Control of shell settling in the swimming sea anemone Stomphia coccinea

In their natural habitat, several species of sea anemone are found living on molluscan shells. The anemone itself appears to be responsible for establishing many of these associations through directed behaviour patterns (see review by Ross, 1967) and it is likely that shell settling may prove to be a widespread phenomenon amongst sea anemones. Recently, electrophysiological techniques have been developed (Josephson, 1966; McFarlane, 1969a) to help overcome the problem of recording electrical activity from these animals, and this has now opened the way for a complete reappraisal of the processes involved in the control of sea anemone behaviour. The existence of multiple conduction systems in sea anemones was first established by McFarlane (1969a), who demonstrated two slowly conducting systems, the SS1 and SS2, in addition to the familiar through-conducting nerve-net (Pantin, 1935).

The present study is concerned with shell settling in the swimming sea anemone Stomphia coccinea. After its swimming response has ceased, the anemone often displays a complex behaviour pattern when it comes into contact with a molluscan shell (Ross, 1965; Ross & Sutton, 1967). This eventually leads to the anemone becoming firmly settled on the shell. An analysis of the control and co-ordination on this important behaviour pattern is presented here.

Specimens of Stomphia coccinea were dredged from the San Juan Channel of Puget Sound at a depth of about 300 ft. They were usually found attached to the shells of the bivalve mollusc Modiolus modiolus. In the Friday Harbor Laboratories, the anemones were maintained in running sea water at temperatures ranging from 14 to 16 °C.

During the experiments, the animals were placed in tanks of running, well-aerated sea water at a temperature of 14 °C. Polythene suction electrodes were used for both recording and stimulation. Recording electrodes were connected to a Tektronix AM 502 differential amplifier. Recorded activity was displayed on a Tektronix 564B storage oscilloscope or recorded with a Racal-Thermonic T3000 instrumentation recorder for later display on a Brush 220 pen recorder. Electrical stimulation was provided by a Devices Neurolog stimulator system. Pulse intensity and duration were varied according to the needs of the experiment (see Results). The recording electrodes were always attached to tentacles as these proved to be the most suitable sites for detecting the diffusely propagated electrical activity in the various conduction systems.

The various stages of the shell-settling response in S. coccinea are depicted in Fig.1. A full description of this process is given by Ross (1965) and Ross & Sutton (1967). Fig. 1 A shows the anemone in the typical posture assumed after ‘swimming’ : resting on its side with the pedal disc detached and withdrawn. During this period, the anemone usually remains motionless and is generally unresponsive to mechanical stimulation. This posture may be maintained for 2–45 min or more. Suddenly, and almost imperceptibly, the animal emerges from its post-swimming torpor and becomes active. The shell-settling response is initiated if the tentacles contact a Modiolus shell. The first obvious response is a tentacle ‘alerting’ reaction wherein one or more tentacles may attach loosely to the shell surface. The tentacular activity at this stage generally results in the oral disc being drawn closer to the shell (Fig. 1B). Within about 1–2 min there is slight constriction of the pedal-disc margin accompanied by slight relaxation of the mid-column circular muscles. The column consequently assumes a ‘barrel’ shape. This form is maintained only briefly, however, for the mid-column circulars soon begin to contract, the margin of the pedal disc relaxes, and the centre of the pedal disc shows the first sign of swelling (Fig. 1B). This continues until the anemone takes on a ‘mushroom’ shape (Fig. 1C-D).

The pedal disc may remain inflated for 1–4 min. In this state it is very sticky to the touch and will adhere to almost any surface it contacts, due possibly to a lowering of the discharge threshold of the nematocysts in the pedal disc (Ellis, Ross & Sutton, 1969). Eventually, the pedal disc starts to deflate until the animal resumes something like its original posture (Fig. 1 E-F). If the pedal disc contacts the shell during any of the foregoing stages, it adheres and moves slowly on to the shell. If not, the whole procedure may be repeated several times until contact is made. Local bending movements of the column (see Fig. 1F), often directed toward the shell, also help to bring the pedal disc into the correct position. The various phases of the shell-settling response do not occur in sharply distinct stages but rather tend to merge into one another. The entire sequence (Fig. 1A-F) may take from 3 to 8 min to complete.

In order to monitor electrical activity successfully during the shell response, it was necessary to select animals that not only responded readily to shells but also generated pulses large enough to be identified on the recording apparatus. Selection was from a batch of about 50 anemones. An anemone was made to swim by placing a starfish, Dermasterias imbricata, against the tentacles (Yentsch & Pierce, 1955). Recording electrodes were attached to the tentacles during the subsequent post-swimming torpor. A Modiolus shell was then placed in contact with the tentacular crown. Recorded pulses could be identified as originating in a particular conduction system simply by comparing their waveforms with those pulses evoked by selective electrical stimulation of the appropriate system. Stomphia possesses three conduction systems (nerve-net, SS1 and SS2) with properties similar to those found in other sea anemones (see McFarlane, 1969a; McFarlane & Lawn, 1972).

During the post-swimming torpor, the electrodes usually record no endogenous electrical activity from the anemone. After about 5 – 15 min in this condition, however, the electrical silence may be broken by a single pulse in the SS2. This is generally followed by further SS2 pulses occurring at a mean interpulse interval of about 60 – 70 s. Occasionally, isolated pulses in the nerve-net or SS1 may be detected, but as a rule this background low-frequency firing in the SS2 is the only electrical activity to be monitored in a resting, detached Stomphia. This is true whether the tentacles are in contact with a Modiolus shell or not.

When the shell-settling response begins, there is a dramatic change in the recorded electrical activity. A complex burst of pulses is usually detected by any electrode near a tentacle that clings to the shell during the tentacle alerting phase. This local activity remains confined to the region around the contacting tentacle and is not detected by an electrode situated on the opposite side of the oral disc (see Fig. 2). It is possible, therefore, that these pulses are associated with the localized movements of the tentacles. Within a few seconds of these localized events the SS2 starts to fire at a much higher frequency than before. In the records shown in Fig. 2, an SS2 pulse appears within 4 sec of the beginning of the local, complex activity. This is followed by further localized complex activity. Less than 1 s later, two more SS2 pulses, about 600 ms apart, were recorded. Note that the diffusely conducted SS2 pulses arrive at recording electrode R1 earlier than at R2. This implies that the pulses originate from a site near R1 (see Lawn, 1975). The fact that R2 is proximal to the tentacle touching the shell suggests that the SS2 activity is being generated at the point of contact. In cases where the only tentacle in contact with the shell is near R2, the SS2 pulses arrive at R2 before R1.

This activity normally continues until the pedal disc starts to inflate (see Fig. 1B-C). Fig. 3 shows electrical activity recorded during a shell response, starting from the tentacle alerting phase and continuing through to inflation of the pedal disc and local bending of the column. SS2 pulses continue to appear until the pedal disc begins to inflate, a period of about 80 s in this particular case. Here, a total of 16 SS2 pulses with a mean interpulse interval of approximately 5 s were evoked, some of which occurred in couplets of 1–2 s, a feature consistently observed in recordings of this type. The constriction of the margin of the pedal disc to form a barrel-shaped column occurs about mid-way through the sequence (after 9 SS2 pulses have been evoked). At the end of the sequence, as the pedal disc starts to inflate, the electrical activity becomes very complex. At this stage, local bending movements of the column and tentacles usually begin, often resulting in the electrodes becoming detached. The complex pulse trains are often recorded in one electrode only and may represent activity related to these localized movements. In cases where the recording electrodes have remained attached during column bending, it was noted that SS2 activity ceases soon after the pedal disc begins to inflate. The complex, localized pulse trains sometimes continue during inflation of the pedal disc but eventually the trace quietens, the pedal disc begins to deflate and the anemone relaxes. If the pedal disc comes into contact with a suitable surface (usually the shell) during these manœuvres, it will adhere. If contact is not made, the whole cycle may be repeated several times until the pedal disc becomes successfully attached.

The number and frequency of SS2 pulses elicited during a shell response may vary considerably from trial to trial. In a single anemone, as few as 14 pulses (mean interpulse interval 3 s) and as many as 22 pulses (mean interpulse interval 4 s) produced successful shell responses. Sometimes, incomplete responses in which the pedal disc became only partially inflated were obtained when only 4 – 8 pulses were monitored. Mean interpulse intervals as short as 2 · 7 s and as long as 7 s were also encountered. These produced successful shell responses, provided a sufficient number of pulses were evoked. Successful shell responses were obtained with pulse trains of varying duration. The shortest recorded lasted for 39 s and had 14 pulses (a mean interpulse interval of 3 s) and the longest lasted for 105 s and had 16 pulses (a mean interpulse interval of 7 sec).

Successful responses were also obtained in cases where only one tentacle was in contact with the shell. This provided the only point of sensory contact for the anemone. If this contact were broken, the train of SS2 pulses would cease. An incomplete shell response would result if only a few pulses had been elicited. Neither a shell response nor an increase in SS2 activity could be recorded from an anemone whose pedal disc was firmly attached to the substrate.

In order to stimulate the SS2 selectively it was necessary to attach the stimulating electrode to a tentacle. A pulse duration of 200 ms was employed at an intensity above SS2 threshold but below nerve-net and SS1 threshold (see McFarlane, 1974). Two recording electrodes were also attached to the tentacles to verify that the correct system was being stimulated.

Fig. 4 shows the effect of stimulating the SS2 in a detached Stomphia. Twenty shocks were given at a frequency of one every 4 s. All the major stages of the shell response were elicited in the correct sequence. Fig. 4A shows the shape of the anemone immediately before stimulation. Within 1 min from the onset of stimulation (15 SS2 pulses elicited), the centre of the oral disc had withdrawn slightly and the margin of the pedal disc was beginning to constrict (Fig. 2 B). About 20 s later, all 20 SS2 pulses had been evoked and the shell response was well under way. Fig. 4C shows the situation 30 s after stimulation had ceased. The margin of the pedal disc is now expanding, the circular muscles of the mid-column are contracting strongly and the centre of the pedal disc is beginning to inflate. The pedal disc continues to inflate (Fig. 4D) until it reaches a maximum (the ‘mushroom’ effect) as shown in Fig. 4E. The inflation process may take up to 3 min to complete. The recovery stage then begins as the pedal disc deflates (Fig. 4F) and the anemone resumes its resting posture. The entire sequence (Fig. 4A-F) took 5 min to complete. The only stage of the shell response that electrical stimulation of the SS2 could not reproduce successfully was the local bending of the column. One would not expect local contractions to be controlled by a diffusely conducting system such as the SS2. It is possible, therefore, that these are co-ordinated by a locally conducting system that could not be detected or stimulated by the techniques used here.

Complete shell responses could be reproduced by stimulating the SS2 with 20 shocks at frequencies as high as 1 shock/s and as low as 1/10 s. Weak or incomplete responses were obtained at 1/500 ms and 1/20 s. Beyond these limits, no obvious response could be detected. The rate at which the response proceeded could be adjusted simply by altering the frequency of stimulation. The response could be completed within 2 · 5 min when the SS2 was stimulated at a frequency of 1 shock/s, whereas this would take up to 8 min at frequencies as low as 1 shock/10 s.

Within the active range of frequencies, a minimum of about 10 shocks was required to produce a complete response. Weaker responses were obtained with fewer shocks. With less than 4 shocks, no response could be detected. If the SS2 was stimulated continually, the shell response would proceed in the usual way. It was often noted, however, that the pedal disc did not become completely inflated until stimulation was stopped. This indicates that activity in the SS2 may partially inhibit those mechanisms responsible for inflating the pedal disc. Conversely, the termination of sensory SS2 activity (see Fig. 3) may be related to inflation of the pedal disc. The process by which the SS2 might be ‘switched off’ in this way remains unknown, but it is possible that some feedback mechanism from mechanoreceptors in the pedal disc may be responsible.

It was not possible to elicit any part of the shell response by electrical stimulation of the nerve-net or SS1. The activities controlled by these systems in Stomphia will be described elsewhere (Lawn, unpublished data). Furthermore, the shell response could not be evoked by electrical stimulation if the pedal disc was already attached to the substrate. In this case, stimulation of the SS2 at frequencies capable of eliciting shell responses in detached anemones would only cause the mouth to open. Detachment of the pedal disc did not occur. Ross & Sutton (1967) described transfers by recently settled Stomphia to shells from other surfaces. In this process, partial lifting of the pedal disc preceded inflation and ‘searching’ for the shell. Apparently, in order for the shell response to proceed in an attached anemone, the detachment mechanism must first be activated. It is believed that this process is controlled by the SS1 (Lawn, unpublished data). A similar mechanism is involved in the detachment of the pedal disc of the sea anemone Calliactis parasitica (McFarlane, 1969 b) during its response to molluscan shells.

It was noted that dropping dissolved or undissolved food material on to the margin of the mouth of an attached anemone caused the mouth to open wide. This may be interpreted as a component of feeding behaviour. This response was always accompanied by an increase in the endogenous SS2 activity. The mean interpulse interval would decrease from a resting level of 60 – 70 s to a value between 4 and 10 s. Dropping dissolved food on to the mouth of a detached anemone evoked a similar increase in SS2 activity. Instead of the mouth opening, however, the anemone carried out a typical shell response. Once again, the only absent component was the localized bending of the column. This indicates that the SS2 in Stomphia may connect with at least two types of chemoreceptor: one responsive to food substances and situated exclusively around the margin of the mouth, the other responsive to the active component in the shell and situated in the tentacle ectoderm. At this stage, however, it is not known how Stomphia responds differently to identical trains of SS2 pulses depending on whether the pedal disc is attached or not. Again, it is possible that some feedback mechanism from receptors in the pedal disc may alter the activation threshold of the effectors involved in these two separate responses.

We have shown that shell settling in Stomphia coccinea is controlled and coordinated, in part at least, by the SS2. This is the first demonstration of a complex behavioural response controlled by this system, although the SS2 in Calliactis parasitica has been shown to co-ordinate mouth opening (McFarlane, 1975). The detached anemone recovers from the torpor induced by swimming and the tentacles become very active. If, during this ‘searching’ activity, a responsive tentacle comes into contact with an active component found in the shells of certain molluscs (Ross & Sutton, 1967), a sensory response is elicited. This involves excitation of the SS2, causing the system to fire at a frequency 10 – 20 times higher than that normally encountered in a resting animal. A direct result of this is that the shell-settling response is activated. When the response has reached the stage where inflation of the pedal disc occurs, the activity in the SS2 declines and eventually ceases. This seems to allow the pedal disc to inflate to a maximum, thereby improving the chances of a successful transfer to the shell.

Electrical stimulation of the SS2 alone is sufficient to reproduce the entire shell-settling response up to the point where local bending movements of the column and adhesion of the pedal disc occur. The number of shocks required to elicit a successful response corresponds closely to the number of SS2 pulses recorded from the anemone during its shell-settling behaviour. Furthermore, successful responses are only obtained when the frequency of the applied stimulus approximates the mean firing frequency of the SS2 during the sensory response. Such a close correlation between the recorded electrical events and the results of electrical stimulation supports the idea that the SS2 is solely responsible for controlling the major components of the response.

The active component in the shell is probably detected by chemoreceptors in the ectoderm of the tentacles. The SS2, however, seems to be confined to the endoderm (McFarlane, 1969 a). It seems probable, therefore, that the chemoreceptors are ectodermal sensory cells that send processes through the mesogloea of the tentacles to synapse with the conducting elements of the SS2. A similar situation may exist around the margin of the mouth, where food-sensitive chemoreceptors also appear to connect with the SS2. Ultrastructural studies are required to confirm the presence of these connexions, but are made difficult by the fact that it is not yet established if the SS2 is a nervous or a neuroid system.

It is not clear how the SS2 co-ordinates the complex muscular activities involved in the shell response. In Calliactis parasitica the SS2 inhibits inherent contractions of endodermal muscles (McFarlane, 1974). A similar action in Stomphia coccinea can, however, provide only a partial explanation for the observed muscular responses. The initial ‘barrel’ shape may result directly from the inhibitory action of the SS2 causing a symmetrical relaxation of the mid-column circular muscles. Likewise, the increased SS2 activity may elicit symmetrical relaxation of the endodermal circulars in the pedal disc and its margin during the subsequent stage where the pedal disc becomes inflated. The accompanying symmetrical contraction of the mid-column circulars cannot, however, be attributed to the inhibitory action of the SS2. In Calliactis, stimulation of the SS2 alone was incapable of eliciting contractions in any muscle group (McFarlane, 1974). Contraction, however, does follow cessation of stimulation. It is possible that in Stomphia contraction of the mid-column circulars may be a direct response to stretching. This would occur during the formation of the ‘barrel’ shape in the initial stages of the response. As these muscles are released from the inhibition imposed on them by the SS2, the mid-column begins to constrict. This forces the coelenteric fluid down into the pedal disc region. Since the muscles in this region will also be relaxed as a result of the increased SS2 activity, the pedal disc will start to inflate. It must be further postulated that the muscles of the pedal disc do not contract in response to stretching. This implies a basic functional difference between these muscles and those found in the mid-column region. The observation that maximal inflation of the pedal disc may be held off until SS2 activity ceases has to be accounted for. It is possible that the constriction of the mid-column, which in turn contributes to the inflation of the pedal disc, cannot reach a maximum until all residual inhibition has been lifted from the mid-column circulars. This would only occur once the SS2 had stopped firing. In Calliactis, a reduction in SS2 activity releases the endodermal muscles from inhibition (McFarlane, 1974). Once the SS2 activity has ceased, therefore, the mid-column circulars become fully contracted and the pedal disc subsequently becomes maximally inflated. The circular muscles of the pedal disc may then contract, perhaps as a result of a much slower escape from the SS2 -mediated inhibition than is shown by the mid-column circulars. This would bring about deflation of the pedal disc during the recovery stage of the response.

The means by which the SS2 is ‘switched off’ at the appropriate stage remain undetermined. Perhaps some type of stretch receptor located in the pedal disc discharges as the inflation process begins and inhibits SS2 activity in some way. The observation that the SS2 can be electrically stimulated while the pedal disc is inflated would suggest that fatigue in the conduction system is not responsible for the decline in activity. The discharge of the presumed stretch receptor, therefore, may raise the activation threshold of the shell-sensitive chemoreceptors in the tentacle and consequently block the sensory activity in the SS2. The pathway by which this negative feedback mechanism might work is not known. Simple sensory adaptation cannot, however, be excluded as a possible mechanism for the observed reduction in SS2 activity.

One problem remains unresolved. How are the local bending movements of the column activated ? Electrical stimulation of the SS2 cannot reproduce them. As mentioned earlier, control may be exerted by a local system, undetected by the techniques employed here. Presumably, mechanoreceptors situated in the tentacle contacting the shell channel sensory activity into this local system. This is one possible explanation for the local electrical events recorded when a responsive tentacle first contacts the shell (see Fig. 2). Unlike the diffusely conducting SS1, SS2 and nerve-net, the local system would confine its activity to a restricted region orientated towards the source of sensory input. The system seems to be polarized in such a way that spread occurs more readily in the longitudinal direction than circularly. This would induce asymmetric contraction of the column musculature, causing the column to bend and bring the pedal disc closer to the surface of the shell. Asymmetric bending of the column is a common phenomenon in actinian behaviour. It occurs in preparatory feeding behaviour of Metridium senile (Pantin, 1950) and Tealia felina (McFarlane, 1970), in shell climbing behaviour of Calliactis parasitica (Ross & Sutton, 1961) and in swimming behaviour of Stomphia coccinea (Sund, 1958). Further information is needed to provide an adequate explanation for these local responses.

The work was carried out at the Friday Harbor Laboratories of the University of Washington where the Director, Dr Dennis Willows, made generous provision of space and facilities. Thanks are also due to Dr D. M. Ross for many helpful discussions relating to this investigation. The following grants made the work possible : operating grant A-1445 of the National Research Council of Canada to Dr D. M. Ross provided a post-doctoral Assistantship for I.D.L. and together with funds from the University of Alberta supplied the equipment ; the Browne Fund of the Royal Society provided travel support for I. D. M.