Long-Term Facilitation in A Swimming Sea Anemone

The experiments were performed during the summer of 1958 at the Friday Harbor Marine Laboratory of the University of Washington.

Sund (1958) gives an illustrated account of the swimming reaction. A brief description is presented here, with a few additional observations based on the differences between the small specimens used in this study and the large ones used by Sund and upon measurements of the duration of the various components of the behaviour.

Following appropriate stimulation Stomphia coccinea successively retracts, closes, elongates, opens widely, detaches itself from the substrate, and bends from side to side violently, propelling itself through the water. During the swimming, and for some time after, the animal usually fails to respond to external stimulation, either mechanical or chemical. The swimming reaction may be initiated by contact with any of several species of starfishes or by appropriate electrical stimulation. The intensity and duration of the response do not depend on stimulus strength, above threshold.

The initial withdrawal is probably not truly a part of the reaction, but only a tactile response. It does not occur in specimens which are induced to swim shortly following one swimming period. Gentle touch by a specimen of the starfish, Dermasterias imbricata (Grube), can result in swimming without initial closure.

The bending movements which effect locomotion usually follow one another at a regular frequency of one or two per second, but the direction of successive movements is often irregular, there being no apparent relationship between one movement and another in space but only in time. In large specimens the bending may at first take the form of a smooth progression of contraction around the circumference of the animal, the result being a swirling of the oral end. Swirling movements were never seen in small specimens. In large specimens the bending (or swirling) movements begin before detachment and probably aid in loosening the foot. This is generally not true for small individuals in which detachment is accomplished by contraction of the basal region only. The frequency of the bending is lower in small specimens so that their reaction is less vigorous and does not result in propulsion.

The period of swimming, here taken as the interval between the first sign of the whole response (extension of the column) and the last swimming movement, is quite variable but typically lasts about 5 min. The duration decreases only after several trials. More immediate effects of repetition are longer latency of the first movement and decrease in the frequency of the swimming movements.

The unresponsive period is even more variable. It may begin during the swimming or some time afterward. Its duration varies from immeasurably short to several hours. On successive trials this duration may either increase or decrease. In one individual, stimulated to fatigue, the period of unresponsiveness grew for the first several trials and then decreased to extinction, and on the next 2 days swimming was not accompanied by any refractoriness. The refractoriness is apparently never complete. A very strong stimulus such as heavy prodding with a rod or cutting the animal in two interrupts the swimming. The cessation is abrupt and may be accompanied by a partial or total withdrawal response. After about half a minute the swimming resumes.

Partial reactions are occasionally observed. Rarely, subthreshold stimulation (see later) produces abortive responses in which the reaction ends after detachment. In the tiniest specimens and in some fatigued ones the whole sequence may occur minus the swimming contractions themselves. Some individuals may swim without opening the oral disk. Others are seen to produce lateral contractions without elongation of the column and detachment.

Parts of animals show the swimming behaviour. The basal halves of animals divided across the oral-aboral axis show all the features of the response normal for them, including long-term facilitation (see later) and resumption of activity after interruption. The oral pieces do not survive well, but the few observations on them suggest that they also retain the ability to make specific response to the starfish.

The sequences in which elements such as elongation, opening, bending, detachment, or refractoriness are missing, or rearranged, as in the ordering of detachment and bending, together with the possibility of resumption of the sequence after interruption, suggest that the swimming of Stomphia is not organized as a simple chain of reflexes such as might explain the walking of Hydra. The normal response seems to be an integrated activity of the whole animal in which functional subdivisions are normally linked by a central cause, but are not sequentially dependent. Furthermore, it appears that during a period of interruption in which some other activity takes place some part of the animal is able to store the information which constitutes the internal stimulus for swimming.

When stimulated by the starfish, Dermasterias, the response is labile and facilitated by previous occurrence of the stimulus. Many individuals, especially small ones, do not react to starfish on the first contact except by an ordinary tactile withdrawal. Stimulation every few minutes (or as soon as the animal opens) usually results in a swimming response on the first, second, or third repetition, but some specimens require more trials and a few do not react with over 20 repetitions. The facilitation is not erased by a response. It decays very slowly and in some cases outlasts the maximum duration of test, which was 7 days. Since the smallest specimens (6—8 mm. in diameter) do not swim and large ones require little facilitation, one can suppose the ontogeny of the behaviour to involve a gradual decrease in the threshold for response. Lacking a knowledge of the experience of individuals it is not possible to decide whether this occurs without encountering the stimulating starfish.

A standard stimulation by means of electric shock allows more precise characterization of the facilitation. The following list summarizes the tests made in order to decide upon a standard stimulus. Pulse stimuli produced by a Grass S4 stimulator were delivered through stimulating electrodes placed against the column. Tests were carried out at about 20° C. Eighteen animals were used in all. The small number of specimens available and the change of reactivity with experience prevented extensive testing. The useful information was gathered with as few trials as judgement allowed:

• There is ordinarily no response to the first shock.

• A retraction and closure occurs following the second shock. This facilitation is presumed to be the same neuromuscular facilitation common to other actinians; it lasts 3 sec.

• In open and not previously stimulated animals a train of precisely 8 shocks results in a full swimming response. The frequency may be any value between 1/2 sec and 4/sec; the threshold is the same as for (b). The figure 8 was observed in all of the sixteen specimens which did swim; fewer shocks never released swimming in one to three widely separated trials on each of 8 fresh or long-rested animals; more were rarely required. In trials in which 8 shocks did not produce swimming additional shocks usually failed as well.

• With a stimulation frequency of 1 /sec. or greater the latency of the response measured from the first shock is nearly constant at 8 sec.

• At frequencies less than 1 /3 sec., the response does not occur even after many repetitions.

• At frequencies greater than 4/sec., more than 8 shocks are required.

The following tests employed a stimulus of 8 shocks at 1/sec., each shock of o-1 msec, duration and just above the voltage threshold for a neuromuscular response on the second shock. It should be stressed that the number of shocks necessary is quite constant in animals which have not been tested previously. This is true even of those which do not react readily to Dermasterias. Some closed animals, however, do not react even to many shocks.

In the main facilitation experiments a burst of 7 shocks was delivered. This usually results only in the closure and retraction responses. But the same stimulus repeated 1-5 min. later causes a full swimming response. With repetition of this sequence the interval between the conditioning and test stimuli can be increased to hours or even days so that a specimen always responds to 7 shocks. An individual which has been thus conditioned may be made to respond to only 6 shocks by similar conditioning. Such an individual may retain the ability to react to 6 shocks after at least 1 week without practice. This decrease in threshold for the swimming response, lasting hours or days, was found in each of 8 specimens. Facilitation lasting 1 week was tested in two specimens only. The maximum period was not determined. The lowered number of stimuli required was probably not due to change of other conditions since the previous stimulation, because 12 animals kept under the same conditions for the same length of time or longer before their first electrical stimulation required a full 8 shocks to release swimming.

For anemones which do not react the first time contact with Dermasterias may be considered a subthreshold stimulus. If this follows, or is followed within a few minutes by, a subthreshold electric stimulus (5 to 7 shocks) the two kinds of stimulus may sum and produce a response. This is true even of anemones which cannot be made to swim after a score of trials with starfish alone. Subsequent to a single response elicited by mixed stimuli such individuals acquire an ability to respond to a starfish alone. These experiments can be arranged to look like Pavlovian conditioning. The difference is that the ‘conditioned’ stimulus is not a neutral one, but a stimulus which ordinarily produces the response in other members of the species without prior conditioning. The simplest explanation of this acquired behaviour is not that there has been a closure of neural paths involved in two different reactions but that there has been a lasting decrease in the threshold for a response that is probably inherent.

A difficulty in describing this behaviour in Stomphia as learning arises in the attempt to assess adaptiveness of the response. Dermasterias is not known to feed upon anemones even when there is opportunity to do so. Even a ‘hungry’ starfish, which subsequently devoured chitons, crept over anemones which were fixed in place without apparent interest in them as food. The distribution and habitat of Stomphia and Dermasterias are such that they would rarely, if ever, contact each other in nature. Sund (1958) discussed other possible significance for the response than predator escape, but he found no evidence to support alternative hypotheses. It remains possible that the adaptiveness of the response must be sought in the history of the species and that at present it is a behavioural atavism. Even this suggestion, however, does not provide escape from the conclusion that the ‘learning ‘has no adaptive significance for the individual. An alternative answer to the question of adaptiveness is that the response is a non-specific one to a predatory group. The negative results of tests made by the previous workers using other animals including many echinoderms from the Puget Sound area weaken the apparent reasonableness of this suggestion. It seems that the response is quite specific to at most a small group of echinoderms. Perhaps the clue to this relationship must be sought in some other portion of the range of the species.

Other cases of ‘learning’ in actinians, both habituation and associative learning, requiring memory lasting for periods similar in length to those found here have been reported (reviewed in Thorpe, 1956). Short-term ‘memory’ perhaps not involving persistence for days is suggested by the results of Batham & Pantin (1950), who found that anemones might walk away hours after a period of intense stimulation. The present results do not add to a knowledge of the mechanism of memory in coelenterates, but they do point to the possibility of such an analysis.

The two kinds of stimulus used in this study probably affect directly different structures in the organism. The identity of intensity threshold for electrical stimulation in both simple withdrawal response and the swimming response suggest that the same structures are initially stimulated. In other species of anemones the withdrawal response is due to stimulation of a through-conducting nerve net. A peculiarity of elicitation of swimming via this through-conducting net is the rather large and constant number of pulses necessary. A similar phenomenon occurs in Calliactis in which slow contractions are produced after several shocks (‘usually not less than six’) by low-frequency stimulation of the through-conducting net (Ross, 1957). These slow contractions are of nearly constant latency regardless of frequency, but are graded in intensity by frequency and number. The present case differs notably in that the frequency must be great enough to produce the fast muscle response. It is possible that the swimming is initiated by slow-type activation of certain muscles.

The great specificity of the response to other organisms shows that stimulation by starfish is routed through sensory levels and that special sensory mechanisms must be present. Stimulation by the starfish presumably does not involve the through-conducting nerve net, for in some cases the swimming response can be elicited without a withdrawal response. The long-term facilitation which occurs must take place along the functional path between this nerve net and the reacting muscles. Since both shock and starfish stimulation produce the response and because subthreshold stimulation by each can summate, the pathways followed by excitation from each must clearly overlap. This overlap may or may not involve more than the last stage, the muscles themselves. The facilitation must occur at the initial point of overlap since electrically induced reactions result in marked decrease in threshold to starfish only when the two are presented in the same trial. These considerations place the probable location of the long-term facilitation at the point of convergence of the two kinds of stimulation used and somewhere between the through-conducting nerve net and the responding muscles. The neuromuscular junction may be named as a possible site.

The decreased latency of response that Wells found in Sepia and the change of behaviour of Stomphia seem likely to be based on similar mechanisms. The first might involve an increased efficiency in the handling of information necessary for recognition of mysids, but the spread in the range of releasing stimuli suggests a decrease in the threshold for attack. The two cases operate in quite different parts of the receptor-effector chain. Wells locates the process in Sepia ‘somewhere between the retina and the motor centres’, whereas it takes place at or near the muscles in Stomphia.

1. Repetition-produced modifications in the behaviour of the swimming sea anemone, Stomphia coccinea, are described. Lowered threshold to number of electrical shocks on successive trials indicates a kind of ‘learning’ called here longterm facilitation.

2. Dissection of the behaviour into its components, both by experimental techniques and observation of atypical cases, shows the swimming reaction not to be simply a chain of reflexes, but to be ‘centrally’ co-ordinated.

3. The conditions for electrical elicitation of swimming are shocks sufficient in intensity to activate the through-conducting nerve net repeated eight times in the frequency range of 1/2 sec. to 4/sec. Fewer than 8 shocks constitute a subthreshold stimulus in fresh animals; more than 8 are rarely required.

4. Repeated subthreshold stimulation by starfish or by 7 electric shocks result in a long-lasting facilitated state in which the same stimulus repeated hours later may produce a full response. The facilitated condition has been observed to last 7 days. Controls kept without stimulation do not show facilitation.

5. The probable site of this long-term facilitation is discussed. It is suggested that this site is at the point of convergence of the two types of stimulation used and between the through-conducting nerve net and the responding muscles.

The author wishes to thank Dr D. L. Ray and Dr A. H. Whiteley for making possible this study, and Dr T. H. Bullock for reading of the manuscript. The work was performed during the tenure of a National Science Foundation predoctoral Fellowship and was aided by a grant (B21) to Dr T. H. Bullock from the National Institute of Neurological Diseases and Blindness.