In the lobster Jasus lalandii, 14 neurones of the stomatogastric ganglion (STG) are organized in a network that produces rhythmic pyloric outputs. In vitro experiments have shown that the STG neurones receive, via the stomatogastric nerve (stn), neuromodulatory inputs that influence the expression of the bursting properties of the neurones and the ability of the network to produce its rhythmic output.

In contrast to these in vitro observations, in vivo transection of the stn does not abolish the pyloric rhythm. Rhythmic output can be recorded by electromyogra-phy immediately after stn transection and for up to 2 years afterwards. We have shown that, under these experimental conditions, the STG appears to be isolated from any neuronal input that might account for the maintenance of the rhythmic output. Experiments carried out in the 2 days after stn transection showed that an in vitro preparation of the isolated STG was unable to produce any rhythmic output, but blood serum added to the system could restore the pyloric output.

These results suggest strongly that the pyloric network receives neural and humoral modulatory influences in parallel and that each type of influence alone is able to maintain the bursting capability of the pyloric neurones.

Central pattern generators (CPGs) are networks of neurones that generate the timing and phasing cues for rhythmic movements (Delcomyn, 1980; Grillner, 1977). The neuronal characteristics and the synaptic connectivity that underlie the endogenous rhythmicity of these networks have been investigated in several preparations, but relatively complete analysis has been achieved in only a few simpler invertebrate circuits (Roberts and Roberts, 1983; Selverston and Moulins, 1985; Getting, 1988). Moreover, recent studies have shown that these circuits are the targets of modulatory influences, with long-lasting effects shaping the final output (Weiss et al. 1981; Harris-Warrick, 1988). These modulatory influences can act directly through neuronal pathways or indirectly through the action of circulating hormones. However, it is difficult to find an experimental situation in which it can be demonstrated that a given neuronal circuit can be the target of neuronal and humoral influences acting in parallel. In the present work, we have used one of the best known invertebrate CPGs, the crustacean pyloric CPG (Selverston and Moulins, 1987) and have designed experimental procedures that strongly suggest that this CPG is subjected to both neuronal and humoral modulatory influences under physiological conditions.

The pyloric CPG is a small neuronal network (14 neurones) in the stomatogas-tric ganglion (STG) that produces a triphasic rhythmic output. The pyloric output can be recorded from in vitro preparations of the stomatogastric nervous system (Maynard, 1972). Using such preparations, it has been demonstrated that the existence of the rhythmic pattern is dependent upon the intrinsic bursting properties of the component neurones (see Miller, 1987). However, all the neurones of the pyloric CPG are conditional oscillators, requiring neuromodulatory inputs from anterior centres in order to express their oscillatory capabilities in vitro (Bal et al. 1988). In other words, neuromodulatory influences are necessary to initiate and maintain the rhythmic output (Moulins and Cournil, 1982; Nagy and Miller, 1987). Consequently, the presence or absence of rhythmic output is a good indication of whether or not the pyloric CPG is receiving permissive modulatory influences at any time.

In vitro, neuromodulatory inputs to the pyloric CPG can be removed by cutting or blocking impulse conduction in the single input nerve to the STG, the stomatogastric nerve (stn). This results in an inactivation of pyloric output. However, in such preparations the pyloric network is devoid of other modulatory factors, such as those carried by the blood in the intact animal. In this study, we have taken a complementary approach: we cut the input nerve of an otherwise intact animal so that the stomatogastric ganglion no longer received neuronal modulatory influences but remained exposed to putative blood-borne influences. Electromyographic recordings (Rezer and Moulins, 1980, 1983) from such preparations have enabled us to test the ability of the pyloric CPG to produce a rhythmic output when exposed only to humoral influences. Our data suggest that, under physiological conditions, the pyloric CPG is the target of both neuronal and humoral modulatory influences able independently to induce the oscillatory properties of the neurones that generate the rhythmic output of this network.

Experiments were performed on 46 (36 operated, 10 intacts) adult male and female specimens of the lobster Jasus lalandii (Milne Edwards), obtained from a local market and kept in large tanks of running sea water. The animals were fed with mussels twice a week.

In vivo electromyographic recordings (EMG) The recording electrodes (Teflon-insulated silver wire, 130 μm in diameter) were implanted in the muscles through small-diameter holes made in the cephalothorax. Electrodes were then attached to the surface of the carapace with tissue glue (Histoacryl glue from Ligatures Peters) and dental cement. Animals were replaced in a small tank of sea water and the free ends of the wires were connected to amplifiers (Grass P5 a.c. preamplifier). Data were recorded onto a Gould ES 1000 electrostatic recorder and simultaneously stored on a Schlumberger tape recorder. This method allowed us to record simultaneously the activity of several pyloric muscles. After experimentation, animals were killed, and the muscles whose activity had been recorded were identified using a marking technique described previously (Rezer and Moulins, 1980). Muscles were classified according to Maynard and Dando (1974).

In most experiments, we recorded from one pyloric dilator muscle (cpvla,b), here referred to as a dilator muscle (D), and two pyloric constrictor muscles (pl and p8-pl2), here referred to as the anterior constrictor (Cl) and posterior constrictor (C2) (Fig. 1A). D muscles were innervated by the two pyloric dilator (PD) motor neurones, Cl muscles by the single latero-pyloric (LP) motor neurone and C2 muscles by the eight pyloric (PY) motor neurones (Maynard and Dando, 1974). These motor neurones form a pattern-generating network, the pyloric CPG, in the stomatogastric ganglion (see Selverston and Moulins, 1987) (Fig. 1B).

Fig. 1.

The pyloric neuromuscular system of Jasus lalandii. (A) Schematic drawing of the stomatogastric nervous system in situ. The complete stomatogastric nervous system consists of the stomatogastric ganglion (STG), the oesophageal ganglia (OG), the paired commissural ganglia (CG) and the nerves that connect them; B, brain. The single input nerve to the STG is the stomatogastric nerve (stn). The pyloric dilator (D) and constrictor (Cl and C2) muscles are innervated by the lateral ventricular nerve (lvn). The arrow indicates where the stn is cut in the operated animals. The marks on the peri-oesophageal connective indicate the portion of the connective kept with the in vitro preparation, avn, anterior ventricular nerve; mvn, medial ventricular nerve. (B) Innervation of the pyloric muscles and circuit diagram of the pyloric network (the black circles represent inhibitory synapses). (C) In vitro preparation of the stomatogas-tric nervous system. The system was dissected from the foregut and pinned in a Sylgard-lined Petri dish. One Vaseline pool was built around the desheathed STG. The circle around the stn indicates where it was cut. Pin electrodes were used to record the motor activity of the pyloric dilator neurone (PD) on the PD nerve (PDn) and LP and PY on the ventral and dorsal branches of the latero-ventricular nerve (vlvn, dlvn). (D) Electromyographic activity of the D, Cl and C2 muscles. Such a recording gives direct access to the suprathreshold activity of the motor neurones (PD, LP and PY) that constitute the pyloric network in the STG (see B).

Fig. 1.

The pyloric neuromuscular system of Jasus lalandii. (A) Schematic drawing of the stomatogastric nervous system in situ. The complete stomatogastric nervous system consists of the stomatogastric ganglion (STG), the oesophageal ganglia (OG), the paired commissural ganglia (CG) and the nerves that connect them; B, brain. The single input nerve to the STG is the stomatogastric nerve (stn). The pyloric dilator (D) and constrictor (Cl and C2) muscles are innervated by the lateral ventricular nerve (lvn). The arrow indicates where the stn is cut in the operated animals. The marks on the peri-oesophageal connective indicate the portion of the connective kept with the in vitro preparation, avn, anterior ventricular nerve; mvn, medial ventricular nerve. (B) Innervation of the pyloric muscles and circuit diagram of the pyloric network (the black circles represent inhibitory synapses). (C) In vitro preparation of the stomatogas-tric nervous system. The system was dissected from the foregut and pinned in a Sylgard-lined Petri dish. One Vaseline pool was built around the desheathed STG. The circle around the stn indicates where it was cut. Pin electrodes were used to record the motor activity of the pyloric dilator neurone (PD) on the PD nerve (PDn) and LP and PY on the ventral and dorsal branches of the latero-ventricular nerve (vlvn, dlvn). (D) Electromyographic activity of the D, Cl and C2 muscles. Such a recording gives direct access to the suprathreshold activity of the motor neurones (PD, LP and PY) that constitute the pyloric network in the STG (see B).

Simultaneous myographic and movement recordings

In several experiments, it was necessary to record simultaneously both the myographic activity and the corresponding movement. This was achieved by using the same wire both as an electromyographic recording electrode and as a movement transducer. A high-frequency electrical field (40 × 103 Hz) was generated (with a Tektronix FG 501 generator) between two steel sheets placed against the walls of the aquarium. An electromyographic (EMG) recording electrode was implanted in a muscle and the displacement of the tip of this electrode in the electrical field (resulting from movement of the muscle) was recorded at the same time as the electrical activity of the muscle. The two signals (EMG and movement) were separated using low- and high-pass filters. The EMG has a frequency of less than 300 Hz, so we used a low-pass (below 300 Hz) filter placed in the first EMG amplifier stage to extract the EMG signal from the movement signal. The movement was obtained by demodulating the high-frequency carrier. We used a capacitative isolation unit to prevent polarization effects. In these experiments the animals had to be restrained.

To block the neuromuscular junction of the D muscle (and suppress its movements), d-tubocurarine (Sigma) was injected into the animal. Enough d-tubocurarine was injected to achieve a final concentration of 10−4moll−1 in the haemolymph, assuming that the total haemolymph volume was 20 % of the weight of the animal (Phillips et al. 1980). The solution to be injected was made up at a concentration that gave a final injection volume of no more than 1.5 ml, and was injected into the body cavity near the D muscle.

In vitro and ex vivo neurographic recordings

The stomatogastric nervous system was dissected from the foregut and pinned out under saline (400 mmol l−1 NaCl, 10 mmol l−1 KC1, 10 mmol l−1 CaCl2, 52 mmol l−1 MgC12, 28 mmol l−1 Na2SO4, 3 mmol l−1 NaHCO3, 0.6 mmol l−1 NaBr) in a Sylgard-lined Petri dish (combined preparation of Selverston et al. 1976) (see Fig. 1C).

We called this combined preparation dissected from an intact animal an in vitro preparation: the STG was still attached to the anterior centres. We called the stomatogastric nervous system dissected from an operated animal an ex vivo preparation. In this case, as the stn had been cut in vivo, the preparation was reduced to a short piece of the stn attached to the STG and the output nerves (see Fig. 1B,C).

Extracellular recordings and stimulation of the various nerves of the preparation were achieved via platinum electrodes according to electrophysiological techniques previously reported (Moulins and Nagy, 1981).

A Vaseline pool built around the stomatogastric ganglion was used to superfuse blood serum. Blood of Jasus lalandii or Cancerpagurus was taken from the base of the fourth leg just before experimentation and serum was obtained by centrifuging it for 2 min at 2000 revs min−1. Serum tests were carried out on ex vivo preparations, i.e. stomatogastric systems isolated from animals previously operated upon.

Operated animals

Animals were anaesthetized by gradual cooling; they were left for 2h at 10°C, then 5h at 5°C. They were then kept on ice throughout the operation. A small piece of carapace was carefully removed from the cephalothorax with sterilized instruments to allow access to the stn. This piece of carapace was placed in a sterile Petri dish during the operation. Under a binocular microscope, the epidermis was removed and the stn was cut anterior to its emergence from the aorta. The natural tension of the nerve was sufficient to cause retraction of the cut ends. The epidermis was then glued back in place with Histoacryl glue. This piece of epidermis is essential for the generation of the new carapace necessary for the survival of the animal at the next moult. The carapace piece was then put back in place and fixed to the rest of the carapace with dental cement. Afterwards, animals were placed in individual tanks of running sea water. More than 90% of the operated animals survived for 3 weeks or longer and some survived for as long as 2 years. All the animals were kept with fresh mussels in their aquarium. The day after the operation, the animals appeared to have recovered completely and EMGs could be recorded from the pyloric muscles. Spontaneous feeding began 10–20 days after the operation (see Table 1). EMG recordings were carried out on 36 animals at different times after the operation. Some animals were used for two recording sessions, giving a total of 42 in vivo experiments (see Table 1). For these animals we used wires of different colour, cut and left in place after the first recording session. The muscles were identified at the end of the second recording session. In a few animals, we recorded the pyloric activity with the same electrodes in place before and after the operation (see Fig. 5).

Table 1.

Rhythmic pyloric pattern tested in vivo in operated animals

Rhythmic pyloric pattern tested in vivo in operated animals
Rhythmic pyloric pattern tested in vivo in operated animals

In vivo transection of stn does not abolish the rhythmic pyloric output

Electromyographic recordings from intact Jasus lalandii show that the pyloric motor activity consists of a regular triphasic rhythmic pattern with an activation in each cycle of the dilator muscles (D), the anterior constrictor muscles (Cl) and the posterior constrictor muscles (C2) (Fig. 1D) (Rezer and Moulins, 1983).

It has been shown, using in vitro recordings from the isolated stomatogastric nervous system of Panulirus, that this activity originates from a network of 14 neurones in the STG (Maynard, 1972; Maynard and Selverston, 1975). Similar results have been obtained subsequently for Jasus lalandii (Nagy and Dickinson, 1983). This network has been extensively studied (for a review, see Selverston and Moulins, 1987) and it has been demonstrated that its triphasic rhythmic output arises both from the synaptic connectivity between pyloric neurones (see Fig. 1B) and from the ability of these neurones to produce bursting pacemaker potentials (BPP) (Miller and Selverston, 1982).

In the Cape lobster Jasus lalandii, Bal et al. (1988) have clearly demonstrated that all the neurones of the pyloric CPG are conditional oscillators, i.e. their abilities to produce BPPs are conditional and can be expressed only under extrinsic modulatory input to the pyloric network. In an in vitro preparation of the stomatogastric nervous system, the suppression of impulse traffic in the stn completely abolishes the ability of pyloric neurones to produce BPPs and, thereby, the ability of the network to produce rhythmic output (Moulins and Cournil, 1982; Nagy and Miller, 1987) (Fig. 2A).

Fig. 2.

Comparison between the effects of deafferentation on the activity of the STG in vitro (A) and in vivo (B). (A) In the control preparation, rhythmic pyloric activity was recorded from an isolated stomatogastric nervous system. The three types of motor neurones fired in successive bursts: PD in the PD nerve (PDn), LP and PY in the vlvn. The neurone that fired with PD in the stn is the anterior burster. Thirty-five minutes after stn section, pyloric activity had completely disappeared. (B) In the control preparation (intact animal), rhythmic pyloric activity was recorded by electromyography. The D muscles (innervated by PD), Cl muscles (innervated by LP) and C2 muscles (innervated by PY) were successively active. After cutting the stn (3-day-old operated animal), the D, Cl and C2 muscles were still rhythmically active, indicating that the PD, LP and PY motor neurones were still bursting.

Fig. 2.

Comparison between the effects of deafferentation on the activity of the STG in vitro (A) and in vivo (B). (A) In the control preparation, rhythmic pyloric activity was recorded from an isolated stomatogastric nervous system. The three types of motor neurones fired in successive bursts: PD in the PD nerve (PDn), LP and PY in the vlvn. The neurone that fired with PD in the stn is the anterior burster. Thirty-five minutes after stn section, pyloric activity had completely disappeared. (B) In the control preparation (intact animal), rhythmic pyloric activity was recorded by electromyography. The D muscles (innervated by PD), Cl muscles (innervated by LP) and C2 muscles (innervated by PY) were successively active. After cutting the stn (3-day-old operated animal), the D, Cl and C2 muscles were still rhythmically active, indicating that the PD, LP and PY motor neurones were still bursting.

The related experiment performed in vivo, however, does not give the same results as above (compare Fig. 2A and 2B). For all the operated animals tested (see Table 1), EMG recordings from pyloric muscles show a rhythmic pattern of activity similar to the pyloric pattern recorded in the non-operated animals (Fig. 2B). This triphasic pattern has been observed in animals just after stn transection, as well as in animals 5 months and even 2 years after stn transection (Fig. 3 and Table 1).

Fig. 3.

Pyloric patterns recorded in operated animals at different times after stn transection. These electromyographic recordings, obtained from five animals at 2h, 24 h, 8 days, 5 months (142 days) and 2 years after stn transection, show that the pyloric rhythmic pattern was present just after the operation and had not disappeared even 2 years later.

Fig. 3.

Pyloric patterns recorded in operated animals at different times after stn transection. These electromyographic recordings, obtained from five animals at 2h, 24 h, 8 days, 5 months (142 days) and 2 years after stn transection, show that the pyloric rhythmic pattern was present just after the operation and had not disappeared even 2 years later.

We isolated the stomatogastric system from operated animals and recorded the pyloric output in the resulting ex vivo preparation. Prior to each ex vivo recording experiment, we had tested for the presence of rhythmic pyloric output in vivo: electromyographic recordings showed that all the animals tested in this way had a rhythmic pyloric motor output. Nevertheless, when these stomatogastric systems were rapidly transferred to Petri dishes, they appeared to be unable to produce rhythmic pyloric output. In recordings from the pyloric motor nerves in nine ex vivo preparations dissected 1 day after stn transection and in seven preparations dissected 2 days after stn transection, no rhythmic activity was seen.

The discrepancy between the results obtained in vivo and ex vivo can be explained by one of two hypotheses: (1) although all the nervous inputs to the STG (and the pyloric network) are generally thought to travel in the stomatogastric nerve, some inputs may use other routes and, thereby, may be responsible for the continuation of a pyloric output in the operated animal; (2) there is a blood-borne factor (normally acting in parallel with nervous influences) that is not present in the in vitro preparations and that is present in sufficient quantity in the operated animal to maintain the ability of the neurones to generate BPPs and thus the rhythmic pyloric output.

The following results strongly support this second possibility.

Startle pyloric responses disappear in operated animals

The first test that can be used to determine whether the pyloric network of an operated animal is effectively isolated from sensory inputs is to compare the responses of the network to different sensory stimuli in the intact animal with those, if any, in the operated one.

Fig. 4 compares the effects of mechanical (Fig. 4A,C) and chemical (Fig. 4B,D) stimuli applied to an intact animal and to an operated animal. The mechanical stimulus used here was a non-specific tapping on the dorsal carapace. Tapping produced a general startle reaction and a transient modification (primarily acceleration) of the pyloric rhythm (Fig. 4A). Although highly evident, this response exhibited habituation in the intact animal, disappearing almost completely after only five repetitive trials. By contrast, the same stimuli never had an effect on the existing pyloric output of operated animals (compare Fig. 4A and 4C).

Fig. 4.

Effects of a mechanical (A,C) and a chemical (B,D) stimulus on intact and operated animals. (A,B) In an intact animal, a mechanical stimulus (A) applied to the dorsal carapace was always associated with an immediate brief modification of the pyloric motor pattern: the frequency of the rhythm increased for several cycles, and the burst duration of the constrictor in these cycles decreased. The activities of D (black triangle) and Cl (black circle) were recorded simultaneously with a single electrode. The stimulus time is marked by a bar under the record. A chemical stimulus (B) (mussel juice introduced into the aquarium) was also always associated with similar transient modifications of the pyloric rhythm. (C,D) In an operated animal, the same stimuli (mechanical in C and chemical in D) were never associated with any change in the pyloric rhythm. Stimulus time is marked by a bar.

Fig. 4.

Effects of a mechanical (A,C) and a chemical (B,D) stimulus on intact and operated animals. (A,B) In an intact animal, a mechanical stimulus (A) applied to the dorsal carapace was always associated with an immediate brief modification of the pyloric motor pattern: the frequency of the rhythm increased for several cycles, and the burst duration of the constrictor in these cycles decreased. The activities of D (black triangle) and Cl (black circle) were recorded simultaneously with a single electrode. The stimulus time is marked by a bar under the record. A chemical stimulus (B) (mussel juice introduced into the aquarium) was also always associated with similar transient modifications of the pyloric rhythm. (C,D) In an operated animal, the same stimuli (mechanical in C and chemical in D) were never associated with any change in the pyloric rhythm. Stimulus time is marked by a bar.

The chemical stimulus used was more specific; juice extracted from fresh mussels was introduced into the aquarium. In the intact animal, this always induced a modification of the pyloric rhythm that lasted for several cycles (Fig. 4B). However, the same stimulus applied to an operated animal never caused a modification of the pyloric rhythm (Fig. 4D). Each of these stimuli was applied to 10 operated animals, and we never recorded a startle response. Moreover, we have been unable to find any sensory stimulus that induces observable effects on the pyloric motor output in operated animals.

The pyloric pattern is not modified at the onset of feeding in operated animals

Feeding is a specific stimulus that alters the activity of the pyloric CPG in intact animals (Rezer and Moulins, 1983). In animals unfed for more than 8 days, the pyloric motor output is highly irregular: the pyloric period is long (mean greater than 2 s) and varies from one cycle to another (left-hand part of Fig. 5Ai). After feeding, the pyloric motor output is very regular, with a shorter (less than 2 s) and almost constant period; this pattern continues without changing for more than 12 h after feeding (right-hand part of Fig. 5Ai). The output of the pyloric CPG switches from one pattern to the other immediately upon feeding (middle part of Fig. 5Ai). In continuous recordings from six operated animals, feeding was never associated with any change in the pyloric pattern (Fig. 5Bi). The pyloric period remained variable and did not decrease after feeding. Thus, the pyloric period had a clearly bimodal distribution when measured in the intact animal during a time interval that included feeding (Fig. 5Aii). When measured in the operated animal under the same conditions, the distribution of the pyloric period was unimodal (Fig. 5Bii).

Fig. 5.

Evolution of the pyloric period at feeding time in intact and operated animals. The evolution of the pyloric period (measured in seconds between the onsets of successive dilator bursts) has been plotted over 14 min during which the animal fed. The electromyographic recording presented under the graph corresponds to the feeding time. In the intact animal (Ai), the pyloric output changed abruptly upon feeding: the pyloric period, which was long and variable, became short and stable. (Aii) Distribution histograms of D period as a percentage of the total cycle period before and after the feeding. D period has a bimodal distribution; for the first mode (1), corresponding to the pattern recorded after feeding, the mean is 0.96±0.079s; for the second mode (2), corresponding to the pattern recorded before feeding, the mean is 2.97±0.79s (N=700) (see text). In the operated animal (B), the pyloric output is not modified at feeding. In Bi, the electromyographic recording is interrupted for 1.30 min, so as to include recordings both before and after feeding. (Bii) Distribution histogram of D period before and after feeding; here the D period has a unimodal distribution (3), with a mean of 5.26±2.81s (N=700).

Fig. 5.

Evolution of the pyloric period at feeding time in intact and operated animals. The evolution of the pyloric period (measured in seconds between the onsets of successive dilator bursts) has been plotted over 14 min during which the animal fed. The electromyographic recording presented under the graph corresponds to the feeding time. In the intact animal (Ai), the pyloric output changed abruptly upon feeding: the pyloric period, which was long and variable, became short and stable. (Aii) Distribution histograms of D period as a percentage of the total cycle period before and after the feeding. D period has a bimodal distribution; for the first mode (1), corresponding to the pattern recorded after feeding, the mean is 0.96±0.079s; for the second mode (2), corresponding to the pattern recorded before feeding, the mean is 2.97±0.79s (N=700) (see text). In the operated animal (B), the pyloric output is not modified at feeding. In Bi, the electromyographic recording is interrupted for 1.30 min, so as to include recordings both before and after feeding. (Bii) Distribution histogram of D period before and after feeding; here the D period has a unimodal distribution (3), with a mean of 5.26±2.81s (N=700).

Although the origin of the modification of the pyloric output in an intact animal is not known, it is clear that its abrupt occurrence at feeding is the mark of extrinsic nervous influences impinging on the pyloric network. This again indicates that the stomatogastric ganglion in the operated animal was completely isolated: its pyloric output was no longer correlated with the behavioural context.

Phasic proprioceptive feedback cannot account for pyloric rhythm generation in operated animals

It has been demonstrated using in vitro preparations that the pyloric network connected to the more rostral ganglia is able to produce organized rhythmic output without proprioceptive feedback (Maynard, 1972; Selverston et al. 1976). However, although proprioceptive inputs are not necessary to produce this output, it remains possible that, under our experimental conditions, they are sufficient to generate the rhythmic pattern.

This hypothesis has been tested with respect to the proprioceptive feedback associated with D muscle movement by (1) pharmacological suppression of the movement and (2) experimental perturbation of existing movement. Such perturbation or suppression should modify any proprioceptive information that may derive directly from contraction of the muscle itself and drive pyloric output.

In intact animals, it has been possible to record simultaneously the EMG activity of the pyloric muscles and the movements they produce via the tip of the same electrode inserted into the muscle (see Materials and methods). This is shown in the experiments of Fig. 6A, in which movements and electrical activity of the D muscle were recorded simultaneously. It is known that the dilator muscle is innervated by two cholinergic motor neurones (Marder, 1987). This property has been used to suppress the movements of the D muscle pharmacologically. This was achieved by injecting d-tubocurarine to a final concentration of about 10−4moll−1 into the haemolymph. Fig. 6B shows that, at this concentration, contractions of the D muscle disappeared.

Fig. 6.

Injection of d-tubocurarine abolished dilator contraction (A,B) but did not abolish the rhythmic pyloric output of an operated animal (C,D). (A,B) In the control (A), the electrode implanted in the D muscles was used to monitor D contractions. Note that this electrode recorded simultaneously the activity in D (black triangle) and Cl (black circle). Injection of d-tubocurarine into the animal (to a final concentration of 10−4 mol 1−1 in the haemolymph) stopped the movement (B). (C,D) Injection of the same concentration of d-tubocurarine into an operated animal did not abolish (or even disturb) the pyloric pattern (compare D with C).

Fig. 6.

Injection of d-tubocurarine abolished dilator contraction (A,B) but did not abolish the rhythmic pyloric output of an operated animal (C,D). (A,B) In the control (A), the electrode implanted in the D muscles was used to monitor D contractions. Note that this electrode recorded simultaneously the activity in D (black triangle) and Cl (black circle). Injection of d-tubocurarine into the animal (to a final concentration of 10−4 mol 1−1 in the haemolymph) stopped the movement (B). (C,D) Injection of the same concentration of d-tubocurarine into an operated animal did not abolish (or even disturb) the pyloric pattern (compare D with C).

If the pyloric pattern recorded from an operated animal was due in some way to proprioceptive feedback, suppression of the feedback coming from the D muscle (by suppression of D muscle contractions) would be expected to stop this pattern or at least to alter it considerably. The experiment shown in Fig. 6C,D indicates that this is not the case. Injection of d-tubocurarine into an operated animal (at the same concentration as in Fig. 6B) does not abolish or modify the pattern: a comparison of Fig. 6C and Fig. 6D shows only a slight increase in the period attributable to direct central effects of the drug on the pyloric network. It should be noted that, although we have shown that d-tubocurarine blocks the neuro-muscular junction (Fig. 6B), electrical activity was still recorded by the electrode implanted in the D muscle. This activity must be of motor neuronal origin, and what we term EMG in such recordings is likely to be primarily a direct monitor of motor neurone terminal activity (neurography).

Another way to test the possible role of proprioceptive feedback from contracting muscles in generating the pyloric output is to perturb existing movement and to look for effects on the motor pattern. To do this, we implanted two electrodes into the D muscle of operated animals; one electrode was used for electromyographic and movement recordings, the other was used to stimulate the muscle electrically. As shown in Fig. 7, a brief stimulation of the muscle was applied at the beginning (Fig. 7A), in the middle (Fig. 7B) and at the end (Fig. 7C) of the cycle. None of these perturbations had any effect on the pyloric period (Fig. 7D,E). The same perturbations imposed by electrical stimulation to either Cl or C2 muscles are also without effects on the pyloric period (not shown). Together with the above pharmacological data, this strongly suggests that proprioceptive inputs cannot be responsible for the organization of the rhythmic output recorded in operated animals.

Fig. 7.

Perturbation of the dilator contractions had no effect on the pyloric period of an operated animal. (A–C) The same electrode was used to record simultaneously the electrical activity of the D muscle and the associated movement. A second electrode was implanted in the same muscle and was used for stimulation. An electric shock was delivered randomly to the D muscle during the pyloric period; stimulations were separated by at least six cycles. The electric shock induced a strong modification of the movement, but no modification of the pyloric period, regardless of the phase [early (A), medium (B) or late (C)] at which the stimulus occurs in the pyloric cycle. (D,E) The observation that the pyloric period does not vary when the dilator muscle is stimulated is shown graphically in D. The formulae used to calculate the points plotted on the abscissa (x) and on the ordinate (y) are shown in E; pl, p2, p3 are pyloric periods without stimulation, pS is the stimulated pyloric period, and IS is the latency of the stimulation in the pyloric period. In the graph, each point represents the mean and standard deviation (vertical bar) of points falling within the corresponding bin. It is clear that there was no variation of the pyloric period when the D muscle was stimulated. This has been confirmed by an analysis of variance, which shows that there are no significant differences among the bins (P=0.3804). Moreover, there is no significant difference (with a confidence level of 99%) between the mean pyloric period of the cycles just before the stimulation (1.549±0.134s), the cycles in which stimulation occurred (1.549±0.151 s) and the cycles just after the stimulation (1.534±0.146s).

Fig. 7.

Perturbation of the dilator contractions had no effect on the pyloric period of an operated animal. (A–C) The same electrode was used to record simultaneously the electrical activity of the D muscle and the associated movement. A second electrode was implanted in the same muscle and was used for stimulation. An electric shock was delivered randomly to the D muscle during the pyloric period; stimulations were separated by at least six cycles. The electric shock induced a strong modification of the movement, but no modification of the pyloric period, regardless of the phase [early (A), medium (B) or late (C)] at which the stimulus occurs in the pyloric cycle. (D,E) The observation that the pyloric period does not vary when the dilator muscle is stimulated is shown graphically in D. The formulae used to calculate the points plotted on the abscissa (x) and on the ordinate (y) are shown in E; pl, p2, p3 are pyloric periods without stimulation, pS is the stimulated pyloric period, and IS is the latency of the stimulation in the pyloric period. In the graph, each point represents the mean and standard deviation (vertical bar) of points falling within the corresponding bin. It is clear that there was no variation of the pyloric period when the D muscle was stimulated. This has been confirmed by an analysis of variance, which shows that there are no significant differences among the bins (P=0.3804). Moreover, there is no significant difference (with a confidence level of 99%) between the mean pyloric period of the cycles just before the stimulation (1.549±0.134s), the cycles in which stimulation occurred (1.549±0.151 s) and the cycles just after the stimulation (1.534±0.146s).

Blood-borne influences may be involved in the generation of rhythmic pyloric output in the operated animals

The preceding results indicate that the pyloric rhythmic pattern recorded in an operated animal is not due to neuronal inputs to the STG, suggesting that blood-borne influences alone are able to maintain the pyloric pattern, possibly by inducing oscillatory properties in the pyloric neurones.

This hypothesis was tested in ex vivo preparations of the stomatogastric nervous system removed from the animal 24 or 48 h after stn transection. The presence of rhythmic pyloric output was first verified by in vivo EMG recordings (Fig. 8A); the stomatogastric system was then removed from the animal, and activity was recorded on the motor nerves. Under these conditions, none of the nervous systems tested (N=16) exhibited any rhythmic pyloric activity (Fig. 8B). To show that humoral factors might be responsible for the activity recorded in vivo, we attempted to restore pyloric cycling by superfusing the STG with blood serum from Jasus lalandii. In only two of nine such ex vivo preparations was rhythmic activity restored. One possible explanation for this low success rate is that the blood may have coagulated during application. Centrifugation of the blood generally does not prevent clotting (Durliat, 1985) and, although such clotting can be prevented with strontium chloride (Durliat and Vranckx, 1981; Vella and Tripp, 1983), this salt strongly affects the pyloric rhythm and thus could not be used. Instead, we tried blood from the crab Cancer pagurus, since clotting of crab blood is effectively suppressed by centrifugation. Bathing the STG of an ex vivo preparation in saline plus crab serum (at least 70% v/v) restored the pyloric rhythm (Fig. 8B) in five of seven preparations tested. Interestingly, the blood serum used in the five preparations in which rhythmic activity was successfully restored was taken from the animal after 20:00 h local time, whereas the serum that gave the two negative results was taken from the animal earlier in the day. These results indicate that blood serum was able to induce rhythmic pyloric activity in an isolated STG.

Fig. 8.

Haemolymph serum can restore rhythmic pyloric output to a silent ex vivo preparation. (A) EMG pyloric activity of a 48 h operated animal (in vivo). (B) The STG (and output nerves) was transferred to a Petri dish (ex vivo) where it was unable to produce any rhythm in the presence of saline alone. Addition of crab blood serum to the saline was sufficient to restore the pyloric output.

Fig. 8.

Haemolymph serum can restore rhythmic pyloric output to a silent ex vivo preparation. (A) EMG pyloric activity of a 48 h operated animal (in vivo). (B) The STG (and output nerves) was transferred to a Petri dish (ex vivo) where it was unable to produce any rhythm in the presence of saline alone. Addition of crab blood serum to the saline was sufficient to restore the pyloric output.

It is now well established that, in several crustacean species, a completely isolated STG in vitro is unable to produce any rhythmic pyloric output. This has been demonstrated in Homarus gammarus (Moulins and Cournil, 1982), in Jasus lalandii (Dickinson and Nagy, 1983, and present paper) and in Panulirus interruptus (Russell and Hartline, 1978; Nagy and Miller, 1987). Under such conditions, the pyloric neurones lose their regenerative bursting properties; these properties are expressed only in the presence of permissive neuromodulatory inputs from the rostral ganglia (Bal et al. 1988).

In the present paper we have shown that, in vivo, an isolated STG remains able to produce the rhythmic pyloric pattern. Under these conditions (1) non-specific sensory stimuli that modify the pyloric pattern in the intact animal are without effect after stn transection; (2) the change in the pyloric pattern that occurs upon feeding in the intact animal does not occur in the operated animal; (3) the suppression or perturbation of putative proprioceptive feedback is without effect on the pyloric activity.

From these results we conclude that the rhythmic pattern observed in the operated animal is organized by the STG alone, i.e. with no extrinsic neuronal influences. The STG of an operated animal becomes electrically silent when removed from the animal and studied in vitro. However, it is possible to restore the pyloric rhythmic activity of such an ex vivo preparation by applying blood serum to the ganglion. We conclude that humoral influences are sufficient to maintain the bursting properties of the pyloric neurones.

Does cutting the stn suppress all STG neuromodulatory inputs?

The stn is the single connection between the STG and the higher centres (see Maynard and Dando, 1974). There is considerable immunohistochemical and pharmacological evidence to suggest that a large population of modulatory fibres travels to the STG from rostral centres via the stn-, these fibres are able to influence the activity of networks in the STG (Marder, 1987). Among these fibres are those such as APM (the anterior pyloric modulator of Jasus lalandii, Nagy et al. 1981), which have a permissive influence on the oscillatory properties of the pyloric motor neurones. The firing of APM restores cycle activity in a quiescent pyloric network (Nagy and Dickinson, 1983) by inducing the oscillatory properties of the neurones (Bal et al. 1988). Such permissive influences are suppressed by cutting the stn. The results obtained in vitro, i.e. cessation of the rhythmic pyloric output after stn conduction has been blocked or after the stn has been cut, are readily explained by such modulation. In contrast, the results obtained in vivo, i.e. continuation of the rhythmic pyloric output after stn transection, are more difficult to understand.

One possible explanation is that the sectioned input fibres of the stn remain able spontaneously to liberate modulatory transmitters in the STG, thus ensuring that the bursting properties of the neurones continue to be expressed. It is well known, especially in Crustacea, that axons can remain alive for a long time after transection (Bittner, 1988). Moreover, this has been specifically noted for the STG input fibres. Based on the morphology of the ganglion more than 200 days after stn transection, Royer (1987) concluded that at this time there were still some input fibres that had not degenerated. Nevertheless, it has not been demonstrated that these axons are still functional. If we are to explain our results with this hypothesis, we must assume that the modulatory sectioned axons are still able to synthesize and liberate transmitter 2 years after axon transection. We must also explain the observation that in vitro, stn blockade or transection resulted in the cessation of the pyloric pattern in less than 15 min.

In the stomatogastric system, modulatory neurones that do not use the stn to project to the STG have been identified in the crabs Cancer borealis and Cancer irroratus. These gastropyloric receptors (GPRs) (Katz et al. 1989) are primary sensory afferent cells that are active in phase with the movements of the gastric mill and that have direct prolonged neuromodulatory effects on neurones of the pyloric and gastric CPGs of the crab STG (Katz and Harris-Warrick, 1989). They project to the STG via the lateral ventricular nerve (Ivn), which can be considered as the output nerve of the STG. Cutting the stn does not suppress the STG input from GPRs and it could be argued that the modulatory activity of the GPR cells can ensure that the pyloric rhythm continues in operated animals. However, it has not been demonstrated that firing of the GPR cells is sufficient to cause rhythmic activity in the pyloric system. In addition, no gastric muscles were active in any of the animals whose stn had been cut less than 1 month before recording. Since the GPR cells are activated by gastric movements, it seems probable that they were silent in the operated animals. Thus, it is unlikely that they are involved in the generation of the rhythmic pyloric activity recorded in vivo for at least the first month after stn transection.

It remains a possibility that movements of the pyloric region might be able to activate the GPR cells, which in turn could induce cycling of the pyloric neurones. Although such a feedback cannot be totally rejected, it must be mentioned that perturbation and suppression of the dilator phase in the pyloric cycle (Figs 6 and 7) are without effect on the activity of the CPG.

Can the pyloric rhythm of an operated animal be the result of phasic proprioceptive feedback?

The work of Maynard (1972) clearly demonstrated that sensory inputs are not necessary for the genesis of a rhythmic pyloric pattern. Nevertheless, these sensory inputs exist, and their activities modify the existing pyloric rhythmic pattern (Dando et al. 1974). These inputs (except those of the GPR) have indirect access to the STG, travelling via the stn (see Moulins and Nagy, 1985; Simmers and Moulins, 1988). Cutting the stn suppresses their influence on the neurones of the STG. After stn transection, pyloric output appears to be produced without any relationship to the behavioural context. For example, the stimuli that normally result in a startle response in the intact animal are without effect on the pyloric rhythm of an operated animal. This loss of responsiveness is clearly demonstrated by the fact that feeding itself, which produces significant changes in the intact animal, does not produce any ‘adaptative’ modification of the pyloric output.

It can also be argued that proprioceptive feedback can introduce timing cues that will be sufficient to maintain a rhythmic pattern of behaviour in the pyloric neurones. Putative candidates for this role are proprioceptors associated with the pyloric muscles and travelling in a nerve other than the stn. However, our results have shown that suppression of the dilator movements (i.e. the suppression of possible feedback) does not perturb the pyloric pattern of an operated animal. Similarly, altering the activity of putative proprioceptors (by electrically disturbing the activity of the pyloric muscles) does not alter the pyloric output.

In the operated animal the bursting properties of the STG neurones are under the control of humoral agents

It could also be argued that deafferentation of the STG results in a profound modification of the membrane properties of the pyloric neurones. In the present case, the conductances involved in the oscillatory properties of the neurones would have to be modified in such a way that these properties became non-conditional (i.e. can be expressed in the absence of extrinsic influences). Although such modifications after denervation have yet to be described, experimental work is in progress to test this hypothesis. However, it seems unlikely that such a modification could be achieved 2 h after stn transection (see Fig. 3). Furthermore, this hypothesis fails to explain why the STG of an operated animal is unable to produce any rhythmic output when tested in vitro.

The best explanation for our results is that, under our experimental conditions (in the operated animal), the bursting properties of the pyloric neurones are induced and maintained by blood-borne modulatory influences. This has been postulated previously, based on pharmacological and histochemical results (Beltz et al. 1984). This hypothesis is supported by the observation that serum is able to restore the ability of an in vitro isolated STG to produce a rhythmic pyloric output. Previous experiments performed in vitro had shown that neuronal modulatory influences from higher centres are sufficient to induce the bursting properties of the pyloric neurones (Russell and Hartline, 1978; Miller and Selverston, 1982; Moulins and Cournil, 1982; Dickinson and Nagy, 1983). The data presented here show that humoral influences are also sufficient to induce the bursting properties of the pyloric neurones. We can consider the pyloric system as a CPG made from conditional oscillators that receive modulatory inputs from two sources, one neuronal and one humoral. Either is sufficient to induce rhythmic activity; neither is necessary.

From results obtained in vitro by bath application, it is clear that numerous amines and peptides could be responsible for the induction of the pyloric cycling (Marder, 1987), and in the intact animal the STG is probably exposed to a large number of active molecules. However, only a few of these substances have been demonstrated to be circulating hormones in Crustacea. Among them we must mention the red pigment concentrating hormone-like peptide, which is known to be a strong activator of the pyloric CPG (Nusbaum and Marder, 1988). Another putative candidate for a humoral inducing factor is the cholecystokinin-like peptide that has been identified in the haemolymph of Panulirus and is known to enhance the oscillatory activity of the pyloric neurones, but is without effect on the frequency of the rhythm (Turrigiano and Selverston, 1989). Interestingly, in Panulirus its level increases after feeding (Turrigiano and Selverston, 1990) but if the same is true in Jasus lalandii this does not induce any obvious modification of the pyloric pattern (see Fig. 5B).

This in vivo preparation will be useful in analyzing the effects of humoral modulatory influences on the pyloric network and in analyzing how these humoral factors interact with the neuronal modulatory influences that also impinge on this network. Moreover, our preparation will now allow us to compare in vivo the activity of a well-known CPG subjected only to humoral influences with the activity of the same CPG subjected to both neuronal and humoral influences.

This work was supported by a Grant from the Human Frontier Science Programme. The apparatus used to record simultaneously, with a single electrode, muscle electrical activity and movement, was built by Pierre Ciret.

Bal
,
T.
,
Nagy
,
F.
and
Moulins
,
M.
(
1988
).
The pyloric central pattern generator in crustacean: a set of conditional neuronal oscillators
.
J. comp. Physiol
.
163
,
715
727
.
Beltz
,
B.
,
Eisen
,
J.
,
Flamm
,
R.
,
Harris-Warrick
,
R. M.
,
Hooper
,
S. L.
and
Marder
,
E.
(
1984
).
Serotonergic innervation and modulation of the stomatogastric ganglion of three decapod crustaceans (Panulirus interruptus, Homarus americanas and Cancer irroratus)
.
J. exp. Biol
.
109
,
35
54
.
Bittner
,
G. D.
(
1988
).
Long term survival of severed distal axonal stumps in vertebrates and invertebrates
.
Am. Zool
.
28
,
1165
1179
.
Dando
,
M. R.
,
Chanussot
,
B.
and
Nagy
,
F.
(
1974
).
Activation of command fibres to the stomatogastric ganglion by input from a gastric mill proprioceptor in the crab, Cancer paguros
.
Mar. Behav. Physiol
.
2
,
197
228
.
Delcomyn
,
F.
(
1980
).
Neural basis of rhythmic behavior in animals
.
Science
210
,
492
498
.
Dickinson
,
P. S.
and
Nagy
,
F.
(
1983
).
Control of a central pattern generator by an identified modulatory interneurone in Crustacea. II. Induction and modification of plateau properties in pyloric neurones
.
J. exp. Biol
.
105
,
59
82
.
Durliat
,
M.
(
1985
).
Clotting processes in crustacean Decapoda
.
Biol. Rev
.
60
,
473
498
.
Durliat
,
M.
and
Vranckx
,
R.
(
1981
).
Action of various anticoagulants on hemolymphs of lobsters and spiny lobsters
.
Biol. Bull. Mar. biol. Lab., Woods Hole
160
,
55
68
.
Getting
,
P. A.
(
1988
).
Comparative analysis of invertebrate central pattern generators
.
In Neural Control of Rhythmic Movements
(ed.
A. H.
Cohen
,
S.
Rossignol
and
S.
Grillner
), pp.
101
127
.
New York
:
John Wiley and Sons, Inc
.
Grillner
,
S.
(
1977
).
On the neural control of movement. A comparison of different basic rhythmic behaviors
.
In Function and Formation of Neural Systems
(ed.
G. S.
Stent
), pp.
197
224
.
Berlin
:
Dahlem Konferenzen
.
Harris-Warrick
,
R. M.
(
1988
).
Chemical modulation of central pattern generators
.
In Neural Control of Rythmic Movements
(ed.
A. H.
Cohen
,
S.
Rossignol
and
S.
Grillner
), pp.
285
331
.
New York
:
John Wiley and Sons, Inc
.
Katz
,
P. S.
,
Eigg
,
M. H.
and
Harris-Warrick
,
R. M.
(
1989
).
Serotonergic/cholinergic muscle receptor cells in the crab stomatogastric nervous system. I. Identification and characterization of the gastropyloric receptor cells
.
J. Neurophysiol
.
62
,
558
570
.
Katz
,
P. S.
and
Harris-Warrick
,
R. M.
(
1989
).
Serotonergic/cholinergic muscle receptor cells in the crab stomatogastric nervous system. II. Rapid nicotinic and prolonged modulatory effects on neurones in the stomatogastric ganglion
.
J. Neurophysiol
.
62
,
571
581
.
Marder
,
E.
(
1987
).
Neurotransmitters and neuromodulators
.
In The Crustacean Stomatogastric System: A Model for the Study of Central Nervous Systems
(ed.
A. I.
Selverston
and
M.
Moulins
), pp.
263
300
.
Heidelberg
:
Springer Verlag
.
Maynard
,
D. M.
(
1972
).
Simpler networks
.
Ann. N.Y. Acad. Sci
.
193
,
59
72
.
Maynard
,
D. M.
and
Dando
,
M. R.
(
1974
).
The structure of the stomatogastric neuromuscular system in Callinectes sapidus, Homarus americanus and Panulirus argus
.
Phil. Trans. R. Soc. Lond. B
68
,
161
220
.
Maynard
,
D. M.
and
Selverston
,
A. I.
(
1975
).
Organization of the stomatogastric ganglion of the spiny lobster. IV. The pyloric system
.
J. comp. Physiol
.
100
,
161
182
.
Miller
,
J. P.
(
1987
).
Pyloric mechanisms
.
In The Crustacean Stomatogastric System: A Model for the Study of Central Nervous System
(ed.
A. I.
Selverston
and
M.
Moulins
), pp.
109
136
.
Heidelberg
:
Springer Verlag
.
Miller
,
J. P.
and
Selverston
,
A. I.
(
1982
).
Mechanisms underlying pattern generation in lobster stomatogastric ganglion as determined by selective inactivation of identifed neurones. 11. Oscillatory properties of the pyloric neurones
.
J. Neurophysiol
.
48
,
1378
1391
.
Moulins
,
M.
and
Cournil
,
I.
(
1982
).
All-or-none control of the bursting properties of the pacemaker neurones of the lobster pyloric pattern generator
.
J. Neurobiol
.
13
,
447
458
.
Moulins
,
M.
and
Nagy
,
F.
(
1981
).
Participation of an unpaired motor neurone in the bilaterally organized oesophageal rhythm in the lobsters Jasus lalandii and Palinurus vulgaris
.
J. exp. Biol
.
90
,
205
230
.
Moulins
,
M.
and
Nagy
,
F.
(
1985
).
Extrinsic inputs and flexibility in the motor output of the lobster pyloric neural network
.
In Model Neural Networks and Behavior
(ed.
A. I.
Selverston
), pp.
49
68
.
New York
:
Plenum Publishing Corp
.
Nagy
,
F.
and
Dickinson
,
P. S.
(
1983
).
Control of a central pattern generator by an identified modulatory interneuron in Crustacea. I. Modulation of the pyloric motor output
.
J. exp. Biol
.
105
,
33
58
.
Nagy
,
F.
,
Dickinson
,
P. S.
and
Moulins
,
M.
(
1981
).
Modulatory effects of a single neurone on the activity of the pyloric pattern generator in Crustacea
.
Neurosci. Lett
.
23
,
167
173
.
Nagy
,
F.
and
Miller
,
J. P.
(
1987
).
Pyloric pattern generation in Panulirus interruptus is terminated by blockade of activity through! the stomatogastric nerve
.
In The Crustacean Stomatogastric System: A Model for the Study of Central Nervous System
(ed.
A. I.
Selverston
and
M.
Moulins
), pp.
136
139
.
Heidelberg
:
Springer Verlag
.
Nusbaum
,
M. P.
and
Marder
,
E.
(
1988
).
A neuronal role for a crustacean red pigment concentrating hormone-like peptide: neuromodulation of the pyloric rhythm in the crab, Cancer borealis
.
J. exp. Biol
.
135
,
165
181
.
Phillips
,
B. F.
,
Cobb
,
J. S.
and
George
,
R. W.
(
1980
).
General biology
.
In The Biology and Management of Lobsters
(ed.
J. S.
Cobb
and
B. F.
Phillips
), pp.
2
82
.
New York
:
Academic Press
.
Rezer
,
E.
and
Moulins
,
M.
(
1980
).
Modalités d’expression du générateur du rythme pylorique chez les Crustacés: analyse électromyographique
.
CR. hebd. Séanc. Acad. Sei. Paris
291
,
353
356
.
Rezer
,
E.
and
Moulins
,
M.
(
1983
).
Expression of the crustacean pyloric pattern generator in the intact animal
.
J. comp. Physiol
.
153
,
17
28
.
Roberts
,
A.
and
Roberts
,
B. L.
(eds) (
1983
).
Neural origin of rhythmic movements
.
Symp. Soc. exp. Biol. 37
.
Cambridge
:
Cambridge University Press
.
Royer
,
S. M.
(
1987
).
Chronic effects of de-afferentation on the stomatogastric ganglion of Panulirus
.
In The Crustacean Stomatogastric System: A Model for the Study of Central Nervous System
(ed.
A. I.
Selverston
and
M.
Moulins
), pp.
251
257
.
Heidelberg
:
Springer Verlag
.
Russell
,
D. F.
and
Hartline
,
D. K.
(
1978
).
Bursting neural networks: a reexamination
.
Science
200
,
453
456
.
Selverston
,
A. I.
and
Moulins
,
M.
(
1985
).
Oscillatory neural networks
.
A. Rev. Physiol
.
47
,
29
48
.
Selverston
,
A. I.
and
Moulins
,
M.
(eds) (
1987
).
The Crustacean Stomatogastric System: A Model for the Study of Central Nervous System
.
Heidelberg
:
Springer Verlag
.
Selverston
,
A. I.
,
Russel
,
D. F.
,
Miller
,
J. P.
and
King
,
D. G.
(
1976
).
The stomatogastric nervous system: structure and function of a small neural network
.
Prog. Neurobiol
.
7
,
215
289
.
Simmers
,
A. J.
and
Moulins
,
M.
(
1988
).
A disynaptic sensorimotor pathway in the lobster stomatogastric system
.
J. Neurophysiol
.
59
,
740
756
.
Turrigiano
,
G. G.
and
Selverston
,
A. I.
(
1989
).
Cholecystokinin-like peptide is a modulator of a crustacean central pattern generator
.
J. Neurosci
.
9
,
2486
2501
.
Turrigiano
,
G. G.
and
Selverston
,
A. I.
(
1990
).
A cholecystokinin-like hormone activates a feeding-related neural circuit in lobster
.
Nature
344
,
866
868
.
Vella
,
F. A.
and
Tripp
,
M. R.
(
1983
).
Strontium chloride inhibition of clotting in blue crab hemolymph
.
J. Invert. Pathol
.
42
,
400
.
Weiss
,
K. R.
,
Koch
,
V. T.
,
Koester
,
J.
,
Mandelbaum
,
D. E.
and
Kupfermann
,
I.
(
1981
).
Neural and molecular mechanisms of food-induced arousal in Aplysia californica
.
In Neurobiology of Invertebrates
, vol.
23
(ed.
J.
Salanki
), pp.
305
344
.
Budapest
:
Pergamon
.