ABSTRACT Free-swimming Sepioteuthis lessoniana hatchlings were videotaped during both feeding and strobe flash sessions. Analysis showed that while feeding the peak instantaneous velocity obtained by the hatchling is twice as great and the peak acceleration is 50% greater than during an escape response to a strobe flash in which only the giant axon is used. Behavioral measurements of mantle radius made with video were used to compute jet thrust and compare performance with varying patterns of motor neuron activity in the restrained hatchling. The coordinated, adultlike pattern of giant and small fiber use resulted in mantle contractions and jet thrusts that were always greater than contractions and thrusts produced by giant-only or small-only responses. Thus, evidence strongly suggests that hatchling Sepioteuthis use the giant fiber in feeding attacks. INTRODUCTION The squid neural system is comprised of two motor pathways, one slow and one fast, that can interact in complex ways. The fast pathway originates in the brain from two first order giant cells which make contact with seven second order giant neuons. One second-order axon projects to the stellate ganglion and synapses there with the third order giant axons that innervate mantle muscles associated with rapid escape responses. The slow pathway runs basically in parallel, starting from the median basal lobe of the brain, connecting with the pallio-visceral lobe, and then continuing on to the stellate ganglion via the pallial nerve. From the stellate ganglion, the slow motor axons emerge, with the third-order giant axons, in the stellar nerves and synapse on circular mantle muscles that are slow to fatigue (Mackie 1990; Otis and Gilly 1990). A major advantage of such a complicated system lies in the the ability to use the two pathways separately or in concert. The giant axons, with much larger diameters (280-718 um in Loligo forbesi ) than the small axons (40-50 um), also have greater conduction velocities (15.0-22.3 m/s vs 3.1-5.6 m/s) (Pumphrey and Young 1938). The two pathways can work together in a coordinated manner, and two electrophysiological escape response patterns have been identified, one for adults and one for embryos. The adult, or delayed-type, pattern shows multiple cycles of small motor activity at long latency (300-500 ms) with the giant spike(s) present, if occurring at all, in the middle of bursts of small motor axon spikes (Otis and Gilly, 1990). The embryonic, or fast-start, pattern characteristically shows a single giant spike preceding any small motor activity in the same cycle, and occurs at short latency (-50 ms) (Gilly et al. 1991). The embryonic pattern is retained by the hatchling, and within a few weeks, the adult pattern develops. Äfter the change to the adult pattern, the fast-start pattern of giant axon firing is seen only following sudden visual stimulation (Otis and Gilly, 1990). The factors responsible for the progressive development from one system to the other are currently unknown. However, one function that may require the use of the adultlike motor patterns is prey capturing. Sepioteuthis hatchlings develop the ability to feed on live prey as early as 4 days posthatching (Hanlon, 1990), and precise timing of a strike would be advantageous in capturing a rapidly moving prey item, such as a copepod or larval fish. Thus, it is of interest to compare in more detail the time course of evolution of the adult-like useage of the giant fiber system with the development of feeding behavior. Additionally, it is possible that the attacks made on live prey may positively reinforce the adult pattern of giant fiber utilization or even be necessary for the changeover. As a first step in exploring these ideas, it must be ascertained whether or not the giant axons are used in feeding by hatchling squid. MATERIALS AND METHODS Animals Late-stage Sepioteuthis lessoniana eggs were obtained from the Marine Biomedical Institute, University of Texas Medical Branch, Galveston, and kept in circular plastic tubs (28 cm in diameter, 15 cm deep) in an aquatron (100 gallon tank with 50 gallon reserve) with gently circulating seawater (22-23 °0). The tubs had two Nytex (0.4 x 0.4 mm pores) panels in the sides and one in the bottom to keep water circulating in the tubs. Each morning, hatchlings born overnight were put in one group and transferred by beaker to either another plastic tub or the open aquatron, so that animals of different ages were separated from each other. Groups of animals were arbitrarily chosen to be "fed" or "unfed". Fed animals were kept in the open portion of the aquatron and were given many small mysid or Gambusia fry immediately after transfer and then at intervals of approximately twelve hours; unfed animals were maintained in the plastic tubs. Experiments Free-swimming escape responses to strobe flashes and feeding sessions of "fed" hatchlings were videotaped from the side of the tank using a Canon Al videocamera set at 1/250 shutter speed. Strobe flashes were made with a Nikon Speedlight SB-15 camera flash. Data were analyzed with a Megavision image analysis system that can store 28 real-time frames (33 ms per frame) and replay them in digitized form. For electrophysiological recordings, an adapted version of a procedure developed by Ötis and Gilly (1990) was used. Hatchlings were first anaesthetized in 0.5% urethane, and the dorsal side of the mantle was glued, with cyanoacrylate cement, to a 30-gauge needle fitted with Tygon tubing. The animals were suspended in a small aquarium with seawater and gently bubbling Ö2. The skin and muscle overlying the left stellate ganglion were surgically removed, and a polyethylene suction electrode was placed on the interface between the stellate ganglion and the second-hindmost stellar nerve. Electric stimulation was applied by two separate electrodes, one to the pallial nerve and one to the head region of the hatchling. Behavioral data were recorded by vidoetaping the profile of the restrained animal with a Canon A1 or LI Hi-8 Camcorder on 1/250 exposure, and data were analyzed by the Megavision. Electrical data were either digitized and stored on a videotape or were stored on the audio track of the behavioral videotape. Calculations In the videotapes of the free-swimming animal, the Megavision system was used to collect the x,y coordinates of the eye in sequential frames. The calculated distance that the squid's eye moved in sequential frames was used to measure the instantaneous velocity. Acceleration is the derivative of the velocity: dy/dt. Experiments were normalized for comparison by dividing all values of velocity and acceleration by the body length of the squid, which was the average measured distance from the squid's eye to its tail in the frames used for data collection. Videotapes of the restrained animal were used to measure the change in radius (r) of the mantle over sequential frames. Volume (Vol) is proportional to r2 for a cylindrical or conical object. The thrust of the water jet produced by the mantle contraction can be calculated based on equations provided by Ö'Dor, Foy, and Helm (1986), and is proportional to (dVol/dt)2. RESULTS Instantanoeus velocity and acceleration during feeding and free¬ swimming escape responses to strobe flashes Strobe-driven escape responses in free-swimming Sepioteuthis hatchlings are very stereotyped, like the case in adult or hatchling Loligo (Otis and Gilly, 1990; Gilly, et al, 1991). In an experiment with an 11 day-old animal that had been actively feeding, the maximum instantaneous velocity reliably occurred between the between the second and third frame (33 ms per frame) after the strobe flash and had a value ranging from 0.45-0.56 body lengths/frame and corresponding accelerations of 0.37-0.44 body legths/ framee. Examples of such responses are indicated by the open symbols in Figs. 1A, B, respectively. Similar responses were also obtained in younger squid (3-8 days; data not illustrated). During an attack on a small fish, the same squid displayed a maximum velocity of 1.1 body lengths/frame and a peak acceleration of 0.57 body lengths/frame2. Thus, the feeding attack of a hatchling squid can be substantially more rapid than an escape resoponse to a strobe flash. Based on the published work on Loligo (Otis and Gilly, 1990; Gilly et al, 1991) and on the data on Sepioteuthis described below, it is assumed that these escape responses are largely, or entirely, diven by a single giant axon spike and that the small motor axon system plays a negligible role. The significantly greater values of both peak velocity and acceleration during the feeding episode suggest that the giant axon may also be utilized during this non¬ escape behavior. In order to pursue this idea, the relationship between motor activity and jetting performance was investigated in restrained hatchlings. Electrophysiological recordings in conjunction with behavior in the restrained hatchling Electrophysiological recordings in restrained animals have shown that escape responses to strobe flashes give a distinctive motor discharge in the stellar nerves in both adult, hatchling, and late-stage embryonic Loligo opalescens (Otis and Gilly, 1990; Gilly et al, 1991). In "simple" escape responses of this type, the giant axon fires first at brief latency (25-50 ms) and may be followed by small unit activity of unidentified origin. Firing of the giant is always accompanied by a sudden jetting reaction. When the giant does not fire in response to the strobe stimulus, no mantle response is detectable with video observations, even though some small motor activity may occur. Sepioteuthis lessoniana hatchlings displayed similar behavior in the present study. Just supra-threshold electrical stimulation of the pallial nerve in a restrained 11 day-old squid (same animal as in behavioral experiment of Fig. 1) results in the fairly selective firing of the giant fiber with a delay of approximately 1 ms, as shown in Fig. 2. A strobe flash stimulus (Fig. 3) also triggers a giant axon spike but at a longer delay (20 ms). The recordings in both Figs. 2 and 3 also show some small unit activity of uncertain origin following the giant spike. If strobe stimulation fails to trigger a giant spike, there is no observable mantle response. In the case of pallial nerve stimulation, proper positioning of the stimulating electrode can yield a similar pattern, but often weak mantle responses (determined by simple visual observation) can be elicited without activation of the giant axons. In either case, pallial nerve stimulation was a reliable means of providing a "template" for the waveform and amplitude of the giant spike to aid in its identification during more complex motor discharges and in obtaining an indication of the mantle response due to preferential activation of the giant axon. Recordings of mantle dynamics and the underlying motor patterns in the stellar nerve were also obtained from a 4 day-old restrained hatchling that had not been fed (Fig. 4). A 5 V shock to the pallial nerve was delivered, and this resulted in a single, abrupt jet response. Firing of the giant axon within one ms after the shock (Fig. 4A) leads to the onset of mantle contraction within one frame (Fig. 4B), and the mantle contraction reaches peak response (the smallest diameter) by 2-3 frames. The rapid onset of this mantle response and its temporal relationship to the giant spike are thus very similar to the results discussed above in conjunction with strobe flash stimuli. Pallial nerve stimulation appears to produce a mantle reponse that is longer lasting than that due to a strobe stimulation (compare Figs. 4B and 1A). However, this may be due to secondary small unit activity that accurs with a sizeable delay only after the nerve shock. Such late activity is apparent in Fig. 2A. Shocks applied to the surface of the head of the same 4 day- old squid give behavioral and neural responses that are more complex than either pallial nerve shock or strobe stimulation. A single 20 V shock to the mouth during frame zero results in the behavioral data shown in Fig. 5A, and the corresponding stellar nerve activity is shown in Fig. 5B. Expanded traces for each of the 3 jetting cycles are shown in Fig. 50. The short latency for the first cycle of neural activity (30 ms) is characteristic of hatchling and embryonic squid (Gilly et al, 1991) and is not seen with adult animals (Otis and Gilly, 1990). From examination of records like those in Fig. 5, it appears that the most powerful jetting responses are those that involve the coordinated ("adult-like") use of the giant and non-giant motor pathways. Thus, contractions associated with small-unit activity alone (Fig 5, cycles 1 and 2) are not as powerful as cycle 3, although they are definitely stronger than the contraction due to a normal breathing cycle (prior to frame zero in Fig. 5A). Finally, a contraction due to the firing of a short-latency, embryonic-type giant spike acting more or less alone (e.g., due to the pallial nerve stimulation) is at best equal to the contraction due to small motor axon activity alone. This can be seen by comparing behavioal data from Fig. 4B with cycle 1 in Fig. 5 (filled symbols in Fig. 5A). Thus, a giant axon spike does not necessarily lead to a marked increase in jetting performance over that provided by the non-giant system. The boosting effect of the giant axon only comes about when its activation is properly timed in relation to the burst of non-giant activity. Differences in the level of behavioral performances achieved in jetting cycles invoking coordinated discharge of giant and non¬ giant motor units and in those cycles generated by small activity alone can also be more subtle than indicated by the data in Fig. 5. Fig. 6 shows the time course of the mantle radius and the corresponding stellar nerve recording of the same 4 day-old unfed squid (as in Fig. 5) in response to a 30 V shock to the mouth. Again, the motor discharge displays a combination of embryonic and adult patterns: An embryonic, giant-type occurs at short latency before the burst of small unit activity in the first cycle; small activity only is present in the second cycle; the adultlike-coordinated giant/non-giant discharge exists in cycles 3 and 4. In Fig. 6A, the corresponding measurements of the mantle radius do not show marked differences in speed or degree of contraction. However, when the corresponding thrust of the water jet (calculated as (dVol/dt)2, where volume is proportional to the mantle radiuse (O'Dor, Foy, and Helm, 1986)) produced by each mantle contraction is compared in Fig. 7, there is a notable correlation between the type of motor activity and the strength of the jet thrust for each cycle: the thrust produced by the embryonic giant (cycle 1) is on the same order as, but slightly greater than, the thrust from the small motor activity alone (cycle 2); adult-like cycles 3 and 4, where the giants fall in the middle of small unit firing, are associated with jet thrusts considerably larger than those of cycles 1 and 2. DISCUSSION The higher values of instantaneous velocity and acceleration in the feeding hatchlings compared with the giant-only escape response produced by a strobe flash seem to indicate that, in order to produce the extra burst of speed, the giant and the small motor systems must be working together to produce an added behavioral effect. In this experiment, behavioral and electrophysiological work on restrained hatchlings shows that the degree of mantle contraction and jet thrust produced are related to the motor patterns of escape response elicited by the hatchling. In considering mantle radii, all escape responses produce stronger contractions than those made during a normal breathing cycle. Small motor activity can produce a contraction as strong, if not stronger than the giant acting alone (overlay, Figure 5A). Small motor activity also gives weaker contractions than when both the small and the giant systems are being used in conjunction (Figures 5A and 6A). While Figure 6A shows that the contraction caused by small activity only is almost equivalent to other cycles in which the giants are involved, it is worth noting that the lowest point in cycle 2 is indicated by a single point, whereas the other cycles have multiple points defining the full contraction. The difference in the position of this low point and the points clustered at a higher level in this contraction could easily be due to noise in sampling measurements. While embryonic and adultlike giant spikes show almost no difference in the degree of mantle contraction, the effect of each response pattern is made apparent when jet thrust is calculated. The jet thrust resulting from the embryonic giant is only about half of the thrust produced by a single adultlike giant spike. Thus, the presence of a giant acting with small motor activity does not necessarily imply that an added behavioral effect will be produced. Instead, the position of the giant spike seems to be the factor that determines if the effects from the giant and the small systems will add or not: a giant fired at long latency, in the middle of a burst of small motor activity will give a summed effect, while a short latency giant that preceeds small motor activity produces behavior that is similar to small motor activity acting alone. Also, the summed effect of the adultlike pattern of giant fiber use consistently produces mantle contractions that are larger than that of the giant alone or the small fiber alone. This would seem to indicate that, in order to attain the high velocity and acceleration seen in feeding, the hatchlings must be using the giant in an adultlike manner. The effect of feeding vs. not feeding on the degree of mantle contraction and jet thrust is yet to be determined. Both fed and unfed squid showed combinations of the adult and embryonic patterns at the age of 4 days. If animals were in the "fed" category, the adultlike response would predominate by the age of 8 days; if "unfed", the hatchling exhibited only embryonic behavior by as early as 6 days old, though they did not survive or were not kept alive beyond 7 days. This might indicate that adultlike-use of the giant fiber must be exercised through an activity such as feeding, otherwise the ability to use the adult pattern is lost once the hatchling is past 4-5 days old. This also supports the idea that the giant fiber is used in an adultlike manner for feeding. Future work may focus on obtaining more video data on hatchling feeding. Due to the difficulty in videotaping the feeding attacks of the Sepioteuthis hatchlings, only a small number of sequences were of use for analysis. The feeding attack/capture sequence of an 11 day-old Sepioteuthis hatchling seemed to be representative of the feeding attacks made by other 7-8 day-old hatchlings, but this is not certain. At the present time, electrical recordings from free-swimming hatchlings are not possible. Therefore, in order to assess whether the adultlike giant is actually being used in feeding, the behavioral data from the restrained animals must be converted to values of free-swimming acceleration and velocity via a mathematical model. Once it is known whether or not the giant fiber is used in an adultlike manner during feeding, it may be determined whether the the development of prey-catching ability is actually responsible for the establishment of the adult pattern of escape response. ACKNOWLEDGEMENTS Many thanks go to Professor Gilly, whose patience, advice, wisdom, and humor were much appreciated; without his help this project would never have been accomplished. I am also indebted to the following people: Professor Baxter for providing the video equipment, Roger Hanlon and the University of Texas for supplying the squid eggs, Bruce Hopkins for advice and help with setting up equipment for experiments, Professor Denny for his infinite knowledge of energetics, and Professor Thompson, director of Megavision Studios. I would also like to thank the 1991 Biology 175H dudes for their support and friendship. LITERATURE CITED Gilly, W. F Hopkins, Bruce, and Mackie, G. O. 1991. Development of giant motor axons and neural control of escape responses in squid embryos and hatchlings. Bio. Bull. 180: 209-220. Hanlon, Roger T. 1990. Maintenance, Rearing, and Culture of Teuthoid and Sepioid Squids. Squid as Experimental Animals. Plenum Press, New York. 35-62. Mackie, G.O. 1988. Giant axons and control of jetting in the squid Loligo and the jellyfish Aglantha . Can. J. Zool. 68: 799-805. O'Dor, R. K., Foy, E. A., and Helm, P.L. 1986. The locomotion and energetics of hatchling squid, Illex illecebrosus. Amer. Malac. Bull. 4: 55-60. Otis, Thomas S. and Gilly, W. F. 1990. Jet-propelled escape in the squid Loligo opalescens: concerted control by giant and non¬ giant motor axon pathways. Proc. Natl. Acad. Sci. USA. 87: 2911-2915. Pumphrey, R. J. and Young, J. Z. 1938. The rates of conduction of nerve fibres of various diameters in cephalopods. J. Exp. Biol. 84: 303-318. FIGURE LEGEND Figure 1A: Instantaneous velocity of an 11 day-old hatchling during a feeding attack and during an escape to a strobe flash Figure 1B: Acceleration of the feeding attack and escape to the strobe Figure 2A: Electrical recording of an 11 day old, "fed" hatchling in response to a 5 V pallial nerve shock Figure 2B: Same as 2A on an expanded timescale Figure 3: Electrical recording of the restrained, 11 day-old, "fed' hatchling escape to a strobe flash Figure 4A: Oscilliscope trace of a 5V shock to the pallial nerve of a 4 day-old unfed hatchling Figure 4B: Corresponding mantle radius, before, during, and after the shock (frame zero) described in Figure 44 Figure 5A: Mantle radius of 4 day, "unfed" hatchling before, during and after a 20 V shock to the mouth at frame zero Figure 5B: Electrical recording of the hatchling's reaction to the shock Figure 50: Cycles 1, 2, and 3 from figure 5B shown on an expanded timescale Figure 6A: Mantle radius and electrical recording of same 4 day, "unfed" hatchling before, during, and after a 30 V shock to the mouth at frame zero. This shock was given approximately 1/2 hour after the shock corresponding with Fig. 5 Figure 6B: Expanded timescale of cycles 1-4 in Figure 6A Figure 7: Calculated thrust of the water jet corresponding with Fig. 6A,B INSTANTANEOUS VELOCITY DURING FEEDING AND ESCAPE 121 0 08 —0— Strobe 06 — — Feeding 0.4 0.2 og 0.0 -10 Frame Figure 1A ACCELERATION DURING FEEDING AND ESCAPE 0.6 0.4 0.2 00 —0— Strobe — —— Feeding -0.2 -0.4 -0.6 + Frame Figure 1B 50 — Pallial Stimulation 10 ms Figure 2A 4 ms Figure 28 Light Flash 10 ma Figure 3 60 - 50 40 Figure 4A Radius of hatchling during pallial nerve stimulation 20 40 60 80 -20 0 Frame Figure 48 500 RADIUS OF HATCHLING DURING ESCAPE JETTING 60 - 50 - Figure 5A 40 -30 -20 -100 10 20 30 40 50 60 70 Frame Figure 5B 3 30 ms Figure 5c 2 effrp THRUST OF WATER JET 8o0000 600000 0000 ooo00 RR R -20 60 80 Frame Figure 7 RADIUS OF HATCHLING DURING ESCAPE JETTING 60 50 40 30 + — -20 40 20 60 80 Frame Figure 6A — o 3 30 ms 2 ygukpe 4 —r Figure 6B