Role of intracellular Ca2+ in stimulation-induced increases in transmitter release at the frog neuromuscular junction.

Under conditions of reduced quantal content, repetitive stimulation of a presynaptic nerve can result in a progressive increase in the amount of transmitter released by that nerve in response to stimulation. At the frog neuromuscular junction, this increase in release has been attributed to four different processes: first and second components of facilitation, augmentation, and potentiation (e.g., Zengel, J. E., and K. L. Magleby. 1982. Journal of General Physiology. 80:583-611). It has been suggested that an increased entry of Ca2+ or an accumulation of intraterminal Ca2+ may be responsible for one or more of these processes. To test this hypothesis, we have examined the role of intracellular Ca2+ in mediating changes in end-plate potential (EPP) amplitude during and after repetitive stimulation at the frog neuromuscular junction. We found that increasing the extracellular Ca2+ concentration or exposing the preparation to carbonyl cyanide m-chlorophenylhydrazone, ionomycin, or cyclopiazonic acid all led to a greater increase in EPP amplitude during conditioning trains of 10-200 impulses applied at a frequency of 20 impulses/s. These experimental manipulations, all of which have been shown to increase intracellular levels of Ca2+, appeared to act by increasing primarily the augmentation component of increased release. The results of this study are consistent with previous suggestions that the different components of increased release represent different mechanisms, and that Ca2+ may be acting at more than one site in the nerve terminal.

Evoked transmitter release at the frog neuromuscular junction in the presence of very low extracellular Ca2+.

We have examined the effect of ouabain, a Na,K-ATPase inhibitor known to secondarily increase intracellular Ca2+ levels, on evoked (phasic) transmitter release at the frog neuromuscular junction. As had been reported previously, we observed an increase in both spontaneous and evoked release with prolonged exposure to ouabain. We also found that following ouabain treatment, evoked release was maintained for a much longer period of time upon removal of extracellular Ca2+ than prior to ouabain exposure. These results indicate that after exposure to ouabain evoked transmitter release can occur in the absence of appreciable Ca2+ entry.

Changes in MEPP frequency during depression of evoked release at the frog neuromuscular junction.

1. Endplate potentials (EPPs) and miniature endplate potentials (MEPPs) were recorded from frog neuromuscular junctions bathed in Ringer solutions containing normal (1.8 mM) or high (3.6 mM) Ca2+. The peptide toxin mu-conotoxin GIIIA was added to the Ringer solution to prevent muscle action potentials and contraction. 2. The nerve was stimulated with conditioning trains of 200-4800 impulses applied at 20 impulses s-1 to characterize the effects of repetitive stimulation on changes in EPP amplitude and MEPP frequency under high quantal conditions. 3. MEPP frequency was dramatically increased during and immediately following repetitive stimulation under high quantal conditions, whereas EPP amplitude was greatly depressed. There was no effect of repetitive stimulation on MEPP amplitude. 4. Following the conditioning stimulation the increase in MEPP frequency decayed back to the control level with a time course that could be described by four exponentials. The time constants of these exponentials were very similar to those that describe the components of stimulation-induced increases in EPP amplitude and MEPP frequency observed under low quantal conditions when depression is absent. 5. The results of this study indicate that depression and the components of stimulation-induced increases in release (facilitation, augmentation and potentiation) can be present at the same time, suggesting that the mechanism of depression involves different underlying factors from the mechanism(s) responsible for increases in release. They also indicate either that depression selectively affects only those quanta destined to be released in direct response to the nerve action potential, which would suggest that EPPs and MEPPs arise from different pools of transmitter, or that depression in some way affects a step in the release process involved only in evoked release, and not asynchronous (spontaneous) release.

Effects of calcium channel blockers on stimulation-induced changes in transmitter release at the frog neuromuscular junction.

We have examined the effects of various calcium channel blockers on stimulation-induced changes in end-plate potential (EPP) amplitude at the frog neuromuscular junction. We found that the addition of small concentrations (1-10 microM) of Cd2+ to the low calcium bathing Ringer reduced both the control EPP amplitude and the increase in EPP amplitude that normally occurs during repetitive stimulation under low quantal conditions. These effects of Cd2+, which developed rapidly following its addition to the bathing solution and were equally rapidly reversed, resulted from changes in the amount of transmitter released from the nerve terminal. The major effect of Cd2+ appeared to be on the facilitation and augmentation components of increased release. Cd2+ had little or no effect on potentiation of release. The other divalent cations tested, Zn2+, Co2+, and Ni2+, also decreased both control EPP amplitude and the stimulation-induced increase in EPP amplitude, but higher concentrations (> 100 microM) of these cations were required. The order of effectiveness in reducing stimulation-induced increases in EPP amplitude was: Cd2+ >>> Co2+,Zn2+ > Ni2+. The organic calcium channel blockers verapamil (20-100 microM) and nimodipine (20-50 microM) had little effect on stimulation-induced increases in EPP amplitude. The results of this study are consistent with previous suggestions that the different components of increased release represent different mechanisms. Furthermore, if Cd2+ is acting by reducing Ca2+ entry into the nerve terminal, then these results suggest that facilitation and augmentation are dependent in some way on Ca2+ entry.

Effect of glycerol treatment on transmitter release at the frog neuromuscular junction.

Exposure of frog skeletal muscle to a Ringer solution made hyperosmotic with glycerol, followed by return of the preparation to a normal Ringer, results in an irreversible uncoupling of the mechanical activity of the muscle. This technique for preventing muscle contraction has often been used to study neuromuscular transmission under normal or high levels of release without interference from muscle contraction. Little was known, however, about the effects of glycerol treatment on the process of transmitter release. In this paper we report the results of a study in which we examined the effects of glycerol treatment on transmitter release during repetitive stimulation at the frog sartorius neuromuscular junction. We found that glycerol treatment altered the stimulation-induced changes in end-plate potential (EPP) amplitude normally observed during repetitive stimulation at this synapse. This effect, which increased progressively with time for up to 8 h following removal of glycerol, could be accounted for by presynaptic changes in the amount of transmitter released.

Use of mu-conotoxin GIIIA for the study of synaptic transmission at the frog neuromuscular junction.

We have looked at the effect of synthetic mu-conotoxin GIIIA, a selective blocker of muscle Na channels, on various parameters of synaptic transmission at the frog sartorius nerve-muscle preparation. We found that 5 microM mu-conotoxin consistently blocked muscle action potentials, but had no effect on nerve action potentials. The toxin also had no effect on the amplitude or frequency of miniature endplate potentials (MEPPs), on the amplitude or time course of endplate potentials (EPPs), or on stimulation-induced changes in EPP amplitude. The lack of an effect of synthetic mu-conotoxin GIIIA on transmitter release makes this toxin an invaluable tool in the study of neuromuscular transmission under conditions of normal levels of release.

Kinetic and pharmacological examination of stimulation-induced increases in synaptic efficacy in the chick ciliary ganglion.

We have characterized the kinetic and pharmacological properties of stimulation-induced increases of synaptic efficacy in the embryonic chick ciliary ganglion. We found what appear to be four components of increased ganglionic efficacy with average time constants of decay of about 60 msec, 400 msec, 30 sec, and 200 sec. These time constants are similar to the those describing the decay of the four components of stimulation-induced increases in neurotransmitter release characterized at other synapses. These components have been termed first and second components of facilitation, augmentation, and potentiation. We found that the addition of small amounts of Ba2+ to the low Ca2+ bathing solution led to an increase in the magnitude of the augmentation-like component, whereas Sr2+ enhanced the magnitude and time course of the component resembling the second component of facilitation. These effects of Ba2+ and Sr2+ are similar to the effects of these same divalent cations on augmentation and the second component of facilitation, respectively, at the frog neuromuscular junction and rabbit superior cervical ganglion. Based on these similar kinetic and pharmacological properties, we conclude that the four components of stimulation-induced increases in release that have been described in other synaptic preparations also appear to be present in the chick ciliary ganglion.

omega-Conotoxin reduces facilitation of transmitter release at the frog neuromuscular junction.

We have examined the effects of the peptide toxin omega-conotoxin GVIA (omega-CgTx), a known calcium channel blocker, on stimulation-induced changes in end-plate potential (EPP) amplitude at the frog neuromuscular junction. We found that the addition of this toxin in submicromolar concentrations reduced both the control EPP amplitude and the increase in EPP amplitude that normally occurs during repetitive stimulation under low quantal conditions. These effects of omega-CgTx developed slowly following its addition to the bathing solution, were concentration-dependent and were essentially irreversible. The effects of omega-CgTx appeared to result from reductions in the facilitation and augmentation components of stimulation-induced increases in release. While the effects of omega-CgTx on EPP amplitude could be reversed by increasing the extracellular concentration of Ca2+, we were unable to reverse the effects of the toxin on stimulation-induced increases in EPP amplitude. Thus it appears that omega-CgTx has a dual effect on neuromuscular transmission, perhaps by acting at two different presynaptic sites.

Plasticity of medial gastrocnemius motor units following cordotomy in the cat.

Experiments were performed in adult cats to determine the effects of lumbar cordotomy on synaptic potentials, motoneuron membrane electrical properties, muscle-unit contractile properties, and whole-muscle histochemical properties of a heterogeneous skeletal muscle. Medial gastrocnemius (MG) motor units were examined 1 wk to 7 mo following complete transection of the lumbar spinal cord (cordotomy). Motor units were classified on the basis of their contractile properties as type FF, FI, FR, or S (8, 68). Muscle fibers were classified as type FG, FOG, or SO on the basis of histochemical staining (59). Motoneuron electrical properties (axonal conduction velocity, action-potential amplitude, rheobase, input resistance, afterhyperpolarization), group I EPSPs, and muscle-unit contractile properties (unpotentiated and potentiated twitch, unfused and fused tetanus, fatigability) were measured. Reduced numbers of type FR motor units and increased numbers of types FI + FF motor units were found in electrophysiological experiments 2 wk to 7 mo following cordotomy. Corroborative data were obtained from histochemical studies of the same MG muscles. Electrical properties of the motoneurons of each motor-unit type were normal following cordotomy. The close correspondence between motoneuron electrical properties and muscle-unit contractile properties found in normal MG muscle (68) was preserved following cordotomy. Contractile strength of muscle units of all types was severely reduced following cordotomy; partial recovery occurred 4-7 mo following cordotomy. Cross-sectional area of muscle fibers was reduced at all times investigated (2 wk to 7 mo). In three cats, homonymous group Ia single-fiber-motoneuron EPSPs were studied 1 or 2 mo following cordotomy at spinal level L4-5 or L5. EPSP amplitude and afferent-to-motoneuron projection frequency were normal. In 12 other cats, composite heteronymous group I EPSPs were studied 2 wk to 7 mo following cordotomy at various levels. Amplitude of these EPSPs was increased, dependent upon level of cordotomy and postoperative time. Hypotheses concerning the influence of motoneurons on muscle, and of muscle on motoneurons, are presented as possible mechanisms whereby the close relation between motoneuron electrical and muscle-unit contractile properties is preserved in the face of redistributed motor-unit populations.

Membrane electrical properties and prediction of motor-unit type of medial gastrocnemius motoneurons in the cat.

Membrane electrical properties [time constant, action potential afterhyperpolarization (AHP), rheobase, input resistance, and axonal conduction velocity] were measured in motoneurons of cat medial gastrocnemius (MG) motor units. Motor units were classified on the basis of their mechanical responses as fast twitch, fast fatiguing (FF); fast twitch with intermediate fatigue resistance (FI); fast twitch, fatigue resistant (FR); or slow twitch, fatigue resistant (S; 11, 22). Motoneuron membrane time constant, estimated from the voltage response at the onset or termination of long (50-100 ms) current pulses and corrected for voltage-response nonlinearities (32), was found to differ significantly among the major motor-unit types, increasing in the order FF less than FR less than S. Afterhyperpolarization magnitude, half-decay time, and duration were all significantly greater for the fast (FF + FI + FR) versus the slow (S) motor units. The AHP half-decay time was correlated with muscle unit twitch time over the entire motoneuron population and within the type S motor-unit population. There was no significant correlation between twitch time and AHP half-decay time among the types FF and FR motor-unit populations. In agreement with previous studies, we found a significant difference in both rheobase and input resistance among the major motor-unit types, with rheobase increasing in the order S less than FR less than FF and input resistance decreasing in that order (S greater than FR greater than FF). The differences in input resistance were present both before and after correcting for voltage-response nonlinearities (32). Also in agreement with previous studies, the mean axonal conduction velocity was significantly faster among the fast (FF + FI + and FR) compared with the slow (S) motor units. These data were used to examine the properties alone to determine motor-unit type, which has traditionally been defined on the basis of the muscle unit's mechanical properties (11, 22). We used a discriminant analysis program to classify 73 mechanically typed motor units for which we had measures of rheobase, input resistance, membrane time constant, and AHP half-decay time. This model was able to properly classify 71 of the 73 motor units of this data set, indicating that the motor units of this data set could be grouped into three categories representing the three major motor-unit types (FF, FR, and S) on the basis of their rheobase, input resistance, membrane time constant, and AHP half-decay time.(ABSTRACT TRUNCATED AT 400 WORDS)

Synaptic and mechanical coupling between type-identified motor units and individual spindle afferents of medial gastrocnemius muscle of the cat.

Experiments were performed to test the possibility that motor unit-muscle spindle pairs that are coupled especially strongly mechanically will also be coupled especially strongly synaptically ("weighted ensemble input": Ref. 4). Synaptic and mechanical coupling between one or two individual muscle spindle afferents and individual motor units of the medial gastrocnemius (MG) muscle were measured in barbiturate-anesthetized cats. Synaptic coupling was assessed by measuring the amplitude of single-fiber monosynaptic excitatory postsynaptic potentials (EPSPs) generated in motoneurons by individual spindle afferents. Mechanical coupling was assessed by measuring the alteration in discharge rate of these spindle afferents caused by tetanic activation of the same motor units. Afferents were classified as primary or secondary on the basis of conduction velocity and response to muscle stretch and contraction. Motor units were classified as slow twitch (S); fast twitch, fatigue resistant (FR); fast twitch, intermediate fatigue resistance (FI); and fast twitch, fatigue sensitive (FF) on the basis of twitch contraction time and resistance to fatigue. In 85% of 138 motor unit-primary afferent interactions tested, tetanic activation of the single motor unit unloaded (i.e., decreased the discharge rate of) the primary afferent. A very weak though significant correlation was found between tetanic contraction strength and primary afferent unloading. In 66% of 155 motor unit-secondary afferent interactions tested, tetanic activation of the single motor unit unloaded the secondary afferent. Again, afferent unloading was but weakly related to tetanic contraction strength. Single-fiber EPSPs generated by primary or secondary muscle spindle afferents were recorded in type-identified motor units. EPSPs generated by primary afferents were significantly larger in oxidative (S + FR) than in glycolytic (FF) motor units. No such differences were seen for EPSPs generated by secondary afferents. The magnitude of the EPSP generated in a motoneuron by a spindle afferent was compared to the magnitude of the unloading of that afferent by tetanic activation of the corresponding motor unit. Overall, no relationship was found between these measures.(ABSTRACT TRUNCATED AT 400 WORDS)

Presynaptic inhibition, EPSP amplitude, and motor-unit type in triceps surae motoneurons in the cat.

1. Composite group Ia excitatory postsynaptic potentials (EPSPs) produced by heteronymous nerve stimulation were recorded from triceps surae motoneurons of barbiturate-anesthetized cats. Motoneuron rheobase, input resistance, and axonal conduction velocity were measured, and motor units were classified on the basis of the mechanical responses of their muscle units. 2. The amplitude of EPSPs recorded from 33 medial gastrocnemius (MG) motoneurons ranged from 0.6 to 4.3 mV. The mean EPSP amplitude differed among the major MG motor-unit types, increasing in the order fast twitch, fast fatiguing (FF); fast twitch, fatigue resistant (FR); slow twitch, fatigue resistant (S) (FF less than FR less than S). The amplitude of EPSPs recorded from 15 soleus motoneurons ranged from 0.3 to 3.4 mV, with a mean of 1.4 mV. 3. Presynaptic inhibition of EPSPs was produced by trains of conditioning volleys in the posterior biceps-semitendinosus (PBST) nerve. In 33 MG cells PBST conditioning stimulation reduced the amplitude of EPSPs by 11-50%, with a mean inhibition of 27%. The amplitude of EPSPs in 15 soleus motoneurons was decreased by 5-84%, with a mean inhibition of 37%. 4. When the magnitude of presynaptic inhibition was expressed as percent inhibition, there was no relation between presynaptic inhibition and either motor-unit type or the amplitude of the EPSP. However, when presynaptic inhibition was expressed as the absolute amount of inhibition in millivolts, the magnitude of inhibition was highly correlated with EPSP amplitude both across the entire triceps surae population (MG, lateral gastrocnemius, soleus) as well as within each muscle population. This correlation was also significant within the MG FF and FR motor-unit populations. 5. We conclude that EPSP amplitude and not motor-unit type is the major determinant of the magnitude of presynaptic inhibition. However, because of the effect of motor-unit type on EPSP amplitude, the net effect is that presynaptic inhibition increases in the order FF less than FR less than S.

Augmentation and facilitation of transmitter release. A quantitative description at the frog neuromuscular junction.

Endplate potentials were recorded from frog and toad sartorius neuromuscular junctions under conditions of greatly reduced quantal contents. The magnitudes of augmentation increased with the duration and frequency of stimulation, often increasing at an accelerating rate during 10-20-s conditioning trains. The magnitudes of the first and second components of facilitation also increased, but reached apparent steady state values within the first few seconds of stimulation. These observations could be accounted for by assuming (a) that augmentation and the first and second components of facilitation arise from underlying factors in the nerve terminal that act to increase transmitter release; (b) that each nerve impulse adds an increment to each of the underlying factors; (c) that the magnitude of the increment typically increases during the train for augmentation but remains constant for the components of facilitation; and (d) that the underlying factors decay with first-order kinetics with time constants of approximately 7 s for augmentation and 60 and 500 ms for the first and second components of facilitation, respectively. The increments of facilitation added by each impulse were about twice as large in the toad as in the frog. Facilitation was described better by assuming a power relationship between the underlying factor and the observed facilitation than by assuming a linear relationship. Augmentation was described by assuming either a linear or power relationship.

A quantitative description of stimulation-induced changes in transmitter release at the frog neuromuscular junction.

Endplate potentials were recorded from frog sartorius neuromuscular junctions under conditions of greatly reduced quantal contents to develop a quantitative description of stimulation-induced changes in transmitter release. Four general models relating potentiation, augmentation, and the first and second components of facilitation to transmitter release were developed. These models were then tested by incorporating equations for the kinetic properties of the four components of increased transmitter release and examining the ability of the resulting sets of equations to predict stimulation-induced changes in transmitter release. Three of the models were essentially consistent with the observation that augmentation had a multiplicative type relationship to facilitation. These models could also predict the effect of frequency and duration of stimulation on endplate potential (EPP) amplitude during and after prolonged (40 s) trains including the response to step changes in stimulation rate. These models extend by about two orders of magnitude the duration of stimulation-induced changes in transmitter release that can be accounted for, and show that the combined kinetic properties of potentiation, augmentation, and the two components of facilitation are generally sufficient to account for these changes.

Changes in miniature endplate potential frequency during repetitive nerve stimulation in the presence of Ca2+, Ba2+, and Sr2+ at the frog neuromuscular junction.

Miniature endplate potentials (MEPPs) were recorded from frog sartorious neuromuscular junctions under conditions of reduced quantal contents to study the effect of repetitive nerve stimulation on asynchronous (tonic) quantal transmitter release. MEPP frequency increased during repetitive stimulation and then decayed back to the control level after the conditioning trains. The decay of the increased MEPP frequency after 100-to 200-impulse conditioning trains can be described by four components that decayed exponentially with time constants of about 50 ms, 500 ms, 7 s, and 80 s. These time constants are similar to those for the decay of stimulation-induced changes in synchronous (phasic) transmitter release, as measured by endplate potential (EPP) amplitudes, corresponding, respectively, to the first and second components of facilitation, augmentation, and potentiation. The addition of small amounts of Ca2+ or Ba2+ to the Ca2+-containing bathing solution, or the replacement of Ca2+ with Sr2+, led to a greater increase in the stimulation-induced increases in MEPP frequency. The Sr-induced increase in MEPP frequency was associated with an increase in the second component of facilitation of MEPP frequency; the Ba-induced increase with an increase in augmentation. These effects of Sr2+ and Ba2+ on stimulation-induced changes in MEPP frequency are similar to the effects of these ions on stimulation-induced changes in EPP amplitude. These ionic similarities and the similar kinetics of decay suggest that stimulation induced changes in MEPP frequency and EPP amplitude have some similar underlying mechanisms. Calculations are presented which show that a fourth power residual calcium model for stimulation-induced changes in transmitter release cannot readily account for the observation that stimulation-induced changes in MEPP frequency and EPP amplitude have similar time-courses.

Facilitation, augmentation, and potentiation of synaptic transmission at the superior cervical ganglion of the rabbit.

The effect of repetitive stimulation on synaptic transmission was studied in the isolated superior cervical ganglion of the rabbit under conditions of reduced quantal content. Excitatory postsynaptic potentials (EPSP) were recorded with the sucrose gap technique to obtain estimates of transmitter release. Four components of increased transmitter release, with time constants of decay similar to those observed at the frog neuromuscular junction at 20 degrees C, were found in the ganglion at 34 degrees C: a first component of facilitation, which decayed with a time constant of 59 +/- 14 ms (mean +/- SD); a second component of facilitation, which decayed with a time constant of 388 +/- 97 ms; augmentation, which decayed with a time constant of 7.2 +/- 1 s; and potentiation, which decayed with a time constant of 88 +/- 25 s. The addition of 0.1-0.2 mM Ba2+ to the Locke solution increased the magnitude but not the time constant of decay of augmentation. Ba2+ had little effect on potentiation. The addition of 0.2-0.8 mM Sr2+ to the Locke solution appeared to increase the magnitude of the second component of facilitation. Sr2+ had little effect on augmentation or potentiation. These selective effects of Ba2+ and Sr2+ on the components of increased transmitter release in the rabbit ganglion are similar to the effects of these ions at the frog neuromuscular junction. Although the effects of Ba2+ and Sr2+ are similar in the two preparations, the magnitudes of augmentation and the second component of facilitation after a single impulse were about 6-10 times greater in the rabbit ganglion than at the frog neuromuscular junction. These results suggest that the underlying mechanisms in the nerve terminal that give rise to the components of increased transmitter release in the rabbit ganglion and frog neuromuscular junction are similar but not identical.

Differential effects of Ba2+, Sr2+, and Ca2+ on stimulation-induced changes in transmitter release at the frog neuromuscular junction.

Endplate potentials (EPP) were recorded from the frog sartorius neuromuscular junction under conditions of low quantal content to study the effect of Ba2+, Sr2+, and Ca2+ on the changes in evoked transmitter release that occur during and after repetitive stimulation. The addition of 0.1-1 mM Ba2+ or Sr2+ to the Ca2+-containing bathing solution, or the replacement of Ca2+ with 0.8-1.4 mM Sr2+, led to a greater increase in EPP amplitudes during and immediately after repetitive stimulation. These changes in release were analyzed in terms of the four apparent components of increased transmitter release that have previously been distinguished on the basis of their kinetic properties. The Ba2+-induced increase in EPP amplitudes was associated with an increase in the magnitude but not the time constant of decay of augmentation. Ba2+ had little effect on potentiation or the first and second components of facilitation. The Sr2+-induced increase in EPP amplitudes was associated with an increase in the magnitude and the time constant of decay of the second component of facilitation. Sr2+ had little effect on potentiation, augmentation, or the first component of facilitation. The selective effects of Ba2+ on augmentation and of Sr2+ on the second component of facilitation were reversible and could be obtained in the presence of the other ion. The addition of 0.1-0.3 mM Ca2+ to the bathing solution had little effect on potentiation, augmentation, or the two components of facilitation. These results provide pharmacological support for the proposal that there are four different components of increased transmitter release associated with repetitive stimulation and suggest that the underlying factors in the nerve terminal that give rise to these components can act somewhat independently of one another.

Transmitter release during repetitive stimulation: selective changes produced by Sr2+ and Ba2+.

The addition of Sr2+ or Ba2+ to the solution bathing the frog neuromuscular junction leads to an increased release of transmitter by each nerve impulse during and following repetitive stimulation. The mechanisms by which Sr2+ and Ba2+ increase release are not the same. Each ion appears to act selectively on a different process involved in transmitter release.

Stimulation-induced factors which affect augmentation and potentiation of trasmitter release at the neuromuscular junction.

1. End-plate potentials (e.p.p.s) were recorded from frog sartorius neuromuscular junctions under conditions of decreased transmitter release to study the effect of repetitive stimulation on augmentation and potentiation of transmitter release. 2. The magnitudes and time constants of decay of augmentation and potentiation were determined both following a primary conditioning train and following an identical secondary conditioning train applied from 30 to 170 sec after the primary conditioning train. 3. The magnitude of augmentation following the secondary conditioning trains was increased over that following the primary conditioning trains even though augmentation, with a time constant of decay of about 7 sec, had decayed to insignificant levels before the onset of the secondary trains. This increase in augmentation was not due to a change in its rate of decay during the secondary trains. 4. The increased magnitude of augmentation can be described as arising from an expression factor which, for conditioning trains of 200 impulses at 20/sec, has an initial magnitude of 1-6 +/- 1-2 (S.D. of observation) (the magnitude of augmentation is increased 2-6 times) and decays approximately exponentially with a time constant of 90 +/- 50 (S.D. of observation) sec. The expression factor thus decays about ten times slower than augmentation. 5. Doubling the number of impulses in the primary conditioning train from 100 to 200 led to a 2-8 +/- 1-0 (S.D. of observation) times increase in the magnitude of the expression factor, estimated by placing a 200 impulse secondary conditioning train 40 sec after the primary conditioning train. 6. The expression factor, while increasing the magnitude of augmentation, had little or no effect on the magnitude of potentiation or on trasmitter release in the absence of augmentation. The expression factor decayed about twice as slowly as potentiation. 7. The time constants characterizing the decay of potentiation were greater following the secondary conditioning trains than following the primary conditioning trains. 8. The increased time constant for the decay of potentiation can be described as arising from a time constant factor which, for conditioning trains of 200 impulses at 20/sec, has an initial magnitude of 1-2 +/- 0-7 (S.D. of observation) (the time constant of potentiation is increased 2-2 times) and decays approximately exponentially with a time constant of 130 +/- 45 (S.D. of observation) sec. The time constant factor decayed about three times slower than potentiation. 9. Doubling the number of impulses in the primary conditioning train from 100 to 200 led to a 1-6 +/- 0-8 (S.D. of observation) times increase in the magnitude of the time constant factor, estimated by placing a 200 impulse secondary conditioning train 40 sec after the primary conditioning train. 10...