival,soma size,muscle recovery,and the influence of testosterone

During the period of synapse elimination, motoneurons are impaired in their ability to generate or regenerate axonal branches: following partial denervation of their target muscle, young motoneurons do not sprout to nearby denervated fibers and after axonal injury, they fail to reinnervate the muscle. In the rat levator ani (LA) muscle, which is innervated by motoneurons in the spinal nucleus of the bulbocavernosus (SNB), synapse elemination ends relatively late in development and can be regulated by testosterone. We took advantage of this system to determine if the end of synapse elimination and the development of regenerative capabilities by motoneurons share a common mechanism, or, alternatively, if these two events can be dissociated in time. Axotomy on or before postnatal day 14 (P14) caused the death of SNB motoneurons. By P21, toward the end of synapse elimination in the LA muscle, SNB motoneurons had developed the ability to survive axonal injury. Altering testosterone levels by castration on P7 followed by 4 weeks of either testosterone propionate or control injections did not change the ability of SNB motoneurons to survive axonal injury during development, although these same treatments alter the time course of synapse elimination in the LA muscle. Thus, we dissociated the inability of SNB motoneurons to recover from axonal injury from their developmental elimination of synaptic terminals. We also measured the effect of early axotomy on motoneuronal soma size and on target muscle weight. Axotomy on P14 caused a long-lasting decrease in the soma size of surviving SNB motoneurons, whereas motoneurons axotomized on P28 recovered their normal soma size. Axotomy on or before P7 caused severe atrophy of the target muscles, matching the extensive loss of motoneurons. However, target muscle recovery after axotomy on P14 was as good as recovery after axotomy at later ages, despite greater motoneuronal death after axotomy on P14. This result may reflect an increase in motor unit size, a decrease in polyneuronal innervation by SNB motoneurons that survive axotomy on P14, or a combination of the two. © 1995 John Wiley & Sons, Inc.     

Brain-derived neurotrophic factor transiently stabilizes silent synapses on developing neuromuscular junctions

A general feature of the developing nervous system is the activity-dependent rearrangement of genetically defined, synaptic connections. A parallel process occurs at the developing neuromuscular junction as activity-dependent synapse withdrawal reduces the initial polyneuronal innervation of individual muscle fibers to a mononeuronal innervation within the first few weeks after birth. Because members of the neurotrophin gene family influence motor neuron differentiation and survival, we examined whether or not they also influence synaptic rearrangements in neonatal muscles. We found that treatment with brain-derived neurotrophic factor (BDNF), neurotrophin-3 (NT-3), or neurotrophin-4/5 (NT-4/5) causes the transient retention of multiple synaptic contacts on neonatal myofibers. However, the combination of both electrophysiological and histological assays revealed that the majority of such supernumerary synaptic contacts are functionally inactive or “silent.” There also occurs an increase in the number of retracting axons. Because BDNF mRNA is expressed in developing muscle and the trkB tyrosine kinase receptor for BDNF is expressed by neonatal motor neurons, our results suggest that BDNF may play an endogenous role in the refinement of synaptic connectivity that occurs in skeletal muscles after birth. © 1996 John Wiley & Sons, Inc.     

Morphometric study of the neuromuscular synapses in the adult rat with special reference to the remodelling concept

There is increasing morphologic evidence that neuromuscular synapses are not rigid structures in the mature muscles of adult animals. On the contrary, they may be submitted to a continuous process of remodelling. In silver-impregnated sternocleidomastoid muscles of the young adult rat, we measured synaptic parameters such as nerve terminal length, the number of branching points of terminal arborization, and muscle fiber diameter, and used a morphometric approach to explore specific questions concerning neuromuscular remodelling. Quantitative data indicate that: (a) The complexity and maturation of the nerve endings in this muscle are very variable and the increase in branching points is not paralleled by an increase in terminal length; (b) Muscle fiber diameter is related only marginally to presynaptic parameters; (c) Accessory ending formation occurs when the original ending does not reach the mean size of endings in singly innervated areas; (d) The complexity of individual endings at dually innervated junctions is smaller than the mean development of singly innervated synapses, indicating the existence of some mutual inhibitory influence between closely spaced endings. Morphometric results suggest a continuous process of synaptic formation in this adult muscle.     

Basal lamina directs acetylcholinesterase accumulation at synaptic sites in regenerating muscle

In skeletal muscles that have been damaged in ways which spare the basal lamina sheaths of the muscle fibers, new myofibers develop within the sheaths and neuromuscular junctions form at the original synaptic sites on them. At the regenerated neuromuscular junctions, as at the original ones, the muscle fibers are characterized by junctional folds and accumulations of acetylcholine receptors and acetylcholinesterase (AChE). The formation of junctional folds and the accumulation of acetylcholine receptors is known to be directed by components of the synaptic portion of the myofiber basal lamina. The aim of this study was to determine whether or not the synaptic basal lamina contains molecules that direct the accumulation of AChE. We crushed frog muscles in a way that caused disintegration and phagocytosis of all cells at the neuromuscular junction, and at the same time, we irreversibly blocked AChE activity. New muscle fibers were allowed to regenerate within the basal lamina sheaths of the original muscle fibers but reinnervation of the muscles was deliberately prevented. We then stained for AChE activity and searched the surface of the new muscle fibers for deposits of enzyme they had produced. Despite the absence of innervation, AChE preferentially accumulated at points where the plasma membrane of the new muscle fibers was apposed to the regions of the basal lamina that had occupied the synaptic cleft at the neuromuscular junctions. We therefore conclude that molecules stably attached to the synaptic portion of myofiber basal lamina direct the accumulation of AChE at the original synaptic sites in regenerating muscle. Additional studies revealed that the AChE was solubilized by collagenase and that it remained adherent to basal lamina sheaths after degeneration of the new myofibers, indicating that it had become incorporated into the basal lamina, as at normal neuromuscular junctions.     

An anatomical and electrophysiological study of synapse elimination at the developing frog neuromuscular junction

Synapse elimination was examined in the developing frog cutaneous pectoris muscle using histological and electrophysiological techniques. Morphological synapse elimination occurred in two phases. The first phase, which began at the time of metamorphosis and continued until the second to third postmetamorphic week, was characterized by a rapid decline in the number of endplates receiving greater than or equal to 3 synaptic inputs. However, 50% of the muscle fibers still remained dually innervated. This dual innervation decreased with a much slower time course; approximately 20% of the muscle fibers were dually innervated in 1- to 2-year-old frogs. During the first phase of synapse elimination no difference was noted between the distribution of acetylcholine receptors or acetylcholinesterase activity associated with the terminal arborizations formed by separate axons at one synaptic site. However, terminal arborizations formed by small diameter axons and consisting of varicosities separated by thin interconnectives became apparent during this period. Such varicose arborizations responded to nerve stimulation and released acetylcholine in proportion to their terminal length as did the nonvaricose arborizations. In addition, the number of morphological and physiological inputs at one endplate site was well correlated throughout the first phase of synapse elimination.     

Synapse elimination from the mouse neuromuscular junction in vitro: A non-hebbian activity-dependent process

The effect of action potentials on elimination of mouse neuromuscular junctions (NMJ) was studied in a three compartment cell culture preparation. Axons from superior cervical ganglion or ventral spinal cord neurons in two lateral compartments formed multiple neuromuscular junctions with muscle cells in a central compartment. The loss of synapses over a 2–7-day period was determined by serial electrophysiological recording and a functional assay. Electrical stimulation of axons from one side compartment during this period, using 30-Hz bursts of 2-s duration, repeated at 10-s intervals, caused a significant increase in synapse elimination compared to unstimulated cultures (p< 0.001). The extent of homosynaptic and heterosynaptic elimination was comparable, i. e., of the 226 functional synapses of each type studied, 111 (49%) of the synapses that had been stimulated were eliminated, and 87 (39%) of unstimulated synapses on the same muscle cells were eliminated. Also, simultaneous bilateral stimulation caused significantly greater elimination of synapses than unilateral stimulation (p< 0.005). These observations are contrary to the Hebbian hypothesis of synaptic plasticity. A spatial effect of stimulus-induced synapse elimination was also evident following simultaneous bilateral stimulation. Prior to stimulation, most muscle cells were innervated by axons from both side compartments, but after bilateral stimulation, muscle cells were predominantly unilaterally innervated by axons from the closer compartment. These experiments suggest that synapse elimination at the NMJ is an activity-dependent process, but it does not follow Hebbian or anti-Hebbian rules of synaptic plasticity. Rather, elimination is a consequence of postsynaptic activation and a function of location of the muscle cell relative to the neuron. An interaction between spatial and activity-dependent effects on synapse elimination could help produce optimal refinement of synaptic connections during postnatal development. © 1993 John Wiley & Sons, Inc.     

Specificity and development of innervation pattern of Xenopus pectoralis muscle

Multielectrode recordings were used to identify and measure the axonal inputs to each end plate on contiguous surface fibers covering about 25% of the Xenopus pectoralis muscle in mature and developing animals. The mature innervation pattern was remarkably precise. Individual axons tended to innervate fibers of similar input resistance (R_in) in compact motor units restricted to only a portion of the region studied. Motor units comprising fibers of similar R_in overlapped mainly near their borders. Most fibers had two end plates. In more than 80% of these fibers, both end plates received input from the same axon. In 57%, this was the only input to both end plates. This implies a powerful mechanism for excluding or eliminating inputs from other axons. About 16% of the mature junctions showed focal polyneuronal innervation, with the weaker end plate potential component often less than 1 mV in noncurarized preparations. However, we have no evidence that the weaker inputs were being eliminated. During development, motor units became more compact, which was associated with synapse elimination; but from the earliest times studied, soon after metamorphosis when many fibers were adding second end plates, a majority of those that had two end plates were innervated at both sites by the same axon. © 1996 John Wiley & Sons, Inc.