ABSTRACT The keyhole limpet Fissurella volcano has a sequential escape response to stimulation by tube feet of the predatory starfish Pisaster ochraceus. The response consists of 1) elevation of the shell at the point of stimulation, 2) extrusion of the mantle over the shell, 3) rotation of the shell around the keyhole axis, and 4) rapid locomotion away (fleeing). Limpets on horizontal surfaces have a stronger response to tube feet than do those on vertical surfaces. Responses to sea water which presumably contains a water soluble starfish scent consist of only the mantle extrusion response. Finally, receptors mediating detection of the starfish substance appear to be located on the tips of the mantle tentacles. INTRODUCTION A variety of gastropods have adaptive avoidance and escape reactions to predatory starfish (Bullock, 1953) and react to their predators both upon contact and at a distance. Detection of the predator is thought to involve chemoreception. A substance which may be involved in detection was isolated from Pisaster ochraceus and is most likely attached to proteins within the tube feet (Feder and Lasker, 1964). Numerous limpets respond to starfish. The Collisella and Acmaea limpets have a stereotyped response to tube foot contact: raising the shell, rocking from side to side, rotation, and fleeing (Bullock,1953). Other limpets such as Fissurella volcano's relative Diodora aspera exhibit a much different response. This response consists of extruding mantle tissue from the outer margin of the shell and out the keyhole to prevent starfish tube feet from fastening to the limpet shell (Margolin, 1964). The present study characterizes the response of Fissurella volcano to contact with Pisaster ochraceus tube feet and to water which had been exposed to a living starfish. Differences in the escape response in limpets tested on horizontal vs. vertical surfaces were also studied. In addition, morphological investigations of the suspected sensory elements were undertaken. MATERIALS AND METHODS Specimens were collected at The Great Tide Pool in Pacific Grove, California during low tides. They were segregated into two groups depending on the orientation of the surface on which they were found (ie. horizontal and vertical). Limpets were placed in tanks and maintained on surfaces in their respective orientations. For scanning electron microscopy the limpet was first stimulated using a tube foot to increase tentacle tip exposure and then immediately submerged in low calcium sea water containing 460mM Nacl, lmM Cacl, 50mM MgCl, 1OmM KCl, 10mM Hepes, and 5mM tetracaine, a solution which prevents contraction of the mantle. The limpet was subsequently de-shelled and fixed in a solution containing 12 glutaraldehyde, 12 para-formaldehyde, 0.22 sodium cacodylate, and 852 sea water at pH 7.3. Pieces of the mantle tissue with tentacles were then dissected and dehydrated in a graded acetone series. The remaining liquid was removed by critical point drying. Behavioral analysis was carried out using a video camera and tape recorder. A magnifying glass placed on the camera lens increased magnification of the subject. Limpets were tested on glass surfaces where they were manually stimulated by a tube foot held by forceps. These surfaces were either horizontal or vertical. A code that characterized the four components of the behavior was devised: L = lifting of the shell, generally at the point of stimulation M = extruding the mantle over the shell R = rotation around the keyhole axis F = fleeing L,M, and R were each preceded by a descriptive symbol. L was accompanied by one of the following descriptions to indicate where the shell lifted up: L = left R = right B = back F = front M was rated on a scale of O to 3 where O was no mantle response, 1 referred to the first row of tentacles sticking out slightly from below the shell, 2 implied that the mantle extended over the shell but the full response was not yet reached, and 3 indicated that the full response took place. Finally, R was preceded by the magnitude of the angle turned which was found by measuring along the longitudinal axis. Above this number an arrow was written indicating the direction turned. Using this code the behavior could be characterized more consistently and accurately. Timing of the behavior was done using a stopwatch and taking an average time from the videotapes. Reaction time of the timer was subtracted for calculations. Detection of the starfish at a distance was tested by first placing a starfish into an elevated bath of running sea water. Rubber tubing was connected to the container and carried the "starfish water" over a ramp and to the limpets. Limpets were placed in a trough with no ends so that water would flow over the limpets unidirectionally at 0.165m/s. The limpets were measured and placed 120mm from the downstream end of the container in alternating directions and subsequently observed for 30 minutes. RESULTS Anatomy of the Mantle Tentacle System Figure 1 illustrates the gross morphology of the mantle tentacles of Fissurella volcano. There are two parallel bands of tentacles. All of the tentacles are highly vascularized by blood sinuses and can be inflated by the pumping of blood into them from the very large heart. A similar single band of tentacles encircles the keyhole region. The picture depicted in Figure 1 is that of the animal after it has displayed the mantle response to starfish tube feet stimulation. The most peripheral band of tentacles extends upwards and over the top of the shell. The second group of tentacles protrudes from beneath the shell and lies parallel to the substrate to which the limpet is attached. The tentacles are irregularly and multiply branched. More detailed morphology of the tentacle tips as revealed by scanning electron microscopy is shown in Figure 2. Clusters of ciliary structures, several microns in diameter, are located on the tentacles and seem to be most common on the tips of the tentacles (Figure 24). Examination of a cluster of ciliary structures at higher magnification (Figure 2B) shows that each cilium is capped by a round bulb. These structures thus appear to be "paddle cilia," which are thought to be chemosensitive in many molluscan tissues (Haszprunar, 1985). Each cluster is seated in a small pit in the microvillar coating of the tentacle epithelium. Upon stimulation of various parts of the mantle and foot with a tube foot, the response was elicited most strongly when touched on the tips of the tentacles. From this information in conjunction with the scanning electron micrographs, it appears that the ciliary structures are probably the chemoreceptors involved in starfish detection. These receptors in Fissurella are arranged very differently from those of Notoacmea scutum, a limpet which runs from Pisaster. Unlike Fissurella, the receptors are much more highly concentrated on the tips of Notoacmea tentacles and are not embedded in pits in the epithelium covering the mantle (Phillips, 1977). Response of Fissurella to Predatory Starfish The entire escape sequence was only elicited upon stimulation of the tentacles with starfish tube feet. Mechanical stimulation of the tentacles resulted in withdrawal of the mantle. The general sequence of the response was derived from the coded videos and is schematized in Figure 3. It consisted of: 1) Lifting (Figure 3B) 2) Mantle extrusion (Figure 30) 3) Rotation (Figure 3D) 4) Fleeing (not illustrated) Lifting involved the upward tilting of the shell at the point of stimulation. This resulted in a broad exposure of the foot to the starfish. Approximately half the surface area of the shell was subsequently covered by mantle tissue emanating from both out of the apical orifice (ie. keyhole) and upwards around the rim of the shell. L and M occur is every limpet stimulated. Although R and F components occurred with less frequency than L and M, they too are important elements of the behavior. Rotation generally turned the limpet away from the tube foot stimulus (See also below). Fleeing served to move the limpet still farther away. Limpets placed in a flow of water which had run over a starfish upstream also invariably display the mantle response. The time scale in which they react is on the order of minutes (Figure 4). The reaction occurs between two and twelve minutes - a much longer time frame than that required for response to tube foot stimulation (See also below). Control animals exposed to an identical flow of sea water which has not flowed over a starfish do not exhibit this behavior. Neither the experimental group nor the control group lift, rotate, or flee. Observations of living starfish in the presence of Fissurella indicate that the limpet's mantle response stimulates an avoidance reaction by the starfish. Pisaster's tube feet appear to be repelled by the mantle tissue upon contact. However, they will grab on to any exposed portion of the shell. Time Course of Mantle Extrusion Response Although all limpets reacted rapidly to contact with a starfish tube foot, the size of the limpet influenced the time it took to begin the extrusion of the mantle following stimulation. Figure 5 indicates that as the length of the limpet increases, latency of the response also increases. Mantle extrusion is not an instaneous process and shows a good degree of variability in its time course. The total time required to complete the mantle response was therefore measured (Figure 6). This time ranged from 1 to 15 seconds. The graph is, however, not adjusted for size of the limpet. Direction and Degree of Rotation Dependence on Stimulus Site Figure 7 illustrates results from a series of experiments indicating that the point of stimulation determines the direction and the magnitude of the angle that the limpet will rotate through. The central oval in the figure represents a keyhole limpet (dorsal view) where the anterior end corresponds to zero degrees. Each numbered line refers to the point of stimulation measured as an angle with respect to the zero mark. The open circles contain information referring to angles turned at each point of stimulation. Arrows on the circles indicate the direction and magnitude of the angle turned by an individual limpet relative to the zero point. The dotted lines correspond to the means of the angles turned for clockwise and counterclockwise rotation. A closed circle refers to an individual limpet which did not detectably rotate. Detection sensitivity was 15'. From this figure it is apparent that the limpet turns in such a way so as to move its head away from the stimulation point especially at the 0, 45, and 90° points. Stimulation at 0° evokes a strong counterclockwise preference for rotation. Rotating seems less consistent and less marked with 135 and 180° stimulation. As shown more quantitatively below, the magnitude of the angle of rotation increases steadily as the limpet is stimulated closer to its head. Responses of Horizontal vs. Vertical Limpets Differences in the intensity of each of the behavioral components of limpets tested on horizontal vs. vertical surfaces were found. Analysis of rotation by horizontally oriented limpets is continued in Figure 8 where the mean absolute value (+ standard error of the mean) of the degree of rotation for horizontal limpets is plotted vs. absolute value of stimulation site (ie. direction of rotation is ignored). Magnitude of rotation decreases smoothly as the point of stimulation approaches the posterior end of the limpet. The magnitude of rotation for vertically oriented limpets is smaller than that of the horizontal limpets for all points of stimulation except at 135° where they are not detectably different. Limpets tested on a vertical surface also exhibit a lesser degree of rotation as the stimulus site moves posteriorly but may show a more abrupt decrease in angle size as stimulation moves from 0° to 45° than do horizontal limpets. The smaller sample size of vertical limpets makes such detailed comparisons difficult though. Fleeing was also studied in horizontal vs. vertical limpets. As shown in Figure 9A, whether or not a limpet on a horizontal surface will flee once stimulated seems random. In comparison, vertical limpets flee far less frequently. Scaling (x 41/27) of the vertical numbers plotted so they represent the same sample size as the horizontal values does not alter this conclusion. This difference in fleeing frequency between horizontal vs. vertical limpets is highly significant by the Chi square test. Magnitudes of the mantle extrusion response for limpets on horizontal vs. vertical surfaces also differ dramatically. Figure 9B shows results on an analysis in which mantle extrusion was estimated with the O - 3 scale (See Methods) for each population. All limpets showed some response to tube foot stimulation, but only 22 of horizontal vs. 452 of vertical limpets stopped the response at the stage 2 level. Again, this difference is still obvious when the vertical values are scaled (x47/35) to correspond to the horizontal sample size (dotted line). DISCUSSION From this research it is evident that Fissurella volcano has a distinct response to the starfish Pisaster ochraceus which undoubtedly serves as an escape mechanism. Once provoked by the starfish, the limpet has a stereotyped behavior which can be broken down into component parts: lifting, extruding, rotating, and fleeing. Diodora aspera, a related keyhole limpet, has a similar response to starfish, however, it only consists of two parts (Margolin, 1964). These components include the mantle response and the lifting of the entire shell as opposed to tilting specifically at the point of stimulation as in Fissurella. In investigating individual parts of the behavior, latency for the mantle response was found to decrease as limpet size decreased. Why this is true is not clear. The times involved are far longer than could be accounted for by neuronal conduction, etc. Possibly the phenomenon is an adaptive one. Because the tenacity of a limpet adhering to substrate must depend on the surface area of the foot, smaller limpets may be much more easily dislodged by a starfish than larger ones. Thus, it may be very advantageous for small animals to react as quickly as possible to prevent tube feet from attaching to their shells. Another method that these limpets have evolved to avoid contact by starfish is rotating the head away from the point of stimulation. Ability to escape the starfish is thus enhanced by this directed turning by allowing the limpet to flee by locomoting forward. Attack by a starfish closer to the anterior end would require a larger degree of turning in order for the limpet to face away from the intruder before fleeing. Precisely this behavior was found experimentally - the closer stimulation is to the head, the greater is the magnitude of the angle turned. When attacked at the posterior end, the direction of rotation (if desirable at all) would not be very critical to the survival of the limpet because the starfish is already behind it. Again, this behavior was found experimentally - nearly all limpets stimulated 180' opposite to the head fail to turn at all. While the response to tube feet is fairly stereotypical, certain differences do exist among limpets stimulated on surfaces with dissimilar orientations (specifically, horizontal and vertical). Limpets on vertical surfaces tend to have less intense responses than those probed on horizontal surfaces. That is, vertical limpets flee less often, extrude their mantle to a lesser degree more often, and nearly always display a smaller angle of rotation than do horizontal limpets. This apparent restraint in the vertical animals may again be adaptive. Limpets on a vertical surface must have a far greater chance of being dislodged by gravitational and water current effects if their ability to grip is significantly decreased, as it must be when they are undergoing the escape response. Falling exposes limpets to many serious dangers such as different predators or not landing upright. Some degree of caution in utilizing the escape response might be valuable for a vertical limpet. Detecting a starfish at a distance is another important aspect of Fissurella's behavioral repertoire. "Scent" detection allows the limpet to prepare itself for a potential attack by extruding the mantle well before contact occurs. The time for the response to begin is, however, much slower than for reaction to a tube foot. The limpet, therefore, would not seem to be constantly in a state of alarm and ready to react to a few water-borne molecules. Similarly, the animal does not have to expend an unnecessary amount of energy in maintaining the extruded mantle in a permanent defensive position. The interplay between water-borne and contact-mediated behaviors suggested here deserves additional study. Unlike several other species of limpets, Fissurella volcano does not flee when exposed to starfish water. Phillips found that Collisella limatula and Notoacmea scutum move downstream when exposed to flowing water which previously passed over a living starfish. In contrast, Fissurella shows no obvious directed movement under identical conditions. Fissurella only flees as a culmination of the entire escape response elicited by contact between mantle tentacles and tube feet. The mantle extrusion response probably more than compensates for this deficiency in their behavior by repelling starfish tube feet. The fraction of the shell that is exposed may not allow the starfish to get a good grip on Fissurella. This combined with movement away from the starfish via rotation and fleeing enhance its ability to escape predation. Despite much superficial similarity in the responses to starfish of Fissurella, Notoacmea, and Collisella limpets, definite differences exist in behavioral details, receptor location and type, and in neural pathways involved. Such specific differences would be expected based on the disparate taxonomic positions of Fissurella and the other genera. Further investigation of these differences may better explain the adaptive significance of the limpets' respective responses in light of the particular environment in which each lives. FIGURE LEGENDS Figure 1. Mantle tentacles. Mantle tentacles are depicted as they appear following extrusion due to stimulation by a starfish tube foot. Figure 2. Scanning electron microscopy of mantle tentacle. 2A) Possible receptors on the tip of a tentacle are shown. Receptors are ciliary structures. 2B) Ten times the above magnification of a receptor group sunk into a pit in the coating of microvilli covering the tentacle tip. Figure 3. Starfish escape response sequence. 3B) Limpet lifts its shell at the point of stimulation, 30) extrudes its mantle, 3D) rotates, and then 3E) flees. Figure 4. Total number of limpets with a full mantle response increases with time of exposure to starfish water or tube feet. The solid line depicts those exposed to water previously run over a starfish. The dotted line refers to the time course involved when touched with a tube foot. Figure 5. Length vs. Time to begin mantle response. As the size of the limpet increases, the time required to begin the mantle response increases. Figure 6. Time vs. Number of limpets that have completed the mantle response. This graph does not account for size of the limpets sampled. All limpets eventually complete the response. Figure 7. Direction and magnitude of angle rotated following tube foot stimulation. See text for explanation. Figure 8. Point of stimulation vs. degrees turned in horizontal and vertical limpets. Differences between horizontal and vertical limpets are significant (f992) at each angle except at 90° where they are marginally significant and at 135°. Figure 9. Differences exist in the escape responses of horizontal vs. vertical limpets. Dotted lines refer to values adjusted for equal sample sizes. 94) Representation of whether animals flee or not when tested on horizontal vs. vertical surfaces. 9B) Degree of mantle extrusion response vs. Number of limpets. The degree of response was measured on a scale of zero to three where three represents the full mantle response. BIBLIOGRAPIY Bullock, T.H. (1953). Predator recognition and escape responses of some intertidal gastropods in presence of starfish. Behaviour 5, 130-140. Feder, H.M. and Lasker, R. (1964). Partial purification of a substance from starfish tube feet which elicits escape responses in gastropod molluscs. Life Sciences 3, 1047-1051. Haszprunar, G. (1985). The fine morphology of the osphradial sense organs of the mollusca. I. Gastropoda, Prosobranchia. Phil. Trans. R. Soc. Lond. B. 307, 457-496. Margolin, A.S. (1964). The mantle response of Diodora aspera. Animal Behaviour 12(1), 187-194. Phillips, D.W. (1975). Distance chemoreception-triggered avoidance behavior of the limpets Acmaea (Collisella) limatula and Acmaea (Notoacmea) scutum to the predatory starfish Pisaster ochraceus. The Journal of Experimental Zoology 191(2), 199-209. Phillips, D.W. (1977). A scanning electron microscope study of sensory tentacles on the mantle margin of the gastropod Acmaea (Notoacmaea) scutum. The Veliger 19(3), 266-271. 16 ACKNOWLEDGMENT Many thanks to Dr. William Gilly for his tremendous help in completing this project and for the great fishing trips which definitely were highlights of the quarter. Thanks also to Chuck Baxter for invaluable advice and to the rest of the Hopkins people for making this a great learning experience. L a E LI 2 O o O S — ------ 2 ) (9 k ta A 31 2 2 . 4 . 2- S 2 5 L — - - - — â —— — 5 + —— . t N —.— — —— — . —— 10 — --- 10 SNOSSSA JILNVN Tn HIIM aSaNON - . — -- . - —— —— —— ——— â —.—— -- 3 —— — u — — — 2 1.6 9 E 1.2 — — — —- O.8 —. * â —. —— — — - O.4 - . — . - - L 2 9 0 —— LENGTH (mm) — — — —— —— --- --------— 1 — :— — —- — ---- — — 0 2 U — ...-- — ——.— 1 — —- L — — —— — — +- d SLAdMIT 4O HaNON — — — — 1 ! - — 0 — - — 0 — ——+ — 7 V S —— 2 — — — 4 — +)- 8 + S — S + — S S 45 5 1— + — —0 — 0 40 2 — — —/ — — 1 90 90 0 — — X — —— — — 4 9 4 — — — 1 1—7 N —— 180 7 — 1 - — - — S +—. + 8 — — —/— —6 — — — — - 4 —— — — — —— — 80 — —0 4 20 1 1 — — —0 30 —14 o — -0 + — 10 — — —— — . — -+ .. — — + 3 HORIZONTAE VERTICALE + S S S S 8 — + S — — — S — 8 — — — — — — L 45 90.—— 3 80 — ka- POINI OTIMULATION + — ... E. — . ---- —— —-— — HORIZONTAL — — ..-— ——— VERTICAL 40 370 6 * — 0 — —— 0.3 —— 2 . — 2 21 2 . O —— t u 5 910 -1 — —— — 77 e 0 â —.—— -- 46 .... .. .. S . .. 40 6.2 .— . O u 020 — — — — —- — 10 ----- — — ——.— — —— : —: - ---- 2 ——.— O — 3 2 - ---.— DEGREE OF SPONSE .. ——— —— —.—— -—- - .. — —— ——