The intervals between miniature end-plate potentials in the frog are unlikely to be independently or exponentially distributed

Probabilities of evoked and spontaneous synaptic transmission at individual active zones: Lessons from Drosophila

Nerve terminals release neuronal transmitters at morphological specializations known as active zones (AZs). Synaptic vesicle fusion at individual AZs is probabilistic, and this property is fundamental for the neuronal information transfer. Until recently, a lack of appropriate tools limited the studies of stochastic properties of neuronal secretion at individual AZs. However, Drosophila transgenic lines that express postsynaptically tethered Ca²⁺ sensor GCaMP enabled the visualization of single exocytic event at individual AZs. The present mini-review discusses how this powerful approach enables the investigation of the evoked and spontaneous transmission at single AZs and promotes the understanding of the properties of both release components.

Structure and Function of the Mammalian Neuromuscular Junction

The mammalian neuromuscular junction (NMJ) comprises a presynaptic terminal, a postsynaptic receptor region on the muscle fiber (endplate), and the perisynaptic (terminal) Schwann cell. As with any synapse, the purpose of the NMJ is to transmit signals from the nervous system to muscle fibers. This neural control of muscle fibers is organized as motor units, which display distinct structural and functional phenotypes including differences in pre- and postsynaptic elements of NMJs. Motor units vary considerably in the frequency of their activation (both motor neuron discharge rate and duration/duty cycle), force generation, and susceptibility to fatigue. For earlier and more frequently recruited motor units, the structure and function of the activated NMJs must have high fidelity to ensure consistent activation and continued contractile response to sustain vital motor behaviors (e.g., breathing and postural balance). Similarly, for higher force less frequent behaviors (e.g., coughing and jumping), the structure and function of recruited NMJs must ensure short-term reliable activation but not activation sustained for a prolonged period in which fatigue may occur. The NMJ is highly plastic, changing structurally and functionally throughout the life span from embryonic development to old age. The NMJ also changes under pathological conditions including acute and chronic disease. Such neuroplasticity often varies across motor unit types. © 2022 American Physiological Society. Compr Physiol 12:1-36, 2022.

The stochastic properties of spontaneous quantal release of transmitter at the frog neuromuscular junction

1. Earlier results showed that it is unlikely that spontaneous quantal release of transmitter at the frog neuromuscular junction is produced by a Poisson process.2. Data sets were tested, by using the u statistic, to see whether if they are assumed to be generated by a Poisson process, the mean interval is changing monotonically with time. By this critieria, some of the data sets are stationary, others are not.3. A variety of mathematical transforms are employed on empirical data sets to characterize the properties of the spontaneous quantal release.(a) The intensity function, which calculates the frequency distribution of all possible combinations of intervals, shows an excess of short intervals, without any sign of periodicity.(b) The variance-time curve, which estimates the accumulated variance of the series as a function of time into the series, lies significantly above the Poisson prediction.(c) The power spectrum, whether calculated on the intervals or on the number of intervals in time bins, deviates significantly from the Poisson prediction at the low frequencies.(d) The ln-survivor curve has two phases: a concave section for the short intervals, and a roughly linear section for the intervals of greater length.These transforms indicate that the min.e.p.p.s are clustered.4. A series of models for spontaneous quantal release were considered.(a) A Poisson model. Rejected because of consistent failure to fit the data.(b) A periodic model. Rejected because the intervals should be ordered rather than clustered.(c) A time-dependent model, in which quantal release is governed by a Poisson process with a mean interval that is oscillating in time. This model will generate clustering; by the transforms the model can be shown to closely fit the data. However, an autocorrelation of min.e.p.p. amplitudes shows that there is a relationship between the amplitudes and their position in the series. This is not predicted by the time-dependent oscillating model.(d) A branching Poisson model, in which a primary release, generated by a Poisson process, is likely to be followed by one or more subsidiary releases from the same site. The parameters of the branching model can be determined from ln-survivor curves. Theoretical curves, created with these parameters, give power spectra, variance-time curves, and ln-survivor curves that strongly resemble those calculated from the data. The model also predicts a significant autocorrelation of amplitudes.5. Min.e.p.p.s recorded with an extracellular electrode also fit well to a branching Poisson model.6. The effects of raised [Ca(2+)](o) on the intervals between min.e.p.p.s were studied. In our experiments the change in extracellular solution did not produce any notable change in release statistics.7. The effects of elevated [K(+)](o) on the intervals between spontaneous releases were studied. Depolarization of the nerve terminal increases the frequency of primary releases and decreases the chance of having subsidiary releases.8. Possible physical mechanisms by which quantal release of transmitter from a nerve terminal would fit a branching Poisson model are described.

Stochastic properties of spontaneous transmitter release at the crayfish neuromuscular junction

1. Miniature excitatory junctional potentials (min.e.j.p.s) were recorded with an intracellular electrode from the adductor muscle of the dactyl of the first or second walking leg of the crayfish, Orconectes virilis.2. The intervals between the min.e.j.p.s were compared to the exponential prediction by five goodness of fit tests. The results indicate that the intervals are not exponentially distributed.3. The autocorrelogram of intervals shows that the intervals are unlikely to be independent.4. A stochastic analysis that includes the power spectrum of intervals, the variance-time curve, and the ln-survivor curve suggest that there is a clustering of min.e.j.p.s. The results are similar to those on the frog neuromuscular junction.5. An autocorrelogram of the min.e.j.p. amplitudes suggests that sizes are not independently distributed.6. These results, which are similar to those previously reported from the frog neuromuscular junction, support the use of the branching Poisson process as a theoretical model for the stochastic properties of spontaneous quantal release of transmitter.

Asynchrony of quantal events in evoked multiquantal responses indicates presynaptic quantal interaction

Asynchrony of quantal events in evoked multiquantal responses indicates presynaptic quantal interaction. We have analyzed the possibility of quantal interactions by inspecting action potential-evoked postsynaptic multiquantal responses recorded extracellularly from the lobster neuromuscular junction. These recorded responses were compared with simulated multiquantal responses constructed from statistically independent quantal events. The simulated multiquantal responses were generated by random superposition of single quantal responses aligned according to the timing of the action potential. The methods of analysis consisted of 1) the comparison of quantal contents obtained from direct counting or by measuring of the size of the responses and 2) the analysis of distributions of quantal latencies. This analysis revealed a large error in the detection of quantal events for responses simulated with no quantal interaction. In contrast, very few errors in quantal detection were made in the analysis of experimental recordings. Latency histograms of recorded responses demonstrate that the proportion of late quantal events (those with latencies of >/=5 ms) increased as a function of quantal content. This shift in latency histograms was not observed for simulated responses. Our interpretation is that quanta interact presynaptically to cause asynchrony of quantal events in evoked responses.

Spontaneous and neurally activated depolarizations in smooth muscle cells of the guinea-pig urethra

1. Membrane potential recordings were made from longitudinal smooth muscle cells of the guinea-pig urethra using conventional microelectrode techniques. 2. Smooth muscle cells of the urethra developed spontaneous transient depolarizations (STDs) and slow waves. Single unit STDs had amplitudes of approximately 5 mV and slow waves seemed to occur as amplitude multiples of single unit STDs. 3. STDs and slow waves were abolished by niflumic acid or low chloride solution and also by cyclopiazonic acid (CPA), BAPTA or high concentrations of caffeine. Lower concentrations of caffeine abolished slow waves but not STDs. Nifedipine inhibited slow waves but not STDs. 4. When stochastic properties of STDs were examined, it was found that the intervals between occurrences were not well modelled by Poisson statistics, instead the STDs appeared to be clustered. 5. Transmural stimulation evoked excitatory junctional potentials (EJPs) and triggered slow waves which were abolished by either alpha,beta-methylene-ATP or tetrodotoxin. Evoked slow waves were also abolished by caffeine, co-application of caffeine and ryanodine or by CPA which left EJPs unaffected. 6. In conclusion, smooth muscle cells of urethra exhibit STDs which are clustered rather than random events, and are the result of spontaneous Ca2+ release from intracellular stores and subsequent activation of Ca2+-activated chloride channels. STDs sum to activate L-type Ca2+ channels which contribute to the sustained phase of slow waves. Stimulation of purinoceptors by neurally released ATP initiates EJPs and also causes the release of Ca2+ from intracellular stores to evoke slow waves.

Quantal neurotransmitter secretion rate exhibits fractal behavior

The rate of exocytic events from both neurons and non-neuronal cells exhibits fluctuations consistent with fractal (self-similar) behavior in time, as evidenced by a number of statistical measures. We explicitly demonstrate this for neurotransmitter secretion at Xenopus neuromuscular junctions and for rat hippocampal synapses in culture; the exocytosis of exogenously supplied neurotransmitter from cultured Xenopus myocytes and from rat fibroblasts behaves similarly. The magnitude of the fluctuations of the rate of exocytic events about the mean decreases slowly as the rate is computed over longer and longer time periods, the periodogram decreases in power-law manner with frequency, and the Allan factor (relative variance of the number of exocytic events) increases as a power-law function of the counting time. These features are hallmarks of self-similar behavior. Their description requires models that exhibit long-range correlation (memory) in event occurrences. We have developed a physiologically plausible model that accords with all of the statistical measures that we have examined. The appearance of fractal behavior at synapses, as well as in systems comprising collections of synapses, indicates that such behavior is ubiquitous in neural signaling.

Quantal neurotransmitter secretion rate exhibits fractal behavior

The rate of exocytic events from both neurons and non-neuronal cells exhibits fluctuations consistent with fractal (self-similar) behavior in time, as evidenced by a number of statistical measures. We explicitly demonstrate this for neurotransmitter secretion at Xenopus neuromuscular junctions and for rat hippocampal synapses in culture; the exocytosis of exogenously supplied neurotransmitter from cultured Xenopus myocytes and from rat fibroblasts behaves similarly. The magnitude of the fluctuations of the rate of exocytic events about the mean decreases slowly as the rate is computed over longer and longer time periods, the periodogram decreases in power-law manner with frequency, and the Allan factor (relative variance of the number of exocytic events) increases as a power-law function of the counting time. These features are hallmarks of self-similar behavior. Their description requires models that exhibit long-range correlation (memory) in event occurrences. We have developed a physiologically plausible model that accords with all of the statistical measures that we have examined. The appearance of fractal behavior at synapses, as well as in systems comprising collections of synapses, indicates that such behavior is ubiquitous in neural signaling.

The effects of synaptic noise on measurements of evoked excitatory postsynaptic response amplitudes

Spontaneously occurring synaptic events (synaptic noise) recorded intracellularly are usually assumed to be independent of evoked postsynaptic responses and to contaminate measures of postsynaptic response amplitude in a roughly Gaussian manner. Here we derive analytically the expected noise distribution for excitatory synaptic noise and investigate its effects on amplitude histograms. We propose that some fraction of this excitatory noise is initiated at the same release sites that contribute to the evoked synaptic event and develop an analytical model of the interaction between this fraction of the noise and the evoked postsynaptic response amplitude. Recording intracellularly with sharp microelectrodes in the in vitro hippocampal slice preparation, we find that excitatory synaptic noise accounts for up to 70% of the intracellular recording noise, when inhibition is blocked pharmacologically. Up to 20% of this noise shows a significant correlation with the evoked event amplitude, and the behavior of this component of the noise is consistent with a model which assumes that each release site experiences a refractory period of approximately 60 ms after release. In contrast with classical models of quantal variance, our models predict that excitatory synaptic noise can cause the apparent variance of successive peaks in an excitatory synaptic amplitude histogram to decrease from left to right, and in some cases to be less than the variance of the measured noise.

Saturation of postsynaptic glutamate receptors after quantal release of transmitter

Miniature excitatory postsynaptic currents (mEPSCs) were elicited from small numbers of release sites after brief microperfusion of Ba2+ and K+ onto proximal dendritic processes of hippocampal neurons in culture. Temporal summation of closely timed mEPSCs deviated significantly from linearity. The number of instances of closely timed mEPSCs that were also closely matched in terms of peak amplitudes was significantly greater than that expected by chance. Amplitude pairing became statistically more significant after prolongation of mEPSC duration and inhibition of glutamate receptor desensitization with cyclothiazide. These results are best explained by postsynaptic receptors that approach saturation after quantal release of transmitter.

Inhibition of quantal transmitter release in the absence of calcium influx by a G protein-linked adenosine receptor at hippocampal synapses

Spontaneous miniature excitatory postsynaptic currents (MEPSCs) were recorded by whole-cell voltage-clamp techniques in cultured rat hippocampal pyramidal neurons. The specific adenosine A1 receptor agonist cyclopentyladenosine (CPA) reduced the frequency of MEPSCs without affecting their amplitude distribution or kinetic properties. This action was blocked by pretreatment of the cells with pertussis toxin. In the presence of divalent cation Ca2+ channel blockers, CPA was still effective in reducing the frequency of MEPSCs. It was shown that this effect cannot be explained by changes in basal Ca2+ influx. These results suggest that neurotransmitters that produce presynaptic inhibition at hippocampal synapses utilize several mechanisms, one of which may involve inhibition of some component of the quantal release apparatus that occurs independently of inhibition of Ca2+ influx.

Characteristics of miniature inhibitory postsynaptic currents in CA1 pyramidal neurones of rat hippocampus

1. Recordings were made in vitro from chloride-loaded CA1 rat hippocampal pyramidal neurones in the presence of tetrodotoxin (TTX) to examine miniature inhibitory postsynaptic currents (IPSCs). 2. Most spontaneous synaptic events recorded before TTX was applied, and all events that were resolved in the presence of TTX, were blocked by the GABAA receptor antagonist bicuculline. 3. At 25 degrees C, averaged miniature IPSCs had time to peak of about 3 ms and in most cases decayed with a single time constant close to 25 ms. 4. With a driving force for chloride ions between 70 and 80 mV, the mean miniature IPSC amplitude was 19.6-27.9 pA, yielding a conductance of 258-326 pS. The mean amplitude of unitary IPSCs recorded before TTX was applied was in the range of 31-73 pA. 5. When intervals between miniature IPSCs were compared with an exponential distribution, there was an excess of events at intervals shorter than 5 ms. Some individual events appeared to represent the nearly simultaneous release of two inhibitory quanta. 6. Miniature IPSC amplitude distributions were better fitted with the sum of two Gaussians than with one Gaussian. The variance in amplitude of a single quantal event exceeded that of the baseline noise. 7. Comparison of the conductance changes corresponding to the first Gaussian distribution with single GABA channel data suggests that one inhibitory quantum opens twelve to twenty chloride channels and that GABA molecules bind once to a postsynaptic receptor.

Estimating the time course of evoked quantal release at the frog neuromuscular junction using end-plate current latencies

The use of end-plate current (e.p.c.) latency measurements to estimate the time course of the stochastic probabilistic process governing evoked release was investigated in the sciatic nerve-sartorius muscle preparation of the frog, Rana pipiens. We also examined the possibility that the release of a quantum depresses or enhances the subsequent release of additional quanta. Muscle end-plates were voltage clamped at 3-4 degrees C. Quantal release was restricted to a short, or localized, region of the nerve terminal using Ca2+-free, EGTA Ringer solution and a Ca2+-filled micropipette. The number of e.p.c.s containing 0, 1, 2, etc. quanta were totalled and compared to numbers predicted using Poisson's theorem. The differences between the actual and predicted numbers of events were not significant at the nineteen junctions studied (P less than 0.05). The latency of the first quantum observed in several hundred e.p.c.s was measured and used to calculate an estimate, alpha 1(t), of the time-dependent, probabilistic process, alpha (t), governing all evoked quantal release (Barrett & Stevens, 1972b). In three experiments, all quantal latencies were measured to obtain the actual alpha (t). The alpha 1(t) function gave an excellent approximation of alpha (t) (P greater than 0.2), in real and simulated latency data. The latency of the second quantum in the e.p.c.s was measured and used to provide another estimate, alpha 2(t), of alpha (t). The alpha 2(t) function was lower (depressed) during the first few milliseconds of the evoked release period, relative to alpha 1(t). The difference was significant (P greater than 0.01) in all experiments. Our measurement procedures were tested using computer-generated 'e.p.c.s' containing randomly occurring 'quanta'. These tests showed that the early depression was due to inadequate detection of the second quantum in the e.p.c.s. The effect of Sr2+ on evoked release was examined using double-barrelled pipettes containing 1 M-SrCl2 and CaCl2 solutions. The major result was that the durations of alpha 1(t) and alpha 2(t) were equally lengthened in Sr2+, relative to Ca2+.

A method for testing an extended poisson hypothesis of spontaneous quantal transmitter release at neuromuscular junctions

A statistical method for testing the Poisson hypothesis of spontaneous quantal transmitter release at neuromuscular junctions has been proposed. The notion of the Poisson hypothesis is extended so as to allow for nonstationarity in the data, since nonstationarity is commonly seen in the occurrence of spontaneous miniature potentials. Special emphasis has been put on the nonstationary analysis of the quantal release. A time scaling technique has been introduced and is discussed for the analysis. Artificially generated data, which simulate three types of nonstationary spontaneous quantal release, i.e., Poisson, non-Poisson-clustered, and non-Poisson-ordered types, were analyzed to demonstrate the effectiveness of the method. Some sets of miniature endplate potentials, intracellularly recorded at frog sartorius neuromuscular junctions in low Ca++ and high Mg++ solutions showing apparent nonstationarities, were analyzed as illustrative examples. The proposed method will extend the range of applicable data for the statistical analysis of spontaneous quantal transmitter release.

Is the MEPP due to the release of one vesicle or to the simultaneous release of several vesicles at one active zone?

Miniature end-plate potentials (MEPPs) recorded at the neuromuscular junction were initially reported to be normally distributed and have been attributed to quantal ACh release. This quantum was later correlated with the release of the content of one clear vesicle. This is the 'classical vesicular hypothesis'. Recent observations of subminiature end-plate potentials (s-MEPPs) and of multimodal distribution of the MEPP amplitudes have led to the formulation of a new 'multivesicular hypothesis'. It attributes the s-MEPP to the release of one vesicle and the MEPP to the simultaneous release of several vesicles at one active zone. The distribution of MEPP intervals, the evaluation of the ACh content of a vesicle and of the ACh necessary to produce a MEPP, estimates of the number of vesicles missing following repeated stimulation, and the freeze fracture studies of the active zone do not permit a definitive rejection of either hypotheses.

Transmitter release from normal and degenerating locust motor nerve terminals

1. An analysis has been made of spontaneous and evoked transmitter release from terminals of 'fast' excitatory motor axons on locust muscle fibres using intra- and extracellular recording together with a Ca-electrode technique for activating transmitter release from single nerve terminals on multiterminally innervated muscle fibres. 2. Spontaneous intracellular miniature excitatory junction potentials (m.e.j.p.s), recorded at active spots on these muscle fibres, occurred non-randomly with frequent bursts of m.e.j.p.s. 3. M.e.j.p.s of subnormal amplitude were also seen but contributed only a small fraction of the minature discharge. 4. The amplitude distribution of intracellularly recorded excitatory junction potentials (e.j.p.s) evoked during ionophoretic application of Ca onto single nerve terminals was adequately predicted by Poisson statistics. 5. During the course of nerve terminal degeneration m.e.j.p.s of subnormal amplitude became more frequent and eventually formed the major part of the miniature discharge. Transmitter quanta responsible for 'small' m.e.j.p.s did not contribute to evoked release either at normal or degenerating terminals. Evoked transmitter release from degenerating axon terminals before excitation-secretion coupling failure conformed to Poisson statistics. 6. It is concluded that more than one release mechanism operates on the transmitter pool or pools in locust motor nerve terminals.

Spontaneous multiquantal release at synapses in guinea-pig hypogastric ganglia: evidence that release can occur in bursts

1. A study was made of some properties of the spontaneous synaptic potentials recorded in cells of the hypogastric ganglia of guinea-pigs. 2. The distribution of the amplitudes of the spontaneous synaptic potentials arising from a single preganglionic fibre was found to be multimodal, with peaks at roughly integral multiples of a unit peak. 3. It was found that the amplitudes of the larger spontaneous potentials were consistent with them being the result of synchronous or near-synchronous release of two or more unit-sized quanta (multiquantal release). 4. The proportion of multiquantal potentials observed was found to be dependent on the extracellular calcium ion concentration. 5. When the stochastic properties of the spontaneous potentials were examined, it was found that the spontaneous release process was not random and independent but appeared to be clustered. Indeed, the probability of occurrence of a unit spontaneous synaptic potential was greatly enhanced during the 40--60 msec immediately following any given spontaneous synaptic potential. 6. When unit spontaneous potentials were excluded from the analysis, the multiquantal potentials were still found to be clustered although the clustering was less marked than that seen in the over-all process. 7. These results suggested that the multiquantal spontaneous potentials arose from the release of unit quanta in short high frequency bursts and it was found that a mathematical model incorporating such a mechanism could describe the spontaneous quantal release process.

Spontaneous subminature end-plate potentials in mouse diaphragm muscle: evidence for synchronous release

1. Miniature end-plate potentials (min.e.p.p.s) were recorded from small muscle cells of mouse diaphragms. Min.e.p.p. amplitude histograms showed successive peaks which were integral multiples of the smallest peak. The smallest potentials (submin.e.p.p.s) averaged 0-3-0-6mV and the mean of the larger min.e.p.p.s averaged 3-7 mV, depending on the muscle cell diameter. There was a positive correlation between time-to-peak and min.e.p.p. amplitude. Time-to-peak of the submin.e.p.p.s fell slightly below the regression line through the larger min.e.p.p.s. 2. Sometimes min.e.p.p. amplitude distributions changed spontaneously such that the mean of the major mode min.e.p.p.s decreased twofold during which time the mean of the submin.e.p.p.s did not change. Spontaneous decreases were most pronounced during low frequencies of release (10/min) achieved at 32 degrees C. 3. Small changes in temperature (2 degrees C steps in the range 32-40 degrees C) greatly altered the number of peaks of min.e.p.p. amplitude histograms without noticeably changing the position of the submin.e.p.p. peak. At 32 degrees C submin.e.p.p.s composed 5-20% of the histograms and the amplitude of the major mode peak was twelve to fifteen-times that of the submin.e.p.p.s. Over-all bell-shaped distributions were obtained at 37 degrees C which showed up to eight peaks with the major peak at the fourth to sixth peak. Temperatures slightly above 37 degrees C gave a flat distribution with the mean amplitude at the third peak. Min.e.p.p. amplitude histograms were initially skewed (mostly small min.e.p.p.s) after a 40 degrees C heat challenge. 4. Two to eight-times the normal concentration of Ca2+ in the saline reversibly increased the min.e.p.p. frequency and also decreased the mean of the major mode min.e.p.p.s (two to nine-times) without noticeably changing the mean of the submin.e.p.p.s. 5. Botulinum toxin A, 10(5) X intraperitoneal median lethal dose (10(5)I.P.LD50)/ml., almost abolished min.e.p.p.s in 30-90 min. The relative proportion of submin.e.p.p.s increased and the mean of the major mode min.e.p.p.s usually did not change during the initial decrease in frequency. Major mode min.e.p.p.s essentially ceased after 200-1000 were generated and remaining min.e.p.p.s of some cells showed skewed distributions with three small peaks that were integral multiples of the submin.e.p.p. peak. Smaller min.e.p.p.s were more resistant to block than the larger min.e.p.p.s and, although frequencies were low, small min.e.p.p.s were recorded after 4 hr of botulinum toxin incubation. 6. Colchicine (5 X 10(-4)M) within minutes reduced the major mode min.e.p.p.s by half (mean of major peak reduced to sixth or seventh peak). Additional colchicine (10(-3)M reduced the major mode min.e.p.p. amplitude to a fifth of that of control (mean of major mode min.e.p.p.s at the third peak) with no change in position of the submin.e.p.p. peak. Min.e.p.p. amplitudes slowly recovered to half control values after washing. 7...

Electrophysiological studies of normal and degenerating mouse neuromuscular junctions

A quantitative study of the functional changes occurring at end-plates of phrenicotomised mouse hemidiaphragms was made. Analysis of the frequency of spontaneous transmitter release revealed large scale deviations from a Poisson based process commencing 6-7 h post-phrenicotomy. Short term bursts of miniature end-plate potentials (minEPPs) lasting 0.5-1.0 sec frequently occurred and also long term bursts of minEPPs which lasted from several minutes to 0.5 h or more. Following the long term bursts there was regularly a lasting silence at the end-plates. MinEPP amplitudes were analysed. Probability analyses suggest that the minEPPs are often normally distributed amongst two or more populations. This relationship was maintained during degeneration of the end-plate. No significant differences between the distribution of mean minEPP amplitudes of degenerating end-plates and normal end-plates were found. This suggests that no change in the quantal unit of transmitter release occurs during degeneration of the end-plate.

The formation of synapses in amphibian striated muscle during development

1. A study has been made of the formation of synapses in developing reinnervated and cross-reinnervated amphibian twitch muscles which receive either a focal (iliofibularis) or a distributed (sartorius) innervation from 'en plaque' nerve terminals using histological, ultrastructural and electrophysiological techniques. 2. During the development of the tadpole through metamorphosis to the adult frog, the sartorius myofibres increased in length at about twice the rate of the iliofibularis myofibres, due to a fast rate of growth at their insertions on to the pelvic tendon. 3. The short iliofibularis and sartorius myofibres of young tadpoles (800 mum long) possessed only a single synapse and the iliofibularis myofibres did not receive any further innervation during development. However the sartorius myofibres received further transient innervation on the new muscle laid down during development at the fast growing pelvic insertion, until the distance between the original synapse formed on the myofibres and the synapse at the pelvic end of the muscle was about 12 mm. 4. During development synapses possessed either skewed, multimodal, or unimodal m.e.p.p. amplitude-frequency distributions; the intervals between m.e.p.p.s. were not distributed randomly according to a Poisson process, as m.e.p.p.s. of similar amplitudes tended to be separated by very short intervals; the unit-size e.p.p. had a similar amplitude-frequency distribution as the m.e.p.p.s. if these had a unimodal distribution. 5. Reinnervation or cross-reinnervation of the sartorius and the iliofibularis muscles in adults or at a late stage of development simply reconstituted the normal focal and distributed innervation patterns of the muscles, as found in the control muscles of the contralateral and unoperated legs. 6. These observations on synapse formation in amphibia are consistent with the hypothesis that during development the axon making the initial synaptic contact on the muscle cells induces a property over a length of muscle membrane adjacent to this site which makes it refractory to synapse formation; thus during reinnervation or cross-reinnervation of adult muscles this refractory property constrains synapse formation to these sites.