Aspects of the Organization of Central Nervous Pathways in Aplysia Depilans

A preliminary account of results of such investigations on the so-called giant cells of Aplysia has already been published (Hughes & Tauc, 1961). In the present paper special attention is given to the general nature of pathways within the central nervous system and to the activity of cells in the abdominal ganglion while this retains its connexion with the rest of the animal. This preparation enables a study to be made of the influence of peripheral stimulation on the spontaneously active cells. Studies have also been made in isolated preparations of the different types of activity found in cells of the pleural ganglion and these are compared with what is already known about abdominal ganglion cells.

During the months of September and October many individuals of Aplysia depilans are found in the Bassin d’Arcachon where they feed on the Zostera and lay their eggs. The animals were freshly caught and kept in well-aerated sea water at the marine station. The specimens used were about io in. long and before any incisions were made the animals were stimulated mechanically to exude most of their purple and white secretions. The animal was then pinned out in a dissecting dish and the foot maintained in a fairly elongated position. The parapodia (or wings) were also pinned out on the two sides. An incision was first made from the base of the reduced mantle cavity anteriorly to the head. The viscera were thus revealed and the alimentary canal was quickly removed to prevent any of the gastric juices coming into contact with the nervous system during the isolating procedure. The dissection was well washed in sea water and the nervous system dissected according to the particular experiment. Isolation of the abdominal ganglia alone was readily accomplished in the usual way (Tauc, 1955). In other experiments the whole central nervous system with the pleural, pedal, cerebral and abdominal ganglia was carefully dissected with all their connectives and more important nerves intact. The isolated ganglia were then spread out in a waxed Petri dish and pinned through the sheath surrounding the ganglia. If micro-electrode recording was to be done this sheath was cut under the microscope by means of a micro-scalpel. The sheath immediately retracts from the cut and reveals the orange-coloured ganglion cells. In whole animal preparations only the first incision was made and the body wall pinned on both sides. Micro-electrode studies were then only possible on the abdominal ganglia because of the difficulty of pinning any others to a firm base, which was necessary if the electrode was to maintain its position during the limited movements of the animal. The great length of the connectives between the abdominal ganglia and the circumoesophageal ring made it possible to pin the abdominal ganglia to a waxed Perspex stage which was rigidly fixed to the dissecting dish.

Conventional electrical recording methods were used. Pairs of silver-silver chloride electrodes were placed on several nerves of the preparation and by means of a bipolar switching arrangement each pair of electrodes could be connected either to a stimulator or to a Grass pre-amplifier. In the case of whole animal preparations as many as four or five pairs of electrodes were placed in position at various places. The glass capillary micro-electrodes were filled with a 2 · 5 molar potassium chloride and connected to a cathode follower. The amplifiers fed into a double-beam oscilloscope and simultaneously to a four-channel pen recorder. The indifferent electrode was a coil of chlorided silver wire placed in sea water bathing the preparation. Changes in the polarization of the cell were achieved through the same electrode as was being used for recording by means of a bridge circuit (Araki & Otani, 1955).

Movements of the parapodia were recorded by means of an RCA 5734 mechanotransducer valve and a light lever hooked to the parapodium.

The descriptions of Mazzarelli (1893), McFarland (1909), and Eales (1921) give very complete accounts of the gross structure of the central nervous system and describe in some detail the nerves from various ganglia, but each is based on a different species of Aplysia. The three species are compared by Hoffmann (1939) who experienced difficulty in homologizing the nerves, partly because of differences in the completeness of these accounts. The description of Mazzarelli, based upon Aplysia depilans, is less complete concerning the precise innervation of periphral structures and in the present work most use was made of Eale’s account of A. punctata. It is apparent, however, that A. depilans differs in several respects from A. punctata and some of these may be related to the greater use which A. depilans makes of its parapodia during swimming. For instance, the third pedal nerve of A. punctata does not innervate the parapodia, whereas it does in A. depilans. Confirmation of this was obtained from both stimulating and recording experiments which showed that both sensory and motor fibres innervating the posterior portion of the wing were present in this large nerve, which in the present paper will be called the posterior parapodial nerve. The other nerves which arise from the pedal ganglia were also studied by recording and stimulating techniques. Two of them which innervate respectively the middle and anterior region of the parapodia were usually identifiable and are referred to as the middle and anterior parapodial nerves. A nerve which arises from the medial and ventral aspect of the pedal ganglia and quite definitely innervates the foot alone is described below as the pedal nerve. These four nerves are the most important in the present work and their position is shown diagrammatically in Fig. 1. where the connectives of the different ganglia are also indicated. The two cerebral ganglia are joined to the pleural ganglia by the cerebro-pleural connectives (C-pl.) and to the pedal ganglia by the cerebro-pedal connectives (C-pe). The pedal commissure (Pe.- pe.) joins the two pedal ganglia. The pleuro-pedal connectives are extremely short and it has not been possible to record from them in the present work. As described above, many nerves radiate from the pedal ganglia to innervate the foot and parapodia on both sides. Few nerves arise from the pleural ganglia, but passing backwards on both sides to the abdominal ganglion is a very long connective. These are referred to in the present work as the right and left (or sometimes pleuro-visceral) connectives. Then-precise homologies are doubtful and depend upon the interpretation of the posterior ganglionic mass, which is described here as the abdominal ganglion. Primitively, the visceral loop has along its length a parietal ganglion on each side and a posterior single or paired visceral ganglion. In Aplysia there seems to be general agreement that the right parietal ganglion has fused with the visceral ganglion and is represented by the right half of the abdominal ganglion. The fate of the left parietal ganglion is not so certain, though most authorities suppose that it is fused with the original visceral ganglion and together they form the left half of the final composite ganglion. These precise homologies are not of great importance here ; but the findings with respect to the pathways of the two giant cells, one in the right side of the abdominal ganglion and the other in the left pleural ganglion, may be of some significance in this respect (Hughes & Tauc, 1961).

Of the nerves leaving the abdominal ganglionic complex the stoutest is the branchial nerve which arises on the right side and innervates the gill and heart region. Another thick nerve on the left side innervates the anal region and also the siphon and is here called the siphon nerve.

Although these preparations were pinned down in a dissecting dish to limit their movements they continued to show motor activity for several hours. In some cases this took the form of waves of contraction passing backwards along one or both parapodia, retractile movements of the gill and siphon, rhinophores, and/or head. In addition peristaltic movements of parts of the gut were observed in some preparations, notably of the gizzard region. With frequent changes in the bathing sea water, some preparations continued to show motor activity, and in some cases micro-electrode recording from single abdominal ganglion cells maintained a constant pattern over periods up to 24 hr. As a general rule, however, the animal ceased to make marked movements after 3 or 4 hr. and most of the observations described below were made during this time.

Electrical activity could be recorded from nerves and ganglion cells even when there was no mechanical or other stimulation applied to the preparation. Such activity in any of the main nerves might be irregular but in some cases intermittent bursts of Varying frequency (o · 5 −5/min.) occurred. Bursts were made up of many units and were frequently but not necessarily accompanied by overt rhythmic movements of the animal notably of the parapodia. Sometimes these were wave-like movements, usually spreading backwards along the parapodia. At the same time bursts of activity were present in the parapodial nerves and often in the right connective. In other cases (Fig. 2 A) such bursts were accompanied by characteristic dorsal movements of the parapodia in which the waves were not so obvious and were associated with retraction of the gill, mantle region and sometimes of the head. Many units were active during such bursts, but one which was very characteristic because of its large size is known to be the axon of the giant cell of the abdominal ganglion. Notice the depression of activity in the right connective after this activity. As can be seen from the records, other units are continuously active but increase in frequency during these bursts. One particular unit, which produces a potential smaller than that of the giant cell, propagates in an anterior direction from the abdominal ganglion and with a velocity greater than that of the giant cell axon. The parapodial movements would bring them over the back of the animal into a position similar to that which is found when Aplysia is at rest or when it is prodded during active creeping. Mechanical stimulation also results in withdrawal of the head and retraction of the siphon. Such movements of the parapodia were noted by ten Cate (1928) when he stimulated the parapodial nerve, and they would seem to be some sort of protective response. Their frequency is usually between 0 · 5 and 1/min.

Intracellular recordings from abdominal ganglion cells in a whole animal preparation have shown phenomena similar to those which have previously been described in isolated ganglia. Tonic and phasic spike activity may be observed, sometimes purely endogenous, sometimes induced or triggered by the synaptic input. The role of this input in the activity of a neurone can be established by artificial hyperpolarization of the cell, which suppresses spike activity but has no effect on the frequency of post-synaptic potentials (Tauc, 1957a, 1960 a). The presence of cells showing endo-geneously generated intermittent bursts of spikes is of considerable interest in relation to the control of intermittent nervous activity (Bullock, 1961). A large number of cells penetrated, however, are apparently inactive but are soon excited when the animal is stimulated either mechanically or electrically. The effects of such stimulation vary considerably according to the nature of the cell and to the position of the stimulation.

Various parts of the animal were stimulated by means of a paint brush which was used to stroke the parapodia, head, gill tail (i.e. the posterior end of the foot), etc. Gentle stimulation wherever applied tends to produce a localized contraction of that part of the animal, but stronger mechanical stimuli evoke a distinct protective response of the whole animal, which involves shortening of the foot, retraction of the mantle cavity and siphon, folding of both parapodial above the body. Stimulation of different parts of the body seems to be equally effective but perhaps the head and mantle region are the most sensitive. In some preparations a correlation has been observed between the type of response and the position of the mechanical stimulation. Fig. 2B, for example, shows the responses recorded in a preparation which was showing rhythmical movements of both parapodia. Mechanical stimulation of the gill or parapodia produced similar though less pronounced dorsal movements of the parapodia. Stimulation of the head, however, elicited a movement of the parapodia in a ventral and anterior direction. Similar movements were also found when the anterior part of the parapodia was stimulated, whereas touching the posterior part gave rise to a dorsal movement similar to that recorded when the gill or siphon was stimulated.

The electrical activity recorded in individual nerves under these conditions showed increases in activity which were related to the position of the mechanical stimulation. For example, the branchial nerve showed a great deal of activity when the gill was stimulated. Some of the impulses recorded were those of sensory fibres and these were followed by an asynchronous discharge along efferent pathways in the same nerve which were involved in the active retraction of the gill. In the case of the parapodial nerves, an increase in activity was obtained even when the part of the parapodium innervated by the nerve was not the part directly excited. This was due to the fact that stimulation, if of sufficient intensity, usually elicits a wave of contraction passing along the parapodium. Such waves sometimes occurred when other parts of the body such as the head were touched (Fig. 3 A). The increase in activity was generally related to the intensity of the stimulus. More localized stimulation produced less increase in activity, and in some instances (e.g. stimulation of one of the rhinophores) did not appear to produce any increase in activity in those nerves which were normally recorded from, unless its intensity was quite considerable. In some preparations quite strong mechanical stimulation to many different parts of the body produced little effect on, for example, the right connective (Fig. 3B). Such preparations often showed intermittent bursts of activity in the connective. Similar phenomena have been observed in preparations with bursts in the post parapodial nerve and neither this nor the siphon nerve showed any marked effect of stimulating various parts of the body, except the siphon.

Experiments in which intracellular recordings were made from single cells in the abdominal ganglion of the intact animal were of special interest as they showed a wide range in the extent to which activity of the cell was affected by peripheral stimulation. Some spontaneously active cells were completely unaffected by mechanical stimulation wherever it was applied to the body and even when this was very intense. At the other extreme were inactive cells which became activated by touching almost any part of the body. All intermediates between these two were found in different cells of the abdominal ganglia. A further variable is the nature of the effect on the rhythm of spontaneously active cells, which might be to accelerate or decelerate the rhythm and in some cases to produce complete inhibition. In fact all the different sorts of influence of the input to this ganglion, which have been observed in isolated preparations using electrical stimulation, have also been observed in the intact animal. Responses similar to those following mechanical stimulation could be shown by electrical stimulation of the appropriate nerve in whole animal preparations, but the electrical stimulation is not selective and in several cases it is complicated by antidromic stimulation of the neurone from which intracellular recording is being carried out. The mechanical stimulation acting through normal sensory receptors provides, therefore, valuable confirmation under more natural conditions.

Fig. 4 was obtained from a cell which showed a complete absence of any effect of mechanically stimulating the animal. In the upper beam (Fig. 4 A), it is apparent that quite marked activity was present in the right connective, and, in the case of stimulation applied to the head or gill, the large spikes indicate activity of the giant cell of the abdominal ganglion. When firing was prevented by hyperpolarization, no synaptic potentials were revealed followed stimulation. This cell is evidently a pacemaker type cell with an endogenous rhythmic activity. It was of interest in this connexion that when the membrane potential was restored to normal by release of the imposed hyperpolarization the cell showed a stage in which the somatic spike did not always occur with each of the pacemaker potentials (Fig. 4B).

Plots of the interval between successive impulses recorded in a cell which is only affected by localized stimulation is shown in Fig. 5. This cell was unaffected by stimulation of the parapodia, the foot or the head, but was significantly affected by touching the gill. The nature of this effect was to produce some inhibition of the spontaneous discharge when the stimulus was lightly applied. Stronger stimulation resulted in an initial increase in the frequency followed by a marked fall in the frequency. Closer inspection of these plots and the original records shows that the response is made up of a lowering in frequency with the occasional dropping out of impulses. This is apparent in Figs. 5 and 6 as an alternation between two or three distinct intervals between individual pulses following stimulation of the gill. The extracellular recordings from the right connective (Fig. 6 upper traces) simultaneous with these intracellular records are of interest because they showed activity in a single unit which was affected in the opposite sense, i.e. excited. This unif was unaffected by stimulation applied elsewhere on the body, but touching the gill resulted in a rise in frequency which was more marked the greater the intensity. The gill usually contracted as a result of this stimulation but the most obvious external feature of the animal’s response in this particular preparation was a distinct dorsal movement of the parapodium when the head was stimulated. Electrical stimulation of the branchial nerve (Fig. 6B) produced excitation followed by inhibition of the cell which was being recorded from intracellularly and these effects appeared to be similar to those produced by mechanical stimulation of the gill. Evidence from later experiments showed, however, that it was being stimulated antidromically as well as synaptically as the axon of this cell runs in the branchial nerve. This is an example where electrical stimulation produced a more complex response than that following a more physiological stimulus. That the axon was passing down the branchial nerve was confirmed by stimulating the cell directly by depolarizing the somatic membrane and by using the spike to trigger the oscilloscope. The recordings from the branchial nerve (Fig. 7C, upper beam) show a regular occurrence of a small spike each time the cell fires and thus prove that this cell does in fact send an efferent axon into the nerve. External recordings from the branchial nerve, during mechanical stimulation of the gill, indicate the presence of afferent fibres in this nerve whose activity precedes the excitation and inhibition of the cell. It is apparent, therefore, that this cell is concerned with localized activity of the branchial region and is excited synaptically via afferents in the branchial nerve which also contains the efferent axon to the gill region. It does not seem to be affected by inputs from mechanical receptors elsewhere, although electrical stimulation of the left connective, for instance, does have some effect on its discharge.

A cell which was responsive to a wider range of peripheral stimulation is shown in Fig. 34, upper beam. The resting discharge of this cell was 15-20/min and this activity became inhibited when the head was lightly touched. In fact stimulation of single rhinophores was sufficient to produce this inhibitory effect. Inhibition was also found when the left parapodium was touched but this was not so constant and the inhibition was slight in response to touching the right parapodium. Touching the gill, however, again produced a distinct inhibitory effect, although this differed from the effect of stimulating head or rhinophore in that it was more delayed. In all cases quite marked discharges were recorded from the right connective and the right posterior parapodial nerve.

A silent type of cell is illustrated in Fig. 35 upper record. Here the cell was quitd| inactive except when the head was touched; it was not affected by touching the tail or parapodia or even the gill of the animal. However, electrical stimulation of nerves produced a quite definite synaptic effect on the cell, although it was lacking in responsiveness to mechanical stimuli. Another example of a silent cell is the giant cell of the abdominal ganglion which is inactive in the unstimulated whole animal preparation, but discharges in response to touch on almost any part of the body (Hughes & Tauc, 1961).

Initially, investigations into the activity of cells in the pleural and pedal ganglia were carried out to discover whether any synaptic connexions existed between them and the axon of the right giant cell. No definite connexions were found, but in the course of this work records were taken with intracellular micro-electrodes from many cells and a study of their properties gave a general picture of the range of cell types found in these ganglia. Recordings from the pedal and cerebral ganglia were made in very few preparations, mainly because the cells in these ganglia are relatively small. As the general picture obtained from the pleural ganglia is essentially similar to that from the abdominal ganglion, a detailed description will not be given except where this has not been done by previous workers on the abdominal ganglion.

Many of the cells in the pleural ganglia are continuously active at frequencies of between one and five per second, but about the same number of cells penetrated were inactive under normal conditions. Some of the active cells showed well-marked saw-tooth pacemaker potentials. The rhythmically active cells varied in the degree to which their rhythmicity could be influenced by electrical stimulation of some of the nerves or connectives in the isolated central nervous system. Most cells were affected by stimulation of at least one of the paths usually used, but in a few cases the resting discharge was completely unaffected even by stimulation of the cerebro-pleural connectives which seemed to have the most striking effect on the vast majority of cells. Some showed a tendency to phasic activity giving trains of spikes separated by a more or less long period of inactivity. Changing the resting potential of such cells changes the frequency of bursts and points to the endogeneous character of this activity. Such a cell (Fig. 8, see also Fig. 4 B) presented a marked tendency to produce intermittent discharges in each of which about six impulses were present (Fig. 8,B). A gradual artificial increase in the polarization of the cell increased the interval between the successive bursts of six impulses and also changed the frequency of these impulses (Fig. 8,C, D) ; artificial depolarization produced the opposite effect (Fig. 8 A) and tended to change the phasic activity into a tonic one. When the cell was sufficiently hyperpolarized to prevent the occurrence of the spikes there was no evidence for the existence of bursts of synaptic potentials. It is concluded, therefore, that this intermittent tendency is an intrinsic property of such cells, the frequency of the bursts being varied by changes in polarization of the membrane itself (see Tauc, 1960a). In other cells, however, phasic activity occurs as a result of synaptic influence. Clearly some of these may result from the effects of inputs derived from cells showing endogenous discharges of the type described above. But many cells respond repetitively with high-frequency bursts to a single shock to one of the nerves (Fig. 9 A). Such repetitive firing is a common feature of the pleural ganglion cells and may often be related either to the endogenous phasic properties of the cell or to the large size of the post-synaptic potentials recorded in this ganglion (Fig. 9B). An input of this type may occur under natural conditions due to the coincidence of impulses arriving from cells firing rhythmically at a low frequency. Each time a sufficient number of these coincided in their convergence on a cell with a tendency to repetitive firing, a brief high-frequency burst would be produced. The intermittent bursts in the cell of Fig. 3 B might have arisen in this way. Convergence is not necessary, however, for a single impulse along a single pathway may result in repetitive firing of a cell. The duration of the repetitive effect of a single shock to one of the nerves may be quite long, as shown, for instance, in Fig. 10. In this hyperpolarized cell (A) synaptic bombardment through several repetitively firing pathways can be seen because of the different sizes of the excitatory post-synaptic potentials (E.P.S.P.) which they produce. The duration of such effects may last for several seconds. Such a network would produce very long intermittent bursts as can be seen in this cell with normal polarization (Fig. 10B). Inputs of this type to a spontaneously firing cell can lead to modulation of its frequency as shown in the preparation of Fig. 10 C. Here a single shock to the cerebro-pleural connective or the right connective resulted in an increase in frequency of the resting discharge. However, if some of the neurones in such networks produce inhibitory effects on subsequent pathways in the chain, intermittency can result from the periodic interruption of such continuous discharges. Fig. 10 D Is from such a cell in the pedal ganglion which is repetitively bombarded by an inhibitory pathway following a single shock, or several maximal shocks, to the C.-pl. In this cell stimulation of the posterior parapodial nerve also produced a slight inhibition of the cell, whereas when the right connective was stimulated the frequency was increased.

As can be seen in Fig. 10 the size of the synaptic potentials recorded in some of the pleural ganglion cells was very large. This could be demonstrated even better by using electrical stimulation to a synaptic pathway. This can be clearly seen in Fig. gB. The size of the synaptic potentials can be increased by increasing the intensity of the shock to the nerve thus bringing in more afferent fibres. Under such conditions of spatial summation the delay before the spike originates becomes reduced as the intensity of the shock is increased and thus produces a decrease in the synaptic delay as is well known in many preparations.

It has been shown (Tauc, 19576; 19606) that when a neurone is stimulated anti-dromically a two-stage invasion takes place. The spike recorded in the soma shows an inflexion in its rising phase; or, when the soma fails to be excited (normally or under fatigue), a small potential remains called the A spike or A potential (figure 7,5). This was interpreted as a blocking of the antidromic spike in the axon at some distance from the soma, which is invaded with some delay. Conduction of the antidromic spike may fail not only at the axono-somatic boundary but also at an axonal ramification; these are numerous which means that a neurone stimulated antidromically from different directions may show several A spikes of different sizes (Tauc & Hughes, 1961; 1962). The A spikes are like E.P.S.P.’s, but may be easily distinguished from them because if the membrane potential of a neurone is artificially increased, the A spike will be blocked earlier and further away in the axon and its size recorded in the soma will diminish. Under the same conditions the amplitude of an E.P.S.P. will increase (Fig. 11) (Tauc, 1958). This is quite important because if a whole nerve trunk is stimulated it is rare for an A potential to be found in the complete absence of any synaptic effect, although the contrary is not uncommon. This technique provides a method for studying the types of branching found in some of the cells of the pleural ganglia. Response to stimulation of the parapodial nerves or different connectives was studied in many preparations, but a systematic survey was only made of the effect of stimulating the cerebro-pleural connective, the viscero-pleural connective, and the posterior parapodial nerve. For each of these possible routes of stimulation there are three possible types of connexion. There may be no influence at all on the cell, or it may be synaptic, or antidromic (usually accompanied by a post-synaptic potential). There are, therefore, twenty-seven different possible patterns of cell type considered from this limited point of view. In a sample of about seventy-five cells by no means all of these possibilities were found. Those most commonly found are set out in Table 1 and of these the six most common types are shown diagrammatically in Fig. 12. About 10% of the cells were quite unaffected by stimulation of these three nerves, whereas a quarter of them were affected synaptically by all three routes. Another quarter of the cells had an axon in one of these nerves as was indicated by the presence of an antidromic potential in the cell. In a very small number of these ceUi antidromic potentials were recorded following stimulation of more than one of the nerves. It was this particular type of branching which was being searched for and used in a study of interaxonal spike propagation presented elsewhere (Tauc & Hughes, 1961). It is noteworthy that most of the antidromic potentials resulted from stimulation of the cerebro-pleural connectives, indicating the presence of many axons from cells in the pleural ganglion running back to the cerebral ganglia. In these cases it was certainly rare to find an antidromic spike alone and without any synaptic potential as the cerebro-pleural connectives also contain many pathways which influence cells of the pleural ganglion synaptically. This work brought out the importance of functional connexions between the cerebral and pleural ganglia.

No attempt has been made to give a complete account of the different pathways within the central nervous system. Many of the results described below were noted during the course of other investigations, but when gathered together they give an indication of the wide variety of neuronal connexions and demonstrate the means by which the far-reaching spread in the effects of stimulation may occur. Most of the information was obtained using external electrodes for stimulating and recording from different nerves. One of the most striking observations was the presence of a large number of direct pathways through the different ganglia. These were recognized by the one-to-one relationship between preand post-ganglionic recording, often when stimulated in opposite directions, and sometimes by the presence of A spikes in a cell when stimulated from different directions. They transmitted with little or no delay up to quite high frequencies and showed little evidence of synaptic properties. In this connexion, however, it must be remembered that ‘pseudo-synaptic’ properties have been observed in the branching axon of the giant cells (Hughes & Tauc, 1961) in addition to the observations of Bullock & Turner (1950) on single axons. Synaptic transmission rarely showed a 1:1 relationship between pre- and post-ganglionic recording, and the presence of a marked after-discharge in response to a single shock was quite common (Fig. g A). In cases where i : i transmission was considered to be synaptic the delay was relatively long, and facilitation, fatigue, and other synaptic properties were clearly recognizable. In some cases the synaptic effects were inhibitory and were recognized by a fall in frequency or complete cessation in the continuous discharge of a neurone in the post-ganglionic nerve.

Many examples of direct transmission across the abdominal ganglia were recognized and are summarized in Table 2. Transmission between all the four nerves studied was obtained in both directions with the exception of a direct pathway from the siphon nerve. In general, these pathways were more numerous between the right connective and the branchial nerve than between the left connective and the siphon nerve, although no quantitative data were obtained. Evidence was sometimes found for direct transmission from one of the nerves to two others, for example, from the left connective to the right connective and branchial nerve.

Direct pathways between different nerves emerging from the pleural-pedal complex of ganglia were shown in both isolated central nervous system preparations and in preparations of the whole animal A great deal of attention was given to pathways between parapodial nerves and the cerebro-pleural and pleuro-visceral connectives during the course of detailed studies of the branching of the giant cell axon. Evidence was obtained in many preparations for the existence of pathways which partially over tapped the extensions of these cells, but in no case was it shown that a single pathway had exactly the same distribution as one of the giant cells. One of the most striking pathways is that between the right connective and the cerebro-pleural connective which conducts impulses much more rapidly than those of the giant cell axon. This pathway would conduct in either direction following electrical stimulation and the delay was the same, confirming it to be in a single axon. A similar pathway exists between the right connective and the posterior parapodial nerve, and here again the single axon (AT) could sometimes be excited at a lower threshold and could conduct with a greater velocity than that of the right giant cell. As can be seen in Fig. 13 A, D, this pathway conducts with the same delay in both directions. Direct pathways from other parapodial nerves to the right connective have also been shown. Pathways from the parapodial nerves to the cerebro-pleural connectives have been found mainly on the right side, although some were present from the posterior parapodial nerve on the left side. On the other hand direct pathways in the opposite direction have not been so well established, mainly because this direction of stimulation (cerebro-pleural connective to parapodial nerve) was not used on so many occasions. Direct pathways from the pedal-pedal commissure to the other nerves were only investigated in one or two preparations but a definite pathway between this connective and the right connective was found. No doubt there are many direct pathways also between the cerebropedal connectives and the parapodial nerves. Only one such pathway was found on the right side to the anterior parapodial nerve. The conduction velocity along mosf of the pathways discussed above range from 0 · 1 to 1 · 0 m./sec.

It must be emphasized that a direct pathway established on the basis of a 1: 1 relationship between impulses in two nerves emerging from a ganglion does not necessarily mean that such a pathway is really physiological. The 1 : 1 response to electrical stimulation may result from the presence in both nerves of branches of a cell located in the ganglion, where spike initiation would normally be expected to occur. In some cases, like the axons of the giant cells, there is good evidence for believing that physiological direct pathways pass through ganglia without any synaptic contact (Hughes & Tauc, 1961).

Synaptic connexions exist between the four main nerves of the abdominal ganglia. These have been found in all directions except that from the siphon nerve to the branchial nerve, although the opposite direction of transmission has been established. These pathways range from simple 1:1 type to the multiplying synaptic type. Both temporal and spatial summation can be seen. These and other properties of these synaptic pathways are similar to those observed when intracellular recordings are made from individual cells of this ganglion. Both excitatory and inhibitory postsynaptic potentials are readily recorded, the height of the potential being usually increased by a more intense shock to the pre-ganglionic nerve (Fig. gB) or by an increase in the frequency of shocks. A very wide variety in the particular type of synaptic behaviour has been described for this ganglion (Tauc, 1958, 1960 a) and similar properties are discussed below for the pleural ganglion.

Synaptic pathways between the parapodial nerves, cerebro-pleural connectives, and pleuro-visceral connectives have been established in many cases, notably between the cerebro-pleural connective, right connective and right posterior-parapodial nerves (see Table 3). Usually there is a considerable after-discharge associated with the response to a single pre-ganglionic shock (Figs. 6B, 9A, 10). Summation, both temporal and spatial, is readily observed. The existence of these pathways follows from the observations on the whole animal preparation in response to mechanical stimulation of different parts of the body with recording electrodes on the parapodial nerves and the right connective.

Longer synaptic pathways have been observed in some instances, for example, stimulation of the branchial nerve elicits V response in the right posterior parapodial nerve. Transmission from one side of the animal to the other has been observed in several instances, but not with such regularity as was expected from the observations of ten Cate (1928) and Turner & Nevius (1951). Stimulation of one of the parapodial nerves on one side does elicit a response on the other side, but it is necessary to use quite a high intensity of shock and 1:1 transmission was not observed. Such transmission was observed, however, between the left connective and the right connective via the pedal-pedal (Pe.-pe.) commissure; similarly, between the left connective and the right posterior parapodial nerve. The presence of such pathways is necessary for the synchronization of the movements of the parapodia on the two sides of the animal. Such synchrony is not invariable, however, especially in preparations such as those used in the present investigations where it was frequently observed that movement of the two parapodia was completely independent.

In recent studies of the organization and connexions within the central nervous system of animals there has been a tendency to make greater use of stimulation of the sensory pathways via the normal receptor mechanisms than was previously the case. The value of such methods over those using electrical stimulation of whole nerve trunks is clearly that information is obtained which is of more physiological significance, as it indicates the particular ways in which more normal patterns of sensory input are combined and sorted out by the central nervous system. In the vertebrates such studies at the spinal level (Frank & Fuortes, 1956; Kolmodin, 1957; Hunt & Kuno, 1959) are relatively new and have made use of intracellular recordings by micro-electrodes, but the recording of activity from the brain in response to stimulation of receptors has been known for a long time, although not at the unit level. Among invertebrates techniques for the analysis of single units have been applied chiefly to arthropods and especially crayfish (Wiersma, 1958; Hughes & Wiersma, 1960a; Wiersma & Hughes, 1961; Kennedy & Preston, 1960) and recently in a dragonfly nymph (Fielden & Hughes, 1962). In many experiments both with vertebrate and invertebrate animals it has been shown that a great deal of convergence takes place from the sensory input to the interneurones of the central nervous system. The many different ways in which this information is combined and sorted out is a general feature of these systems and provides a good beginning in our search for integrative mechanisms.

Among molluscs similar studies appear to be lacking and it is in this respect that the present observations described from the abdominal ganglion of Aplysia are of significance. Moreover, recording from a neurone with intracellular electrodes has not previously been done in many invertebrates under conditions in which the effects of stimulation through the normal sensory pathways could be studied. In this instance this technique has provided a more sensitive method of estimating the breadth of sensory input which affects a given neurone, because it is possible to discern the synaptic input even when it does not produce a propagated action potential in the axon. In addition, important confirmation has been obtained, that the different types of response to electrical stimulation of isolated abdominal ganglion preparations (Tauc, 1958, 1960 a) also occur under more normal conditions of stimulation. Excitatory and inhibitory effects of different sorts have been observed and their effects studied in both silent cells and those which show rhythmic activity in the intact animal. The latter cells have been shown to vary a great deal in the extent to which their rhythmic activity can be influenced by peripheral stimulation. The frequency of some of these seems not to be altered by any afferent stimulation. Such ‘spontaneously’ firing cells may be considered as truly pacemaker in function, their rhythm being endogenous and not due to bombardment of the cell by other central or sensory neurones. Finding of such cells within the central nervous system with most of its afferent nerves intact may be compared with observations already made on single units in the crayfish, (e.g. 50 and 20/sec. fibres; Wiersma, 1960) and insects (Fielden & Hughes, 1962), and indicates that this type of activity must be taken into account when discussing the mechanism of integration of the central nervous system. It is always possible, however, that cells of this type are only affected by more complex patterns of sensory input and it may be their function to give specific responses to such inputs. The presence of intermittent bursts of impuses in some of these cells, often in constantly repeating patterns, is also noteworthy. While in some cases this property is endogenous and does not require any input there are others in which the arrival of a single impulse along a given pathway is necessary to trigger off the sequence. Commonly impulses along several pathways are required and a burst results from the spread in their time of arrival (cf. Horridge, 1961). The relationship between temporal spread of the input and discharge of a given neurone is a further variable depending in part upon the size of the cell. Large cells have longer time constants and longerlasting discharges may result from impulses which are not so temporally dispersed (Fessard & Tauc, 1957). Examples were also found of intermittent discharges from pleural ganglion cells, evidently through some interneurones, when stimulation at constant frequency was given to a cerebro-pleural connective. This is similar to the observation, at a still more complex level of integration, of intermittent patterned bursts in many motor fibres of the swimmeret nerves of the crayfish when a single fibre of the circumoesophageal commissure was stimulated at 50/sec (Hughes & Wiersma, 1960 b).

The use of micro-electrodes for the tracing of anatomical pathways within invertebrate central nervous systems has been possible here because of the knowledge gained previously (Tauc, 1957 b) of the antidromic potentials recorded within Aplysia ganglion cells. In this way we have shown that it is possible to study the different types of branching of gastropod neurones and the ways in which they are influenced synaptically. The value of this method is exemplified by the description of the giant cell axons in Aplysia (Hughes & Tauc, 1961) although it was confirmed by more conventional stimulation and recording methods. These methods were also applied in the present study and have shown the considerable number of direct pathways which pass through individual ganglia of the central nervous system in the absence of any synapse. This, of course, enables an impulse to pass more rapidly, between one part of the body and another and, in addition, prevents impulses in such fibres from having any direct effect on other neurones at this particular ganglion. Turner & Nevius (1951) and Turner (1953) showed many synaptic pathways between fibres in the pedal nerves on one side and those on the opposite side of the central nervous system, but such pathways were not numerous in the present investigations. In Aplysia, transverse pathways certainly occur but they are often direct ones.

This study has served to give an account of the general layout of the central nervous system of Aplysia. It does not claim to be complete by any means, but it is hoped that it shows the usefulness of this animal for such studies, especially the way it can be used with micro-electrode techniques for the establishment of neuronal connexions. During the present work it was possible to record only from the abdominal ganglia in situ, but it is hoped that in future work it might be possible to record from some of the other ganglia, which are more involved in the basic mechanisms which control locomotory movements and the major activities of the organism.

1. The organization of the central nervous system of Aplysia depilans has been investigated in whole animal and isolated ganglion preparations using mechanical and electrical stimulation.

2. Intracellular micro-electrodes have been used to record activity in nerve cells of the abdominal ganglia in situ. Some cells are spontaneously active and quite unaffected by mechanical stimulation, whereas others show varying degrees of responsiveness. Those which are unaffected may exhibit regular rhythmic activity or intermittent bursts which are intrinsic to the cells themselves but in other cases are due to synaptic input from other central neurones.

3. In isolated central nervous system preparations a special study of the pleural ganglion has revealed many types of cell with electrical activity similar to that shown in isolated abdominal ganglion preparations. A notable feature of the pleural ganglion cells was the large size of the excitatory post-synaptic potentials recorded in response to stimulation of pre-synaptic fibres.

4. Different types of branching of cells of the pleural ganglia were investigated. By observing the somatic potential it was possible to decide in which nerve a particular cell sent collateral branches and which nerves contained fibres affecting the cell synaptically. By this means it was clear that a large number of pathways connect the cerebral and pleural ganglia on each side.

5. A large number of direct pathways were found of nerve fibres passing through ganglia without any synapse.

6. Synaptic pathways varied in the number and intrinsic properties of the individual synapses along their route. Synapses between fibres in the nerves innervating the foot and parapodial lobes of the two sides were not as common as has been described for Ariolimax.

7. In general the results have shown a great variety in the extent to which afferent stimulation may affect the whole or part of the central nervous system. They have also revealed the great multiplicity in the pathways whereby this is achieved.

One of us (G.M.H.) wishes to express his thanks to Prof. A. Fessard for giving him the opportunity to perform this work. Part of his personal expenses were defrayed by a grant from the Travelling Expenses Fund of the University of Cambridge, for which he is grateful.