The Electrical and Mechanical Events of Neuro-Muscular Transmission in the Cockroach, Periplaneta Americana (L.)^*

Adults of the American cockroach, Periplaneta americana, were used throughout. Though the same results were obtained with both sexes, males were preferred owing to the relative absence of fat within the thorax. The insects were decapitated and pinned ventral side up in a shallow Petri dish which had been half filled with wax. The metathoracic legs were widely spread and held in a position of retraction by means of pins. The metathoracic ganglion, nerves, and the extrinsic leg muscles were exposed by removal of the ventral cuticle, surrounding fat and tracheae (Pl. 1). Nerve 3 (Pringle, 1939) was prepared for placement of stimulating electrodes by cutting branch B which supplies muscles within the coxa, and by tracing branch A to the tergal muscles of the trochantin (Carbonell, 1947). The latter consist of three long (5-6 mm.) and narrow muscles which originate from the anterior part of the metathoracic tergum and pass downward and backward to a common insertion on the median end of the trochantin close to the anterior median end of the coxa. From the ventral approach they are overlapped by the broad episternal pro-motor muscle, which must be removed in part before they can be completely exposed. For some studies the whole group was employed, though the effect of stimulation of nerve 3 could be limited to the second tergal muscle of the trochantin (muscle 162 of Carbonell) by careful section of small branches from nerve 3 A which pass to the other muscles. This muscle, which will be designated as muscle 162, is the most medial member of the group. It was the subject of most of the experiments reported below.

Reflex responses were prevented by cutting other nerves leaving the ganglion, and by crushing the ganglion with forceps. Nerve 3 was allowed to remain attached to the remains of the ganglion, since its small size made handling difficult and survival short after it had been severed.

Sufficient insect Ringer was added so as to cover the wax at the bottom of the dish. Seepage of Ringer up through the cut surfaces of the cockroach kept the preparation in excellent condition for several hours with little or no addition of fluid to the surface of the muscle. Stimulation of nerve 3 was accomplished while it lay just below the surface film of liquid collecting between muscle and ganglion. If raised above the liquid the nerve dried within a few minutes, while immersion in a pool of mineral oil caused shrinkage and deterioration.

Stimulation was accomplished by fine tapered and hooked silver electrodes, which were brought into position below the nerve by means of a manipulator (Pl. 1). Muscle potentials were recorded from similar electrodes placed directly on the exposed surface of the muscle. For monopolar recording, an ’indifferent’ electrode was obtained by placing a ring of silver wire, 5 cm. in diameter, in the saline surrounding the animal.

The onset of contraction was detected by a fine glass stylus placed on the muscle surface near the electrodes used to record potential change. The stylus was inserted in the needle socket of a piezo-electric pick-up of the type ordinarily used in phonograph play-back equipment. The pick-up was connected to one of the amplifiers and cathode-ray deflexion system. It served as an extremely sensitive indicator of rapid and small movements because of the high gain of the amplification used. It was used only to determine the onset of mechanical change in the muscle, hence its linearity and frequency characteristic are of no consequence.

Recording equipment consisted of two Grass P3 resistance-capacity coupled amplifiers with a flat response up to 15 kcyc./sec. After further amplification, potential changes provided the Y deflexion for two beams of a du Mont 5 SP11 cathode-ray tube. The X deflexion of both beams was accomplished by the same sweep circuit throughout. The stimulus was a square pulse of variable duration triggered by the sweep pulse. A stimulus to nerve 3 of 1-2 V. (at the stimulator) and o.2 msec, duration ordinarily caused a stepless response of muscle 162.

All experiments were carried out at room temperature.

Temporal relations of the muscle potential. The sequence of electrical and mechanical events following indirect stimulation of muscle 162 are shown in Text-fig. 1. Potential changes on the surface of the muscle were recorded between electrode R₁ placed in the middle of the muscle opposite the point of nerve entrance, and electrode R₂ 2 mm. distant on the anterior end of the muscle.

Following the stimulus artifact by 0-5 msec, is a small upward deflexion produced by the action potential in the motor nerve as the latter enters the muscle substance. The muscle potential appears as a downward deflexion (positivity of R₁) about 1.0-1.3 msec, after the nerve spike. It reaches a maximum in 13msec., falling to about one-half of its peak value in a similar period. This is followed by a slower decline, leaving in some cases a small potential which persists throughout the contraction. The major part of the muscle potential occupies 4-5 msec., and muscle shortening begins abruptly during the latter part of its declining phase. The latent period (from onset of muscle potential to onset of contraction) lasts 2-5-3-2 msec. Using mechanical recording, Solf (1931) gives a latent period of 9-0 msec, for the same muscle in Decticus verrucivorus. The shorter value given here is undoubtedly due to the relatively inertia-free method of registration, and the fact that nerve conduction and end-plate delay are not included in the measurement. In some piezo-electric recordings of the muscle contraction, a small degree of relaxation appears to precede the active shortening. This may be an example of latent relaxation, though it could be due to muscle movement taking place at some point distant from the recording stylus.

Sign of the muscle potential. When the muscle potential is recorded between two electrodes on the muscle surface, that nearest to the point of nerve entrance (Text-fig. 1, R₁) becomes positive to an electrode (R₂) placed at either end. Under these circumstances the nerve-action potential has the usual negative sign. In most cases active muscle becomes positive to an ’indifferent’ electrode. In the experiment recorded in Text-fig. 2 an active electrode (R₁) on the muscle is pitted against an electrode (R₀) in contact with the saline surrounding the roach. The lower tracing records the positive potential of R₁ against R₀ as the former is moved to various points along the muscle. The maximum deflexion is produced when R₁ is on the middle of the muscle opposite the point of nerve entrance. Rr is also pitted against another electrode R₂ also on the muscle surface. Rr and Ra are separated by a constant distance of 2 mm. as the pair is moved together to different points on the muscle surface. It can be seen that the R₁R₂ potential reaches a minimum and changes sign when the pair is near the point of nerve entry. This shows that a fairly steep potential gradient develops along the muscle surface during activity, the surface near the point of nerve entry becoming positive to any other point on the surface, which is in turn positive with respect to the ’indifferent’ electrode.

Attempts were made to remove the nerve and muscle entirely from the body. The handling and injury caused by this procedure resulted in a rapid loss of irritability, and it was impractical to cut the muscle away from its origin and insertion on the cuticle. In one experiment both nerve and muscle were removed from the insect leaving some of the cuticle attached to the origin and insertion. This preparation was placed in a shallow dish containing saline and an indifferent electrode. On stimulation of the nerve, the muscle electrode became negative with respect to the saline electrode, both nerve spike and muscle potential appearing as upward deflexions. When two electrodes were placed on the muscle surface of this preparation the central region became positive to either end though the nerve spike retained its negative sign.

Insertion of an electrode tip into the muscle substance made no difference to the recorded sign of the potential, which increased up to a point, only to decrease as the electrode passed right through the muscle. Placement of an electrode on the cut ends of a severed muscle did not give satisfactory results owing to rapid loss of irritability. In one or two cases where a few observations could be made, the surface electrode became positive to the cut surface during activity.

Though active muscle in situ was most frequently positive, the sign of the muscle potential was often quite capricious. In one experiment the anterior part of the muscle became positive to a saline electrode while the posterior part became negative. Wiersma & Wright (1947) noted that many active crustacean muscles developed a positive potential, and suggested that in these cases innervation of each fibre from one side or from within the muscle substance made it impossible to bring a surface electrode into contact with the active region. They concluded that the external cuticle afforded a better electrical connexion through the muscle insertion with the cathodally depolarized muscle surfaces than did an electrode placed on the body of the muscle, hence a distant or ’indifferent’ electrode could become negative to an electrode placed on the muscle surface. It seems probable that the same explanation applies to insect-muscle potentials.

Form of the potential. Provided that not more than one muscle was excited by motor-nerve stimulation, the muscle potential invariably took a monophasic form. This occurred irrespective of stimulus frequency, interelectrode distance, or form of recording (monopolar or bipolar). At first it was thought that the monophasic curve obtained with bipolar recording (Text-fig. 1) was due entirely to the local or nonpropagated nature of the muscle potential. Though this was partly responsible, observation of a muscle contracting rhythmically under stroboscopic illumination shows that all parts of the muscle participate in the shortening process. Therefore, excitation must be propagated throughout the muscle, whether it be through the agency of nerve or muscle fibres.

Recordings were made of the time relations of stimulus artifact, nerve spike, and muscle potential at an electrode placed at measured distances along the muscle and pitted against a saline electrode (Text-fig. 3). Comparison of the oscillograms shows that the time interval between shock artifact and nerve spike remains constant at all points on the muscle, though the nerve potential becomes vanishingly small as the exploratory electrode moves away from the point where the nerve is seen to disappear within the muscle substance. This suggests that the nerve potential originates from the motor nerve at this point, and that spike potentials in motor-nerve branches approaching their terminations cannot be recorded with a surface electrode. Therefore, the neuro-muscular delay cannot be measured directly, though it must have some value less than fo-i’3 msec., the interval between nerve spike and muscle potential.

As a single exploratory electrode is moved from the point of nerve entrance toward either end of the muscle (Text-fig. 3), the muscle potential becomes reduced in height and shows a delay in onset. Comparison of oscillograms C and F in Text-fig. 3 shows that the muscle potential begins 0-4-0-5 msec, later in F, which was recorded 2-3 mm. distant from C. Though measurements over these short distances have a large error, this would give a rate of spread of the muscle potential of 5-6 mm./msec. Since the duration of the muscle potential is 4-5 msec., its wavelength would be 20-30 mm. or 8-10 times the possible interelectrode distance for two electrodes placed on one-half of the muscle (Text-fig. 1). Therefore, the onset of potential change under the distal electrode of a pair on the muscle is only slightly displaced relative to change under the proximal electrode, and the monophasic potential is due almost entirely to the potential gradient developed along the muscle, owing little or nothing to propagation.

The motor unit and summation. Providing the response to a single indirect stimulus was limited to a single muscle, electrical and mechanical responses always occurred together and showed no stepwise increase with changes in stimulus strength. It is concluded that the response of muscle 162 was due to stimulation of a single motor fibre which innervates all the muscle fibres. This has been demonstrated by Pringle (1939) for other muscles in the cockroach. He also showed that each muscle was supplied by a ’quick’ fibre which was responsible for a muscle potential and maximal twitch showing no summation at any stimulus frequency (up to 150 per sec.), and a ’slow’ fibre which operated only through summation and caused a gradual shortening of the muscle on repetitive stimulation.

A few experiments were carried out with pairs of indirect stimuli separated by a variable interval (Text-fig. 4). When the stimulus interval was less than i-8 msec. (Text-fig. 4, A, B), the second shock evoked no electrical response. When it was lengthened to 1-9-2-0 msec., the muscle potential due to the second shock appeared as a small hump on the decaying phase of the first muscle potential (Text-fig. 4, C). Further separation of the stimuli up to a 4-0 msec, interval caused rapid growth of the second potential. Slower growth continued up to a stimulus interval of 10 msec, when the second potential approached the first potential in size (Text-fig. 4, G). Since the whole muscle appears to be innervated by a single ’quick’ fibre, changes in size of the muscle potential must be due to changes at the neuro-muscular junctions. The interval of 2-10 msec, occupied by the recovery of the muscle potential could then be occupied by the progressive return to activity of individual neuro-muscular junctions from a refractory residue. The whole picture is one of depression rather than summation.

The mechanical response of the muscle to two stimuli runs parallel to the electrical response. It must be noted that registration of shortening by means of the piezo-electric pick-up introduces serious distortion during the later phases of contraction. This is seen as a dip below the base-line (Text-fig. 4, upper trace in each record). In addition, the amplifier coupling prevents registration of any sustained shortening so that the mechanical records must be viewed with some reservations. However, no prolongation of the muscle contraction is evident until the stimulus interval approaches 5-0 msec. (Text-fig. 4, F). While the stimulus interval lies between 5-0 and io-o msec, the twitches remain fused, though when allowance is made for the distortion mentioned above, the second peak appears to be much lower than the first. With an interval of 10.0-20.0 msec, the twitches separate (Text-fig. 4, G, H), and no mechanical summation is evident.

Direct stimulation of the muscle. Placement of stimulating electrodes directly on muscle 162 caused an electrical response with similar threshold and latency to that produced by stimulation of nerve 3, and there is every reason to suspect that the response was due to stimulation of small branches of the motor nerve within the muscle substance.

Curare (b tubocurarine Squibb, 10-3) failed to alter the indirect response of muscle 162. Therefore, in order to eliminate the motor nerve as a source of muscle stimulation, nerve 3 was severed surgically in a number of adult male cockroaches. After some practice, this could be done through a small incision next to the ganglion with no damage to the latter or to other nerves. Approximately 90% of the insects survived, and appeared to be entirely normal except for a motor defect in one metathoracic leg. They were kept in a container with food and water, and individuals were withdrawn at intervals for experimentation. Muscle 162 was exposed on the denervated and on the intact side of each insect, and both muscles were subjected to direct electrical stimulation.

For 3-5 days after nerve section, there was little difference in the response to electrical stimulation of denervated and normal muscles. Quite abruptly at the end of this period, the denervated muscles failed to show any response whatever, and a similar situation existed up to 14 days after nerve section, which is the longest period over which operated insects have been followed. Stimuli up to 50 V. (at the stimulator) and 10 msec, in duration failed to stimulate the denervated muscles, though a stimulus of 1-3 V. and 01-0-2 msec, duration elicited a maximal response from the intact muscle 162 in the same insects. With the thought that perhaps the chronaxie of denervated muscle was very different from that of nerve, direct current with a rapid rate of rise was applied through electrodes placed on the muscle surface. Though the maximum current strength used caused violent electrolysis at the electrode site and marked contraction of normal muscles in other parts of the body of the roach, there was no detectable response in the denervated muscles. The preparations were examined closely through a binocular dissecting microscope during stimulation, and a check was made by placing the piezo-electric pick-up close to the point of stimulation. In not one out of twelve successfully denervated muscles could any movement be attributed to the electrical stimulus.

There appears to be no published work on the structural changes in insect muscle and nerve after motor-nerve section, and a histological study is in progress. However, within 14 days of nerve section, fresh denervated muscle showed no gross or microscopic changes in colour, size, appearance, or presence of striations. Visible traces of the peripheral nerve stump were lost after 3 days. It seems to be established that cockroach muscle becomes electrically inexcitable after denervation. This suggests that the muscle can be excited only through the agency of the motor-nerve impulse.

Potential changes in crustacean muscles are similar in many ways to those described for insects. In addition to reporting that muscle potentials in Cambarus had a positive sign, Wiersma & Wright (1947) noted that they were generally monophasic. These authors conclude that the recorded potential represented the sum of local muscle potentials developed at many innervation sites along the muscle fibres, van Harreveld (1939) concluded that crustacean muscle fibres could not be directly excited and that a muscle conduction process was absent. From experiments on neuro-muscular transmission in various marine crabs and fresh-water crayfish, Katz & Kuffler (1946) concluded that single stimuli applied directly or through the motor nerve caused local graded electrical and mechanical responses in crustacean muscle. These negative local potentials were analogous to the end-plate potential of vertebrate muscle and readily summed on repetitive stimulation to produce a potential several times the unit height, from which propagated muscle action potentials took off to produce total muscle twitches.

The experiments reported in this paper suggest a similar conclusion to that reached by van Harreveld (1939) and Wiersma & Wright (1947), namely, that in arthropod muscles conduction is accomplished over nervous rather than muscular pathways. Observations which bear out this conclusion are :

1. The presence of only one form of quick fibre muscle potential at a variety of stimulus strengths and frequencies. No evidence of a transition from local potential into propagated spike, or of local potential summation as reported by Katz & Kuffler (1946).

2. A marked potential gradient developed during excitation between the muscle surface near the point of nerve entry, and the muscle surface at either end. This suggests greater depolarization in those regions of the muscle where motor endings might be expected to be most dense. The monophasic potential recorded from two electrodes on the muscle surface is due to this gradient rather than to later arrival of excitation at the ends of the muscle.

3. Rapid spread (4-6 m./sec.) of excitation in normal muscle from the point of nerve entry toward either end. All parts of the muscle appear to be involved within 0-5 msec. In contrast, there is a complete loss of muscle excitability to electrical stimulation after the motor nerve has been allowed to degenerate.

If this evidence is accepted, the muscle potential in the cockroach is analogous to the local (end-plate) potential of vertebrate and crustacean muscle, the potential change recorded at the muscle surface being the sum of local potentials developed simultaneously in several fibres and sequentially at several points on each fibre. At this point the similarity ends, since in normal vertebrate muscle the local potential invariably grows into a muscle-propagated wave of depolarization, in crustacean muscle it may (Katz & Kuffler, 1946) or may not (Wiersma & Wright, 1947) develop into muscle-propagated excitation, while in insects it remains as a large number of local responses, excitation and propagation throughout the entire muscle being accomplished by the motor nerve. Since curare prevents the development of local into conducted responses in vertebrates, its lack of action on cockroach, and in other arthropod neuro-muscle preparations, could be due to the absence of muscle conduction in arthropods.

It should be pointed out that although the muscle potential appears to be the sum of local responses throughout the muscle, the response of muscle 162 to single indirect stimuli is strictly all-or-none. Hence an impulse in the motor nerve must reach every part of the muscle capable of depolarization, giving a total electrical and mechanical response as far as the quick fibre is concerned. The lack of summation and reduced response to the second of a pair of stimuli indicates relative refractoriness at motor endings in the muscle for 2-10 msec. It is not clear whether this is due to reduced responses at all the innervation points, or to failure of a certain percentage of endings to respond to the second nerve impulse.

Attempts to examine the structural relations of motor nerve and muscle fibres in muscle 162 have not met with much success. Methylene blue and Golgi methods have not differentiated the finer nerve branches, though profusely branched segments of nerve have been seen in one or two teased preparations. However, many early workers have noted the presence of several motor endings or Doyere cones on each insect muscle fibre. Marcu (1929) records 12 and 20 Doyere cones on a millimetre length of thoracic muscle fibre from Geotrupes stercorarius and from Musca domestica respectively. This author describes each Doyere cone as an abrupt and brush-like branching of each small division of the motor nerve as it reaches the surface of the sarcolemma. The extremely fine fibres composing the brush penetrate the sarcolemma and pass into the sarcoplasm among the myofibrillae. He noted that the number of Doyere cones varied in different regions of the same muscle fibre, though in some places they were so closely packed that the ’ brushes ’ overlapped and intermingled. In thoracic muscle of Phyllodromia germánica, Marcu noted that compact Doyere cones were absent, though the motor nerve branched profusely, and its terminal fibres made contact with the sarcolemma of each muscle fibre at many points. Foettinger (1880) made an examination of muscle fibres from several species of insects which had been rapidly fixed while in a state of partial contraction. He concluded that local contractions could occur only in the vicinity of motor endings. This anatomical picture makes it difficult to see how propagated excitation in insect muscle is either necessary or possible, and the motor nerve appears to provide the sole means of conduction.

Some light on the cause of the positive potential in insect muscle is shed by the observations of Marcu (1929). He noted that Doyere cones were distributed mainly along one side of each muscle fibre, and in one of his figures consisting of a cross-section of eight muscle fibres, the Doyere cones are represented as lying only upon the inner surface of each muscle fibre of the group. This is in accord with the explanation proposed by Wiersma & Wright (1947) for positive muscle potentials in Crustacea. They suggested that the locally active and cathodally polarized muscle surfaces were arranged in such a manner as to make better electrical connexion with the muscle insertion and external cuticle than with the overlying muscle surface. In a cylindrical muscle this could be realized if the asymmetrically innervated muscle fibres were grouped in cylindrical bundles with the motor nerve running in the centre of the bundle and innervating each fibre at several points along its inner surface. The catelectrotonus developed on excitation would be limited in the absence of muscle conduction to the inner side of each fibre, and the negativity developed in the core of the bundle might cause inward current flow from the muscle insertions if the fibres were closely packed. Hence, the muscle surface would become a source of current which would flow towards muscle origin and insertion.

The complete lack of response of denervated muscle to any form of electrical stimulation remains very puzzling. Absence of propagated excitation in denervated muscle is not unexpected if the motor nerve normally performs this function, but the absence of a local response of any kind cannot be explained. Either the muscle surface has such high electrical resistance that externally applied current fails to reach the depolarizable regions of muscle fibres arranged as described in the last paragraph, or cockroach muscle is electrically inexcitable. In order to examine the last possibility studies on the chemical sensitivity of denervated muscle are in preparation. No way can be seen by which the first possibility could be tested.

1. A nerve-muscle preparation in the metathorax of the cockroach is described. It consists of the second tergal muscle of the trochantin (muscle 162 of Carbonell) innervated by a branch of nerve 3 A. Electrical changes are recorded from electrodes on the muscle surface, and the onset of contraction is registered by the stylus of a piezo-electric pick-up.

2. With low (3-5 per sec.) stimulation rates at room temperature the neuromuscular delay is less than 1.2 msec., and the latent period of contraction about 3.0 msec. The muscle potential is 4-5 msec. in duration, positive in sign at the muscle surface, and monophasic in form with either monopolar or bipolar recording. During excitation a potential gradient develops along the muscle, the greatest positivity being in the middle near the point of nerve entry.

3. Neither electrical nor mechanical response show gradations with changes in stimulus strength or frequency. No facilitation is evident, and the response appears to be due to stimulation of a single quick motor nerve fibre.

4. In order to study the effects of direct stimulation nerve 3 was sectioned and allowed to degenerate. All trace of the peripheral nerve stump was lost after 3 days, when the muscle became completely inexcitable to all forms of electrical stimulation. There were no gross structural changes which would account for this loss of excitability.

5. It is concluded that the recorded muscle potential in the cockroach is analogous to the vertebrate end-plate potential, being the sum of local muscle potentials developing simultaneously in several fibres, and sequentially at several innervation points along the same fibre. Conduction within the muscle is carried out entirely by motor nerve fibres.

6. Possible causes of the positive sign of the muscle during activity are discussed.

Fresh dissection of the metathoracic ganglion, nerves and tergal muscles of the trochantin. Nerves are numbered according to the usage of Pringle (1939). On the left stimulating electrodes have been placed below nerve 3 and recording electrodes rest on the tergal muscles of the trochantin. Muscle 162 lies just to the right of the electrode tips. During experimentation branch B and several smaller branches of nerve 3 would be severed.