UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The fast-start swimming of angelfish, Pterophyllum eimekei Domenici, Paolo 1993

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1993_spring_phd_domenici_paolo.pdf [ 4.39MB ]
JSON: 831-1.0086375.json
JSON-LD: 831-1.0086375-ld.json
RDF/XML (Pretty): 831-1.0086375-rdf.xml
RDF/JSON: 831-1.0086375-rdf.json
Turtle: 831-1.0086375-turtle.txt
N-Triples: 831-1.0086375-rdf-ntriples.txt
Original Record: 831-1.0086375-source.json
Full Text

Full Text

THE FAST-START SWIMMING OF ANGELFISH, Pterophyllum eimekeibyPAOLO DOMENICILaurea, University di Milano, 1988A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIESDepartment of ZoologyWe accept this thesis as conformingto the rquired standardTHE UNIVERSITY OF BRITISH COLUMBIAJanuary, 1993© Paolo Domenici, 1993In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)Department of  2-0 ( m 77The University of British ColumbiaVancouver, CanadaDate  4/0"/ / 20^r3DE-6 (2/88)ABSTRACTThe^kinematics^of turning^manoeuvres^and^the^distance-timeperformance in escape responses of startled angelfish (Pterophyllumeimekei) are investigated employing high speed cinematography (400 Hz).All escape responses observed are C-type fast-starts, in which the fishassumes a C shape at the end of the initial body contraction (stage 1).Subsequent kinematics (stage 2) allows for classification of theresponse into two types; Single Bend (SB), in which the tail does notrecoil completely after the formation of the C, and Double Bend (DB),in which it does.The two types of response result in different total escape angles(measured considering subsequent positions of the center of mass, SB0^0120.0 ; DB 73.3: p<0.005), different stage 2 turning angles (in the0same direction as stage 1 for SB, 11.0 ; in the direction opposite to0stage 1 for DB, -21.9: p<0.0005), and different maximum angularvelocity in the direction opposite to the initial one (SB -8.08 rad/s;DB -56.62 rad/s: p<0.001). There are no significant differences instage 1 kinematics. Stage 1 turning angle is linearly correlated tostage 2 turning angle for DB only (p<0.01; 12=0.60), and to totalescape angle for both types of response (p<0.0001; 1 2 =0.80). Stage 1duration is linearly correlated to stage 1 turning angle (p<0.0001;r2=0.83) and to total escape angle (p<0.0001; r2=0.72) for both typesof escape.Distance-time performance is also different between the two responseiitypes, mainly due to differences in stage 2 (maximum velocity, SB 0.99m/s; maximum velocity DB, 1.53 m/s: maximum acceleration SB, 34.1 m/s2 ;maximum acceleration DB, 74.7 m/s2 : p<0.0001 in both cases). As aresult, significant differences in the performance throughout the wholeresponse (maximum velocity 1.02 m/s and 1.53 m/s for SB and DBrespectively; maximum acceleration 63.2 m/s2 and 91.9 m/s2 for SB andDB respectively) as well as within a fixed time (0.03 s) are found.Overall, higher distance-time performance associated with smallerangles of turn are found in DB when compared to SB.Comparison with previous studies reveals that angelfish have goodfast-start performance despite specializations for low speed swimmingin the labriform (pectoral fins) mode. In addition, angelfish turningradii (0.065 ± 0.0063L: Mean ± 2 s.e.) are lower than those previouslyreported for any fish.The swimming trajectories of angelfish escape responses tomechanical stimulus are determined. Although fish escape trajectoriesare linearly related to stimulus direction, they vary considerablyafter the initial turn away from the stimulus. Past studies of escapetrajectories in fish and other animals have been analysed by employinglinear plots of stimulus angle versus body turning angle. Here, Idefine escape trajectories as a circular variable, with 0 ° as stimulusdirection.Angelfish escape in non-random trajectories when the stimulus ispresented laterally, within an angular zone of approximately 30 ° -120 °(discriminating zone). The circular plot of escape trajectories shows abimodal pattern that cannot be revealed by linear analysis. Angelfishiiiescape preferentially at 180° and 130° away from the stimulus,maximizing the distance covered from the stimulus and escaping at thelimit of their discriminating zone, respectively. P. eimekei correcttheir responses when turning towards the stimulus, . suggesting thatescape trajectories are modulated by sensory feedback.Reanalysis of published work on other animals by employing circularhistograms of escape trajectories, reveals multimodal patterns whichare also not apparent from the linear plots. I suggest that thepresence of multiple preferred trajectories may be adaptive inpreventing predators from learning any single fixed pattern of responseand compensating for it.The effect of size on angelfish escape responses is alsoinvestigated. Size has an effect on the type of response performed.Small angelfish employ DB responses for all escape angles, whereaslarger fish commonly show SB responses. This is due to a decrease inmaximum DB escape angles for larger fish. Escape angles for pooled SBand DB responses are size-independent. Distance covered within a giventime, maximum velocity and maximum acceleration for pooled SB and DBresponses are size-independent. With the exception of the smallestfish, angelfish show sub-maximal performance when making large turns. Isuggest that angelfish employ a particular type of response dependingon their predators' strike angle. Extremely small and large fish useprincipally DB and SB responses, respectively. The employment of eitherDB or SB responses in small and large fish, respectively, may berestricted to a narrow range of predator strike angles.ivTABLE OF CONTENTSAbstract^ iiTable of Contents^ vList of Tables viiiList of Figures^ ixAcknowledgements xiChapter 1: Introduction^ 1Chapter 2: General Materials and MethodsA. Studies Species^ 5B. Fish Maintenance 6C. Eliciting the Escape Responses^ 7D. Filming and Film Analysis 8Chapter 3: The Kinematics and Performance of Angelfish (P.eimekei)escape responsesA. Introduction^ 10B. Materials and Methods^ 121. Morphometrics 122. Analysis^ 123. Sources of error in maximum^18acceleration analysisC. Results^ 181. Turning kinematics^ 182. Distance-time performance 303. Correlations between turning^33kinematics and performanceVD. Discussion^ 381. Turning kinematics^ 382. Distance derived performance^423. C-start types^ 464. General conclusions 47Chapter 4: Escape Trajectories in AngelfishA. Introduction^ 52B. Materials and Methods^ 53C. Results^ 641. Effect of fish initial orientation^642. Escape trajectory frequency 67distributions3. Steering and angular compensation^79D. Discussion^ 821. Effect of initial orientation^822. Frequency distribution of 84escape trajectories3. Steering and angular compensation^864. Escape trajectories in other animals^88Chapter 5: The Effect of Size on Angelfish Escape ResponsesA. Introduction^ 94B. Materials and Methods^ 95C. Results^ 981. Fast-start types^ 982. Turning parameters 103vi3. Distance-time parameters^ 103D. Discussion^ 1131. The occurrence of single bend^113and double bend responses2. Turning kinematics^ 1203. Distance-time performance^122Chapter 6: Summary^ 127References^ 130vi iLIST OF TABLESTable 3.1 Mean morphometric characteristics^ 14Table 3.2 Turning angles, angular velocityand turning radius for Single Bend 22and Double Bend responsesTable 3.3 Results of distance derived parameters^ 32Table 3.4 Summary of previous fast-start studies 45Table 4.1 Turning angles, angular steering and^ 63angular compensationTable 5.1 Summary of morphometric characteristics 97Table 5.2 Initial orientation, stage 1, stage 2^ 102and total escape angles for singleand double bend responses of differentsize groupsTable 5.3 Stage 1 duration and distance time^ 109parameters up to the end of stage 2Table 5.4 Distance travelled and maximum 112velocity attained within 0.045 sfrom startviiiLIST OF FIGURESFigure 3.1 Turning radius of an escaping fish^ 17Figure 3.2 Tracing of a Single Bend and 20a Double Bend C-startFigure 3.3 Angular velocity, velocity and^25acceleration of Single Bend andDouble Bend C-startsFigure 3.4 The relationship of s 1 turning angle^ 27with s2 turning angle and withtotal escape angleFigure 3.5 The relationship between reverse^ 29angular velocity and s2 turning angleFigure 3.6 The relationship of s 1 duration 35with s 1 turning angle and total escape angleFigure 3.7 The relationship of Maximum total^ 37velocity with total escape angleFigure 3.8 Range of s 1 turning angles and total escape^51anglesFigure 4.1 Experimental set up and time delay in 56stimulus signal as perceived by two hydrophonespositioned along the axis of stimulus directionFigure 4.2 Diagram illustrating angular variables^ 59Figure 4.3 The sign of steering and angular 61compensationFigure 4.4 Angular deviation of s 1 and escape^ 66trajectories for differentorientation sectorsFigure 4.5 Percentage of away responses for^ 69different orientation sectorsFigure 4.6 Circular frequency distribution of 71sl trajectories and escape trajectoriesFigure 4.7 Fitting a von Mises distribution^ 74to frequency polygons of s 1 andescape trajectoriesixFigure 4.8 Percentage of responses within^ 76each preferred escape trajectoryfor different orientation sectorsFigure 4.9 Linear analysis of angelfish escape angles^78Figure 4.10 Relationship between angular steering 81and s 1 trajectories in away responsesFigure 4.11 Circular frequency distribution of^ 91escape trajectoriesFigure 5.1 Percentage of double bend responses for 100the different size groupsFigure 5.2 The relationship between double bend maximum^105total escape angle and fish sizeFigure 5.3 The range of single bend and double bend^107total escape angles for differentsize groupsFigure 5.4 The range of single bend and double bend^115distance travelled within 0.045 s fordifferent size groupsFigure 5.5 The range of single bend and double bend^117maximum velocity within 0.045 s fordifferent size groupsxACKNOWLEDGEMENTSThis study could not have been completed without the advice andsupport of many people. First, I would like to thank my family forproviding unconditional moral and personal financial support. I wouldlike to thank my supervisor Bob Blake, for all the things that he hastaught me and all the support that he has given me throughout theseyears. I also thank the members of my committee J. Gosline, J.Matsubara, D. Jones and D. Randall, for their support and advice. Ithank M. Smith, R. Frith. D. Harper, H. de la Cueva, M. Kshatriya, B.Kolotylo from the Animal Locomotion Laboratory at UBC for their helpand suggestions. I thank A. Blachford from the Biosciences Data Centerfor his assistance with computer analysis. I also thank E. Denton, D.Ludwig, P. Matthews, M. Hodgson, L. Barrett-Lennard and S. Wainwrightfor valuable discussions. Many friends helped me and gave me their warmsupport: K. Duff, C. Galindo, G. Zuleta, M. Kasapi, D. Bouras and F.Kroon.This thesis is dedicated to my parents, my brothers and my sister.xiCHAPTER 1IntroductionSwimming functions in fish may be divided into three maincategories, cruising, manoeuvring, accelerating (Webb, 1984a). Cruisingby means of body/caudal fin propulsion is utilized in activitiessustained from seconds to several weeks (e.g. in migrations).Manoeuvring by means of median or paired fin propulsion is typical oflow speed, fine manoeuvre activities. Accelerating by means ofbody/caudal fin propulsion is employed in fast-starts, such as thoseperformed by striking predators or escaping preys.Escape reactions involving fast-start responses allow fish to avoidsudden actual or potential danger in their environment. Fish lacking afast-start response possess other structural or behaviouralantipredator adaptations such as toxins, spines and burrowing habits.Escape responses in fish have been studied from biomechanical (e.g.Weihs, 1973; Webb, 1976, 1978; Harper and Blake, 1990; physiological(e.g. Eaton et al., 1981, 1988, 1991; Faber et al., 1989; Covell etal., 1991) and behavioural (e.g. Dill, 1974a,b; Hurley and Hartline,1974; Blaxter et al., 1981) perspectives. Adaptations for effectivefast-start performance are thought to include a large proportion ofwhite muscle (fast-starts are fueled anaerobically) relative to red,large caudal fin and body depth for producing thrust, and high body1flexibility (Weihs, 1973; Blake, 1983; Webb, 1984a).The ability of fish to escape predators may depend upon linearperformance (velocity, acceleration) (Webb, 1976; Harper and Blake,1990), accurate timing (Eaton and Hackett, 1984), and turningcapabilities (Howland, 1974; Webb, 1982; Nissanov and Eaton, 1989).The escape response has been shown to be similar in kinematics fordifferent species of fish (Eaton et al. 1977; Webb, 1978). It has beendescribed as a fixed action pattern consisting of a strong unilateralcontraction of the body musculature which bends the fish in a C shape(stage 1), followed by a strong propulsive stroke of the tail in thedirection opposite to the initial contraction (stage 2) (Gillette,1987). The result is an extremely rapid acceleration of the animal. Athird kinematic stage can be present, in which the fish continues toswim or coasts (Weihs, 1973; Webb & Blake, 1985). Previous studies havefocused on the first two kinematic stages (see Webb, 1976, 1978; Eaton& Hackett, 1984, for reviews).The mechanisms mediating stage 1 are relatively well understood(DiDomenico et al. 1988; Nissanov & Eaton, 1989). The initialcontraction during stage 1 of a C fast-start is usually initiated by asingle, pair of prominent neurons (Mauthner cells), although alternativecircuits may exist (Eaton et al. 1984; DiDomenico et al. 1988). Aparallel^network^of neurons interacting^with^the^Mauthner cellscontrols the extent of stage 1 contraction (DiDomenico et al. 1988).Nothing is known about the mechanisms triggering stage 2.Fish escape responses were thought to be stereotypic, characterizedby fixed angles of turn and maximum acceleration performance (Webb,21986b). Daniel and Mayhtifer (1989) showed a negative relationshipbetween rotation and translation in carridean shrimp escaperesponses. In Chapter 3, the range of turning angles for angelfishescape responses is investigated and the relationship betweenacceleration performance and turning angle is assested.Weihs (1973) hypothesized that body morphology would be a keydeterminant of maximum acceleration in fish. Previous studies havefocused on fusiform fish (Dubois et al. 1976; Webb, 1976, 1978, 1986a;Eaton et al. 1977; Harper & Blake, 1990). Chapter 3 addresses thequestion of how a disc-shaped fish specialized for paired fin swimmingat low speed (Blake, 1979) performs in fast-starts. Specialization forpaired fin swimming has been hypothesized to impair fast-startperformance (Webb, 1984a). However, since axial locomotion and pairedfin locomotion are decoupled systems, adaptation for one may notcompromise the other.Escape paths tend to be away from the stimulus, since the axialmusculature contralateral to the stimulus contracts during stage 1,forming the C-bend (Blaxter et al. 1981; Eaton et al. 1981). It hasbeen suggested that turning angles are preprogrammed at the onset ofthe escape response in fish (Nissanov and Eaton, 1989; Eaton et al.1991) and other animals (Camhi and Tom, 1978; Camhi and Levy, 1988).Recently, Eaton and Emberley (1991) showed that escape trajectories arelinearly related to stimulus direction. However, they vary considerablyafter the initial turn away from the stimulus (Eaton et al. 1981;Eaton and Emberley, 1991).The analysis of escape trajectories in fish (Eaton et al., 1981;3Eaton and Emberley, 1991) and other animals (Camhi and Tom, 1978; Comerand Dowd, 1987; Nalbach, 1990) has been performed by employing linearplots of stimulus angle versus body turn (defined as the angle betweenthe animal's initial orientation and its orientation at the end of theresponse). In Chapter 4, escape trajectory is considered as a circularvariable, and is defined as the angle between the stimulus directionand the animal's swimming path at the end of the response. The mainresult is that the circular distribution of escape trajectories ofangelfish appears to be bimodal, not unimodal as linear analysissuggests.The effect of size on the angelfish escape response is investigatedin Chapter 5. Size is known to have an effect on distance covered andmaximum speed attained by fish during fast-starts (Wardle, 1975; Webb,1976). This is because the duration of fast-starts increases with size,and acceleration is size-independent (Webb, 1976). When consideredwithin a fixed time, however, both distance covered and maximumvelocity are size-independent (Webb, 1976). In Chapter 5, the effect ofsize on both linear performance and turning kinematics is investigated.The linear performance of angelfish was shown to be submaximal at largeturning angles (Chapter 3). I hypothesized that fixed escapetrajectories may compromise the linear performance of angelfish(Chapter 4). Chapter 5 investigates the effect of size on the trade-offbetween turning angle and swimming performance in angelfish escaperesponses.4CHAPTER 2General Materials and methodsA. STUDY SPECIESPterophyllum eimekei (synonymous to P. scalare, Schultz, 1967) hasbeen chosen as an extreme example of specialization for low speedswimming. Its fineness ratio (Length/Depth) is amongst the lowest ofall fishes (Aleev, 1969). Its fins are very specialized. The triangulardorsal and ventralpectoral fins arefacilitating independence ofcomplicated waveform to bevery large and directed backward. Thetruncated planar form, their broad basemotion of individual fin rays, allowing apassed down the pectoral fins (Blake, 1981).fins areof aThe pelvic fins are narrow and elongated, probably specialized astactile receptors (Aleev, 1969).Because of its disc-shaped body, P.eimekei can be considered achaetodontiform (sensu Webb, 1984a) specialized for low speed swimmingand high manoeuvrability. These characteristics are required in complexenvironments (Webb, 1984a) such as coral reefs or weedy rivers whereangelfish live. However, it escapes predators by performing fast-startresponses. No study has yet investigated its fast-start swimming,although it has been observed (Bergmann, 1968; Blake, 1979). Its paired5fin swimming kinematics has been fully investigated by Blake (1983, forreview).Although commonly kept in aquaria, P.eimekei has been little studiedin its natural habitat (Burton and Burton, 1975). It belongs to thefamily Cichlidae and inhabits the Amazon Basin (Sterba, 1966). It hasbeen speculated that it may have few enemies due to its camouflageamong weeds and the somewhat cloudy waters it inhabits (Burton andBurton, 1975). To my knowledge, there are no literature reports on thepredators of angelfish. P.eimekei, like most cichlids, possesseshighly developed parental behaviour (Bergmann, 1968). However, thejuvenile stage lasts for a very short period, the young assuming theadult stage when only 36 day old and approximately 2 cm long (Burtonand Burton, 1975). During parental care, male P.eimekei displayaggressive sounds (Myrberg et al. 1965) and fighting postures(Bergmann, 1968) in reaction to conspecific intruders. During ontogenicdevelopment, different types of social organization were found:schooling, territoriality, hierarchy-formation (Bergmann, 1968).B. FISH MAINTENANCEAngelfish^(Pterophyllum eimekei)^were^obtained^from^a local(Vancouver, B.C.) dealer and held in a glass tank (75x40x30 cm)supplied with aerated, dechlorinated water and equipped with arecirculating filter. The water temperature was maintained between 24and 26 C. The fish were fed pelleted food.6C. ELICITING THE ESCAPE RESPONSESSingle fish were transferred to the experimental glass tank (60cm X32cm X 30cm), placed on four rubber supports (diameter 5cm, height2cm), one at each corner, and situated in the middle of a largerPlexiglas tank (240cm x 120cm x 45cm) surrounded by a black plasticscreen. The fish could see neither the approaching stimulus nor theinvestigator. Fish were left in the experimental tank for at least halfan hour prior to being startled. A mirror angled at 45 ° over the tankallowed the top view of the fish to be filmed. A plastic containerfilled with 1 liter of water was suspended by a 1.7 m string at a fixedheight of 1.5 m over the edge of the tank. This container was lifted 1m away from the tank and swung against the side of the external tank toelicit the escape response.A 2.5 cm reference grid was placed on the bottom of the experimentaltank. In all the escape responses analysed, the center of mass of anygiven fish was within a 7.5cm x 7.5cm square in the center of theexperimental tank (60cm x 32cm x 30cm) and at least 1 body length awayfrom the walls. Therefore, stages 1 and 2 of the escape responses wereunobstructed by walls since angelfish escapes cover about half a bodylength (Chapter 3).Eaton and Emberley (1991) showed that the proximity of walls caninfluence the escape trajectories of goldfish when less than 0.6-0.77body length away from the fish's center of mass. Eaton and Emberley(1991) suggested that escape responses in which the initial position ofthe fish was approximately 8.8 cm (0.7-0.9 body length) from the sideof the tank were influenced by wall effects, when compared to escapesin which the initial position averaged 13.2 cm (1.0-1.3 body length)from the nearest wall. In our experiments, fish were startled whenbetween 12.25 cm and 16 cm from the nearest wall, which corresponded toa minimum of 1 body length (for the largest fish) and maximum of about3 body lengths (for the smallest fish). Although to be certain of theabsence of any wall effects one should ideally operate in open water,we have only analysed trials in which wall effects should be minimal.D. FILMING AND FILM ANALYSISThe experimental tank was illuminated by two 650 W photographiclights and escaping fish were filmed with a high speed cine camera(Locam model 51-0002) on Kodak 7277 4X 400 ASA cine film at 400 Hz.Processed films were projected on a white panel (55 x 80 cm), allowingthe image to be magnified, in order to minimize measurement error(Harper & Blake, 1989). The position of the "stretched straight" fishcenter of mass for the film analysis was determined by aligning a wire,marked at 0.37 body length (Chapter 3), along the midline of the imageof the fish. The center of mass, tip of the head and tail were recordedframe by frame. These points were later digitized on a digitizing pad8(GTCO type, 24x36 inches) connected to a computer (80286 ATcompatible). Data were then transferred to an Olivetti M24 PC forfurther analysis.9CHAPTER 3The kinematics and performance of angelfish^(P. eimekei) escaperesponsesA. INTRODUCTIONWebb (1984a) classified fish swimming styles into three broadcategories: body-caudal fin (BCF) periodic (cruising) propulsion; BCFtransient (fast-start) propulsi on; and median and paired fin (MPF)propulsion, used in slow an d precise manoeuvres. Webb (1984a) arguesthat "specialist fish" whi ch excel in one of these categories,sacrifice performance in the others. Alternatively, generalist fishperform moderately well in all functions but have superior performancein none. According to Webb (1984a, b), morphological and physiologicalcharacteristics optimizing one particular swimming function "trade-off"with the others.Webb's idea seems to hold for cruising and accelerating, when fishpropel themselves utilizing one locomotory system (Body/caudal finlocomotion). Adaptations for accelerating clearly contrast those forcruising. High acceleration performance requires large tail and bodydepth, enhancing thrust, body flexibility allowing large amplitudepropulsive movements, and high percentage of anaerobic musculature, topower burst swimming. On the other hand, cruising adaptations includelunate tail, minimizing drag, stiff body, minimizing recoil and1 0therefore drag, and high percentage of aerobic musculature forendurance. However, trade-offs are less obvious in other cases.Accelerating and manoeuvring imply axial locomotion and paired finlocomotion respectively. These are decoupled systems, and adaptationfor one may not necessarily impair the other. The idea ofdecoupled systems is valid from a morphological and kinematicpoint of view. However, it may be not adequate at aphysiological and biochemical level, particularly when asequence of escape responses is considered. Such a high levelof activity would cause lactate accumulation in the anaerobicmusculature. This lactate may be taken up and oxidized by pectoral finaerobic muscles. It was shown that vertebrate non-working muscles mayoxidize significant amounts of lactate produced by heavy exercise ofdifferent muscles (Lindinger et al. 1990).Manoeuvre "specialists" such as angelfish propel themselves at lowspeed utilizing their pectoral fins. High manoeuvrability is due tolateral insertion of pectoral fins, extended anal and dorsal fin anddeep, laterally flattened body (Webb, 1984a). Despite relatively lowpercentage of muscle mass expected because of their body shape,angelfish can be expected to perform well in escape responses,possessing two of the three main adaptations for fast-start locomotion.Angelfish have a very deep body, with a fineness ratio (L/D, where L istotal length and D is maximum body depth) amongst the lowest of allfishes (Aleev, 1969). Lateral compression of their body should allowhigh flexibility (Aleev, 1969).Although the kinematics and performance of fish in escape responses11has received much attention in the past two decades (Webb, 1976, 1978;DuBois et al. 1976; Taylor and McPhail, 1986; Harper and Blake, 1990;Frith and Blake, 1991; Webb et al. 1991; Gamperl et al. 1991; Kasapi etal. 1993), the fast-start capabilities of disc-shaped fish are notknown, since most studies have focused on fusiform fish. Here, Iinvestigate how a disc-shaped fish specialized for paired-fin swimmingat low speed (Blake, 1979) performs in fast-starts.B. MATERIALS AND METHODS1. MorphometricsFive fish (total length, L = 7.26^0.4 cm; Mean ± 2 s.e.) were usedin the experiment. Morphological parameters are listed in Table 3.1.The position of the center of mass (CM) was determined by hanging thedead fish on one point along its body profile. The procedure wasrepeated for two different points; the crossing point of the twostraight lines descending from the hanging points indicated CM. Totalwetted surface area (body and median fins) was determined byoverlapping a piece of plastic sheet on both sides of the fish andcomparing its weight to a standard of known area. Percentage of musclemass was determined by dissection of freshly killed fish (Scale Mettlermodel PK 300).2. AnalysisTwenty sequences were analysed. Processed films were projected on awhite panel (55 cm X 80 cm), allowing the image to be magnified fivetimes, to minimize measurement error (Harper and Blake, 1989).^Thevelocity and the acceleration data were derived from^the raw12TABLE 3.1: Mean morphometric characteristics. Errors are given inbrackets as ±2 s.e.13Total^ Muscle^Center of mass^WettedLength(L)^Mass(M)^Mass(mM) Distance from nose(CM) S.area(Sw)•(cm)^(g) (g) (cm) (cm)7.26(0.4)8.550.0223L 3(0.86)3.330.389M(0.45)2.690.37L(0.2)53.781.02L 2(4.9)514distance-time data employing a five-point smoothing regression(Lanczos, 1956). Stage 1 (sl) and stage 2 (s2) durations weredetermined based on the change in the direction of displacement of thehead. Two types of response were observed: Single Bend (SB) in whichthe body did not completely straighten after the initial C bend, andDouble Bend (DB) in which a full return flip of the tail was observedafter the initial contraction. The end of stage 2 was not observablefor SB responses; in this case, s2 parameters were calculated up to 8frames after the end of s 1.Stage 1 turning angle for the anterior part of the body wasdetermined by measuring the angle between the straight lines passingfrom the center of mass and the tip of the head at frame 0 and at theend of stage 1. Stage 2 turning angle was measured employing the samecriterion, considering the angle occurring between the end of sl andthe end of s2.A total escape angle of the CM path about the end of s2 was alsodetermined. This angle was measured between the fish initialorientation and the regression line considering seven positions of thecenter of mass about the end of s2 (s2 ± 3 frames).The angular velocity of the center of mass was determined, employinga 5 point smoothing regression (Lanczos, 1956) of the originalcumulative angle data. The total turning radius (T.R.) for each escapewas calculated employing the mean instantaneous distance moved (d) andthe mean instantaneous angle of turn (y) of the center of massthroughout stage 1 (Fig.3.1). The turning radius is given byT.R. = d/(2 cos (n-y)12).15FIGURE 3.1: Turning radius (T.R.) of an escaping fish. Numbers indicatepositions of the center of mass.^A straight  line passing from position1 and 2 is drawn in.^The dotted line indicates the path of the centerof mass. (d) indicates distance between subsequent positions of thecenter of mass.16173. Sources of error in maximum acceleration dataFilm-derived acceleration data are subject to sampling frequencyerror (S.F.E., the error due to over-smoothing at low frame rates) andmeasurement error (M.E., the error involved in measuring the distancemoved) (Harper & Blake, 1989). Harper & Blake (1989) conclude thatsubcutaneous microaccelerometry should be used to attain the mostaccurate measurement of maximum acceleration. The fish employed in thisstudy were too small (7.26±0.4 cm) to implant microaccelerometers.Here, filming rate of 400 Hz and 5X magnification were employed.According to Harper & Blake (1989), the resulting S.F.E. is about 8%.The highest measurement error is below 8%, estimated for the lowestvalue of maximum acceleration (s2 SB 10.9 m/s2 ), and is estimated to beless than 5% for all the other maximum acceleration values recorded.C. RESULTSAll fast-starts analysed were C-type. Nevertheless, fast-starts canbe classified into two main types. Single Bend (Fig. 3.2A), in whichthe tail does not recoil completely after the formation of the C, andDouble Bend (Fig. 3.2B) showing a clear full return flip during stage2. The two escape types differ in both turning kinematic anddistance-time performance.1. Turning KinematicsTurning kinematic parameters are compared between the two types ofresponse in Table 3.2. There is no significant difference between the18FIGURE 3.2: Tracing of a Single Bend (A) and a Double Bend (B) C-start.The midline of the fish and the center of mass are shown. Numbersindicate time (in msec) and can be matched with figures 3.3A, 3.3B, 3.3C.19SINGLE BEND^DOUBLE BENDA^ BTABLE 3.2: Turning angles, angular velocity and turning radius forSingle Bend and Double Bend responses. T-test is used for allcomparisons except * (Mann-Whitney). Errors are given in brackets as ±2s.e.21MaximumnegativeStage 1^Stage 2^angular^TurningTurning Turning^Escape^Velocity^Radius/angle^angle angle^(radians/^Length^N(degrees)^(degrees)^(degrees)^second)SB 109.8 11.0 120.0 -8.08 0.067 9(16.7) (7.9) (17.5) (7.75) (0.0075)DB 84.6 -21.9 .73.3 -56.62 0.063 11(16.9) (10.7) (22.1) (17.98) (0.0098)Test n.s. P<0.0005 P<0.005 P<0.001* n.spooled 95.9 -7.1 94.3 -36.42 0.065 20(13.0) (10.1) (17.7) J15.60) (0.0063)220for^SB is^positive (11.0)same direction as duringangle for^DB is negativeanglein theturnings 1 turning angles of the two fast-start types. Stage 2 turning anglesare statistically different (p<0.0005). The mean value of s2 turningindicating that the turn is continueds 1. However, the mean value of s20(-21.9 ), meaning that the directionof turn is opposite to that in stage 1. This results in the two typesshowing a significantly different total escape angle (p<0.005), with a0^0^ 0mean value larger than 90 for SB (120 ) and smaller for DB (73.3 ).Likewise, the trajectory of the center of mass changes direction ofturn during the response in DB. This is reflected in the angularvelocity profiles (Fig. 3.3A). These show a clear change of directionduring stage 2 in DB, and a gradual decline towards zero in SB. Themaximum negative angular velocity for SB (-8.08 rad/s) is statisticallydifferent from DB (-56.62 rad/s).The mean value of the turning radius relative to body length, (L),is not statistically different between SB (0.067 L) and DB (0.063 L).The pooled mean value is 0.065 ± 0.0063 L (Mean ± 2 s.e.). Stage 1turning angle is linearly correlated with s2 turning angle only for DBresponses (p<0.01; Fig. 3.4A) and sl turning angles are linearlycorrelated with the total escape angle for both types. Since the twoslopes and elevations are not statistically different, the data havebeen pooled (Fig.3.4B, p<0.0001). Si turning angle and total escapeangle are correlated linearly with maximum angular velocity in thedirection opposite to the initial one for DB only (p<0.005 and p<0.0005respectively), whereas the correlation between s2 turning angle andangular velocity holds for both types, although with different slopes23FIGURE 3.3: Angular velocity (A), velocity (B), and acceleration (C) ofthe center of mass in Single Bend (dotted line) and Double Bend(continuous line) C-starts. Time at completion of stage 1 (s 1) andstage 2 (s2) is indicated.24■■A\■N.\N. ,...- - - SB..\DB......^..... %.... .„).BI// . . - „,.. - - - . - - -.._/////C/Ai \ /A-.4.4r\1 AII/\/1 1 \ i.../I 1 N.,1 11 g%)... I"C1 400L••■•••300L)La-1200(710000.4E030 --30 -1.3 -1.00.71^I II0^10^20^30^40^50TIME (msec)I t tsl DB sl SB s2 DB^s 2 SB25FIGURE 3.4: The relationship of sl turning angle with s2 turning angle(A) and with total escape angle (B) for Single Bend (open diamonds) andDouble Bend (closed circles) responses.^The linear regression for A(y=0.49x-63; r2=0.6; p<0.01) is valid for Double Bend only.^In B thedata are pooled to give the regression line (y=1.27x-28; r2=0.87;p<0.0001). Dotted line represents line of identity.26O20AO O OO020Cr/Li/LLJCCL.L11JJ-JCDCCCC -4082060000w1J-1CCCDL.1.11-4.1-JCDCC1110-CC1.1.1-JCCO4-FIGURE 3.5: The relationship between reverse angular velocity and s2turning angle for Single Bend (open diamonds) and Double Bend (closedcircles) responses. Linear regressions for Single Bend (y=0.80x+18;r2=0.64;^p<0.01)^and^for^Double^Bend^(y=0.56x+12;^r2=0.92;p<0.0001) are shown.2820• 441b-..^ ,•7O..^. .-120^-80^-40 0MAXIMUM NEGATIUE ANGULAR VELOCITY (rad/s)29(Fig .3.5 ; SB p<0.01; DB p<0.0001).The relative turning radius (T.R./L) is not related to any of theperformance parameters measured.2. Distance -rime performanceFigures 3.3B and 3.3C show velocity and acceleration profiles of anSB and a DB escape. The tracings of these two fast-starts are shown infigures 3.2A and 3.2B respectively. The Double Bend fast-start showstwo acceleration peaks of similar magnitude (one each stage), whereasthe Single Bend fast-start shows high acceleration during sl, followedby lower acceleration during s2 (Fig. 3.3C). Therefore, after the endof s 1 , velocity is maintained but not increased in SB responses (Fig.3.3B).Comparison of the distance-time parameters between escape types ismade considering the stages separately (Table 3.3). Stage 1 differssignificantly for duration (SB 0.023 s; DB 0.017 s; p<0.005), distancecovered (SB 0.015 m; DB 0.011 m; p<0.05), mean acceleration (SB 28.6m/s2 ; DB 47.0 m/s2 ; p<0.0005) and maximum acceleration (SB 62.8 m/s 2 ;DB 89.1 m/s2 ; p<0.05). Mean and maximum velocity are not significantlydifferent during s 1.Stage 2 parameters are compared in Table 3.3. Since no clear end ofs2 could be established for SB escapes, their s2 duration wasconsidered a fixed value (20 ms; 8 frames). This is reasonable since,for DB, s2 is on average about 80% of s 1 duration; this corresponds to7-8 frames for SB. Except for distance covered, all the otherdistance-time parameters are statistically different (p<0.0001 in all30TABLE 3.3: Results of distance derived parameters during stage 1, stage2, the total response, and during the first 0.03 s of the response.T-test is used for all comparisons. ** Stage 2 duration is assumed 0.02s for Single Bend escapes. Errors are given in brackets as ±2 s.e.31STAGE 1averageduration distance velocity(s)^(m)^(m s^)maximum^average^maximumvelocity^acceler!tion acceleration(ra s^)^ (m s^) (m s^)NSB 0.023 0.015 0.65 0.93 28.6 62.8 9(0.002) (0.002) (0.05) (0.08) (3.7) (11.1)DB 0.017 0.011 0.64 0.98 47.0 89.1 11(0.002) (0.002) (0.06) (0.10) (6.3) (13.9)test P<0.005 P<0.05 n.s n.s. P<0.0005 P<0.05pooled 0.020 0.013 0.64 0.96 38.7 77.3 20(0.002) (0.002) (0.04) (0.07) (5.6) (10.7)STAGE 2SB 0.020** 0.019 0.92 0.99 -1.8 34.1 9(0.002) (0.09) (0.08) (5.6) (12.7)DB 0.014 0.018 1.36 1.53 37.9 74.7 11(0.002) (0.002) (0.12) (0.15) (10.0) (7.7)test n.s. P<0.0001 P<0.0001 P<0.0001 P<0.0001pooled 0.017 0.018 1.16 1.29 20.0 56.4 20(0.001) (0.001) (0.13) (0.15) (10.8) (11.6)TOTAL (81+32)SB 0.043** 0.034 0.77(10.8L)^1.02(14.4L)^14.9 63.2 9(0.002) (0.003) (0.05) (0.09) (3.9) (11.2)DB 0.031 0.029 0.93(12.6L)^1.53(20.7L)^43.2 91.9 11(0.003) (0.004) (0.05) (0.15) (7.3) (12.8)test n.s. 2<0.0005 P<0.0001 2<0.0001 P<0.005pooled 0.036 0.031 0.86(11.7L)^1.30(17.8L)^30.4 79.0 20(0.003) (0.003) (0.05) (0.15) (7.8) (10.7)!FIXED TIME (0.03 a)SB 0.022 0.71 1.01 24.0 63.0 9(0.001) (0.04) (0.08) (5.2) (11.2)DB 0.028 0.92 1.52 43.1 91.9 11(0.002) (0.06) (0.15) (6.4) (12.8)test 2<0.0001 2<0.0001 P<0.0001 2<0.0005 2<0.005pooled 0.025 0.83 1.29 34.5 78.9 20(0.002) (0.06) (0.14) (6.0) (10.7)32cases) during s2.Performance values for s 1 and s2 taken together are summarized inTable 3.3. Again, except for distance covered, all distance-timeparameters are statistically different. Overall the two types ofresponse show similar performance during stage 1, followed by largedifferences in stage 2.In addition, when considering a fixed time interval (30 ms), all thedistance-time parameters for the two fast-start types are statisticallydifferent (Table 3.3).3. Correlations between turning kinematics and performanceStage 1 turning angle and the total escape angle are linearlycorrelated with the duration of s 1 for both types of response. The dataare pooled in Figs 3.6A and 3.6B respectively (p<0.0001 in both cases),since the slopes and elevations of the regression lines are notstatistically different for the two escape types.No other distance-time parameter measured showed any relation toturning parameters within each type of response. However, when turningkinematic data are plotted against distance-time data, the two types ofresponse occupy different regions of a graph. An example is given inFig.3.7, where total maximum velocity is plotted against total escapeangle. Single Bend and Double Bend responses occupy the upper left andthe lower right of the graph respectively.33FIGURE 3.6: The relationship between sl duration and sl turning angle(A; y=6.0x-22; r2=0.83; p<0.0001) and total escape angle (B; y=7.7x-56;r2=0.72; p<0.0001). The regression lines are for Single Bend (opendiamonds) and Double Bend (closed circles) pooled data.34208040Cf)wwCCCDwCDCCCDCCF-02008400w1-1-1CCwwCDCCwa-CCwCCCDFIGURE 3.7: The relationship between Maximum total velocity and totalescape angle. Mean ± 2 s.e. for Single Bend (open diamonds) and DoubleBend (closed circles) responses are shown.364)---.1MAXIMUM UELOC I TY^(M S1)D.DISCUSSION1. Turning KinematicsEscape responses in fish are described as consisting of two stages:an initial body bend, in which the fish assumes a C shape, and asubsequent return flip of the tail (Webb, 1976; Eaton & Hackett, 1984).Here, these stages are observed as described in previous studies(Weihs, 1973; Webb, 1976; Eaton & Hackett, 1984) in DB responses. After0the initial turn during s 1 (mean angle 84.6 ), the head turns in the0opposite^direction^(mean^angle^-21.9)^as^a^result^of^thecontralateral body bend. Therefore, the center of mass undergoes areversal of direction of motion (Fig.3.3A). The relationship between s2turning angles and the magnitude of the reversal of angular velocity isshown in Fig. 3.5.In SB escapes the first stage consists of a C bend in which the fish0head turned at a mean angle of 109.8 , not statistically different fromDB. Subsequently, the fish straightens, but without bending in the0opposite direction. S2 angle is positive on average (+ 11.0 ). As aresult, the angular velocity profile of the center of mass does notshow the abrupt decrease typical of DB, but slowly decreases towardszero as the center of mass goes along the tangent of the spiraltrajectory (Fig.3.3A). The mean total escape angles of the two types of^0^0response are different (SB 120.0 ; DB 73.3 ; p<0.005) and they are38respectively larger and smaller than their corresponding sl turning0^0angles (SB 109.8 ; DB 84.6 ). This is because DB s2 turning angle isopposite to sl turning angle, resulting in a more linear trajectory ofswimming, whereas SB s2 turning angle tends to continue the initialturn, producing an overall turning manoeuvre.Stage 1 angle is related to s2 angle only for DB escapes (Fig. 3.4A;p<0.01; r2 0.60). Eaton et a/. (1988) in a study on goldfish found asimilar correlation, although their analysis does not discriminatebetween the two C-start types. This might explain why our slope value(0.49) differs from Eaton et a/. (1988) (1.23), although the Yintercepts are similar (-63.3 present study; -59 Eaton et al., 1988).Another possibility is that the results of Eaton et a/. (1988) includeboth fast-start types. Pooling the data from the two fast-start typesthe correlation is still significant (p<0.01), although with a lower r2value (0.40).The relationship between s 1 and s2 angle is important, because ittells us something about how to predict the escape direction. Eaton eta/. (1988) pointed out that because s 1 and s2 angles are correlated,the neural commands for escape trajectory could be organized by the endof stage 1. However,  here, the relationship does not hold for SBresponses.^If^instead^the^total^escape^angle^is^considered,significant correlation with the sl turning angle of both types isfound. The slopes and elevations of the regressions for the two C-starttypes are not significantly different, so the data have been pooled toobtain the regression line of Fig. 3.4B. A line of identity and the^0^ 0regression line cross over at 102 . Therefore, below 102 , the total39escape angle tends to be smaller than the s 1 angle and vice versa above0^0^102 . Interestingly, 102 lies between the mean^escape angle ± 2 s.e.0^0for the two types (SB 120 ±17.5; DB 73.3 ±22.1). Therefore, in SBresponses the escape angle increases after stage 1, whereas in DB itdecreases, resulting in a more linear trajectory of escape. Theseconsiderations suggest that sl turning angle is a better predictor ofthe actual swimming escape path than it is of s2 turning angle. This isimportant because the total escape angle, measured considering theswimming path, is probably the most biologically significant in termsof predator-prey interactions.Eaton et al. (1988) observed that s1 angles were related to sl EMGduration. My observations based on film analysis (Fig. 3.6A) show asimilar correlation between sl angle and sl duration. Therefore, atleast indirectly, sl duration can predict the total escape angle (Fig.3.6B). Since the escape path tends to be away from the stimulus(Blaxter et al. 1981; Eaton et al. 1981) and the former is correlatedto sl duration, it might be that some stimulus characteristicsinfluence s 1 duration.It has been suggested that turning radius is a relevant parameter inpredator-prey interactions (Howland 1974, Webb 1976, Weihs & Webb1984). Turning radius is hypothesized to be independent of velocity butproportional to length (Howland, 1974). Webb (1976) found this to bethe case for trout. Turning radius is independent of velocity forangelfish and is not correlated with any other parameter measured.Specific turning radii are 0.067L for SB response and 0.063L for DB.These values are not statistically different, with an overall mean of400.065L, a value significantly lower than those reported by previousstudies (Webb, 1976, 1983; Webb & Keyes, 1981). The difference is notsurprising because minimum turning radius depends on body flexibility(Aleev, 1969), which in turn is a function of the degree of lateralcompression. The angelfish is extremely compressed laterally as can beinferred from the relationship between its total wetted surface areaand its length, Sw=1.02 L 2 (Sw=0.41 L2 and Sw=0.5 L2 for trout and bassrespectively; Webb, 1983).Howland (1974) suggests that the relative turning radii of two fishcan be predicted assuming turning moments are generated by lift forceson the body and the caudal fin. Then turning radius should beproportional to the ratio of the mass (M) and the projected lateralarea (Aw, sagittal section) and the ratio of the turning radii of twospecies should be predicted by the ratio of their M/Aw values. ForWebb's (1983) trout and bass, we obtain 1.63 for the ratio of the M/Awof the two species and 1.77 for the ratio of their turning radii. Thesevalues are close, confirming Howland's prediction. However, values of5.65 for the ratio of the M/Aw values and 9.7 for the ratio of theturning radii are found when comparing the angelfish with trout, and3.46 and 5.5 respectively when comparing it with bass. Thesediscrepancies suggest that other force components, especiallyacceleration reaction (Daniel, 1984) may be more important.A low value of turning radius can be beneficial in complexenvironments such as coral reef or weedy rivers where angelfish live.Although the angelfish keeps its body rigid during routine low speedlocomotion, it is well designed for tight turning manoeuvres during41escapes.2.^Distance derived performanceThe results indicate the presence of two types of C-start associatedwith different performance levels. Figure 3.3C shows that bothfast-start types reach high acceleration during sl. Double Bendresponses show a second peak in acceleration of similar magnitudeduring s2 and a corresponding increase in velocity (Fig. 3.3B), whereasfor SB subsequent acceleration is lower and sl maximum velocity isbarely maintained. As a consequence, although not all distance-timeperformance values during sl are significantly different between thetwo fast-start types, the differences in s2 are such that the overallvalues of performance throughout the whole response differsignificantly (Table 3.3). In addition, all the parameters measuredwithin a fixed time are significantly larger for DB; this includesdistance covered, a parameter suggested by Webb (1976, 1978) as anindicator of fast-start performance.The difference in velocity and acceleration values during s2 can bereconciled with the difference in kinematics of the two C-start types.According to Weihs (1973), the highest forward acceleration shouldoccur in stage 2 as a result of the return flip. Since here SBresponses do not show a clear return flip nor a body bend in theopposite direction to the initial one, it is not surprising thatvelocity and acceleration profiles differ from DB responses, whichcorrespond kinematically to Weihs (1973) description of a fast-start.Although average acceleration is negative during SB s2 (mean -1.842m/s2±5.6), a small peak in acceleration is observed shortly after theend of s 1 (Fig. 3.3C). The mean value of this peak is 34.07 mis 2 ± 12.7(mean Maximum acceleration s2 SB), which corresponds to about half themagnitude of DB s2, SB s1 and DB s1. This acceleration shows that thereis some thrust being produced after the end of s 1.This study provides the first fast-start performance data on apaired fin propulsion specialist. It is of interest to consider if thisspecialization impairs performance in body-caudal fin fast-startswimming. Most previous studies have not made distinctions amongC-start types. Therefore, DB and the pooled data (Table 3.3), arecompared with previous studies (Table 3.4). Velocities are given inactual values and as specific velocities (body length/sec), sincefast-start velocity has been shown to increase with size (Webb, 1976).However, according to Webb (1976), acceleration performance isindependent of size and so absolute values for acceleration rate arecompared here. Maximum and mean values for specific velocity andacceleration rate of the angelfish fall in the high range, consideringboth DB and pooled performance values (Table 3.3, 3.4). Film rates of200-250 hz give values most useful for comparison with mine. Accordingto Harper and Blake (1989) these film rates should underestimate theinstantaneous maximum acceleration by 20-30%. Given my estimated errorof 8%, the angelfish maximum acceleration performance remains high. Therelatively high velocity observed in angelfish is not surprising forthe comparison with larger fish (Wardle, 1975), however some of thefish listed in Table 3.4 are very similar in length to the ones Itested.43TABLE 3.4: Summary of previous fast-start studies; all values reportedrefer to C-type fast-start unless unspecified by the authors (1) ormean of pooled C and S starts (2).44Meanmaximumacceleration(ms -2 )Meanaccelerationrate(ms -2 )Meanmaximum velocity Mean velocityDistance(m)Time(s) Species Common nameLength(m) MethodRate(Hz)(m s - 1 ) (L^1 ) (m^I ) (L^)Weihs (1973) 40.0 Salmo trutta Trout F -50.0 Esox sp. Pike - F25.5 20.6 0.57t 1.7 0.113$ 0.200$ Salmo gairdneri Rainbow trout 0.330 F 40Webb (1975) 42.1 12.1 1.21 8.5 0.72t 5.0 0.056 0.078 Salmo gairdneri Rainbow trout 0.143 F 6415.7 8.1 0.67 8.4 0.36t 4.5 0.029 0.079 Lepomis cyanel/us Green sunfish 0.080 F 6495.0 Salmo gairdtzeri Rainbow trout* F 64Dubois et al. (1976) 23.5 2.80 4.4 0.210 Pomotamus saltatrix Bluefish (1)* 0.630 AWebb (1976) 40.6 17.6 2.85 7.4 1.63t 4.2 0.163 0.100 Salmo gairdneri Rainbow trout (2) 0.387 F 64Webb (1977) 26.6 8.6 1.44 8.3 0.70t 4.0 0.076 0.109 Salmo gairdneri Rainbow trout 0.174 F 250Webb (1978a) 41.0 1.71 12.6 1.13t 8.3 0.113 0.100 Salmo gairdneri Rainbow trout 0.136 F 250Webb (19780 39.5 10.4 1.56 7.2 0.73 3.4 0.084$ 0.115$ Esox sp. Tiger musky 0.217 F 25032.6 10.6 1.58 8.1 0.75 3.8 0.085$ 0.114$ Salmo gairdneri Rainbow trout 0.195 F 25023.9 9.3 1.15 7.4 0.50 3.2 0.051$ 0.103$ Perca flavescetzs Yellow perch 0.155 F 25028.8 12.3 1.30 8.5 0.67 4.4 0.059$ 0.088$ Lepomis macrochirus Bluegill 0.153 F 25028.7 11.0 1.14 10.7 0.49 4.6 0.038$ 0.078$ Notropis corn was Common shiner 0.107 F 25022.7 6.1, 0.77 9.4 0.43 5.2 0.035$ 0.081$ Coitus cognatus Slimy sculpin 0.082 F 25032.3 10.3 0.89 14.4 0.43 6.9 0.024$ 0.056$ Etheostoma caerzzlem Rainbow darter 0.062 F 250Webb (1983) 80.0 2.50 9.7 Salmo gairdneri Rainbow trout 0.257 F 60110.0 2.50 10.6 Micropterus dolomieu Smallmouth bass 0.236 F 60Webb (1986a) 16.0f Esox sp. Tiger musky 0.065 F 6015.0f 0.96 18.8 Micropterus salmoides Largemouth bass 0.051 F 6014.5f 1.01 15.8 Lepomis macrochirus Bluegill 0.064 F 6011.5# 0.81 14.0 Pimephales promelas Fathead minnow 0.058 F 60Harper and Blake 56.6 21.4 2.79 8.8 1.16 3.6 0.141 0.134 Salmo gairdneri Rainbow trout 0.318 A(1990) 157.8 54.7 4.70 12.4 2.27 6.0 0.194 0.085 Esox lucius Pike 0.378 A* Single event; $ calculated from data; $ read from figures.A, accelerometer; F, film.Salmo gairdneri=Oncorhynchus mykiss.3. C-start typesEaton et al. (1981) recognized a "fast forward C-start", as opposedto C-starts in which the turn continued in the same direction throughthe response. Their criterion for discriminating between the two typesis based on the change of direction during s2, whereas here DB s2 angle0approaches zero as s1 angle increases beyond 90 (Fig. 3.4A); however,these DB large turns differ from SB, showing a higher degree ofcontralateral bending and bimodal distance-time profiles.Previous studies (Eaton et al. 1977, Webb, 1976, 1978; Dubois etal. 1976) have focused on fusiform fish, that employ axial locomotionfor routine swimming. The angelfish is specialized for low speedswimming employing pectoral fin rowing (Blake, 1979). Axial locomotionis only employed for fast-starts and rapid turning manoeuvres. Thismight account for its fast-start types and explain the presence ofSingle Bend and Double Bend C-starts, which highlight manoeuvre andhigh speed respectively.The main difference between the two types of escape is in theintensity of the stage 2 return flip, which suggests that, in SB, stage2 might not be a purely active process. Eaton et al. (1977) documentedthe occurrence of an escape response where stage 2 was absent ingarfish (Xenetodon cancila). In this case the fish bent into a C duringstage 1 and did not straighten the body subsequently. They suggestedthat this observation supports the idea that the return flip (s2) is anactive process and not a passive mechanical consequence of rapid bodilybending (s 1). However, the garfish is rather elongate and probably46shows a passive recoil to a lesser extent than the laterally compressedangelfish. In the absence of EMG data, I can only speculate that SBand DB responses might be the result of a differential contralateralmuscular contraction, and passive recoil might play a different role inthe two types. Eaton et al. (1988) suggested that a purely mechanicaleffect cannot entirely explain the s2 propulsion and observed two EMGsignals, one associated with the s1 contraction and the second with thepropulsive stroke during s2. However the EMG data were not matched withdistance-time data, and it is not possible to relate differences inacceleration profiles to the EMGs. In addition to differentialcontralateral contraction, the two types of C-start might differ in therelaxation phase of the initial contraction. Covell et al. (1990)observed that, during s2, the deformation curve of ultrasonic dimensiongauges implanted on the side of the initial contraction varied bothwith the location of the gauge and the nature of the response.4. General conclusionsWebb (1984a) classified fish swimming styles into three broadcategories: body-caudal fin (BCF) periodic (cruising) propulsion; BCFtransient (fast-start) propulsion; median and paired fin (MPF)propulsion, used in slow and precise manoeuvres. It has been suggestedthat specialization for locomotion performance in any one area isusually associated with reduced performance in one or more of theothers (Webb, 1984a). This study suggests that optimal morphology forMPF propulsion does not impair performance at the BCF transient level.Optimal design for BCF transient propulsion has been suggested to47involve a large body depth (especially caudally), flexible body, largemuscle mass relative to body mass (Webb, 1984a). The angelfishpossesses two of these specializations. Its fineness ratio (LID) isamongst the lowest of all fishes (Aleev, 1969) and its body surfacearea in relation to its length is higher (Sw = 1.02 L2) than in anyother species studied so far. High flexibility associated with extremelateral compression allows the angelfish to perform very tight turnswhen escaping. Although the value of muscle mass relative to body massis lower (0.39M) than fast-start specialists (e.g. pike, 0.55M; Webb,1978), it is comparable to many generalist fish (Webb, 1978).The angelfish is well designed for two different locomotor modes,which are employed in different situations (feeding at low speed,acceleration to escape predators). During low speed swimming the bodyis kept rigid and the pectoral fins are moved by aerobic musculature.In contrast, during fast-starts, the body is bent employing theanaerobic axial musculature. Webb's (1984a) ideas on form and functionare based on locomotory modes employed when feeding. However, fish thatroutinely swim at low speed in the labriform mode, often maintain adecoupled system (axial locomotion) that allows them to perform highacceleration burst when attacked by predators. Arguably, optimaldesigns for decoupled systems do not necessarily trade-off.Within its fast-start capabilities, the angelfish has two options.Escape responses of SB type highlight manoeuvre, by allowing the bodycontraction to continue in only one direction, resulting in turns,whereas in DB responses the initial contraction is compensated by onein the opposite direction, resulting in a more linear trajectory. In48the latter case a higher velocity, given by a double acceleration, isachieved (Fig. 3.7). Further studies would be required in order tounderstand the mechanism underlying this differential pattern ofbehaviour and its biological significance. Perhaps there arebehavioural trade-offs between high distance-time performance and largeturns, such that the prey would employ a particular type of responsedepending on the predator's strike tactics.Fig. 3.8A shows the ranges of s 1 turning angle observed for bothtypes of response. Combined, the two ranges would allow a total range0of angles of turn between 37 and 145. These values might not be theactual limits of the angelfish sl turning angles. However, a lowerlimit, set by the sl contraction, and an upper limit, set by limitedbody flexibility, are likely. Since survival depends on the fishcapability to escape dangers from all directions, it would be importantto have the widest range of turning trajectories available on eachside. This might be achieved only when employing SB escapes for smallturns and DB escapes for large turns (Fig. 3.4B). The new limits when0^0considering the total escape angle are much wider (14 -159 ; Fig .8.8B)and^result^respectively^from^contralateral^bending^(DB)^andcontinuation of the turn (SB) during s2. Arguably, the choice ofC-start type might be important in determining the escape trajectory toavoid predators.49FIGURE 3.8: (A) Range of s 1 turning angles for Double Bend (continuousline) and Single Bend responses (dotted line). (B) Overall range oftotal escape angles given by a Double Bend (continuous straightarrow) and a Single Bend (dotted straight arrow) response. Curvedarrows indicate the change in direction between s1 and s2 turningangles.50B37oACHAPTER 4Escape trajectories in angelfish Pterophyllum eimekeiA. INTRODUCTIONEscape responses are central to the survival of fish and otheranimals, and determine important aspects of their behaviour, physiologyand ecology. Many animals have evolved sensory systems tuned to detecttheir predators (Alcock, 1989), along with fast-conductingneuromuscular systems associated with rapid escape manoeuvres (Bullock,1984). Some animals have a limited range of escape responses, forexample crayfish escape either upward or backward (Wine and Krasne,1972). Others, such as certain insects and fish, are able to escape inmany directions (Chapter 3; Eaton et al. 1991).In^fish,^a^C-start^implies^short^response^time^and^highacceleration. In addition, the prey fish must perceive the direction ofattack, it must escape at the right time and it must turn in thecorrect direction so that it does not collide either with the predatoror with nearby obstacles (Eaton et al. 1991). It has been suggestedthat the direction of escape is preprogrammed at the onset of thestartle response (Nissanov and Eaton, 1989; Eaton et al. 1991).However, escape trajectories relative to a stimulus vary considerablyafter the initial turn away from it (Eaton et al. 1981). Here, I52investigate^this^variability^using^circular^analysis^of^frequencydistribution.The analysis of escape trajectories in fish (Eaton et al,^1981;Eaton and Emberley, 1991) and other animals (Camhi and Tom, 1978;Corner and Dowd, 1987: Nalbach, 1990) has been performed by employinglinear plots of stimulus angle versus body turn (defined as the anglebetween the animal's initial orientation and - its orientation at the endof the response). Here, escape trajectory is considered as a circularvariable, and is defined as the angle between the stimulus directionand the animal's swimming path at the end of the response.B. MATERIALS AND METHODSAngelfish (Pterophyllum eimekei) of four group sizes (total length,L=4.9±0.4 cm; mean±2S.E., N=4; L=7.3±0.4, N=5; L=10.9±0.4, N=4;L=13.5±0.6, N=2) were employed. Fish were kept as described in Chapter2. A total of 62 escape responses (14, 20, 18, 10 for each size grouprespectively) were analysed. Experimental conditions are thosedescribed in Chapter 2.In^order^to^ascertain^the directionality of the^stimulus,^twohydrophones (Sparton 60 CX 123, omnidirectional, with a flat response± 3 dB from 50 Hz to 80 kHz) separated by 18 cm along the axis of thestimulus direction (5 cm and 23 cm away from the wall closest to thestimulus, respectively) were placed in the middle of the experimental53tank. Therefore, the separation line between the two hydrophones wasbisecting the area of 7.5 cm x 7.5 cm in the center of the experimentaltank, in which the center of mass of any given fish was located whenstartled. The recorded time interval between the onsets of the soundsignal for the two hydrophones was 0.12 ms (sampled at intervals of0.02 ms; Fig.4.1). The calculated delay for the speed of sound in fresh-water at 20 °C (1481 m s 1 ; Kinsler et al. 1982) is 0.124 ms for an 18cm separation. The agreement of my measurement and the time delaycalculated for a sound stimulus travelling along the distance betweenthe two hydrophones, showed that the stimulus presented to the fish wasindeed directional along the separation line between the twohydrophones. This confirmed the stimulus direction that I have used inmy analysis. If sound was travelling in any other direction, the delaybetween the signals received by the two hydrophones would have beenshorter or reversed.The orientation of the fish relative to the stimulus direction wasrandom (Watson U2 test; P>0.1; N=62). Responses to stimuli from theleft and right were pooled as if the stimulus was always on the rightside of the animal. Stage 1 (s1) angle for the anterior part of thebody was determined by measuring the angle between the straight linespassing from the center of mass to the tip of the head at frame 0 (oneframe before the first detectable movement) and at the end of stage 1.A total escape angle was measured between the midline of the fish atframe 0 and the regression line considering seven positions of thecenter of mass about the end of stage 2 (stage 2 ± 3 frames) (Chapter3). Here, these angles were redefined as s 1 turns and escape turns when54Fig. 4.1:^A:^Diagram showing the experimental set up (top view). Theouter rectangle represents^the external^tank.^The^arrow^indicatesstimulus direction. The inner rectangle represents the experimentaltank. The two circles represent the two hydroph ones used to ascertainthe direction of the stimulus. The hydrophones were not in the tankwhen the fish were startled. The inner square represents the area inwhich the center of mass of the fish was at the onset of the stimulus.Numbers indicate lengths in cm. Relative dimensions are not in scale.B: Time delay in stimulus signal as perceived by the two hydrophonespositioned along the axis of stimulus direction. The distance betweenhydrophones was 18 cm. The bottom trace is for the hydrophone closer tothe stimulus. Arrows show the first inflection of the sound signal foreach hydrophone. Sampling intervals indicate 0.02 ms. Six points (0.12ms) separate the two arrows. The Y axis represents amplitude inarbitrary units since the two hydrophones had different sensitivities.550^0.1^0.2^0.3^0.4^0.5^0.6Btime (ms)0.856calculated relative to the midline of the fish at frame 0, and s 1trajectories and escape trajectories when relative to the stimulusdirection (Fig.4.2). Directionality of the response was indicated by theorientation of the C-bend relative to the stimulus (B lax ter et al.1981). Therefore, escape responses were divided into 'away responses'and 'towards responses' when the escaping fish was oriented away andtowards the stimulus, respectively.Initial orientation was defined as the angle between the midline ofthe fish anteriorly and stimulus direction at the onset of the response(Fig.4.2). Since responses to left and right stimuli were pooled,initial orientation spanned from 0 ° to 180° . An angle of 0° indicatedthe stimulus direction, therefore, a fish positioned at 180 ° was facingdirectly away from the stimulus. Angular measurements were made to thenearest degree. The discriminating zone was defined as the angular zoneof initial orientation within which the fish responded to a stimulus byescaping in non-random directions.Steering was defined as the difference between escape turn and s1turn. Therefore, steering was positive if the fish continued the turnafter sl, and negative if it reversed the turn direction (Fig.4.3).Angular compensation was calculated as having the same absolute valueas steering, however, its sign was negative if steering was directedtowards the stimulus, and positive if away from it (Fig. 4.3). Steeringand angular compensation have been determined for the semicircles0 ° -180 ° and 180 ° -360° separately (based on the fish orientation at theend of stage 1), since the signs of the two variables within eachsemicircle did not always coincide (see Table 4.1 and Fig. 4.3).57Fig.4.2: Diagram illustrating angular variables. Solid arrowindicates stimulus direction, dashed arrow indicates escape direction.The straight fish indicates position at time 0, the fish bent into a'C' indicates position at the end of stage 1, and the swimming fishindicates position at the end of the escape response. Black dotsindicate the fish 'stretched straight' center of mass. 1) indicates theangle representing initial orientation. 2) represents the angle of s 1turn. 3) represents the angle of escape turn. 4) represents the angleof s1 trajectory. 5) represents the angle of escape trajectory.584II5 9Fig.4.3: The sign of steering and angular compensation. Fish positionis shown at end of s 1. The top diagrams show positive steering (A) andpositive angular compensation (B, C). The bottom diagrams show negativesteering (D) and negative angular compensation (E, F). The angularorientation relative to the stimulus is shown in (E). Steering (curvedarrow) is positive when the escape turn (straight solid arrow) is inthe same direction of the C-bend, and negative if vice versa. The signof angular compensation (curved arrow) depends on the direction ofsteering relative to the stimulus (dashed arrow). Angular compensationis positive when steering is away from the stimulus, and negative whentowards. Since all the escapes were analysed as if the stimulus was onthe right side of the fish at the onset of the response, the fish'sinitial orientation (not shown) is always between 0° and 180° .6090270360; 0Steering^Angular CompensationTable 4.1: Turning angles, angular steering and angular compensation in2 semicircles for towards and away responses. Escape responses areassigned to a semicircle based on the fish orientation at the end ofstage 1. A t-test is used for all comparisons except * (Mann Whitneytest). ± 2 S.E. of the mean are given in brackets.62TOWARDSRESPONSESSemicircle^al^Escape^Angular^Angular^Initial^Nturn turn steering compensation orientation(degrees) (degrees)^(degrees)^(degrees)^(degrees)0^o0 -180 62.3 45.7 -16.5 16.5 134.4 11(20.7) (29.8) (11.3) (11.3) (17.0)180 o-360 o 116.4 127.4 11.0 11.0 34.0 5(20.4) (22.8) (13.9) (13.9) (34.6)p<0.01 p<0.005 p<0.05 n . 3 . p<0.05*pooled 79.2 71.2 -7.9 14.8 103.1 16(20.0) (28.8) (10.9) (8.7) (28.6)AWAYRESPONSESo -1800 81.0 73.7 -7.3 -7.3 60.2 29(12.2) (14.4) (4.9) (4.9) (13.0)1B0 ° -360 ° 84.5 78.2 -6.2 6.2 124.5 17(13.8) (14.1) (7.6) (7.6) (15.0)n.s n.s n.s p<0.005 p<0.001*pooled 82.3 75.4 -6.9 -2.3 84.0 46(9.2) (10.4) (4.3) (4.7) (13.1)63Circular statistics (B atschelet, 1981) were employed to analyse sland escape trajectories, since they could vary from 0 ° to 360° . Linearregression was used to analyse the relationship between sl trajectoriesand steering, since each regression (A and B in Fig. 4.10) includedonly sl trajectory values within 40° . When a circular variable isrestricted to such a narrow interval, linear analysis is appropriate(B atschelet, 1981). All other angular variables were considered angulardistances and were analysed employing linear statistics.C. RESULTS1. Effect of fish initial orientationThe^effect^of the^initial^orientation^on^randomness^of thetrajectories has been analysed by considering six arbitrarily defined"orientation sectors" of 300 eachfrom 0 ° to 180 ° ).^For initial(N=10,9,14,10,11,8, for each sectororientations of 0 ° -30° ,^120° -150 ° ,150 ° -180°, both sl and escape trajectories are randomly distributed.For the remaining three orientation sectors (30 ° -60° , 60° -90° ,90 °-120 ° ), sl and escape trajectories are not random (p<0.01; p<0.005;p<0.005 for each sector respectively; Watson's U 2 test). Therefore, thediscriminating zone spans from 30 ° to 120° . Angular deviations of bothsl and escape trajectories in each orientation sector are shown inFig .4.4.Initial^orientation^has^an^effect^on^directionality^(away^vs.64Fig.4.4: Angular deviation of sl and escape trajectories for different30°^orientation^sectors.^S^indicates^r non-random^distribution^oftrajectories.^Dark bars indicate escape trajectories and white barsindicate sl trajectories.658060 inI^I^I0-30^30-60^60-90^90-120^120-150 150-180initial orientation (degrees)66towards responses) of escapes. Away responses occur significantly moreoften than towards responses only for initial orientation sectors of30 ° -60 ° and 60 ° -90 ° (p<0.05 in both cases; Binomial test; Fig.4.5). Whenthe initial orientation is near 180 °, a towards response at a smallangle of turn may actually mean that the fish is effectively swimmingaway from the stimulus. Therefore, escape trajectories within 2semicircles^('away^semicircle'^vs.^'towards^semicircle')^have^beenconsidered (Fig.4.5). There are significantly more escape trajectories inthe semicircle away from the stimulus (90 0 -270° sector, see inset inFig.4.5) than in the opposite semicircle when the initial orientation is90 °4 20 ° and 120 ° -150 ° (p<0.05 in both cases; Binomial test; Fig.4.5).Circular-linear correlation of away responses shows that escapetrajectories are independent of initial orientation (N=46; p>0.1; Rankcorrelation test), whereas sl trajectories are positively correlated toit (N=46; p<0.01; Correlation coefficient D n =0.4; Rank correlationtest).^Initial^orientation has^no^effect on^steering^(N=46;^p>0.1;linear regression test).2. Escape trajectory frequency distributionsFish size has no effect on sl and escape trajectory distributions(N=62; p>0.05; Mardia-Watson-Wheeler Test). Away responses occursignificantly more often than towards responses (46 and 16,respectively; different from random; p<0.001; Binomial Test). Thecircular frequency distributions of both s 1 and escape trajectories arenon-random in away responses (Watson's U2 test; p<0.005 in both cases;Figs. 4.6A,B) and random in towards responses (Watson's iY test; p>0.167Fig.4.5: Percentage of away responses for different 30 ° orientationsectors. S means significantly different from random. Dark barsindicate away responses and white bars indicate responses in thesemicircle away (90 ° -270 °) from the stimulus. Insert shows theorientation of the C bend of the fish at the end of stage 1 for awayresponses (dark square), and the position of the fish for escapetrajectories that are within the semicircle away from the stimulus(white square). Arrow indicates stimulus direction.686 9Fig.4.6: Circular frequency distributions of sl trajectories (A; awayresponses, C; towards responses) and escape trajectories (B; awayresponses, D; towards responses). Responses to left or right stimuliare plotted as if the stimulus was always on the right side of thefish. Frequency histograms (away responses) and scatter diagrams(towards responses) are shown. The frequency interval is 1(P. In awayresponses, each concentric circle represents a frequency of 2. In B,two main modes, separated by 50° , are present: numbered arrows indicatetrajectory 1 (180°) and trajectory 2 (130°).70aio00£^ 0££00£0906OZ 1OLZOt7Z9OS1^01Z091i 10S1 01Z06 OLZ09100£^ 0££09^ 00£OZ 1^ Ot7Z0910 1 ZOS1in both cases; Figs. 4.6C,D). The frequency distribution of the sltrajectories of away responses is not statistically different from avon Mises distribution (Fig. 4.7A; N=46; p>0.25; X 2 test), whereas thatof the escape trajectories is (Fig. 4.7B; N=46; p<0.05; X 2 test), withtwo peaks at 130 ° and 180°. Although I have not tested the significanceof the two peaks (there is no standard method for this procedure), abimodal pattern is apparent. To test whether the presence of the twomodes is related to the initial orientation, I have separated escapetrajectories into two groups (110 ° -150° range and 150° -190° range) thatcan be considered as representative of the two modes. Figure 4.8 showsthe percentage of responses within each 40 ° range of escapetrajectories for different initial orientations. Initial orientationhas no effect on the choice of escape trajectory (Mann-Whitney test;P>0.1).In addition to circular statistics, I have employed a linearregression analysis of stimulus angle versus escape angle, followingCamhi and Tom (1978) and later studies (e.g. Eaton et al, 1981;Nalbach , 1990). I define stimulus angle as the angle between themidline of the fish posteriorly and stimulus direction at the onset ofthe response (see inset in Fig.4.9); escape angle corresponds to mydefinition of escape turn. Therefore, if escapes were always directlyaway from a stimulus, they would appear as a line passing through theorigin at 45 ° in Fig.4.9. The bimodal nature of the data visible whenemploying circular analysis is not apparent from the graph, althoughthe linear regression is significant (P<0.0001; r 2 =0.3). Arguably,linear regression analysis is not appropriate because it assumes that72Fig.4.7: Fitting a von Mises distribution to frequency polygons of s I(A) and escape (B) trajectories. Frequency intervals are 10 ° . (A) doesnot differ from a von Mises distribution, whereas (B) does (P<0.05).73'166 ' 1 46 '180 '^'260^'300 1s1 trajectory (degrees)3-2-1-0 Purr340 20B5-0 rumor „340^20^60^'166 ' X140 ' Ilao^I2c's^260 ' 360escape trajectory (degrees)g 4-E374Fig.4.8: Percentage of responses within 110° -150° (dark bars) and150°-190° range of escape trajectories (white bars) for differentorientation sectors.755040i vill ^0_30^30-60^60-90^90-120^120-150 ,0.0302010 -150-180initial orientation (degrees)76Fig.4.9: Linear analysis of angelfish escape angle, followingconventions established by Camhi and Tom (1978). Points in the righthalf of the graph show escape angles of trials in which the stimuluswas on the right side of the animals. Points on the upper half of thegraph indicate trials in which the fish's escape angle is on the rightside of its midline. Therefore points in the lower right sector or inthe upper left sector indicate responses in which the escape angle wasaway from the stimulus. N=62; r 2= 0.3; Y=-0.44X-12; P<0.0001. Insertshows the position of the fish at time 0 (straight fish) and itsmidline (dashed line), and at the end of the escape response (curvedfish). Arrow indicates stimulus direction. Curved lines (X and Y)indicate stimulus angle and escape angle respectively.77stimulus angle (degrees)for any given value of X, the Y's are normally distributed (Zar, 1984,Sokal and Rohlf, 1981). I have shown that this is not the case as thecircular distribution of the data does not fit a normal circular (vonMises) distribution.3. Steering and angular compensationSteering in away and towards responses has been determined forsemicircles 0 ° -180° and 180° -360° separately. Steering in the twosemicircles is not statistically different for away responses and thevalues have the same sign, whereas they have opposite signs and arestatistically different in towards responses (Table 4.1). Pooledangular compensation values for away and towards responses differ(t-test; p<0.001) as do their absolute values (t-test; p<0.05). Inaddition,^the^initial^orientation,^sl^and^escape^turn^of^towardsresponses^differ for the two^semicircles^(Table^4.1).^In^awayresponses,^only^initial^orientation^and^angular^compensation^arestatistically different for the two semicircles (Table 4.1).^The relationship between steering and sl^trajectory of awayresponses is shown in Fig.4.10. The two regression lines include onlyresponses that have escape trajectories within 110 ° -150° (line A) or150 0 -190° (line B). These two ranges were chosen as representative ofthe two modes. The slopes of the two lines are not significantlydifferent from each other, whereas their elevations are (p<0.001).79Fig.4.10: Relationship between angular steering and sl trajectories inaway responses. The two regression lines include only escapes whosetrajectory is within 110°-150° (line A; closed circles) or 150°-190°(line B; open circles). For (A) y=94-0.75x; r2= 0.59; p<0.001; N=15.For (B) y=132-0.78x; r 2 = 0.60; p<0.001; N=17. The slopes of the twolines are not significantly different from each other, whereas theirelevations are (p<0.001). Open boxes represent responses whosetrajectory is <110° or >190° .80s 1 trajectory (degrees)D. DISCUSSIONI. Effect of initial orientationAngelfish discriminate the direction of a stimulus by escaping innon-random trajectories when the stimulus is presented laterally,within an angular zone (discriminating zone) that extends approximatelybetween 30 ° and 120 ° of initial orientation (Fig.4.4). These results areconfirmed by the left-right choice data (Fig. 4.5), both fordirectionality (left-right C-bend) and semicircle chosen.Canfield and Eaton (1990) have shown that swimbladder acousticpressure^transduction^initiates^Mauthner-mediated^responses,^andsuggest that the detection of particle motion ensures thedirectionality of the response. Although mechanical sensitivity ismaintained for 360° around the fish (Hawkins and Homer, 1981),left-right discrimination should decrease when the stimulus is more inline with the longitudinal axis of the fish, due to limits in theangular discrimination between two sound sources (Schuijf, 1975).The influence of initial orientation on the escape response is animportant factor in the predator-prey interactions of fish. Webb(1986b) suggests that a predator should attack in line with theanticipated escape trajectory of the prey. Since escape responses werethought to be fixed turns of 90 °, Webb (1986b) suggests that the strikeangle should also be about 90° . However, escape turns are not fixed,but cover a wide range of angles (Chapter 3; Eaton et al. 1991). Inaddition, the prey may optimize its positioning relative to a stalking82predator.To maximize directional mechanical sensitivity and visual acuity,the prey should orient itself perpendicular to a stalking predator.However, if readiness to escape is to be maximized, the prey should beoriented away from the predator, in order to minimize the time forturning away, while keeping the predator within its angular regions ofdirectional mechanical discrimination and visual field. Thiscorresponds to an orientation of about 130 ° away from the predator(Hall et al. 1986).Webb and Skadsen (1980) report values of strike angle for pikeattacking minnows. Although calculated as the angle subtended betweenthe prey body axis and the strike path of the predator's snout, thesevalues should approximate the prey's orientation, since strike pathswere more or less straight. The frequency distribution of strike anglesdoes not deviate significantly from normality (Webb and Skadsen, 1980).The mean angle is 82° and the highest mode is the 80°4 00° sector. Thiscorresponds to a position approximately perpendicular to the predator.A second peak occurs at 120 ° -140° (Fig. 3, Webb and Skadsen, 1980)regardless of strike pattern. This suggests an alternative preystrategy, corresponding to the position maximizing readiness to escape.Initial orientation has no effect on escape trajectory of awayresponses and a positive effect on s1 trajectory. This may explain whythe distribution of sl trajectories does not show the distinct modesseen in escape trajectories (Fig.4.7). With steering, swimmingtrajectories are brought to fixed directions away from the stimulus.832. Frequency distribution of escape trajectoriesAs in previous studies, (Blaxter et al., 1981; Eaton et al., 1981),away responses occur significantly more often than towards responses.In addition, towards responses are uniformly distributed around acircle (Figs. 4.6C,D), whereas away responses are not (Figs. 4.6A,B).Towards responses have been suggested to be due to errors of symmetryin the left-right direction (Eaton and Emberley, 1991).Eaton and Emberley (1991) reanalysed the data of Eaton et al. (1981)by pooling left and right responses and found that the magnitude ofboth s 1 and escape turns decreases linearly as the angle of the initialorientation increases. Arguably, comparing my data to those of Eatonet al. (1981) and Eaton and Emberley (1991) is confounded bydifferences in methodology. In particular, in both studies on goldfish(Eaton et al., 1981; Eaton and Emberley, 1991), trajectories weremeasured at a fixed time after the initial movement. However,fast-starts can differ in duration by as much as 100% reflectingdifferences in turning angle (Chapter 3). For my measurement of escapeduration (Chapter 3), like that of Camhi and Tom (1978), Corner and Dowd(1987) and Nalbach (1990) for other animals, the escape response isconsidered over when the animal's turn stops.When employing traditional linear methods of analysing escape anglesversus stimulus direction (Camhi and Tom, 1978; Eaton et al, 1981;Nalbach, 1990), I obtain a significant linear regression for angelfishescapes (Fig. 4.9), suggesting a unimodal distribution of escapetrajectories. However, a circular plot reveals a bimodal pattern thatdiffers from normal circular (Fig. 4.6B).84Multiple preferred trajectories may be adaptive in preventingpredators from learning any fixed single pattern of response andcompensating for it. Eaton et al. (1977) suggest that fish escapetrajectories should be unpredictable for this same reason. However, arandom distribution of escape trajectories would include, for instance,escapes directed straight towards the predator's mouth. Also, linearanalysis (Fig. 4.9) suggests a normal unimodal distribution of escapetrajectories, whereas circular analysis (Fig. 4.7) shows that angelfishdo not employ a single (unimodal) direction of response. Multiplepreferred trajectories may maximize the distance from the stimulus(trajectory 1; 1800; Fig. 4.6B) and by following trajectory 2 (Fig.4.6B; 130 ° ), fish may maximize this distance while swimming just withintheir discriminating zone. Interestingly, the direction of visualavoidance responses in fish are just within their visual zone (Hall etal. 1986). Sensory feedback at these particular orientations(trajectories 1 and 2) may control the escape turns and angularsteering. The choice of one preferred trajectory is independent of thefish initial orientation.Eaton and Emberley (1991) found that the linear relationship betweeninitial orientation and sl turn was stronger than that between initialorientation and escape turn. They suggested that this is due to thevariability of steering (stage 2 in Eaton and Emberley, 1991), and thatthis variability is due to the influence of nearby walls. My resultsshow that steering subsequent to s 1 contributes to the bimodal pattern(Fig. 4.7 and Fig. 4.10). In my experiments, wall effect was minimal.Although there were differences in relative distance (body lengths)85from the wall due to differences in fish size, the latter had no effecton the escape trajectory distribution. Therefore, varying the relativedistance from the wall had no effect on the escape trajectory.Fish size has no effect on s 1 and escape trajectories. However, sizehas an effect on turning (Domenici and Blake, 1991) and flexibility(Aleev, 1969). Smaller fish are more flexible about their center ofmass than larger fish. For large fish, the angle required to set aparticular trajectory from a given orientation may be too large for asingle body contraction. A fish with limited flexibility may employ acontinuation of the turn which extends sl angle ('single bend'responses; Chapter 1). This type of turn, however, compromises velocity(Chapter 1) and, although bigger fish can achieve higher absolutevelocity than smaller ones (Wardle, 1975), preferred trajectories maylimit their escape performance in terms of velocity and acceleration.3. Steering and Angular compensationA sensory feedback mechanism is suggested by the behaviour of fishduring towards responses. When the towards responses are triggered at alarge angle of initial orientation (Table 4.1), fish minimize sl turnsremaining within the 0 °-180 ° semicircle, and swim away from thestimulus by steering. On the other hand, at a small angle of initialorientation (Table 4.1), correction of the turn does not occur beforethe fish orientation is 0 ° (facing the stimulus). After reaching anorientation of 0 ° , fish continue their turn within the semicircle180 0 -360 ° and move away from the stimulus. The resulting s1 turns arelarge and are further increased by steering (Table 4.1).86In away responses, steering allows fish to reach the preferredtrajectories as shown in Fig .4.10. Although I was unable to makepredictions as to which of the preferred trajectories the fish wouldfollow, the animals never reached trajectory 2 (130 °) when sltrajectories were >180 °. This may reflect biomechanical constraints onstage 2 turning.Angelfish are able to adjust their trajectories by steering awayfrom the stimulus once they have committed an apparent tactical errorby making a towards response. In addition, the different values of s 1towards turns for the two semicircles suggest that feedback may startbefore the end of s 1. Although inhibition of the lateral linemechanosensory input lasts for approximately the duration of sl(Russell, 1976), this would still allow sensory feedback to influencesteering. However, Eaton et a!. (1988) suggest that stage 2 ispreprogrammed and not dependent on movement-induced feedback. UsingEMGs, they show that the contralateral muscular activity can beginbefore the escape response has started. Although preferred trajectoriescould be achieved without sensory feedback, it is difficult to explaincorrection of 'wrong turns' in its absence. Although the stage 2contraction starts before the movement of the fish, its extent maypartially be controlled by a feedback system. However, thedirectionality of the stimulus during stage 2 could have beeninfluenced by echoes in the tank, following the initial stimulus.Further behavioural analysis of towards responses conducted togetherwith physiological measurements are needed in order to establish if andto what extent sensory feedback plays a role in determining fish escape87trajectories.4. Escape trajectories in other animalsCockroaches evade wind puffs simulating a toad's tongue protrusion(Camhi and Tom, 1978; Corner and Dowd, 1987). Their escape is triggeredby a threshold value of wind acceleration (Camhi et al. 1978), detectedby sensory cercal hairs arranged in 14 rows (Dagan and Camhi, 1979).Hairs of a particular row have a preferred axis of pliancy which isexcitatory in one direction and inhibitory in the other (Dagan andCamhi, 1979). Sensory information from these hairs is conveyed to giantneurons which are also directionally sensitive (Westin, 1979).Reanalysis of four published data sets suggests that the circulardistribution of the escape trajectories appears multimodal (Fig.4.1113).I suggest that preferred escape directions correspond to particularorientations of the sensory organs relative to the stimulus. Possiblecandidates controlling the cockroach escape trajectories are thepreferred directions of pliancy of its sensory hairs (Fig.4.11B) andthe directional responses of the giant neurons (Westin, 1979). Themechanisms underlying the onset of the cockroach escape response areunderstood (Ritzmann, 1984; Camhi and Levy, 1989), but little is knownabout what determines the magnitude of its turn. The sensory hairs arenot completely phasic receptors (Westin, 1979) and, in addition todetecting wind acceleration, should also be able to detect a moreconstant wind velocity as occurs later during the response (Camhi etal. 1978). As for fish, escape trajectories cannot be predicted fromthe initial orientation.88Fig.4.11: Circular frequency distributions of escape trajectories inresponse to a stimulus (arrow on the bottom of the graph). Only awayresponses are plotted. Responses to left or right stimuli are plottedas if the stimulus was always on the right side of the animal. Thefrequency interval is 10°. Each concentric circle represents afrequency of 2 (A,C) or 10 (B). None of the frequency distributions arerandom (U2n test; p<0.005) and all of them are significantly differentfrom normal circular (von Mises) distributions (X 2 test; (A), p<0.05;(B), p<0.001; (C), p<0.05).A) Angelfish (Pterophyllum eimekei). Escape trajectories in responseto a mechanical stimulus. Responses away from the stimulus represent74 % of the total (N=62; significant at p<0.001; Binomial test). Twomain modes, separated by 50°, are present, at 180 ° (arrow 1) and 130°(arrow 2).B) Cockroach (Periplaneta americana). Escape trajectories in responseto wind puffs. Data from Camhi and Tom (1978) and Corner and Dowd(1987). Four data sets are pooled together (figs 5A and 6A from Cornerand Dowd, 1987 and figs 5 and 6 in Camhi and Tom, 1978) (notstatistically^different;^p>^0.05;^Mardia-Watson-Wheeler^test).Responses away from the stimulus represent 83% of the total (N=408;significant at p<0.001; Binomial test).^The inhibitory directions ofhairs (Dagan and Camhi, 1979) from both cerci match closely with themodes and are shown as stars in the graph.89Fig.4.11 (Continuation)C)^Soldier crab (Mictyris longicarpus). Escape trajectories inresponse to visual stimulus represented by a walking person. Data fromNalbach (1990). Responses away from the stimulus represent 71 % of thetotal (N=31; significant at p<0.05; Binomial test).90180210 _■-r----, 150^120^24090^270000tB tSoldier crabs walk forward unlike most shore crabs, and their escaperesponses are visually mediated (Nalbach, 1990). It has been suggestedthat they do not escape in a fixed direction relative to a stimulus,but rather within a fixed range of angles relative to their initialbody orientation (Nalbach, 1990). However, a circular plot shows abimodal pattern of escape trajectories, with modes either side of 180°(Fig.4.11C). As in the previous examples ; I suggest that the twopreferred escape directions correspond to particular limits imposed bythe crab's sensory organs. Soldier crab eyes have a limited field ofview, which may be extended by the rotation of their eye stalks(Nalbach and Nalbach, 1987). By following the two preferredtrajectories, soldier crabs may be able to maximize the distance awayfrom the stimulus while it is within the limits of their visual field.The common escape response, described above, involves a turn awayfrom the stimulus (away responses). Escape responses may involve a turntowards the stimulus (towards responses), as discussed previously forangelfish. I have shown that the fish tend to correct their 'wrongturns', by steering away from the stimulus.The cockroach's mean body turn in towards responses is small (29° )compared to the away responses (52° ). Comer and Dowd (1987) found thatthe magnitude of the turn was linearly related to the initialorientation in both towards and away responses. They suggest that thisreinforces the hypothesis of absence of sensory feedback during theresponse. However, a linear relationship is significant only if towardsresponses ending in both semicircles (0° -180° and 180° -360° ) areconsidered. As discussed for angelfish, when the stimulus is delivered92almost frontally, the turn may continue away from the stimulus (in the180 0 -360 ° semicircle) after the animal reaches a frontal position. Themean value of the body turns in this case is 90° . When these responsesare excluded, the mean body turn for towards responses is even smaller(17 °) and has lower variance, making them highly predictable. Perhapsthe animals are aborting these 'wrong' turns. This supports mysuggestion of a sensory feedback mechanism occurring during the escaperesponse. In soldier crabs, escape responses to a stimulus presentedposteriorly (between 160 ° and 180°) can elicit a towards response whichallows the animal to reach its preferred trajectory of 150 ° -160° .The existence of preferred escape trajectories in animals hasimportant implications for animal behaviour, physiology and ecology.Behaviourally, the presence of multiple preferred escape trajectoriesmay be adaptive in preventing predators from learning any single fixedpattern of prey response and compensating for it. Physiologically,preferred escape trajectories imply mechanisms controlling themagnitude of the turn relative to a stimulus. This may be achievedthrough sensory feedback. Ecologically, the choice of a specificturning angle (regardless of size) may constrain important escapeparameters such as velocity (Chapter 3). For example, although biggerfish can achieve higher absolute velocities than smaller ones (Wardle,1975), they have lower turning capabilities (Domenici and Blake, 1991)and are less flexible at their center of mass (Aleev, 1969). This maybe a determinant of size specific differential survival. In addition,the consideration of preferred escape trajectories may have practicalimplications, for example, in fishing gear design (Wardle, 1986).93CHAPTER 5The effect of size on angelfish (P. eimekei) escape responsesA. INTRODUCTIONThe effect of size on fish swimming performance is widely documented(Bainbridge, 1958; Wardle, 1975, 1977; Webb, 1976). During escaperesponses, larger fish attain higher velocities than smaller ones. Thisis because fast-start durations are longer in larger fish, whileacceleration is size-independent (Webb, 1976). When considered within afixed time, however, both distance travelled and maximum velocity aresize-independent (Webb, 1976).Daniel and MayhOfer (1989) point out that past studies on fishswimming have considered the problems of either acceleration orturning, but no study has considered how these act simultaneouslyduring escape manoeuvres. Daniel and MayhOfer's (1989) study oncarridean shrimp escape locomotion is the first example of suchapproach in studies of aquatic locomotion. They show a trade-offbetween rotation and translation which imposes size limits in theescape locomotion of carridean shrimp.Fish escape responses were thought to be stereotypic, characterizedby fixed angles of turn and maximum acceleration performance (Webb,941986). In Chapter 3, I show that acceleration can be submaximal (in"single bend" escape responses). In addition, although the trajectoriesaway from a stimulus are fixed, the turning angles may varysubstantially when the fish's initial orientation relative to thestimulus is random (Chapter 4). The performance of angelfish issubmaximal at large turning angles. Here, I investigate the effect ofsize on the trade-off between turning angle and swimming performance inthe away responses of angelfish.B. MATERIALS AND METHODSAngelfish (Pterophyllum eintekei) of four size groups (total length,Group 1, L=4.9±0.4 cm; mean±2S.E., N=4; Group 2, L=7.3±0.4, N=5; Group3, L=10.9±0.4, N=4; Group 4, L=13.5±0.6, N=2) were employed.Morphological variables are listed in Table 5.1. Experimentalconditions are described in Chapter 2. Data analysis is described inChapter 3. Performance and kinematic variables are defined in Chapter3. Initial orientation is defined in Chapter 4. Here, processed filmswere magnified between 3 and 7 times (for the largest and the smallestfish, respectively). This corresponds to an apparent size of 32 to 38cm. The purpose of image magnification was to minimize measurementerror (Harper and Blake, 1989). Escape responses were classified asaway responses (N=46) and towards responses (N=16), based on theorientation of the C-bend relative to the stimulus (Blaxter et al,1981; Chapter 4). Only away responses were analysed, since towards95Table 5.1: Summary of morphometric characteristics of the four groupsof angelfish.9 61 2 3 44.9 7.3 10.9 13.50.39 0.37 0.36 0.372.0 8.5 22.7 48.6M = 0.0172^L3.05 (r 2 =0.98)0.7 3.3 9.2 17.8mM = 0.0063 L3.08 (r 2 =0.98)GROUPLENGTH, L (cm)position ofcenter ofmass along Lbody mass, M(g)muscle mass, mM(g)wetted surface^22.2^53.8^97.0^143.0area, Sw(cm 2 )N^4^5^4^2Sw = 1.31 L 1.81 (r2 =0.98)97responses are rare and should not be pooled with away responses, due todifferences in escape trajectories (Chapter 4).C. RESULTS1. Fast-start typesAll fast-starts analysed were C-type, classifiable into two types,following Chapter 3: single bend (SB) and double bend (DB). Thesmallest fish (Group 1) performed exclusively DB responses. Fish ofintermediate size (Group 2 and Group 3) executed DB responses 62% and67% of the total, respectively. The largest fish (Group 4) showed DBresponses 27 % of the total (Fig. 5.1). A log-likelihood ratio goodnessof fit test (G test) shows that the occurrence of DB responses differssignificantly among the size groups (P<0.005). A "Tukey-type"non-parametric multiple comparison for proportions (Zar, 1984; P. 401)shows that the proportion of DB responses for the smallest (Group 1)and the largest (Group 4) fish are different from all the others andfrom each other (Fig. 5.1). Only the proportions of DB responses of thetwo intermediate size groups are not significantly different (Fig. 5.1).In summary, the occurrence of DB responses is highest in fish of Group1, intermediate in fish of Group 2 and Group 3, and lowest in fish ofGroup 4. For both intermediate size groups and the pooled data, theinitial orientation of SB responses is lower than that of DB responses(Table 5.2). Initial orientation does not differ significantly amongsize groups for both escape types and for the pooled data (DB plus SB;98Figure 5.1: Percentage of double bend responses for the different sizegroups.^Horizontal^line^underscores^percentages^that^were^notsignificantly^different at^the^5%^level^(Tukey-type^non-parametricmultiple comparison for proportions).991 0080IMFMI6040MNIMN200MINM4.9 Ii7.3^I^10.9n. s.Size group (cm)13.5100Table 5.2: Initial orientation, stage 1, stage 2 and total escapeangles for single and double bend responses of different size groups.Lines underscore means that were not significantly different at the 5%level (Tukey-type non-parametric multiple comparison). P values for thecomparisons between SB and DB responses within Group 2 (7.3 cm), Group3 (10.9 cm), and the pooled data are shown (Mann-Whitney non-parametrictest).101SIZE^(cm) 4.9 7.3 10.9 13.5 TOTALsl angle^SB(degrees)104 107 89 99DB 76 81 70 42 73test n.s. P<0.05 P<0.005pooled 76 89 80 77 82s2 angle^SB(degrees)8 0 14 8DB -27 -27 -15 -22 -23test P<0.005 P<0.05 P<0.001pooled -27 -15 -11 5 -13escape^SBangle(degrees)^DB 6110561110 108 9866 59 56test P<0.05 P<0.01 P<0.001pooled 61 82 73 87 76initial^SBorientation(degrees)^DB 9467 23 59 529998 92 166test P<0.001 P<0.005 P<0.001pooled 94 87 72 86 84N^SB 5 4 6 15DB 10 9 10 2 31pooled 10 14 14 8 46102Table 5.2).2. Turning parametersFigure 5.2 shows a negative linear correlation between size groupsand maximum total escape angle for DB responses (i = 0.98; p<0.05).Turning kinematic parameters (sl angle, s2 angle and total escapeangle) are shown in Table 5.2. Within each escape type, there is nosignificant difference among size groups (Kruskal-Wallis non-parametricanalysis of variance by ranks). For the pooled data (DB plus SBresponses), there is no significant difference among sizes for slturning angle and total escape angle. Pooled stage 2 turning angle issignificantly different among the size groups.Turning parameters for SB and DB responses are compared within eachsize group for Group 2 and Group 3, and for the pooled data(Mann-Whitney non-parametric test). Statistical comparisons of responsetype within Group 1 and Group 4 were not possible, due to the rareoccurrence of SB and DB responses in each size group, respectively.Stage 1 turning angles of SB and DB responses differ for Group 3 andthe pooled data, but are not significantly different for Group 2 (Table5.2). Values of s2 turning angles and total escape angles for SB and DBresponses differ in both size groups and in the pooled data (Table5.2). Figure 5.3 shows the ranges of total escape angles observed in SBand DB responses for the four size groups.3. Distance-time parametersTable 5.3 shows values of distance-time parameters within stage 1103Figure 5.2: The relationship between double bend maximum total escapeangle and fish size. Y=178-7.95x; r 2 =0.98; p<0.05.104180 =I^I^I^I^I^15 7 9 11 13 15Average size of each group (cm),—.u)a)^135 1.1C-o)a)-o......a)90 owcr)Ccoa)0coN 45 =Icna)CZ4-,0F-03105Figure 5.3: The range of single bend (broken arrows) and double bend(solid arrows) total escape angles for different size groups.106VVV180 mi.03I^I^I^I^I^I5 7 9 11 13 15Average size of each group (cm)107Table 5.3: Stage 1 duration and distance time parameters up to theend of stage 2. Lines underscore means that were not significantlydifferent at the 5% level (Tukey-type non-parametric multiplecomparison).108SIZE^(cm) 4.9 7.3 10.9 13.5 TOTALsl duration(s)tot. duration(s)tot. distance(m)max. velocity(m/s)max.acceleration(m/s 2)0.016 0.019 0.022 0.025 0.0200.0360.0321.3898.20.030 0.034 0.036 0.0460.026 0.030 0.033 0.0411.34 1.38 1.48 1.2693.2 78.9 114.7 109.4N 10 14 14 8 46109(sl duration) and for the whole response (total duration, totaldistance, maximum velocity, maximum acceleration) for each of the foursize groups; single bend and double bend responses are pooled. Valuesare compared across size groups (Kruskal-Wallis non-parametric analysisof variance, and Tukey-type non-parametric multiple comparison when theANOVA showed differences among size groups). Both stage 1 duration andtotal duration are significantly different among size groups. Stage 1duration is considered the best indicator of the duration of theresponse, since total duration includes fixed values for the s2duration (0.02 s) of SB responses in all size groups. This isreasonable, since the s2 duration of DB responses is not significantlydifferent among size groups. Total distance travelled is significantlydifferent among size groups, whereas both maximum velocity and maximumacceleration are not significantly different among size groups.Fast-start performance within a fixed time (0.045 s) is shown inTable 5.4. Following Webb (1976), the fixed time chosen correspondsroughly to the escape duration of the largest fish. Therefore, it isassumed that acceleration continues after s2 in smaller fish whoseescape responses are shorter than the fixed time. Distance travelledand maximum velocity within a fixed time were chosen as the mostmeaningful parameters of performance, since acceleration is independentof size whereas escape duration increases with size (Webb, 1976). Meanvalues are given for the pooled data as well as for SB and DB responsesseparately. Within each escape type, there is no significant differenceamong size groups for the maximum velocity of SB and DB responses, andfor distance travelled of SB escapes. Distance travelled for DB110Table 5.4: Distance travelled and maximum velocity attained within0.045 s from start. Lines underscore means that were not significantlydifferent at the 5% level (Tukey-type non-parametric multiplecomparison). P values for the comparisons between SB and DB responseswithin Group 2 (7.3 cm), Group 3 (10.9 cm), and the pooled data areshown (Mann-Whitney non-parametric test).111SIZE^(cm) 4.9 7.3 10.9 13.5 TOTALdistance^SBtravelled(m)^DB 0.0460.035 0.040 0.038 0.0380.0510.053 0.053 0.053test P<0.001P<0.005 P<0.05pooled 0.046 0.047 0.049 0.041 0.046Maximum^SBvelocity(m/s)^DB 1.490.99 1.18 1.15 1.101.661.73 1.79 1.64test P<0.005 P<0.01 P<0.001pooled 1.49 1.46 1.61 1.27 1.48N^SB^ 5^4^6^15DB^10^9^10^2^31pooled^10^14^14^8^46112responses is significantly different among size groups (Table 5.4). Forthe pooled data (SB and DB responses), there is no significantdifference among size groups in either distance travelled or maximumvelocity.Performance values within 0.045 s for SB and DB responses arecompared within each size for Group 2 and 3, and for the pooled data(Mann-Whitney non-parametric test; Table 5.4). In all cases,significant differences are found between SB and DB distance travelled,and SB and DB maximum velocity, with values for DB responses beinghigher than SB responses (Fig. 5.4 and 5.5). A Siegel-Tukey,non-parametric test (Gibbons, 1976) was performed to test fordifferences in the non parametric index of variability between theperformance of size groups in which both DB and SB responses occurred(Group 2, 3, 4) versus that in which only DB escapes occurred (Group1). For both distance travelled and maximum velocity, the variabilityof Group 1 is smaller than that of the other size groups combined(P<0.001 and P<0.0005, respectively).D. DISCUSSION1. The occurrence of Single bend and Double bend responsesThe occurrence of two types of response in angelfish of intermediatesize has been fully described in Chapter 3. Overall, higherdistance-time performance associated with smaller angles of turn were113Figure 5.4: The range of single bend (broken arrow) and double bend(solid arrow) distance travelled within 0.045 s for different sizegroups.114I65 —,a.5545-oC)35 NoCa)UC4—,^25cn0^3I5 7^9 13^15Average size of each group (cm)115Figure 5.5: The range of single bend (broken arrow) and double bend(solid arrow) maximum velocity within 0.045 s for different sizegroups.1162.5 NWEL()'cr^2.0 .110Oc4-,1.5 I1.0Tv^ttvIII^:^I1 4-10.53 5^7^9^11^13^15Average size of each group (cm)117found in DB than in SB responses. In Chapter 3, I suggest that thereare behavioural trade-offs between high distance-time performance andlarge turns, such that angelfish would employ a particular type ofresponse depending on their predator's strike tactics.Webb (1976) reports that the occurrence of two types of responses(S-starts^and^C-starts)^triggered^by^electrical^stimulation^wassize-dependent in trout. No significant difference in performancebetween these two fast-start types was found. S-starts are fast-startscommonly used by predators during strikes, and are also employed asescape responses by larger fish (Webb, 1976). S-starts were notobserved in angelfish. Both double bend and single bend responses areC-starts (Chapter 3).Here, startled fish were randomly oriented relative to the mechanicalstimulus. It was found that DB responses occurred at larger initialorientations than SB responses, for the two intermediate size groupsand the pooled data. Small fish showed exclusively DB responses, andlarge fish showed DB escapes only in a small percentage, such thatdifference in orientation between SB and DB responses could not betested in these two size groups.Since angelfish of different sizes escape away from a stimulusfollowing preferred trajectories (130° and 180° relative to stimulusdirection; Chapter 4), the turning angle required to achieve thesetrajectories will be large at small initial orientations, and small atlarge initial orientations, when fish are already facing away from thestimulus. Intermediate size fish employ SB responses for escape angle>90 ° (Fig.5.3). For these intermediate sizes, the initial orientation118dictates the magnitude of the escape angle and, consequently, theresponse type. Although the former is true for all sizes, since escapeangles (Table 5.2) and escape trajectories (Chapter 4) aresize-independent, the smallest fish are capable of employing DBresponses in all cases, regardless of initial orientation. In contrast,large fish showed mainly SB responses. Although the small sample sizeof DB responses for large fish did not allow for statistical testing ofthe influence of initial orientation on the occurrence of DB over thetotal number of responses, it is apparent that DB responses areemployed by large fish only at small escape angles (about 40 ° -60° ),beyond which SB responses are employed.I suggest that angelfish employ a particular type of responsedepending on their predators' strike angle. Extremely small and largefish (15 cm is considered to be the upper limit of Pterophyllum spp.total length; Sterba, 1966) use principally DB and SB responses,respectively. The use of either SB or DB responses in small and largefish, respectively, may be restricted to a narrow range of predator'sstrike angles.Here, I analyse only away responses, with escape trajectoriesrelative to a stimulus that are fixed (Chapter 4). In towardsresponses, escape trajectories are randomly distributed relative to thestimulus (Chapter 4). In this case, escape angles are not constrainedby preferred escape trajectories, and fish may always be able toperform maximally by employing DB responses. However, other kinds ofconstraints such as angular compensation (defined as the angularcorrection in the direction away from the stimulus after completion of119stage 1; Chapter 4) may apply to towards responses. Therefore, towardsand away responses should be treated separately. Unfortunately, beingrelatively rare (Blaxter et a/.1981; Eaton and Emberley, 1991; Chapter4), towards responses do not constitute large sample sizes and are notanalysed here.2. Turning kinematicsThe principle kinematic difference between DB and SB responsesoccurs in stage 2 (Chapter 3). During stage 1, angelfish assume a Cshape in both types of responses. In DB responses, stage 2 consists ofa return flip of the tail while the head turns in the oppositedirection (mean DB s2 angle for all sizes -23°) to that of stage 1. InSB responses the body straightens during stage 2 without a completerecoil of the tail, and the head continues the turn initiated in stage1 (mean SB s2 angle for all sizes +8°). As a result of the differentbehaviours during stage 2, total escape angles differ greatly betweenthe two escape types (mean SB total escape angle for all sizes 610 ;mean DB total escape angle for all sizes 105° ).Within each response type, no significant difference in s1 and totalescape angles is found among different size groups (Table 5.2). This isexpected, since initial orientations are random and not different amongsize groups (Table 5.2). Since fish of different sizes move towardssimilar escape trajectories (Chapter 4), similar stimulus directionsshould trigger similar escape angles. Significant differences are foundin the pooled (SB plus DB responses) s2 turning angle of different sizegroups. The smallest fish achieve larger negative s2 turning angles120(mean -27°) than the largest fish, whose mean s2 turning angle ispositive (+5°). Since no significant differences among sizes are foundwithin each escape type, this is likely due to the different percentageof response types (DB/Total) observed at the two extremes of the sizerange. The smallest fish employ only DB responses, which produce largenegative s2 turning angles, whereas the largest fish use mainly SBresponses which, on average, produce positive s2 turning angles.It is apparent that the ranges of total escape angles within whichSB and DB responses are employed vary with size (Fig.5.3). Inparticular, the upper limit of total escape angles for DB responses(corresponding roughly to the lower limit of total escape angles for SBresponses) decreases with size (Fig.5.2). This may explain thedifferent percentage of DB/Total observed for different size groups. Asthe upper limit of escape angles for DB responses decreases, the rangeof DB escape angle gets smaller and the incidence of DB responses inlarger fish decreases.The question remains as to whether large fish are incapable ofachieving larger DB escape angles than observed because ofmorphological and/or physiological constraints, or whether they simplydo not display DB responses at large turning angles for behaviouralreasons. Previous work on the flexibility of dead fish of various sizessuggest that larger fish are less flexible at their center of mass thansmaller fish of the same species (Aleev, 1969). However, this may notbe sufficient to explain differences in the DB response escape anglesfor fish of different size. The bending properties of dead fish may notbe the same as live fish. The fast glycolitic portion of the fish121myotome is complex (Wainwright, 1983), and fish contracting theirmusculature at a high rate in water experience forces that cannot bereproduced in simple bending experiments.Alternatively, large fish may be capable of large DB escape angles,but employ SB responses instead, as an antipredator tactic. How couldemploying a lower velocity response (SB response) be advantageous? Highvelocity is regarded as a major anti-predator adaptation during escaperesponses (Weihs, 1973; Webb, 1976; 1986b; Harper and Blake, 1990).However, velocity in escape responses can be submaximal. Recently, itwas shown that high performance in some kinematic parameters such asturning (Chapter 3) and pitch (Kasapi et al. 1993) may be preferredover high distance-time performance when the two trade-off. Highvelocity may therefore not always be beneficial. It may be that forfish escaping at large turns, it is better not to continue swimming butto stop abruptly, thereby confusing the predator. This could beparticularly true for complex environments such as those inhabited byangelfish. Angelfish have very good deceleration performance (-50 to-80 m s - 2; Personal observations), probably due to the action of largepectoral fins which can be employed for braking. However, why SBresponses occur mainly in larger fish is more difficult to explain froma behavioural perspective. Perhaps angelfish habitat selection issize-related. More complex environments would require manoeuvringwithin shorter distances than in open waters.3. Distance-time performanceAngelfish are specialized for low speed swimming in the labriform122(pectoral fin) mode. However, their fast-start performance isrelatively high when compared to other species (Chapter 3). Maximumvelocities of angelfish fall within the range of values observed introut of similar size (Webb, 1976). However, the accelerationperformance of "generalist" forms (e.g. trout) is lower than that forangelfish.Employing excised muscle tissue from a variety of fish species,Wardle (1975) reported durations of single contractions (which I assumeto be equivalent to stage 1 duration) that are shorter than thosedetermined here for angelfish of similar size at comparabletemperatures (approximately 20 °C ). This may imply that the forcenaturally experienced by intact fish during a fast-start is differentfrom that of stimulated excised muscle blocks.Fish escape responses were thought to involve maximum acceleration(Webb, 1986b), as opposed to predators' strikes which can be submaximal(Webb, 1984b; Harper and Blake, 1991). Here, I document the presence oftwo response types with different performance levels (Fig.5.4 and 5.5).In addition, the occurrence of the two response types issize-dependent, and this can affect the overall performance of eachsize group. When examining the performance within the whole response(Table 5.3), size has a positive effect on both distance travelled andduration. Although total duration has been calculated including fixedvalues of SB s2 duration (the end of s2 is not determinable for SBresponses; Chapter 3), sl duration also increases with size. Thisconfirms Webb' s (1976) findings on the effect of size on troutfast-starts. Acceleration is also size-independent, as found by Webb123(1976) for trout. Although acceleration should theoretically vary withLength - (Daniel and Webb, 1987), Webb and Johnsrude (1988) suggestthat this may not be so because of summation of muscle twitches. Here,maximum velocity is size-independent, likely the result of a largerproportion of low performance (SB) responses observed in largerangelfish. This contrasts with previous studies on fusiform fish(Wardle, 1975; Webb, 1976). Whether this is due to morphologicaldifferences or differences in methodology (previous studies onfast-start^performance^do^not^report^if^the^fish^initialorientation relative to the stimulus was random) cannot be determined.Since maximum acceleration is independent of size and the durationof fast-starts increases with size, comparison of performance betweengroups should be made for a given time, as suggested by Webb (1976).Within each escape type, both maximum velocity and distance travelledare size-independent, with the exception of DB escapes, where distancetravelled is significantly different between Group 1 and Group 2 (Table5.4). The lower performance of small fish in DB response distancetravelled may be due to the fact that calculating performance within afixed time (0.045 s) assumes that acceleration continues after s2 insmaller fish whose escape responses are shorter than the fixed time(Webb, 1976). Nevertheless, the distance travelled and maximum velocityperformance for both response types pooled are size-independent (Table5.4). This means that the higher percentage of SB responses observed inthe largest fish does not compromise their overall performancesignificantly. It is important to notice that the smallest fish canemploy high performance (DB) responses at all the escape angles124observed, whereas larger fish utilize an increasing proportion of lowperformance (SB) responses. The performance of the smallest fish istherefore more consistent than that of larger fish that utilize both SBand DB responses.Low performance responses at large escape angles by large fish mayhave consequences for their interactions with predators. Predators maylearn to attack large fish at small angles of initial orientation, whenSB responses are likely to occur (Table 5.2). On the other hand, largeangelfish could avoid these initial orientations relative to theirpredators. In addition, large angelfish may select particularly complexenvironments, where the absence of a full return flip during stage 2may not be detrimental.Prey selectivity by predators has been assumed to depend on theratio of gain to cost of capturing a prey, where both cost and gainincrease with prey size at different rates (Harper and Blake, 1988).Although linear performance has been assumed to be a major factordetermining the prey's ability of escaping predators (Harper and Blake,1988), it is apparent from my results that turning ability or otherkinematic variables (Kasapi et al. 1993) must be at least as important.If we assume that both high linear performance and escape trajectoriesare important in angelfish escape responses, it follows thatcompromising the former at large angles of escape (in SB responses)must be due to morpho-physiological constraints rather than behaviouralchoice.Concluding, linear performance is not sufficient to evaluate escapeperformance. Linear performance must be considered together with other125kinematic parameters such as turning angle (Daniel and Mayhtifer, 1989),turning radius (Howland, 1974; Webb, 1976; Chapter 3), pitch and roll(Kasapi et al. 1993). In addition, biomechanical considerations are notalways sufficient to explain escape behaviour. Here I show that abehavioural factor (i.e. initial orientation) may have importanteffects on swimming performance. Other behavioural factors such asresponse latency (Eaton and Hackett, 1984), reaction distance (Dill,1974a), and sensory responsiveness (Blaxter and Fuiman, 1990) may playmajor roles in the ability of fish to escape predation, and should betaken into consideration as integral parts of escape responseperformance.126CHAPTER 6SummaryThis work investigates the effect of specialization for low speedswimming on fast-start performance (velocity, acceleration). For thepurpose, angelfish (Pterophyllum eimekei) were chosen as an example offish that routinely swim at low speed, employing labriform locomotion.Overall, higher distance-time performance associated with smallerturning angles were found in Double Bend escape responses (DB) whencompared to Single Bend escape responses (SB). Contrary to previouslypublished theories (Webb, 1984a, 1986b), I found that angelfish performrelatively well in fast-start when compared to other fish previouslystudied. Specialization for low speed swimming does not compromisefast-start performance. Therefore angelfish are well designed for twodifferent locomotory modes. During low speed swimming, the body iskept rigid and the pectoral fins are moved by aerobic musculature. Incontrast, during fast-starts, the body is bent employing the anaerobicaxial musculature. Webb's (1984a) ideas on form and function are basedon locomotory modes employed when feeding. However, fish that routinelyfeed swimming at low speed, often maintain a "decoupled system" (axiallocomotion) that allows them to perform high accelerative bursts whenattacked by predators. Optimal designs for decoupled systems do notnecessarily "trade-off".127Secondly, I have investigated the turning pattern of angelfishreacting to a mechanical stimulus. Past studies of escape trajectoriesin fish and other animals have employed linear plots of stimulus angleversus body turning angle. Here, I defined escape trajectories as acircular variable. My main finding is that angelfish escape in fixeddirections away from the stimulus. Reanalysis of published work revealsthat other animals also show multiple escape trajectories. It hasbeen suggested that turning angles are preprogrammed at the onsetof the escape response (Eaton et al. 1991). However, escapetrajectories relative to a stimulus vary considerably after theinitial turn away from it (Eaton et al. 1981). Here, I show thatthis variability reflects multiple preferred trajectories. This mayprevent predators from learning any single fixed prey response andcompensating for it. In addition, angelfish correct their responseswhen turning towards the stimulus, suggesting that escape trajectoriesmay be modulated by sensory feedback.As fish escape trajectories are fixed relative to the stimulus, theturning angle that the fish has to achieve from a particular initialorientation will vary. Using fish of different sizes, I found thatsmall fish perform equally well (in terms of distance travelled andvelocity) at any turning angle. However, the performance of large fishis impaired when large turning angles are employed. Therefore, althoughlarge fish are expected to swim faster when escaping in a linearfashion (Wardle, 1975), this may not hold true when a turn is involved.The low performance responses at large turning angles performed bylarge angelfish may be due to morpho-physiological constraints rather128than behavioural choice. I suggest that angelfish employ a particulartype of response (SB or DB) depending on their predator strike angle.129REFERENCES:Alcock, J. (1989). Animal Behavior. Sinauer Press, Massachusetts.Aleev, Yu G. (1969). Function and Gross Morphology in Fish. KaterPress, Jerusalem.Bainbridge, R. (1958). The speed of swimming of fish as related to sizeand to the frequency and amplitude of tail beat. J. Exp. Biol. 35:109-133.Batschelet, E. (1981). Circular Statistics in Biology. Academic Press,New York.Bergmann,^V.H.H.^(1968).^Eine^descriptive^Verhaltenanalyse^desSegelflossers (Pterophyllum scalare). Zeit. Tierpsych. 25: 559 -587.Blake, R.W. (1979). The mechanics of labriform locomotion I. Labriformlocomotion in the Angelfish (Pterophyllum eimekei): an analysis of thepower stroke. J.Exp.Biol. 82: 255-271.Blake, R.W. (1981). Influence of pectoral fin shape on thrust and dragin labriform locomotion. J. of Zool., London, 194: 53-66.Blake, R.W. (1983). Fish Locomotion. Cambridge University Press.Blaxter, J.H.S. & Fuiman, L.A. (1990). The role of the sensory systemsof herring larvae in evading predatory fishes. J.Mar.Biol.Ass. U.K. 70:413-427.Blaxter, J.H.S., Gray, J.A.B. & Denton, E.J. (1981). Sound and startleresponses in herring shoals. J.Mar.Biol.Ass.U.K. 61: 851-869.Bullock,^T.H.^(1984).^Comparative^neuroethology^of^startle,^rapidescape, and giant fiber-mediated responses. In Neural mechanisms ofstartle behavior. Eaton, R.C. (ed.), pp 1-13. Plenum Press, N.Y.Burton, M. & Burton, R. (1975). Encyclopedia of fish. Octopus Book,London.130Camhi, J. M. & Tom, W. (1978). The escape behavior of the cockroachPeriplaneta americana. I. Turning response to wind puffs. J. comp.Physiol. A 128: 193 -201.Camhi, J. M. & Levy, A (1988). Organization of a complex movement:fixed and variable components of the cockroach escape behavior. J.comp. Physiol. A 163: 317 -328.Camhi, J. F. & Levy, A. (1989). The code for stimulus direction in acell assembly in the cockroach. J. comp. PhysioL A 165: 83-97.Camhi, J. M., Tom, W. & Volman, S. (1978). The escape behavior of thecockroach Periplaneta americana. II. Detection of natural predators byair displacement. J. comp. Physiol. A 128: 203-212.Canfield, J. G. & Eaton, R. C. (1990). Swimbladder acoustic pressuretransduction initiates Mauthner-mediated escape. Nature 347: 760-762.Corner, C. M. & Dowd, J. P. (1987). Escape turning behavior of thecockroach. Changes in directionality induced by unilateral lesions ofthe abdominal nervous system. J. comp. Physiol. A 160: 571-583.Covell, W.C., Smith, D.M., Harper, D. & Blake, R.W. (1990). Skeletalmuscle deformation in the lateral muscle of the intact rainbow troutSalmo gairdneri during fast-start maneuvers. J.Exp.BioL 156:^453 -466.Dagan, D. & Camhi, J. F. (1979). Responses to wind from the cercal nerveof the cockroach Periplaneta americana. II. Directional selectivity ofthe sensory neurons innervating single columns of filiform hairs. J.comp. Physiol. A 133: 103 - 110.Daniel, T.L. (1984). Unsteady aspects of aquatic locomotion. Amer.Zool.24: 121-134.Daniel, T.L. & Mayhtifer, E. (1989). Size limits in escape locomotion ofcarridean shrimp. J.exp.Bio1.143: 245-265.Daniel, T.L. & Webb, P.W. (1987). Physical determinants of locomotion.In Comparative Physiology: Life in Water and on Land (ed. P. Dejours,L. Bolis, C.R. Taylor and E.R. Weibel), pp. 343-369. Padova: LivianaPress.131DiDomenico, R., Nissanov, J. & Eaton, R.C. (1988). Lateralization andadaptation of a continuously variable behavior following lesions of areticulospinal command neuron. Brain Res. 473: 15-28.Dill, L. M.^(1974a). The escape response of the zebra danio(Brachydanio rerio) I. The stimulus for escape. Anim. Behay. 22:710-721.Dill, L. M.^(1974b). The escape response of the zebra danio(Brachydanio rerio) II. The effect of experience. Anim. Behay. 22:723-730.Domenici, P. & Blake, R. W. (1991). The escape response in theangelfish, Pterophyllum eimekei. J. mar. biol. Ass. U.K. 71: 712.Dubois, A.B., Cavagna, G.A. & Fox, R.S. (1976). Locomotion of bluefish.J.Exp. Zool. 195: 223 -226.Eaton, R.C., &^Hackett J.T. (1984). The role of Mauthner cell infast-starts^involving escape in teleost fish. In Neural mechanisms ofstartle behavior. Eaton, R.C. (ed.), pp. 213-266. Plenum Press, N.Y.Eaton, R.C., Bombardieri, R.A. & Meyer, D.H. (1977). The Mauthnerinitiated startle response in teleost fish. J.Exp.Biol. 66: 65-81.Eaton, R.C., Lavender, W.A. & Wieland, C.M. (1981). Identification ofMauthner-initiated response pattern in goldfish: Evidence fromsimultaneous cinematography and electrophysiology. J.Comp.Physiol. 144:512-531.Eaton, R.C., Nissanov, J. & Wieland, C.M.(1984). Differential activationof Mauthner^and^non-Mauthner^startle^circuits^in^the^zebrafish:Implication^for functional substitution. J.Comp.Physiol.(A)155: 813-820.Eaton, R.C., DiDomenico, R. & Nissanov, J. (1988). Flexible bodydynamics^of^the^goldfish^C-start:^implication^for^reticulospinalcommand mechanisms. J. Neurosci. 8: 2758 -2768.Eaton, R.C. & Emberley, D. S. (1991). How stimulus direction determinesthe trajectory of the Mauthner-initiated escape response in a teleost132fish. J. exp. Biol. 161: 469-487.Eaton, R.C., DiDomenico, R & Nissanov, J. (1991). The role of theMauthner cell in sensorimotor integration by the brainstem escapenetwork. Brain, Behay. Evol. 37: 272-285.Faber, D. S., Fetcho, J. R. & Korn, H. (1989). Neuronal networksunderlying the escape response in goldfish: general implications formotor control. Ann. N. Y. Acad. Sci. 563: 11-33.Frith, H.R. & Blake, R.W. (1991). Mechanics of the startle response innorthern pike, Esox lucius. Can.J.Zool. 69: 2831-2839.Gamperl, A.K., Schnurr, D.L. & Stevens, E.D. (1991). Effect of asprint-training protocol on acceleration performance in rainbow trout(Salmo gairdneri). Can.J.Zool. 69:578-582.Gibbons, J.D. (1976). Nonparametric Methods for Quantitative Analysis.Holt, Rinehart and Winston Press, New York.Gillette, R. (1987). The role of neuronal command in fixed actionpattern of behavior. In Aim and methods in neuroethology. D.M.Guthrie(ed.), pp. 46-79. Manchester Univ. Press, Manchester.Hall, S. J., Wardle, C. S. & MacLennan, D. N. (1986). Predator evasionin a fish school: test of a model for the fountain effect. Mar. Biol.91: 143-148.Harper,^D.G.^& Blake,^R.W.^(1988).^Energetics^of pescivorouspredator-prey interactions. J.theor. Biol. 134: 59-76.Harper, D.G. & Blake, R.W. (1989). On the error involved in high-speedfilm when used to evaluate maximum acceleration of fish.Can.J.Zoo1.67: 1929-1936.Harper, D.G. & Blake, R.W. (1990). Fast-start performance of rainbowtrout Salnw gairdneri and northern pike Esox lucius, during escapes.J.Exp.Biol. 150: 321-342.Harper, D.G. & Blake, R.W. (1991). Prey capture and the fast-startperformance of northern pike Esox lucius. J.exp.Biol. 155: 175-192.133Hawkins, A. D. & Horner, K. (1981). Directional characteristics ofprimary auditory neurons from the cod ear. In Hearing and soundcommunication in fishes (ed. Tavolga, W.N., Popper, A.N. & Fay, R.R.) pp. 311-327. New York: Springer-Verlag.Howland, H.C. (1974). Optimal strategies for predator avoidance: Therelative importance of speed and manoeuvrability. J.Theor.Biol. 47:333-350.Hurley, A. C. & Hartline, P. H. (1974). Escape response in thedamselfish Chromis cyanea (Pisces: Pomacentridae):^A quantitativestudy. Anita Behay. 22: 430-437.Kasapi, M.A., Domenici, P., Blake, R.W. & Harper, D.G. (1993). Thekinematics and performance of the escape response in the knifefish,Xenonzystis nigri. Can.J.Zool. (In press).Kinsler, L. E., Frey, A. R., Coppens, A. B. & Saunders, J. V. (1982).Fundamentals of Acoustics. J. Wiley and Sons, New York.Lanczos, C. (1956). Applied Analysis. Englewood Cliffs, New Jersey:Prentice Hall.Lindinger, M.I., Heigenhauser, G.J.F., McKelvie, R.S. & Jones, N.L.(1990). Role of nonworking muscle on blood metabolites and ions withintense intermittent exercise. Am. J. Physiol. 258: R1487-R1494.Myrberg, A.A.Jr., Kramer, E. & Heineke, P. (1965). Sound production bycichlid fishes. Science 149: 555-558.Nalbach, H.-0. (1990). Discontinuous turning reaction during escape insoldier crabs. J. exp. Biol. 148: 483-487.Nalbach, H.-O. & Nalbach, G.^(1987). Distribution of optokineticsensitivity over the eye of crabs: its relation to habitat and possiblerole in flow-field analysis. J. comp. Physiol. A 160: 127-135.Nissanov, J. & Eaton, R.C. (1989). Reticulospinal control of rapidescape turning manoeuvres in fishes. Am.Zool. 29: 103-121.Ritzmann, R. E. (1984). The cockroach escape response. In Neural134Mechanisms of Startle Behavior (ed. R. C. Eaton), pp. 93-131. (PlenumPress, New York).Russell, I. J. (1976). Central inhibition of lateral line input in themedulla of the goldfish by neurones which control active bodymovements. J. comp. Physiol. A 111: 335-358.Schuijf, A. (1975). Directional hearing in the cod under approximatefree field conditions. J. comp. Physiol. 98: 307-332.Schultz, L.P. (1967). Review of South American Freshwater Angelfish -Genus Pterophyllurn. Proc. of U.S. Nat. Museum 120: 1-10.Sokal, R. R. & Rohlf, F. J. (1981). Biometry. W. H. Freeman & Co., SanFrancisco.Sterba, G. (1966). Freshwater fishes of the world. Studio Vista Press,LondonTaylor, E.B. & McPhail, J.D. (1985). Prolongued^and burst swimming inanadromous^and^freshwater^threespine^stickleback,^Gasterosteusaculeatus. Can. J. ^64: 416-420.Wainwright, S.A. (1983). To bend a fish. In Fish Biomechanics. (ed.P.W. Webb and D. Weihs), pp. 68-91. Praeger Press, New York.Wardle, C.S. (1975). Limit of fish swimming speed. Nature, London255: 725-727.Wardle, C.S. (1977). Effect of size on the swimming speeds of fish. InScale effects in animal locomotion. (ed. T.J. Pedley), pp. 315-331.Academic Press, London.Wardle, C.S. (1986). Fish behaviour and fishing gear. In The Behaviourof Teleost Fishes (ed. T. J. Pitcher), pp. 463-495 (Croom Helm,London).Webb, P.W. (1975). Acceleration performance of rainbow trout S.gairdneri and green sunfish Lepomis cyanellus. J.Exp.Biol. 63: 451-465.Webb, P.W. (1976). The effect of size on fast-start performance of135rainbow^trout^S.gairdneri,^and^a^consideration^of^piscivorouspredator-prey interaction. J.Exp.Biol. 65: 157-177.Webb, P.W. (1977). Effects of median fin amputation on fast-startperformance of rainbow trout (S.gairdneri). J.Exp.Biol. 68: 123-135.Webb, P.W. (1978). Fast-start performance and body form in sevenspecies of teleost fish. J.Exp.Biol. 74: 211-226.Webb, P.W. (1982). Avoidance responses of fathead minnow to strikes byfour teleost predators. J.Comp.Physiol. 147: 371-378.Webb, P.W. (1983). Speed, acceleration, and manoeuvrability of twoteleost fishes. J.Exp.Biol. 102: 115-122.Webb, P.W. (1984a). Body form, locomotion and foraging in aquaticvertebrates. Am. Zool. 24: 107-120.Webb, P.W. (1984b). Body and fin form and strike tactics of four teleostpredators^attacking^fathead^minnow^(Phimapales^promelas).Can.J.FishAq.Sci. 41: 157-165.Webb, P.W. (1986a). Effect of body form and response threshold on thevulnerability of four species of teleost prey attacked by largemouthbass. Can.J.FishAq.Sci. 43: 763-771.Webb, P.W. (1986b). Locomotion and predator-prey relationship. InPredator-Prey Relationship. G.V. Launder and M.E.Feder (eds.), pp.24-41. Univ. of Chicago Press, Chicago.Webb, P. W. & Skadsen, J. M. (1980). Strike tactics of Esox. Can. J.Zool. 58: 1462-1469.Webb, P.W. & Keyes, R.S. (1981). Division of labour between median finsin swimming dolphin (Pisces: Coryphaenidea). Copeia 1981: 901-904.Webb, P.W. & Blake R.W. (1985). Swimming. In Functional VertebrateMorphology. M.Hildebrand (ed.), pp. 110-128. Harvard Univ. Press,Cambridge, Mass.Webb, P.W. & Johnsrude, C.L. (1988). The effect of size on the136mechanical properties of the myotomal-skeletal system of rainbowtrout (Salmo gairdneri). Fish Physiol. Biochem. 5: 163-171.Webb, P.W., Sims,D, & Schultz, W.W. (1991). The effects of an air/watersurface on the fast-start performance of rainbow trout (Oncorhynchusmykiss). J.exp.Biol. 155: 219-226.Weihs, D. (1973). The mechanism of rapid starting of slender fish.Biorheology 10: 343-350.Weihs, D. & Webb, P.W. (1984). Optimal avoidance and evasion tactics inpredator-prey interactions. J.Theor.Biol. 106: 189-206.Westin, J. (1979). Responses to wind recorded from the cercal nerve ofthe cockroach Periplaneta americana. I. Response properties of singlesensory neurons. J. comp. Physiol. A 133: 97-102.Wine, J.J.& Krasne, F.B. (1972). Organization of escape behaviour inthe crayfish. J. Exp. Biol. 56: 1-18.Zar, J. H.^(1984). Biostatistical Analysis. Prentice Hall, EaglewoodCliffs, New Jersey.137


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items