Abstract Neurological research on escape response in the squid Loligo opalescens has raised the possibility that active feeding behavior post hatching stimulates development of adult-like control over the use of giant axons in both escape and attack jets. This study provides behavioral evidence to support that hypothesis, by comparing the attack behavior developed by hatchlings raised on prey which required an active pursuit and capture strategy with that of hatchlings raised on slow-moving prey. 2 groups of squid were given slow-moving prey post hatching and introduced to fast-moving prey at different stages in the life cycle to determine (i) whether attack behavior varied with prey type; (ii) if so, whether feeding behavior could be successfully adapted to new prey; and (iii) if there was an optimal time for this transition to take place. The squid developed strikingly different capture techniques for slow and fast prey. Animals given fast-moving copepods on day 10 modified their prey capture behavior sufficently for adequate survival on the fast prey. Squid given copepods on day 34 failed to alter their attack behavior, were unable to capture enough prey and died of starvation within 10 days. It is proposed that effective prey capture behavior involves the use of the giant axon system in an adult-like manner, and can only be stimulated by exposure to fast-moving prey. Exposure must occur during a short critical period after hatching (less than 33 days) for normal behavioral and nervous development to take place. INTRODUCTION Squid are active predators that from birth depend on high-speed, jet-propelled attacks to capture their prey. Precise aiming and timing of these jets are probably necessary for maximum efficiency, since the bulk of a young squid's prey consists of fast-moving organisms such as copepods and larval fish. Another high-speed, jet- propelled behavior that requires precise control is the escape response. As these typically backwards-directed escape jets are comparable in thrust to the forward- directed jets made during prey attacks (Packard, 1969), it is reasonable to assume that the same neuromuscular systems underlie both behaviors and that their ontogeny could be linked. The giant axons of loliginid squid are a vital link in the motor system producing escape jetting, a pathway first described in detail by Young (1938). Recent in vivo recordings (Otis and Gilly, 1990) have shown that although strong escape jets can be generated by an "ordinary" small motor axon system (Wilson, 1960) acting alone, the most powerful escape responses are elicited by delayed activation of the giant axon system in coordination with bursts of activity in the small motor fibers. Optimal performance thus requires careful control of the timing of giant axon excitation relative to non-giant motor activity. It is not yet clear precisely when after birth this dual motor control in escape response behavior emerges, or what factors influence its development. Gilly, Hopkins and Mackie (1991) found that in embryonic and newly hatched squid, short-latency firing of the giant axons preceded excitation of any small motor neurons in escape responses. This pattern is thus the reverse of the delayed firing pattern found in adult squid. These workers also discovered that the delayed recruitment pattern of the giant axon begins to appear shortly after birth. More recent experiments (Gilly, in preparation) have revealed that by 10 days post¬ hatching, most animals that had been feeding on plankton (mostly copepods routinely employed delayed, adult-like recruitment of the giant axons. Animals that were not provided with this prey, however, remained at a developmental stage more closely resembling the embryonic stage characterized by short-latency activation of the giant axon. A developmental connection thus appears to exist between the ontogeny of the escape response and the nutritional state or prey capture activity of hatchling squid. This study explores this possible connection through behavioral observations of squid hatchlings' prey capture behavior, guided by the hypothesis that tight, adult- like control over the giant axon system is necessary for the successful capture of fast moving prey organisms. Squid were raised on both fast- and slow-moving prey in order to address the questions: Can any type of feeding activity stimulate development of effective prey capture technique or does prey type determine the style of attack behavior? Can squid accustomed only to slow-moving prey adapt their capture technique to fast prey? Is there an optimum time period for such modifications to occur? METHODS Culture Conditions Eggs of Loligo opalescens were obtained by divers in April-May 1993 in Monterey Bay. The eggs were held in flow-through (50 micron filtering) natural seawater tanks at ambient temperature (-15 C) in the Monterey Bay Aquarium and Hopkins Marine Station until hatching, after which they were divided into the experimental groups described in this paper. Most eggs were hatched in their own rearing tank to minimize handling damage, but some Group 3 hatchlings were transferred by beaker. The egg masses hatched over a period of 4 days before they were removed, and the major hatch-out occurred on day 3 of this period which is hereafter defined as culture "Day 1". The two main experimental groups (Groups 2 and 3) were kept in black, conical tanks with a capacity of 320 liters (1 meter diameter). Water level was maintained by a Hartford Loop system. The perforated central drainpipe was covered with pleated paper filters (50 micron) to avoid any spots of high velocity of the effluent water leaving the tank. These filters were in turn covered with a nylon stocking. Both stocking and filters were cleaned twice per week by scrubbing and soaking in bleach solution. Detritus and uneaten food were dipped off the water daily with a small net. The tanks were exposed to natural light during the day but were otherwise covered. When feeding or behavioral observations took place in the evening, a 60 watt incandescent bulb was held over the tank. Mortality and Survival Each morning the bottom and sides of all tanks were cleaned using a siphon and the dead and moribund squid recovered were transferred to a separate container to be counted and examined. Daily mortality was calculated as a running percentage by dividing the number of dead squid collected by the number alive on the previous day. The number of live squid on day 1 was determined at the end of the study by summing the daily death tolls. Group 2 began with 1688 squid and Group 3 with 1031 Feeding Zooplankton used for feeding was obtained daily when possible (see below) from Monterey Bay, either by tows or by a near-shore plankton pump. Many different species made up these samples, but the prey most often pursued by the squid was the small, fast-moving copepod Arcatia tonsa. Larger calenoid copepods were sometimes also present. The plankton was sieved and rinsed with fresh seawater before it was fed to the squid. Little attempt was made to regulate or measure the daily amount provided, because availability of samples was sometimes erratic and limited because of weather conditions. Generally a fairly high plankton density was maintained in the tanks, but occasionally one feeding had to last for 2 or even 3 days. Prey items were present at all times, however. During the last week of feeding (days 35-43) large numbers of copepods were available and feeding was at a consistently high density Artemia (brine shrimp nauplii) were cultured in the lab and were thus constantly available. Äfter hatching, they were enriched by a period of feeding on cultured phytoplankton to increase their nutritional value. 20-30 ml of high density Artemia was given in each daily feeding; Artemia were always abundant in the rearing tanks. All squid were fed each morning. Observations were always made less thar two hours later, and the majority occured immediately after feeding. Before evening observations, squid on a plankton diet were given a second feeding to increase prey density. Behavioral Observations of Feeding Attacks As a first gauge of the level of feeding activity, the number of attacks observed per minute during each observation period was recorded. These periods were usually 30 minutes long, varying between 25 and 50 minutes. Squid were sampled by "dividing" the tank into four observational parts and visually scanning a selected volume until an animal was observed orienting towards a prey item or positioning itself for an attack. This individual was then followed for as long as it continued to engage in attack behavior (see below and Fig. 4) in order to record the number and outcome of each attack episode. When the squid captured or abandoned its prey another squid was chosen and the observation process repeated. If no animals in the volume under scrutiny initiated an attack within 2-3 minutes, observations shifted to another section of the tank. Because water in the tank slowly circulated, many of the squid revolved with it, enabling a better picture of group activity Obviously, when very high numbers of attacks or attacks involving multiple squid occurred, the method of observation and recording employed could not keep pace with the activity level. During such high activity periods, attack rates were underestimated. In order to quantitate feeding attack outcomes, a scheme for a typical attack sequence was defined based on detailed original observation in conjunction with previous work on Loligo opalescens (Hurley, 1979) and other species (Messenger 1968; LaRoe, 1970). This behavioral sequence includes both successful and unsuccessful behaviors and is discussed below in conjunction with Fig. 4. All attack outcomes reported in this paper were classified according to this plan. On any given day, the experimental groups of squid were fed and observed consecutively. There was no obvious difference in the rate of feeding after morning versus evening feedings, and both sets of data were pooled into a set for that day. To compensate for unavoidable variation in prey density and duration of observation from day to day, the various outcomes of each group's feeding attacks were calculated as fractions of the total number of attacks observed that day for the group. RESULTS Mortality and Survival With Different Prey Each of the two experimental groups of squid hatchlings were provided with different prey types at different stages after birth and closely observed during feeding behavior. Two questions of interest are whether prey capture behavior varies with prey type and whether previous feeding experience affects a squid's ability to master a new type of prey item. Group 2 squid were fed slow-moving Artemia nauplii from culture days 1-33, and on day 34 were introduced to wild plankton and fed on this until the end of the study (day 45). Group 3 was given Artemia over days 1-9 and switched to plankton on day 10. They fed exclusively on planktonic prey (mostly copepods) thereafter, except on days 31-33 when plankton was unavailable and Artemia was substituted. Figure 1 shows the daily mortality rate for Group 2 hatchlings over the entire experiment, and Fig. 2 illustrates analogous data for Group 3. In both cases there are two main "peaks" of mortality over the 45 day period. This survival pattern has been routinely found in attempts to culture Loligo (Hurley, 1979, Hanlon et al, 1980, 1982). The first peak typically occurs during the first 5-10 days post-hatching (days 8-9 in Groups 2 and 3; see Figs. 1,2) and reflects the hatchlings' difficulty in making the transition from living on internal yolk reserves to capturing live prey. Once this transition is made, the surviving squid display much lower mortality until a period around days 40-60. This second major peak is generally attributed to accumulated skin abrasions, fin damage, bacterial infection and possibly malnutrition. Rate of Attack on Different Prey Types Figure 3 compares the number of attacks per minute that squid in Groups 2 and 3 made on slow- and fast-moving prey introduced at different times in culture. Group 2 was exposed only to Artemia from day 1, but no attacks were observed until day 6 (Fig. 3B) At this time, the attack rate jumped to 1.9 per minute and continued to fluctuate around this value for the next 28 days. As shown in Fig. 3A, this low attack rate persisted for several days after Group 2 was switched to planktonic prey on day 34, but the attack rate then suddenly escalated to 4.5 per minute on day 37 The same pattern was found when Group 3 was switched to plankton on day 10 (Fig. 3A). Such a sharp rise in attack rate presumably reflects the squids' increasing recognition of copepods as desirable prey. Äfter this early peak, the number of attacks made per minute by both Groups 2 and 3 began to decline, accompanied by heavy mortality in both groups. Group 2's attack rate decreased steadily until there were only two survivors at the end of the 10 day period (day 44). Although the Group 3 attack rate also fell drastically until day 6 of plankton exposure, it then increased again to a high level by day 10 of the experiment (Fig. 3A). These squid continued to display high attack rates of 4-5 per minute throughout the next 20 days of observation (data not illustrated), at the end of which 22 squid survived. Definition of the Feeding Attack Sequence It thus appears that exposure to planktonic prey either early (Group 3) or late (Group 2) after hatching stimulates initial interest in the new faster-moving prey. It is important to note, however, that the number of attacks per minute does not indicate a squid's ability to capture prey. To take this disparity into account, attack behavior was classified as indicated in Fig. 4. There are several steps in a successful prey capture sequence, and at each step one or more alternative behaviors result in an unsuccessful attack, i.e. the prey item is not captured. An attack begins as "orientation" when the squid positions itself so that its arms point straight at the prey item. The squid then swims slowly forward, directly towards the prey, maneuvering close enough to attack. At this point, the pursuing squid may actively swim off in another direction, or simply drift away, apparently losing interest in the prey. In the results to be described, these outcomes were defined as "incomplete" attacks. Squid prepared for attack sometimes also effectively abort the attempt by making a rapid backwards escape jet that is apparently stimulated by sudden movement of the prey item. Such escape behavior was commonly seen in squid with the fast-moving planktonic prey, whereas animals fed on Artemia most often abandoned the pursuit. When a squid does make a committed attack, it approaches the prey with arms and tentacles pointed together in front of the body. This leads directly into a final attack jet during which the arms open to seize the prey. Sometimes the attack jet is improperly aimed, and the squid misses the prey entirely. Often a squid contacts the prey during the attack jet but then immediately gives a strong backwards jet before the arms can be closed, a necessary step in subduing the prey Both of these patterns were defined as a "missed" attack. Squid would sometimes successfully execute a well-aimed attack jet, hit the prey, and enclose it in the arms only to then suddenly release it, either swimming away or making a backwards escape jet. These unsuccessful attacks were labeled "releases." An attack was considered successful if the squid subdued the prey in its arms and held onto it while jetting backwards. The squid then proceeded to eat the prey during normal slow swimming or while simply drifting. Only these outcomes were counted as "eaten" attacks. Although the strong backwards jet after prey capture is very common, it is not always executed. As this behavior quickly removes both squid and prey from the capture site, it may have adaptive value as a means of avoiding larger predators that might be attracted to the struggle, or other hatchling squid which commonly engage in fighting and prey¬ stealing (unpublished observations). In summary, feeding attacks involve a complex sequence of behaviors, including backwards and forward jets which require precise control and direction. Inappropriate escape jets (marked by * in Fig. 4) must be suppressed in the early stages of the sequence and forward jets must be properly timed and directed. A central hypothesis of this paper is the idea that both inhibition of the escape response and its critically timed usage are essential for maximum-efficency capture of fast-moving prey. Analysis of Feeding Attack Outcomes Squid provided with fast zooplankton or with slow Artemia nauplii followed the same basic plan of attack as that described above, but failed in different ways to capture prey. Daily percentages of attacks defined as incomplete, missed, released and eaten are presented in Figs. 6-8. Figure 6 shows the outcomes of feeding attacks made by Group 2 on Artemia over the first 21 days in culture. No feeding attempts were observed until day 6. When hatchlings finally initiated attacks, they were capable of successfully progressing through the capture sequence on the first day of feeding. Roughly a third of all attacks made on day 6 were successful, and the success level remained around 20% for the rest of the 21-day observation period. The percentage of missed attacks was similar during this time, but the level of incomplete attacks was quite a bit higher, generally over 40%. Thus, slow-moving Artemia apparently did not often stimulate hatchling squid to a degree that led to complete attacks, but if the squid did attack they had a good rate of success. Figure 7 shows how these patterns changed upon introducing Group 2 to plankton on day 34 in culture. For the first 3 days (34-36), the majority of squid continued to feed on residual Artemia and seemed to ignore the copepods as possible food. Relatively few squid even oriented towards the new type of prey and only 4 copepods were seen to be captured and eaten during these 3 days. By the fourth day of plankton exposure, the Artemia had been depleted, and the frequency of attacks on copepods rose rapidly (See Fig. 3A). Many squid began active pursuit and attack, and the level of incomplete attacks fell to less than 10% during days 37-42. However, the percentage of missed attacks during this period was extremely high (-90%). Very few animals initially made contact with their prey, and by days 40-43 fewer than 5% of attacks ended in successful capture. By day 44 all the Group 2 squid had died, presumably of starvation and exhaustion from continual fruitless pursuit and attack. Figure 8 shows results over the entire 33 day period of Group 3 exposure to planktonic prey after 9 days on Artemia. The feeding pattern immediately after exposure to copepods was similar to that of Group 2. There was a two day period of apparently low interest, characterized by a high percentage of incomplete attacks. On the third day of exposure the frequency of attacks rose sharply (See Fig. 3A). Like Group 2, the Group 3 squid also showed a high level of missed attacks, a reduced frequency of incomplete attacks, and a low percentage of successful captures for the next few days. From day 17 onward, the surviving squid showed a high attack rate coupled with a low success rate, although the Group 3 squid reached higher success ates and survived well up to day 42. During the Group 3 trial, plankton were unavailable for 3 days, and Artemia were substituted. During this period (days 31-33, open bars in Fig. 8), there was a noticeable change in feeding behavior. The frequency of incomplete attacks temporarily increased, the level of missed attacks simultaneously fell, and the success rate was higher. Analysis of Errors in Prey Capture: Inappropriate Escape Jets Although the two groups of squid in this study were all exposed to the same fast-moving prey, their behavior during attacks was significantly different, suggesting that the animals' age and previous feeding experience may affect their ability to capture an unfamiliar fast-moving prey item. When naive squid are provided with plantktonic organisms such as copepods right after birth, they can successfully capture them with rapid attack jets, sometimes even on the first day of exposure (Hurley, 1976; Yang et al, 1983, 1986; unpublished observations). In contrast, the squid in both Groups 2 and 3 that were given Artemia nauplii after birth did not initiate any feeding for several days. When feeding behavior did commence, little rapid movement was involved on the part of either squid or prey The squid moved slowly into position and made a short jabbing attack with open tentacles in an attempt to grasp the prey and pull it into the mouth. Squid seldom made backwards escape jets from Artemia following a failed attack, and a second or third attempt was usually successful. When Group 2 squid were confronted with copepods on day 34, they essentially could not catch these fast, jerky animals. On days 34-36 very few attacks were made and these consisted of several futile attempts (maximum 6). On days 37- 42, after longer exposure to copepods, the squids’ efforts at pursuit increased considerably, and one individual made 25 fruitless attacks on the same copepod. The method of attack in these cases was clearly inappropriate for this type of prey. The squid could not approach the copepods very closely, because any movement by the prey would lead to a backwards escape jet by the squid (as discussed in conjunction with Fig. 4). Very often an attack was initiated from too great a distance, the squid making a short stabbing movement with open tentacles. Although this type of attack is effective with Artemia, it inevitably failed with copepods, because these organisms would respond with their own extremely rapid escape response. In most instances, the squid then reacted to the darting copepod with a rapid backwards escape jet, and the prey capture encounter would end in a missed attack and failure. Filled symbols in Figure 8A display the high frequency of Group 2 missed attacks on copepods that were followed by such inappropriate escape jets (over the first 9 days of their exposure to plankton). Plotted values are the percentage of the total number of attacks observed on each day (numbers indicated in Fig. 8A). Figure 8B illustrates the complementary data for unsuccessful attacks that were no followed by escape jets, that is the squid either made another attack attempt or remained stationary. For Group 2 squid, these values are low (filled symbols). Data plotted as open symbols in Figs. 8A, B reveal that Group 3 squid, which had been previously exposed to plankton for 23 days, made far fewer attacks followed by escape jets during the period of days 35-43 in culture and followed through more often with multiple attacks on the same copepod (maximum 60). Although the day 35-43 performance of the Group 3 squid was better than that of Group 2 in suppressing inappropriate escape jets, their behavior after their initial exposure to copepods on day 9 was much like that of Group 2. Attack-related escape jets commonly occurred at this time, both in response to movement of the prey and following attack jets whether or not contact with the prey was made (data not illustrated). Although careful analysis of Group 3 failures was not made during this period, it was obvious that significant improvement had occurred by day 34 when the 10 day comparison with Group 2 was initiated. DISCUSSION Numerous rearing experiments on Loligo opalescens (Hanlon, 1979; Hurley 1976; Yang et al., 1983, 1986; personal unpublished observations) have demonstrated the squids' ability to capture copepods immediately after hatching. Pursuit and attack seem to be triggered by the copepods' fast, jerky movements, which make them an attractive prey item compared to other planktonic organisms. In contrast, when the hatchlings in this study were exposed only to Artemia at birth, they did not respond for several days. On the day they began feeding, they too were capable of successful captures, but their feeding attacks were lethargic and 10 squid frequently abandoned their prey altogether when it did not move. Hanlon (1986) reports similar results: his group "curtailed the feeding of brine shrimp since these were found to be unattractive to the hatchlings." Nevertheless, if an attack on Artemia was completed, it usually succeeded, and survival on this prey was good (see Fig. 1). Group 3 squid, introduced to copepods on day 10, made very few successful captures on the first days of feeding despite energetic pursuit. As exposure continued, they modified their attack behavior noticeably, making longer, well¬ timed attack jets and increasingly suppressing premature escape responses. The proportion of attacks that ended in prey consumption was still much lower on plankton than on Artemia. However, although mortality for this group was higher than that for Artemia-raised squid (compare Figs. 1 and 2), the daily death rate also stabilized after an early peak. More squid were obviously capturing enough food to survive, using an attack strategy very different from the method by which they had caught Artemia. Group 2 was not exposed to copepods until day 34 of the study. When active attack began, the success rate was extremely low for two reasons. First, escape responses to movements by the copepods hindered many squids' approach to the prey. Second, most squid within striking distance of a copepod would attack as if it were Artemia. Attack jets were typically too short, too slow and incorrectly timed, often cut short by a backward jet. As Group 2 did not alter this behavior sufficiently to catch enough food and avoid starvation, this study suggests that Loligo's ability to capture fast-moving prey depends on its exposure to this prey early in the life cycle After the change in prey on day 34, the entire group died so quickly (see Fig. 1) that can not attribute the majority of deaths to factors such as skin abrasion or infection. It is reasonable to hypothesize that young squid with slow-moving prey do not need to use their giant axon system in an adult-like manner during prey capture. Animals pursuing fast-moving copepods would presumably benefit greatly from the ability to critically time giant axon pulses to boost attack jets initiated by small motor neurons, in the normal adult pattern (Otis and Gilly, 1990). Close control over giant axon excitation would also enable the squid to delay powerful backwards jets until it had securely captured the prey. However, a typical attack on Artemia does not require precise timing of fast jets or suppression of premature escape responses to the prey. Establishment and development of adult-like giant axon control may require the active stimulation of vigorous prey pursuit. Both published and unpublished work by Gilly et al. support this hypothesis (Gilly, Hopkins and Mackie, 1991; Gilly, 1992-3, unpublished). Gilly et al.'s neurological and behavioral experiments on restrained Loligo and Sepioteuthis hatchlings found that many Sepioteuthis, which had been actively feeding and growing post-hatching, displayed the adult-like pattern of giant axon use in delayed escape response behavior. Loligo, unfed during these experiments, demonstrated fast-start escapes typical of embryonic animals. Although unknown developmental differences between the two species could underlie this disparity, the different feeding experiences of the two groups might have been responsible for Sepioteuthis fürther progression. More recent work by Gilly on Loligo hatchlings found that a majority of fed individuals gave adult-like delayed escape responses to direct electrical stimulation whereas most unfed hatchlings reacted in the embryonic manner (Gilly, 1993, unpublished). In this study, a large group of healthy animals were maintained in approximately "unfed" conditions for 33 days: that is, they were given no food that required active pursuit and capture. The squid presumably had no need to utilize an adult-like pattern of maximum-efficiency giant axon recruitment in order to eat and survive. They gave no behavioral indication of having adopted such a patterr either during exposure to slow-moving prey or after fast-moving copepods were introduced on day 34. Many animals exhibited fast, powerful escape behavior during encounters with the copepods. These could be the fast-start, giant axon-driven escape jets characteristic of embryonic Loligo or young, unfed hatchlings (see Gilly et al., 1991). When copepods were first substituted for Artemia, the numerous attacks that Groups 2 and Group 3 made on the new prey may have represented attempts to refine the timing of powerful forward jets and suppress premature backward escape jets while feeding. The discovery that animals 10 days old had some difficulty perfecting this behavior, and that 33-day old squid failed entirely, suggests that development of adult-like giant axon control requires active stimulation during a very short critical period of time. It is unlikely that attacks among the squid themselves contribute to the establishment of sophisticated giant axon control. Even under crowded rearing conditions, very few fights were observed between squid compared with the total number of feeding attacks made. Active feeding experience appears to be the crucial means of early neurological stimulation. To promote effective development of the adult nervous system without sacrificing the high survival rates achieved by feeding Artemia, ideal rearing conditions might include a diet of both fast and slow prey from birth. Copepods would stimulate the development of tight control over giant axon boosts during feeding attacks, while Artemia would provide more easily accessible calories. The nauplii could be gradually replaced with larger fast-moving prey such as mysids, to stimulate further growth. The results of this study allow speculation about squid development in the wild. Development of adult-like control of prey capture and possibly escape behavior seems dependent on early exposure to high density, fast-moving prey. Without this experience, squid might suffer behavioral consequences such as loss of ability to capture valuable fast-moving prey, failure to recognize such prey, or impaired ability to escape from predators. If animals in a controlled environment have difficulty adapting to changing prey conditions during the 6-7 weeks post hatching, wild squid must be very vulnerable to fluctuations in marine ecosystems. Experience during the first weeks of life would determine the viability of their nervous systems and behavioral repertoire for adulthood. Acknowledgments I would like to thank Dr. Bill Gilly for advice, patience, criticism and the squid house. I am very grateful to Gil van Dykhuizen and the Husbandry Staff of the Monterey Bay Aquarium for allowing me to participate in the squid culture project, making this study possible. The generosity of Sam Wang and the Stuart Thompson lab with their time and equipment was much appreciated, and I am also grateful for Freya Sommer and John Lee's assistance with construction of the pseudo-Kreisel. Finally, many thanks to Sarah Gilman for invaluable help with data processing and figure preparation. 13 LITERATURE CITED Gilly, W.F., B. Hopkins and G.O. Mackie. 1991. Development of giant motor axons and neural control of escape responses in squid embryos and hatchlings. Biol. Bull. 180: 209-220. Hanlon, R.T. 1979. Rearing experiments on the California market squid Loligo opalescens Berry 1911. The Veliger 21 (4): 428-431. Hurley, A.C. 1976. Feeding behavior, food consumption, growth and respiration of the squid, Loligo opalescens raised in the laboratory. Fishery Bulletin 74 (1): 176- 182. LaRoe, E.T. 1970. The rearing and maintenance of squid in confinement with observations on their behavior in the laboratory. Ph.D. Thesis, University of Miami, Coral Gables, Fla., 136pp. Messenger, J.B. 1968. The visual attack of the cuttlefish, Sepia officinalis. Anim. Behav. 16: 342-357. Otis, T. S., and W. F. Gilly. 1990. Jet propelled escape in the squid Loligo opalescens: concerted control by giant and non-giant motor pathways. Proc. Natl. Acad. Sci. USA. 87: 2711-2915. Packard, A. 1969. Jet propulsion and the giant fibre response of Loligo. Nature 221: 875-877. Yang, W.T., R.T Hanlon, M.E. Krejci, R.F. Hixon, and W.H. Hulet. 1983. Laboratory rearing of Loligo opalescens, market squid of California. Aquaculture 31: 77-88. Yang, W.T., R.F. Hixon, P.E. Turk, M.E. Krejci, W.H. Hulet, and R.T. Hanlon. 1986. Growth, behavior and sexual maturation of the market squid, Loligo opalescens,cultured through the life cycle. Fishery Bulletin 84 (4): 771-798. Young, J.Z. 1938. The functioning of the giant nerve fibres of the squid. J. Exp. Biol. 15: 170-185. Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 FIGURE LEGENDS Mortality for Group 2 squid was high over days 5-13, fell and remained constant until day 32, and increased dramatically after day 40. Slow prey was given on days 1-33 (shaded bars); fast prey was introduced day 34 (marked with arrow) and fed until day 43 (dark bars). Mortality is plotted as a daily percentage. Mortality for group 3 peaked between days 7-12 and days 15-16. Mortalities were then low until a third rise on days 42-44. Slow prey was given from days 1-10 (shaded bars) and fast prey from days 10-44 (dark bars). An arrow marks the introduction of fast prey. Comparison of attacks made per minute on fast and slow prey, introduced to squid of different ages. Figure 3A compares the number of attacks made by Groups 2 and 3 during their first 10 days of exposure to fast prey. Both groups made increasing numbers of attacks until days 3-4 of exposure. Group 2's attack rate then declined permanently (circle symbols) but Group 3's rose steadily after reaching a minimum on day 6 (square symbols). Figure 3B shows that Group 2 made no attacks on days 1-5 of exposure to slow-moving prey, and that the attack rate fluctuated around a low level thereafter. Stages of a typical feeding attack. Successful steps leading to prey capture are shown at left. Alternative behaviors which would abort the attack are described on the right. Stars indicate the points at which backwards escape jets may be made. Group 2 made high percentages of successful attacks on slow-moving prey from days 6-21; percentages of incomplete attacks were even higher. The percentages of missed and "release" attacks were low. 15 Figure 6 Figure 7 Figure 8 Group 2 made extremely low percentages of successful attacks on fast prey (given from day 34-day 43) and very high percentages of "miss" attacks. Äfter day 36, few attacks were incomplete. Group 3 made low percentages of successful attacks on fast prey (given from day 10) and high percentages of missed attacks. Proportions of incomplete attacks were generally low, except immediately following the heavy mortality which occurred on day 15 (marked with arrow). When slow prey was given on days 31-33 (indicated by white bars), percentage consumption was higher and the proportion of missed attacks lower. More attacks were incomplete, however. 8 A: A high proportion of Group 2's unsuccessful attacks on copepods were followed by inappropriate escape jets. Over the same period of time, a much lower proportion of Group 3's unsuccessful attacks were followed by escape jets (Group 3 had been previously exposed to copepods for 23 days). The total number of attempts observed on each day is shown below the data point. 8 B: Complementary data for the frequency of unsuccessful attacks not followed by an escape jet. DAILY MORTALTTES FOR GROUP (SLOW PREY TILL DAY 33, FAST PREY GIVEN ON DAY 34) DAILY MORTALITIES (%) FOR GROUP 2 100 - 90 - 80 70 40 30 - 20 10 — AA o DAY FIGURE 1 100 90 80 70 + 60 50 40 30 20 10 A DALY MORTALTTES FOR GROUP 3 (SLOW PREY TIL DAY 9,FAST PREY FROM DAY 10) DAILY MORTALITIES (%) FOR GROUP 3 LE aaa- DA FIGURE 2 FIGURE 3 COMPARISON OF ATTACKS MADE PER MINUTE ON DIFFERENT PREY TYPES, FED TO SQUID OF DIFFERENT A A. GROUP 2 GROUP 3 Days 10-19 m Days 34-43 O — 1 2 3 4 5 6 7 8 9 10 DAY OF EXPOSURE TO FAST PREY B. GROUP 2 o0-0-O0 oowooc DAY OF EXPOSURE TO SLOW PREY FIGURE 4 FEEDING ATTACK SEQUENCE SUCCESSEUL UNSUCCESSEUL ORIENT TOWARDS PREY FOLLOW — LOSE INTEREST, MOVE AWAY — ESCAPE JET — "INCOMPLETE" FORWARD JET ATTACK WITH OPEN ARMS — MISS PREY CONTACT PREY "MISS" — BACKWARDS JET CLOSE ARMS ON PREY SUBDUE — RELEASE PREY— RELEASE" BACKWARDS JET WITH PREY RETAIN AND EAT PREY—EAT FEEDING ATTACK OUTCOMESFOR GROUP2 GIVEN SLOW PREY DAY 1-33, NOEEEDNGSEENTILDAYE % INCOMPLETE % RELEASE 100 100 90 90 80 1 80 70 4 70 60 60 50 50 40 40 30 30 20 20 10 10- MILI — kakoaa- - " 0 DAY DAY % MISS % EAT 100 100 90 90 80 80 70 70 60 § 60 50 50 40 40 § 30 30 20 20 10- 10 unn DAY DAY FIGURE 5 100 90 80 70 60 50 40 30 20 10 100 90 80 70 60 50 40 30 20 10 FEEDING ATTACK OUTCOMES: GBOUP 2 SQUID GIVEN FAST PREY DAY 34 (AFTER SLOW PREY SINCE DAY 1) % INCOMPLETE % RELEASE 100 90 80 70 60 550 40 8 30 20 10 LHAHaE c onoo aa aa- DAY DAY %MISS % EAT 1001 90 80 70 60 50 40 § 30 20 10 E oo aaaaaaava- DAY DAY FIGURE 6 100 90 - 80 70 560 50 30 20 10 100 90 80 70 60 50 40 30 20 EDING ATTACK OUTCOMES GROUP 3: FAST PREY DAY 10-44 BUT SLOW PREY DAYS 31-33 HEAVYMORIAIHYDAY 15 % INCOMPLETE % RELEASE 100 - 90 80 70 60 50 40 30 20 10 IEMI HLL LE IRL a - A DAY DAY % MISS % EAT 100 90 80 70 60 50 40 30 20 T LER ELLL I DAY DAY FIGURE 7 FIG. 84 Frequency of "missed" attacks followed by escape jets 5 189 it 262 0.8 - 0.7 13 125 27 g0.6 — Group 2 0.5 Group 3 + 0.3 0.2 0.1 tavaa- ktatataa DAY Frequency of "missed" attacks not followed by escape jets 09 0.8 ss 0222089 2 0.7 —03 991 9131 F 0.6 —• Group 2 isa 4g L 0.5 Group 3 5 0.4 0.3 0.2 0.1 35 36 37 38 39 40 41 42 43 DAY FIG. 88 FIGURE 8