Locomotor Biomechanics and Behaviour in the Ocellate River Stingray

Home , Cowtail stingray, Ocellate river stingray, Round stingray

University of Calgary PRISM: University of Calgary's Digital Repository

Graduate Studies The Vault: Electronic Theses and Dissertations

2020-04-22 Locomotor biomechanics and behaviour in the ocellate river stingray

Seamone, Scott G.

Seamone, S. G. (2020). Locomotor biomechanics and behaviour in the ocellate river stingray (Unpublished doctoral thesis). University of Calgary, Calgary, AB. http://hdl.handle.net/1880/111920 doctoral thesis

University of Calgary graduate students retain copyright ownership and moral rights for their thesis. You may use this material in any way that is permitted by the Copyright Act or through licensing that has been assigned to the document. For uses that are not allowable under copyright legislation or licensing, you are required to seek permission. Downloaded from PRISM: https://prism.ucalgary.ca UNIVERSITY OF CALGARY

Locomotor biomechanics and behaviour in the ocellate river stingray

Scott G. Seamone

A THESIS SUBMITTED TO THE FACULTY OF GRADUATE STUDIES IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

GRADUATE PROGRAM IN BIOLOGICAL SCIENCES

CALGARY, ALBERTA

APRIL, 2020

Abstract

Stingrays are fishes that are dorsoventrally flattened in the same plane as the substrate, similar to a hydrofoil, with long thin tails that have an absent or reduced caudal fin, and anterior to the pelvic girdle the longitudinal body axis is relatively rigid. These characteristics would appear to constrain or preclude many of the locomotor behaviours that are employed by fishes that typically swim via undulations of the longitudinal body axis and caudal fin, and which tend to dominate descriptions of fish swimming in the literature. In contrast, stingrays exhibit a variety of locomotor behaviours powered via enlarged and flexible pectoral fins that wrap around the body and head (i.e. the pectoral disc), yet an in depth understanding of the biomechanical mechanisms that permit these behaviours has not been formed. Potamotrygon motoro, the ocellate river stingray, lives along the substrate in a benthic environment, and possesses an extremely rounded pectoral disc, from the dorsal view. It is used in these studies to represent the flattened shaped, low profile, and relatively rounded disc common to benthic stingrays, to better understand how these animals achieve different locomotor behaviours. The studies described in this thesis offer insight into how the shape of P. motoro is employed to accomplish behaviours exhibited by many benthic stingrays such as fast- start maneuverability, station holding and burying. Chapter 1 reviews our current and somewhat limited understanding of how shape impacts swimming behaviour in fishes that are flattened in the same plane as the substrate, described here as foil fishes, and explores relationships of shape and ecology observed in stingrays. Chapter 2 describes studies where video analysis was used to reveal that flexibility in the movements of the pectoral fins around the flattened and nearly symmetrical disc shape permits fast-start escape in all directions across the benthic plane with similar performance, regardless of initial orientation of the fish, which appears to challenge the conventional description of maneuverability typically used to evaluate fishes. Chapter 3 describes studies where recordings of changes of pressure beneath the pectoral disc, and video observations of movements of dye, are used to argue that stingrays can exercise movements of the body and fins to flush water from beneath the ventral surface to create and maintain a seal between the pectoral disc and benthos, to achieve suction pressures via a vacuum

ii and possibly Stefan adhesion, that can resist an upwards displacing force to hold station along the benthos. Chapter 4 describes studies that used video analysis and particle image velocimetry to explain how rapid and vigorous movements of the body and fins in stingrays fluidize and suspend vortices of sediment below the ventral surface of the fins, which are then directed up and over onto the dorsal surface to cover the fish with sediment and effect burying, and that the fish appear to direct and control these vortices to modulate the extent and pattern of burying. Chapter 5 describes studies that used time-lapse photography and video analysis to reveal that in the presence of sediments that differ in grain size, stingrays mostly choose to inhabit and bury in finer grained sediments when threatened, and this appears to reflect these fishes being more effective at burying in finer sediments, such that the rate of coverage of the dorsal surface is faster for a given finbeat speed. Chapter 6 provides a summary of what has been revealed by these studies, conclusions and future directions. These studies advance our understanding of how a flattened and rounded disc shape in P. motoro might find success in a benthic environment, and might inspire engineers interested in fish for the design of underwater robotics.

Keywords: biomechanics, video, locomotion, behaviour, stingray, particle image velocimetry

iii

Acknowledgements

This research was supported by a Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery Grant to DAS, and an NSERC Alexander Graham Bell and University of Calgary Killam Scholarship to SGS. Thank you to the support of my supervisor DAS, my committee members JEAB and JRP, the staff at UofC, my lab mates JR, TM and NT, and my family and friends APR, BJS, CAS, CBS, CDS, CKM, CMLU(TB), JDW, JNH, MSDS, MTA, MES, MS, NPFD, RWH, SAS and WRN.

Table of contents

Title page…………………………………………………………………………………………i Abstract……….…………………………….…………………………….……………………...ii Acknowledgements…………………………….………………………….…………………...iv List of figures and tables…….....……………….…………………………………………….vii Chapter 1: Swimming behaviours in stingrays and other foil-shaped fishes……………..1 - Abstract…………………………………………………………………………..1 - Introduction………………………………………………………………………2 - Shape and swimming in fish………………..………………………………….3 - Foil-shaped fishes ..…….……………………………………..………………..6 - The pectoral disc of stingrays ………………………………..………………17 - Thesis chapters………………………………...……….……..………………22 Chapter 2: Disc starts: the pectoral disc of the ocellate river stingray (Potamotrygon motoro) promotes omnidirectional fast starts across the substrate..…..……...…………25 - Abstract…………………………………………………………………………25 - Introduction …………………………………………………………………… 26 - Materials and methods…….………………………………………………….29 - Results………………………………………………………………………….36 - Discussion……………………………………………………………………...45 Chapter 3: The ocellate river stingray (Potamotrygon motoro) can generate suction to station hold along the substrate……………………………………………………………...56 - Abstract…………………………………………………………………………56 - Introduction……………………………………………………………………..57 - Materials and methods………………………………………………………..60 - Results………………………………………………………………………….64 - Discussion……………………………………………………………………...72 Chapter 4: The ocellate river stingray (Potamotrygon motoro) exploits vortices of sediment to bury into the substrate………………………………………………………….77 - Abstract…………………………………………………………………………77

- Introduction……………………………………………………………………..78 - Materials and methods………………………………………………………..83 - Results………………………………………………………………………….89 - Discussion…………………………………………………………………….105 Chapter 5: Sediment selection by the ocellate river stingray (Potamotrygon motoro) is associated with constraints in burying……………………………………………………..114 - Abstract………………………………………………………………………..114 - Introduction……………………………………………………………………115 - Methods……………………………………………………………………….118 - Results………………………………………………………………………...124 - Discussion…………………………………………………………………….134 Chapter 6: A rounded disc-shaped foil can be an effective shape for a benthic environment…………………………………………………………………………………..139 - Summary……………….……………...…………………………..………….139 - Conclusions and future directions……………………………………..…...146 References……………………………………………………………………………………150

List of figures and tables

Chapter 1: Swimming behaviours in stingrays and other foil-shaped fishes - Fig. 1. The ocellate river stingray (Potamotrygon motoro)………………..24 Chapter 2: Disc starts: the pectoral disc of the ocellate river stingray (Potamotrygon motoro) promotes omnidirectional fast starts across the substrate - Fig. 1. Disc starts: forwards, sideways and backwards escape…………..37 - Fig. 2. Body is not restricted to following the orientation of the head..…..39 - Fig. 3. Impacts of prod speed on translational performance…..……..…...42 - Fig. 4. Impacts of prod speed on yaw-rotational performance…………....46 Chapter 3: The ocellate river stingray (Potamotrygon motoro) can generate suction to station hold along the substrate - Fig. 1. Impacts of lifting force and duration of suction event on suction pressure……………..……………………………………..65 - Fig. 2. Relationships of body thrust and finbeat motions……………….....66 - Fig. 3. Impacts of finbeat and body thrust number and frequency on suction pressure………………………………….…………..…... 67 - Fig. 4. Impacts of body thrust velocity and acceleration on suction pressure…………………………………………………....68 - Fig. 5. Lateral view of fins flushing dye from beneath the pectoral disc....70 - Fig. 6. Dorsal view of fins and branchial jets flushing dye from beneath the pectoral disc…………….……………..…..………71 Chapter 4: The ocellate river stingray (Potamotrygon motoro) exploits vortices of sediment to bury into the substrate - Fig. 1. Particle image velocimetry and anterior view of a burying event…90 - Fig. 2. Particle image velocimetry and dorsal view of a burying event…. 92 - Fig. 3. Sediment coverage of head, body, fins and tail locations……..….96 - Table 1. P-values for coverage of head, body, fins and tail locations...…97 - Fig. 4. Average sediment coverage of four individual P. motoro…………98 - Fig. 5. Relationships of number and frequency of body pumping and finbeat motions……………………………………………….... 100

vii

- Fig. 6. Impacts of speed and displacement of body pumping on sediment speed…………………………………………….……..102 - Fig. 7. Impacts of speed and displacement of finbeat motions on sediment speed………………………………...…………………103 - Fig. 8. The extent of burying……………………………………….……..…104

Chapter 5: Sediment selection by the ocellate river stingray (Potamotrygon motoro) is associated with constraints in burying - Fig. 1. Percentage of time spent in different grain sizes……………..…..125 - Fig. 2. Percentage of threat events in different grain sizes…………..….126 - Fig. 3. Percentage of burying events in different grain sizes…………....127 - Fig. 4. Sediment coverage in different grain sizes………………………..129 - Fig. 5. Duration of burying events in different grain sizes………..………130 - Fig. 6. Rate of sediment coverage in different grain sizes...... 131 - Fig. 7. Finbeat speed while burying in different grain sizes……………..132 - Fig. 8. Effectiveness of burying in different grain sizes……………..……133

viii

Chapter 1

Swimming behaviours in stingrays and other foil-shaped fishes

Abstract

A foil is a flattened object that when placed in a moving fluid at a suitable angle of attack the lifting force is substantially larger than the force due to drag. In fishes, a flattened body shape in the same plane as the substrate, like a foil, has convergently evolved in elasmobranchs (e.g. batoid rays, skates, sawfish, guitarfish and electric rays, and some species of shark) and pleuronectifomes (e.g. flatfish such as flounder, sole and halibut). In this dissertation, I refer to these fish collectively as foil-shaped fishes, because 1) they generate low drag when oriented parallel to the substrate, and 2) they generate, and are reliant on, lifting forces to achieve vertical equilibrium in the water column when swimming. Despite a relatively small number of species, foil-shaped fishes are quite common in benthic environments worldwide, and they demonstrate considerable diversity in swimming behaviour. These fishes power swimming via axial bending of the body and caudal fin, or propulsive waves passed along median and paired fins while the body remains relatively rigid. Moreover, fins with low aspect ratios that achieve relatively high drag and low lift to power drag-based propulsion are most common in benthic species, whereas fins with high aspect ratios that achieve relatively high lift and low drag appear to power lift-based propulsion, which is performed by a limited number of species that tend to be pelagic (e.g. eagle rays, pelagic stingrays and cownose rays). Although a foil shape parallel to the substrate appears to give rise to intriguing behaviours in an aquatic environment, such as gliding, punting and walking, ground effect, fast-start maneuverability, station holding, and burying, we lack an in-depth understanding of the biomechanical mechanisms that promote these behaviours in most foil fishes, such as in stingrays.

Keywords: foil fishes, pleuronectiformes, batoids, stingrays, glide, punt, ground effect, fast starts, station hold, bury

Introduction

Fish demonstrate exceptionally diverse mechanisms to interact with water, to accomplish a wide range of locomotor behaviours that promote survival in nearly all aquatic environments, including benthic, demersal and pelagic ecosystems in lakes, ponds, rivers, and the ocean, making these animals an intriguing model to advance our understanding of the interrelationship between shape, biomechanics, behaviour and habitat in water. Water is nearly incompressible, denser and more viscous when compared to air, evoking the potential for greater forces of drag and buoyancy that counter forces of thrust and gravity, respectively (Webb, 1974), and fishes can effectively swim through water by using their bodies and fins to disturb the fluid that surrounds them, creating action reaction forces and differences in pressure to generate thrust (Webb, 1988). Now, with the accessibility of high-speed video and particle image velocimetry, we can visualize and quantify the biomechanics of locomotor behaviours in fish, such as the rapid kinematics of the body and fins, and the resultant dynamics of fluid around these animals, to describe processes that might link shape to ecology.

Accordingly, fish locomotion seems to interest more than just biologists, but also engineers finding inspiration for the design of underwater vehicles via biomimicry.

A flattened object, that when placed in a moving fluid at an angle of attack generates substantially greater lift (i.e. a force generated perpendicular to the fluid flow) than drag (i.e. the retarding force generated parallel to the fluid flow), is commonly referred to as a foil in fluid mechanics. A relatively small percentage of fishes, such as

2 stingrays (suborder Myliobatoidei), are flattened like a foil in the same plane as the substrate. Our understanding of the consequences to the biomechanics of locomotion in these foil-shaped fishes has not been formed. This dissertation stems from an interest in how a disc shaped foil common to stingrays that live in a benthic environment is used to power various facets of locomotion, and with the recognition that using biomechanics to describe mechanisms of many of the behaviors that these stingrays exhibit along the benthos of an aquatic environment, such as escape, station holding and burying, has not been attempted. Here, I will review concepts and terms that are important in understanding fish swimming, the diversity and versatility of a flattened foil shape in fish, shape and locomotion in stingrays specifically, and then a summary of the chapters in this dissertation.

Shape and swimming in fish

We tend to categorize fish shape, mechanisms of movement and resultant swimming behaviours by making generalizations about each. This results in the number of such categories being much less than the actual diversity and complexity in the shape and function of locomotor behaviours observed in these aquatic animals.

While this may misrepresent, and perhaps mislead our understanding of the true diversity in form and function present in fish, attempts to categorize and generalize about the impacts that shape can have on locomotion can be beneficial when attempting to comprehend and communicate about how shape and movement might influence hydrodynamics, energetics and performance in water, particularly in an effort to identify mechanisms of behaviours that might link shape to ecological success.

Swimming movements in fishes are commonly divided into two categories, one driven primarily by propulsive movements of the body and caudal fin (BCF swimming), and the other driven primarily by movements of single or multiple sets of median or paired fins

(MPF swimming) (Blake, 2004; Sfakiotakis et al., 1999; Webb, 1984a). These movements may power drag-based or lift-based thrust. Lift-based thrust is driven via asymmetries in flow around a surface, that generates a pressure difference across the surface normal to the incident flow, which can create thrust at an appropriate angle of attack (Fish, 1996; Webb, 1988). In contrast, drag-based thrust is induced by a surface essentially dragging and pushing backwards through the water, and hence, exerts a backwards force on the water, which generates an action-reaction force that thrusts the fish forwards (Fish, 1996; Webb, 1988). A transition from drag-based propulsion to lift- based propulsion is observed across fish species, and often within species as fishes move at different speeds, and this transition appears to be closely associated with the aspect ratio of the propulsive surface and ecology of the fish (Blake, 2004; Lighthill,

1969; Webb, 1974; Webb, 1984a).

Aspect ratio defines the span of the fin squared divided by the surface area of the fin. Thus, fins with lower aspect ratios have relatively small spans with larger surface areas that can accelerate larger masses of water (high drag) to promote high thrust forces, and therefore high accelerations, but consequently achieve lower sustained speeds due to relatively high resistance in the retarding drag force as speed increases

(Webb, 1974; Webb, 1984a). Fishes with low aspect-ratio fins most commonly live along the substrate in benthic environments, routinely maneuvering around structures

4 and exhibiting transient (i.e. unsteady) predator-prey behaviours (Fontanella et al.,

2013; Franklin et al., 2014; Lighthill, 1969; Webb, 1984a). In contrast, fishes that continuously swim in pelagic environments at higher speeds and often exercise periodic

(i.e. continuous) predator-prey behaviours, tend to be better shaped to overcome the resistance of drag in swimming than benthic fishes, and therefore, they tend to have fins with higher aspect ratios (Fontanella et al., 2013; Franklin et al., 2014; Lighthill, 1969;

Webb, 1984a). As aspect ratio increases, the coefficient of lift increases and drag decreases, and hence, lift-based propulsion as the mechanism to generate thrust eventually becomes a more effective approach to maximize thrust for a given finbeat

(Blake, 2004; Fish, 1996; Lighthill, 1969; Webb, 1974; Webb, 1988). Of note, fishes that commonly exhibit transient and periodic predator-prey behaviours, often referred to as generalist swimmers, tend to have fins with moderate aspect ratios and might inhabit benthic, demersal and pelagic environments (Webb, 1984a). The shape of a fish, and its fins, will thus have a large impact on the mechanisms available to swim, constrain locomotion at some levels, but provide opportunities at others. This forms the basis of my interest in swimming in fishes that are flattened like a foil in the same plane as the substrate, such as stingrays, which are fishes with body and fin shapes notably different than the majority of fishes, and the implications of which are not well understood.

Foil-shaped fishes

In fluid mechanics, a foil is a flattened shape that generates considerably more lift than drag, when placed at an angle of attack in the direction of a flowing fluid. A foil may be referred to as an aerofoil in air, or as a hydrofoil in water, and they are commonly used to promote thrust and maneuverability in locomotion. For example, planes, bats, birds and many insects often have aerofoils as wings to help fly through air, while submarines, fish and marine mammals often have hydrofoils as fins to aid in swimming through water. Fishes might even have bodies that are notably shaped like a foil, whereby species might be laterally or dorsoventrally flattened. When swimming forwards, the flattened width of the fish may be oriented parallel to the surface or perpendicular. These fishes might be collectively referred to as foil-shaped fishes because they are expected to have a substantially higher capacity to generate high lift- drag ratios along the body than fishes that are more cylindrical in body shape, which is likely to impact swimming behaviour. Moreover, critical differences in how these foil- shaped fishes are expected to utilize lift, suggests that the orientation of the foil fish should be categorized into two subgroups. Foil shape fishes that are flattened perpendicular to the substrate (e.g. reef fish, such as angelfish, butterfly fish, batfish) are expected to rely on high magnitudes of lift during left and right yawing turns

(Domenici and Blake, 1993a). Furthermore, they tend to have a hydrostatic swim bladder to maintain neutral buoyancy in the water column. In contrast, most, if not all foil-shaped fishes that are flattened parallel to the substrate, seem to lack a swim bladder and are negatively buoyant. Accordingly, these fishes pitch the body into a positive angle of attack to generate lift when swimming, to maintain vertical position in

6 the water column (Blevins and Lauder, 2012; Rosenberger, 2001; Webb, 2002). These differences in lift generation are expected to have profound impacts on the swimming behaviour and ecology of foil-shaped fishes.

The research chapters in this thesis explore aspects of the mechanics of different locomotor behaviours in a benthic stingray, which is a dorsoventrally flattened fish in the same plane as the substrate. Hence, the remainder of this introduction will focus on the diversity in shape and locomotor behaviours promoted by foil-shaped fishes that are oriented parallel to the substrate. A flattened body, oriented parallel to the substrate plane, has convergently evolved in a relatively small number species of fishes, including elasmobranchs (e.g. batoid rays, skates, sawfish, guitarfish and electric rays, and some species of shark) and pleuronectifomes (e.g. flatfish such as flounder, sole and halibut).

Other than a small number of species of pelagic rays, a foil shape that is parallel to the substrate is most common to species that inhabit benthic environments, suggesting that a foil in this orientation might be particularly effective for locomotion along the substrate boundary. The following descriptions aim to formulate our understanding of the range of locomotors behaviours promoted by a foil-parallel shape in water, and consequently highlight that we know very little about the benefits and constraints of a disc-shaped foil in water, which is a common shape to benthic stingrays.

Diversity in the foil shape and swimming mechanics

Foil-shaped fishes that are flattened in an orientation parallel to the substrate, give rise to a range of shapes and mechanics to power swimming via a foil in water.

Flatfish (i.e. pleuronectiformes such as flounder and sole) are benthic species that are laterally flattened and flipped on their side, such that when they are sedentary along the benthos, they rest on a lateral surface. The shape of these fishes often resembles an arrowhead. Flatfish employ BCF drag-based propulsion via lateral undulations that are very similar to the anguilliform mode of swimming, except that the undulations are oriented vertical to the substrate (Webb, 1984b; Webb, 1988; Webb, 2002). The anguilliform swimming mode is named after the genus Anguilla (i.e. eel) (Breder, 1926).

This mode describes fishes of different species that swim by exercising propulsive wavelengths that are less than the body length; hence, the number of propulsive waves observed along the length of the flatfish and other anguilliform swimmers is greater than one, and furthermore, the wavelengths tend to steadily increase in amplitude from the snout towards the tip of the caudal fin (Breder, 1926; Lighthill, 1969; Lighthill, 1970;

Webb, 1974; Webb, 2002). These fish possess bodies with high flexibility, and both the body and fins constitute a propulsive surface with a low aspect ratio. Little is understood about the sustained swimming behaviours Anguilliform fish which promote the capacity for high acceleration and maneuvering, but comes at a cost of lower sustained speeds, and accordingly, anguillifom fishes tend to perform transient predatory strategies in benthic and demersal environments (Domenici and Blake, 1997;

Webb, 1981; Webb, 1984b; Webb, 1984a; Webb, 1988).

Electric rays, sawfishes, guitarfishes, and foil-shaped sharks are benthic species that are dorsoventrally flattened with pronounced tails and caudal fins, and have enlarged pectoral fins of various sizes that are fused to the body. These fishes most often exhibit a BCF drag-based swimming referred to as a subcarangiform mode. In contrast to anguilliforms, the anterior portion of a subcarangiform fish is relatively motionless, such that the number of propulsive waves observed along the length of the body appears to be less than one, and they rapidly increase in amplitude over the posterior half of the fish (Archer and Johnston, 1989; Donley, 2003; Rosenblum et al.,

2011; Sfakiotakis et al., 1999; Webb, 1974; Webb, 2002; Webb and Keyes, 1982)).

Median (dorsal and anal) and paired (pectoral and pelvic) fins are usually very well defined, and may be used for secondary thrust, stability and maneuvering (Drucker,

2003; Maia and Wilga, 2013; Rosenblum et al., 2011; Standen, 2008; Standen and

Lauder, 2007; Tytell et al., 2008; Wilga and Lauder, 2001). Body flexibility and the aspect ratio of the fins tends to be low to moderate in subcarangiforms, and these fishes tend to be generalist swimmers in demersal environments that exhibit moderate performance in acceleration and speed (Domenici and Blake, 1997; Domenici et al.,

2004; Seamone et al., 2014; Webb, 1984a), but some species exhibit high acceleration and transient predation (Domenici and Blake, 1997; Hale, 2002; Schriefer and Hale,

2004; Webb, 1984a; Webb and Skadsen, 1980).

Skates and rays are dorsoventrally flattened fishes that power MPF mobuliform oscillation or rajiform undulation via enlarged pectoral fins that bend in the vertical plane, with long thin tails that appear to be constrained in swimming due to a reduced or

9 absent caudal fin. Mobuliform oscillation defines lift-based propulsion via wing-shaped pectoral fins with higher aspect ratios (Fontanella et al., 2013; Rosenberger, 2001). Up to one-half of a wavelength is present on the fin at a given time, presenting the appearance of an oscillating fin (Fish et al., 2016; Rosenberger, 2001). This swimming mode is most commonly found in pelagic species that exhibit periodic propulsion

(Fontanella et al., 2013; Franklin et al., 2014; Webb, 1984a). In contrast, rajiform undulation defines drag-based propulsion via undulations that pass along rounded or diamond shaped pectoral fins with a low aspect ratio (Blevins and Lauder, 2012;

Fontanella et al., 2013; Franklin et al., 2014; Rosenberger, 2001; Webb, 1974).

Undulations of the pectoral fin appear similar to anguilliform motions, whereby more than one wavelength is present on a fin at a given moment giving a rippling or undulating effect to fin movement, and amplitude of the propulsive wave increases from the anterior of the fin towards the posterior (Blevins and Lauder, 2012; Rosenberger,

2001; Rosenberger and Westneat, 1999). This swimming mode is most common to benthic skates and stingrays. Our understanding of high performance swimming in mobuliform and rajiform swimmers is largely incomplete.

Gliding

Lift and negative buoyancy are key parameters that enable objects to glide

(unpowered motion) through fluid, and foil fishes possess both. When they sink through water, the foil shape generates lift as water flows over it, enabling these animals to harness kinetic from potential energy and travel forward without moving their fins

(Tucker, 1988; Tucker and Parrott, 1970; Vernes, 2001; Webb, 1974). Gliding

10 behaviours are commonly observed in a number of foil fishes, including electric rays, pelagic and benthic stingrays and flatfish (Braun et al., 2014; Kawabe et al., 2004; Olla et al., 1972; Parson et al., 2011; Rosenberger, 2001; Rosenblum et al., 2011; Stickney et al., 1973; Takagi et al., 2010). The diversity of aspect ratios observed in these fishes would likely impact the lift-drag ratio and hence the glide ratio (distance travelled forward per distance of drop) of foil fishes, but this has not been explored (Fontanella et al., 2013; Franklin et al., 2014; Henningsson et al., 2014; Webb, 1974). Weihs (1973) suggested that actively swimming upwards followed by powerless gliding decent is an efficient mode of migrating in fishes that are negatively buoyant, potentially leading to energy savings of over 50% for a given horizontal distance travelled, but it’s not clear if foil-shaped fishes benefit from this behaviour.

Benthic locomotion: walking and punting

Mechanics of short distance, benthic locomotion, entails protraction of the fin towards the anterior of the animal, then planting the fin onto the substrate and retraction in the caudal direction to thrust the animal forward; this is followed by a glide and recovery phase before the next cycle (Bilecenoglu and Ekstrom, 2013; Fox et al., 2018;

Koester and Spirito, 2003; Macesic and Kajiura, 2010; Macesic et al., 2013). Skates, stingrays, and electric rays use their pelvic fins for benthic locomotion, sharks use both the pectoral and pelvic fins, whereas flatfish use their dorsal and anal fins (Fox et al.,

2018; Goto et al., 1999; Holst and Bone, 1993; Macesic and Kajiura, 2010). The term walking is used to refer to alternate motion of the fins, while punting refers to the synchronous motion of the fins; the benefits or limitations of punting versus walking are

11 not clear (Lucifora and Vassallo, 2002; Macesic and Kajiura, 2010). Furthermore, augmented punting, commonly observed in stingrays, refers to undulations being passed along the length of the pectoral fins simultaneously to movement of the pelvic fins, so that swimming with the pectoral fins might supplement punting locomotion, whereas punting solely with the pelvic fins is referred to as true punting, which has been described in electric rays and skates (Macesic and Kajiura, 2010; Macesic et al., 2013).

True punters were found to have a slight edge over augmented punters in the speed and distance they can travel, which has been attributed to larger muscles found in the pelvic fins of true punters (Macesic and Kajiura, 2010; Macesic et al., 2013). In augmented punters, pelvic fins appear to mostly determine punt performance, while the role of the pectoral fins is not fully clear (Macesic et al., 2013). Macesic et al. (2013) suggest that slip calculations indicate that pectoral fin undulations observed during augmented punting do produce thrust, however, they found no significant relationship between increasing pectoral fin kinematics (frequency, amplitude and wave speed) and an increase in punt performance. Thus, any supplementary thrust from the pectoral fins appears to only partially compensate for the reduced mass in the pelvic girdle region of augmented punters compared to true punters, such that augmented punters do not outperform true punters despite also employing their pectoral fins. A reduction in the overall mass of an animal would most likely reduce the energetic costs of swimming through the water column (Ekstrom and Kajiura, 2014; Macesic and Kajiura, 2010;

Macesic et al., 2013), and so perhaps might benefit augmented punters. Furthermore, it has been suggested that augmented punting may be a transitional mode of locomotion from punting into swimming, however, this has not been explored (Macesic et al., 2013).

Given that punting is exercised for small scale locomotion, it is quite possible that the augmentation of the pectoral fins is a mechanism to maintain a layer of water between the pectoral disc and the substrate, enabling stingrays to swim at slower speeds while preventing the body from dragging along the benthos, although this has not been explored. Nevertheless, further research is needed to understand the limitations to these different modes of benthic locomotion.

Ground effect

When fishes swim close to the benthos, the energetic costs of locomotion can be lessened due to decelerated flow of water between the benthos and ventral surface of the fish. This ultimately increases the pressure on the underside of the body producing lift, in addition to decreasing drag through the reduction of three-dimensional vortices shed off the fins as a result of close proximity to the substrate (Rozhdestvensky, 2006).

The magnitude of ground effect is determined by the ratio of the distance between the structure that is moving over the ground and the benthos itself (i.e. gap) to the width of the structure parallel to the ground (i.e. span); ground effects greatly decrease as the gap-span ratio increases, becoming insignificant at a ratio of 3 (Blake, 1979; Blake,

1983; Blevins and Lauder, 2013). Thus, wide structures moving close to the substrate benefit the most from ground effect, and hence, foil fishes are anticipated to possibly benefit in performance and energetics from ground effect when swimming along the benthos. While swimming parallel to the substrate, flatfish modulated their swimming kinematics as the distance between the ground and the ventral surface of the fish changes, suggesting they are experiencing ground effect (Webb, 2002). As opposed to

13 swimming out of ground effect, flatfish decrease the amplitude of the propulsive wave and slightly increase the frequency when swimming along the substrate. In result, total mechanical power required to swim decreased as flatfish swam closer to the substrate

(Webb, 2002). Blevins and Lauder (2013) used a robotic stingray model, and found that undulating fins do not necessarily result in faster swimming under most kinematic conditions when in ground effect, and power requirements and the robotic cost of transport were found to increase as well. They suggested that living stingrays, as opposed to the robotic model used in their studies, might modulate their swimming kinematics to minimize locomotor penalties and incur benefits from swimming near the substrate. Hence, more research is required to understand the conditions that may enhance the performance or reduce energetics from ground effect in foil fishes when swimming along the substrate.

Fast-start maneuverability

Fast starts are high performance maneuvers involving high translational and rotational accelerations, which are commonly employed during predator-prey interactions. While fast-start swimming has been studied quite extensively in BCF swimmers that undulate the body and caudal fin in the same plane as the substrate

(Domenici, 2010a; Domenici and Blake, 1997; Domenici and Hale, 2019), we lack an understanding of how the diversity in form and swimming of foil fishes may impact these behaviours. C-starts are a form of fast-start maneuvers common in many BCF swimming fishes, where a unilateral contraction of the axial muscles bends the animal into a c-shape, changing the direction of swimming from the initial orientation; this is

14 followed by a contralateral contraction associated with a return flip of the tail, which accelerates the animal in the new direction (Domenici and Blake, 1997; Domenici and

Hale, 2019; Domenici et al., 2004). Flatfish also perform C-starts, but the bending of the body occurs perpendicular to the substrate, rather than in the same plane as the substrate as occurs in most BCF swimming fishes (Brainerd et al., 1997; Webb, 1981).

In contrast to flatfish that are laterally flattened and BCF swimming, stingrays are dorsoventrally flattened fishes with long, thin tails that do not have a clear role in swimming. Rather, they have enlarged and flexible pectoral fins that wrap around the head and body, the latter being essentially rigid and thus limit dorsoventral and axial bending. Hence, it might be expected that the unusual shape of stingrays when compared to BCF fishes might promote a different approach to fast start maneuverability, although this has yet to be explored.

Burying

Most foil fishes, including stingrays, skates, electric rays, guitarfish and angel sharks, that live along the benthos have been observed to bury, in the wild and in aquariums (Seamone, personal observation). However, the mechanics of burying have only been explored in flatfish, and appear closely related to the mechanics of their swimming behaviours (Corn et al., 2018; McKee et al., 2016). To bury, flatfish pass undulations along the length of their body while in close proximity to the substrate, which fluidizes sediment beneath the fish that is then moved up and over onto the exposed upper surface (Corn et al., 2018; McKee et al., 2016). Increasing the frequency of undulations appears closely associated with an increase in sediment

15 coverage of the exposed surface (Corn et al., 2018; McKee et al., 2016). Larger flatfish use lower cycle frequencies and take longer to bury, but sediment coverage was not impacted by size (Corn et al., 2018). Furthermore, grain size of the sediment does not affect undulation frequency or time to burial, but as grain size increased, sediment coverage of the exposed surface decreased (Corn et al., 2018). Moreover, these fishes appear to spend more time in sediment sizes that promote increased sediment coverage (Gibson and Robb, 1992; Gibson and Robb, 2000; Moles and Norcross,

1995). Whether other species of foil fishes, such as stingrays that have large and flexible pectoral fins with a relatively rigid body axis, are reliant on fluidizing sediment to bury and inhabit sediments that promote enhanced burying performance is not known.

Station holding

Station holding refers to a behaviour that fishes use to maintain position against a displacing force, such as currents or predators that attempt to lift their prey off the substrate (Arnold et al., 1991; Webb et al., 1996). The flattened profile of foil fishes promotes relatively less drag than non flattened fishes, which would assist with station holding, but a foil shape would promote lift in current, and while it has been revealed that these fishes rely on mechanisms to reduce lifting forces in station holding (Arnold and Weihs, 1978; Webb, 1974; Webb, 1989), whether they also rely on mechanisms to resist lifting forces, such as adhesion to the substrate, is not known. For example, as flow increases, flatfish and rays pass waves of bending along their propulsive surfaces to increase the flow underneath the fish relative to that of the upper surface, and consequently, reduce the pressure gradient, and hence, the lifting force (Webb, 1989).

The flattened ventral surface of these fishes has also been suggested to potentially act like a suction disc, enabling fishes to adhere to the benthos (Arnold and Weihs, 1978;

Fontanella et al., 2013), but this has not yet been measured.

The pectoral disc of stingrays

There are about 220 species of stingrays belonging to the suborder

Myliobatoidei, categorized into 10 families, all found in temperate to tropical waters.

Stingrays have cartilaginous skeletons, with long thin tails that often have a stinger used for defence, while the tail appears to be limited in its role in swimming due to a reduced or absent caudal fin. Rather, swimming is powered via enlarged pectoral fins that wrap around the head and body (Blevins and Lauder, 2012; Rosenberger, 2001;

Rosenberger and Westneat, 1999), and the pectoral fins and the body are collectively referred to as the pectoral disc. The body axis of the pectoral disc is essentially rigid along the dorsoventral and lateral axis (Parson et al., 2011), but the pectoral fins are otherwise exceptionally flexible in three dimensions (Blevins and Lauder, 2012), and they appear to function in a range of behaviours such as routine swimming, fast-start maneuvering, burying, station holding and possibly more (Seamone, personal observation). Hence, whereas most species of fish rely on a range of surfaces to interact with water and to locomote, stingrays appear to accomplish a diverse range of locomotor behaviours using a single broad surface of the pectoral disc.

The number of undulatory waves present on the pectoral fin at any given moment was the first measurement used to distinguish different swimming modes in stingrays and other pectoral swimming batoids (e.g. skates), and has been used to place stingrays along a locomotor continuum between rajiform undulation and mobuliform oscillation (Rosenberger, 2001). Locomotor style was later associated with phylogeny, body and fin shape, and ecology (Fontanella et al., 2013; Franklin et al., 2014).

Variation in the aspect ratio (span2/area) appears to dominate variation in pectoral fin shape across stingrays and other batoids, followed by variation in the location of the fin tip relative to the body to a lesser extent (Fontanella et al., 2013; Franklin et al., 2014;

Rosenberger, 2001). Aspect ratio and location of the fin tip are related to phylogenetic genera, and aspect ratio can predict locomotor style (rajiform undulation versus mobuliform oscillation) and the ecology of the fish (benthic versus pelagic) (Fontanella et al., 2013; Franklin et al., 2014). Hence, there appear to be links between shape and ecology in stingrays, but we lack a clear understanding of how shape impacts many aspects of movement such as fast-start maneuverability, suction holding and burying behaviours in these fishes.

Pelagic stingrays and mobuliform oscillation

Mobuliform oscillation of high aspect-ratio pectoral fins that are wing-shaped is a more derived form of MPF locomotion (Franklin et al., 2014), most common to pelagic rays that tend to exercise periodic propulsion, although benthic species that employ mobuliform oscillation exist, such as butterfly rays (Fontanella et al., 2013; Franklin et al., 2014; Webb, 1984a). It appears that high aspect-ratio fins in a pelagic lifestyle have

18 emerged at two independent evolutionary nodes within Myliobatoidei, connecting eagle rays (Myliobatidae) with stingrays (Myliobatoidei), and connecting the pelagic stingray

Pteroplatytrygon violacea (Bonaparte 1832) with stingrays of the genus Dasyatis

(Franklin et al., 2014). This swimming mode is defined by having less than half of a complete propulsive wave passing along the pectoral fin at a given point in time

(Rosenberger, 2001). Hence, the propulsive wavelengths passing along the fin are at least twice as great as the chord length of the fin, promoting the appearance of an oscillating (flapping) fin. The higher aspect-ratio and greater tapering of the fin tips than rajiform undulators, is anticipated to increase the magnitude of lift while reducing drag experienced by these fins (Fontanella et al., 2013; Webb, 1974). Therefore, this swimming mode has been suggested to generate thrust as a component of lift-based propulsion (Fontanella et al., 2013; Rosenberger, 2001), but this has not yet been supported by direct measurements of hydrodynamics in living fish.

As opposed to rajiform swimmers, mobuliform fishes exercise relatively higher amplitude and lower frequency finbeats (Rosenberger, 2001). To increase speed, the pelagic cownose ray (Rhinoptera bonasus) increases wavespeed and fin-tip velocity, whereas the benthic butterfly ray (Gymnura micrura) decreases wave number and increases fin-tip velocity (Rosenberger, 2001). High amplitude oscillations of the pectoral fins promote substantial medial stress, and in result, an increase in calcification along with cross bracing between fin rays in central areas of the fins has been observed

(Schaefer and Summers, 2005). The centre of mass is anteriorly located relative to the maximum fin span and position of the apex of the fin tip (Fontanella et al., 2013). This

19 might promote maneuverability or a passive mechanism to obtain gliding stability that promotes maximum glide ratio, as also observed in flatfish (Takagi et al., 2010; Webb,

1974). These fishes can exercise both gliding and powered turns, and have been suggested to be poor at maneuvering due to the rigid body-axis (Parson et al., 2011), although their fast start performance has not been explored. Mobuliform oscillation have been mimicked in an autonomous underwater vehicle called the mantabot (Fish et al., 2016; Liu et al., 2015).

Benthic stingrays and rajiform undulation

Rajiform undulation of low aspect-ratio pectoral fins that are more rounded in shape, common to benthic skates and rays, appears to be the basal form of locomotion in batoids and is defined as having more than one wave passing along the pectoral fin at a given point in time (Aschliman et al., 2012; Blevins and Lauder, 2012; Fontanella et al., 2013; Franklin et al., 2014; Rosenberger, 2001; Rosenberger and Westneat, 1999).

These fishes employ drag-based propulsion with relatively high frequency undulations with a smaller amplitude than mobuliform oscillators (Rosenberger, 2001). Different species appear to exercise different kinematic mechanisms to increase swimming speed (Rosenberger, 2001; Rosenberger and Westneat, 1999). The amplitude of the propulsive wave increases across both the anteroposterior and mediolateral fin axes in the smooth back river stingray, and furthermore, this stingray has been observed to cup the water with the lateral edge of the pectoral fins. This behaviour is believed to enhance stream-wise momentum, enhancing thrust forces and reducing wingtip vortices, thus reducing induced drag (Blevins and Lauder, 2012). Hence, the distal fin

20 edges of these animals experience substantial bending, and in result, many species have increased calcification in the rays of the fins with joint staggering towards the lateral edge of the fins to accommodate the potential stress (Schaefer and Summers,

2005).

The center of mass is located along the anteroposterior axis at the position of maximum fin span, which has been suggested to provide stability for rotational movement (Fontanella et al., 2013). The peripheral placement of nearly symmetrical fins that wrap around the centre of mass of the fish, and can function independently, enables these fish to rotate around the vertical axis with minimal turning radius, promoting slow precision maneuvering when feeding on non evasive prey along the substrate, which is a common foraging strategy in these fishes (He and Zhang, 2013;

Webb, 1984a; Wilga et al., 2012). However, like mobuliform oscillators, rajiform undulation has been attributed to poor maneuverability during routine swimming due to the possession of a rigid body axis (Parson et al., 2011), and this might also impact fast- start swimming in these fishes, although this has not been explored. The foil shape of rajiform undulators shows cambering along the anterioposterior and medial axis of the dorsal surface, whereas the ventral surface is essentially flat. This shape has been suggested to decrease the magnitude of drag experienced by the fish during station holding, and to promote the ability to generate suction to prevent lifting off the substrate

(Fontanella et al., 2013; Webb, 1989), although this has not been measured.

Furthermore, these fishes show a remarkable ability to rapidly bury, in the wild and in aquariums, but the mechanisms relating shape, sediment dynamics and burying

21 performance in these behaviours have not yet been described. Understanding the mechanisms of behaviours such as fast-start maneuverability, station-holding and burying might help advance the design of robotics mimicking the rounded disc-shape of benthic stingrays, such as the tissue-engineered, soft-robotic ray (Umbanhowar et al.,

2016).

Thesis chapters

This dissertation explores aspects of locomotion in benthic stingrays, dorsoventrally flattened fishes with a foil-shaped body and a relatively rounded pectoral disc. I used the ocellate river stingray, Potamotrygon motoro, as an experimental subject because it is a representative model for MPF rajiform undulation, and hence drag-based locomotion, via its nearly symmetrically rounded pectoral fins with a low aspect-ratio (Fig. 1). The nearly symmetrical rounded disc-shape of this species is very representative of the shape of about 35 river stingrays belonging to the family

Potamotrygonidae that have migrated into the river systems throughout South America, and closely resembles the shape of many species throughout Myliobatoidie, such as in families Urolophidae and Urotrygonidae. Hence, the mechanisms revealed here can likely be generalized to other rounded disc-shaped stingrays. Each chapter describes studies on a different aspect of locomotion. Chapter 2 describes, for the first time, mechanisms employed for fast starts and the associated maneuverability of stingrays, which have previously been designated as poor maneuverers because they possess a rigid-body axis and a tail with low surface area. This study has been published in the

Canadian Journal of Zoology (Seamone et al., 2019). Chapter 3 is an account of the

22 changes in pressure between the pectoral disc and the substrate during station holding, and use of video and dye movement to describe body, fin and water movements around the stingray during these behaviours, to question whether the disc shape of P. motoro can generate suction to adhere to the benthos in resistance to an upwards displacement force. Studies described in Chapter 4 used video analysis and particle image velocimetry to measure how stingrays use their body and fins to rapidly transport sediment from beneath the ventral surface onto the dorsal surface to bury. Chapter 5 used high-speed videography and video time-lapse photography to question if stingrays demonstrate preference for inhabiting sediments of different grain size, and how grain size impacts burying performance. Chapter 6 is a summary and conclusion regarding the outcome of this work, and how it contributes to the mounting evidence that the flattened disc-shape and drag-based propulsion via low aspect ratio and flexible pectoral fins of river stingrays permits versatility in function for locomotor behaviours in a benthic environment.

Fig. 1. The ocellate river stingray (Potamotrygon motoro) in a benthic environment.

Chapter 2

Disc starts: the pectoral disc of the ocellate river stingray (Potamotrygon motoro)

promotes omnidirectional fast starts across the substrate

Abstract

This study explored how the flattened and rounded pectoral disc of the ocellate river stingray (Potamotrygon motoro) enables them to use the benthic plane during fast-start escape. Escape responses were elicited via prodding different locations around the pectoral disc and were recorded using video. Modulation of pectoral fin movements that power swimming enabled omnidirectional escape across the substrate, with similar performance in all directions of escape. Hence, translation of the body did not necessarily have to follow the orientation of the head, overcoming the constraint of a rigid body axis. An increase in prod speed was associated with an increase in initial translational speed and acceleration away from the prod. As prod location shifted towards the snout, yaw rotation increased, eventually reorienting the fish into a forward swimming position away from the prod. Furthermore, P. motoro yawed with essentially zero turning radius, allowing reorientation of the head with simultaneous rapid translation away from the prod, and yaw rate during escape was substantially greater than reported during routine swimming for stingrays. We conclude that stingrays employ a distinctive approach to escape along the substrate, which we have termed disc starts, that results in effective maneuverability across the benthic environment despite limited longitudinal flexibility of the body, and that challenges the concept of maneuverability typically used for fishes.

Key words: stingrays, fast starts, escape responses, benthic, maneuverability, rigid body axis.

Introduction

Fast starts are high-performance maneuvers, involving high accelerations and speeds from rest or routine locomotion, employed during predator–prey interactions such as in attack and escape responses (Domenici and Blake, 1997). Some of the most impressive maneuvers in biology occur in predator–prey interactions. Hence, fast starts during escape and attack may reveal diversity in the capacity for locomotor performance promoted by physiology and morphology (Harper and Blake, 1991;

Higham et al., 2017; Martin et al., 2005; Sharp, 1997; Tucker, 1998), and accordingly, may provide valuable insight into shape and behaviour that is effective for survival in different environments (Higham et al., 2016). Body shape is expected to have a large impact on swimming mechanics during fast starts. Fishes are often categorized as either body and caudal fin (BCF) or median and paired fin (MPF) swimmers, based on the propulsive surfaces that are used to power routine swimming (Blake, 2004; Breder,

1926; Sfakiotakis et al., 1999; Webb, 1974; Webb, 1984a; Webb, 1984b). In BCF swimmers, contractions along the axial musculature power swimming, passing a propulsive wave through the body and caudal fin (Altringham and Ellerby, 1999; Bone,

1966; Donley, 2003; Gray, 1933). In these fishes, two types of fast-start behaviour are exhibited (C-starts and S-starts), which are characterized by the axial deformation of the body that occurs during the event. During a C-start, a unilateral contraction yaws the animal into a C-shape, changing the angular orientation of the head (Borazjani et al.,

2012; Domenici et al., 2004; Webb, 1981), followed by a return sweep of the tail, which accelerates the fish in the new orientation away from the threat. S-starts involve simultaneous contractions on both sides of the axial musculature, bending the animal

26 into an S-shape (Hale, 2002; Schriefer and Hale, 2004; Webb, 1976), followed by contralateral contractions, as the fish accelerates forwards with little angular displacement. Hence, maneuverability in BCF-powered fast starts is dominated by yaw rotation and forward translation. MPF swimmers encompass a diverse range of fishes that may use pectoral, dorsal, anal, and pelvic fins, or a combination of these median and paired fins, to power swimming (Blake and Chan, 2011; Jagnandan and Sanford,

2013; Korsmeyer et al., 2002; Rosenberger, 2001; Ruiz-Torres et al., 2014; Walker and

Westneat, 2002). Many MPF swimmers display a decoupled locomotor strategy, whereby they switch from MPF swimming during routine activity to BCF swimming during fast starts (Borazjani et al., 2012; Domenici and Blake, 1993b; Kasapi et al.,

1993; Westneat et al., 1998). This strategy appears common to MPF swimmers with body shapes that closely resemble the profiles of BCF swimmers. Whether all MPF swimmers rely on a decoupled locomotor strategy for fast starts is not clear, yet this is unlikely due to the substantial diversity in morphology in these fishes.

Stingrays (suborder Myliobatoidei) are dorsoventrally flattened fishes that employ

MPF swimming, but possess a striking difference in morphology compared with most

BCF and MPF swimming fishes (Aschliman et al., 2012; Fontanella et al., 2013; Franklin et al., 2014). The caudal region is fashioned into a whip-like appendage with a barbed stinger used primarily for defence, while appearing poorly suited to power swimming due to low surface area. Stingrays have very large and flexible pectoral fins that join to the head and the body, and together, the head, body, and fins are referred to as the pectoral disc (Blevins and Lauder, 2012; Fontanella et al., 2013; Franklin et al., 2014).

Drag-based undulations of the low aspect ratio (i.e., rounded) pectoral fins are used to power a rajiform mode of MPF swimming (Blevins and Lauder, 2012; Fontanella et al.,

2013; Rosenberger, 2001; Rosenberger and Westneat, 1999). Anterior to the pelvic girdle, the body axis is relatively inflexible. Hence, a decoupled locomotor strategy, where stingrays could employ axial undulations and BCF swimming during fast starts, seems unlikely. This morphology has been associated with poor maneuverability, apparently due to the rigid body axis (Parson et al., 2011). That said, stingrays can apparently yaw with next to no turning radius during routine swimming (He and Zhang,

2013). Furthermore, stingrays are preyed upon by predators such as crocodilians, sharks, and marine mammals (Cliff and Dudley, 1991; Nifong and Lowers, 2017; Strong et al., 1990; Visser, 1999), and so we might expect them to have impressive escape

(i.e., maneuvering) abilities. Yet there seem to be few studies that explore escape behaviours in stingrays (Semeniuk and Dill, 2005; Semeniuk and Dill, 2006), and none have explored behaviour in the context of morphology and fast-start performance where we might expect to observe a fuller demonstration of the implications of a flattened disc shape for maneuverability.

Whether MPF or BCF, a fish suspended in an aquatic environment has the potential to accelerate along three translational freedoms (i.e., fowards–backwards, side-to-side, and up–down) and three rotational freedoms (i.e., yaw, pitch, and roll)

(Dudley, 2002; Webb, 2004). Surprisingly, fishes studied thus far do not seem to use all freedoms of motion during fast starts, although whether this is a physical or behavioural limitation, or a failure to invoke such responses in the experimental design, is not clear.

The abilities of stingrays to exploit these freedoms in a benthic environment is also unclear. In the present study, we evoked and described escape responses in the ocellate river stingray (Potamotrygon motoro). Potamotrygonidae stingrays characterize an extreme form of benthic stingrays such that the pectoral disc is nearly symmetrical from a dorsal view, with rounded pectoral fins that generate high drag (Blevins and

Lauder, 2012; Fontanella et al., 2013; Franklin et al., 2014). It was hypothesized that this morphology may promote flexibility in escape behaviour in a benthic environment, allowing the fish to use the entire plane along the benthos for escape from rest.

Materials and methods

Animals and housing

All procedures were approved by the University of Calgary Animal Care

Committee, following the guidelines of the Canadian Council on Animal Care. Four

Potamotrygon motoro (pectoral disc width of 14.9 ± 0.410 cm (mean ± SE); mass of 155

± 10.2 g (mean ± SE)) were purchased from a licenced supplier and transported to facilities at the University of Calgary. P. motoro were housed in a cylindrical holding tank

(180 cm diameter × 70 cm height, approximately 1400 L) with flow-through freshwater at

27° C, pH of 6.5, bubbled with air. Room lighting was provided with a 12 h ON : 12 h

OFF cycle. The P. motoro were provided 2.5 cm3 of frozen bloodworms per fish twice a day and were monitored to ensure feeding. Non-abrasive substrate (fish-tank gravel) was used to cover the bottom of the tank, with a grain size greater than 1 cm to reduce

29 burying behaviours. P. motoro were acclimated for 2 weeks prior to experiments, water chemistry was measured daily, and the fish had regular veterinary oversight.

Data collection

A separate tank with the same dimensions, water chemistry, and substrate as the holding tank was used for experimentation. An individual P. motoro was transferred to the experimental tank via a rubber mesh net and given 24 h to acclimate without being fed. The P. motoro was then lured to the middle of the tank using bloodworms wrapped in a small mesh net. An escape response was then induced by quickly prodding the pectoral fins with a flexible, plastic tube (2.5 mm diameter), with a tip made of soft silicone tube. P. motoro were not prodded on the body or the tail because these areas seem to be more prone to a reflex defence response by the tail (Campbell 1951). The escape response was recorded by five submerged video cameras (Hero3 Black; GoPro

Inc., San Mateo, California, USA) at a frame rate and a video resolution of 240 frame/s and 848 pixels × 480 pixels, respectively. One camera filmed the dorsal view 60 cm above the P. motoro to capture its movements and the prod across the substrate, discussed below. Four cameras filmed along the substrate, separated by 90° around the tank and placed approximately 20 cm from the fish, to capture movements of the P. motoro and the prod above the substrate; of these four cameras, images from the camera for which the prod moved perpendicular across the field of view were selected for analysis. A measuring stick was placed in the tank to calibrate the cameras before experimentation commenced, and was used to confirm that the calibration was the same across the field of view in the X and Y axes for the camera filming the dorsal view

30 and in the Z axis for each camera filming in the horizontal view. A pair of light-emitting diodes attached to the prod was used to synchronize the cameras at the onset of recording. Two 1000W halogen lights were attached to the perimeter of the tank to illuminate the filming area (Southwire Tools, Carrollton, Georgia, USA).

A total of 24 responses was collected. Each of the four animals were prodded, and the escape response recorded, six times (one prodding event = one trial). An effort was made to distribute the prodding events evenly around the pectoral fins for each animal. Furthermore, we exercised a range of prodding speeds to investigate the effect of stimulus intensity on escape performance. The four P. motoro were rotated into the experimental tank throughout the study such that only a single escape response was recorded from an animal before it was placed back into the holding tank and another animal moved into the experimental tank. This provided more than a 96 h resting period between each trial on a given animal.

Data analysis

Since none of the fish were prodded in the same location more than once, all trials from the four fish were treated as independent observations. For each trial, the tip of the prod, the body midline, the centre of the disc, and 11 points (30° intervals) around the periphery of the pectoral fins were digitized and their locations tracked for successive frames using Image J (bundled with 64-bit Java 1.8.0_112; Wayne Rasband

Developers 1997). The location of the tip of the prod was used to quantify the impacts

31 of prod speed and prod location on different aspects of the escape behaviour, defined below. The movements of the prod and the P. motoro were measured in three dimensions; X and Y coordinates were obtained from the camera filming from the dorsal view, whereas Z coordinates were obtained from a camera filming along the horizontal view. Digitizing of the prod was initiated from the onset of the thrusting motion towards the fish and terminated when the plastic tube bent from contact with the animal. Escape movement of the P. motoro was minimal in the Z axis (i.e., the fish escaped extremely close along the substrate). Hence, only the X and Y coordinates of the centre of the disc were used to assess translational kinematics of the P. motoro. The line between the centre of the disc and the snout was quantified to assess angular rotation (i.e. yawing) kinematics.

To power escape, P. motoro roll their pectoral fins up and over towards the dorsal midline. Thus, points around the pectoral fins, every 30° relative to the centre of the disc (starting from the snout (0°) and not including the tail base, which does not possess any pectoral fin), were digitized to explore activation and termination of the first finbeat of the escape response (Fig. 1). Digitizing of the kinematics of the pectoral disc

(fins, centre, and midline) was initiated at the onset of the prod being thrust towards the animal and was terminated after the P. motoro clearly stopped yawing (i.e., the P. motoro reached its preferred swimming orientation away from the prod). Data were smoothed using a quintic spline (Walker 1998) in R (R Development Core Team 2008) and were transferred to Excel to calculate escape parameters, discussed below.

Escape responses were divided into phases 1 and 2 (Fig. 1). Phase 1 corresponds to the first finbeat of the escape response. Initiation of the first finbeat was defined as the onset of the movement of a peripheral point of the pectoral fin upwards and towards the centre of the disc; onset of movement was defined as a fin edge moving more than 5% of the distance between the centre of the disc and the fin edge at rest. Termination of the first finbeat corresponded to the time that all peripheral points of the fin involved in the response had reached the shortest distance from the centre of the disc (i.e., maximum fin roll). Phase 2 started with the termination of phase 1 and ended when the fish finished rotating and reached its preferred swimming orientation away from the prod. Of note, phases 1 and 2 of the escape responses in stingrays are not intended to be comparable with stages 1 and 2 of a C-start or S-start, which have entirely different kinematics; the phases are only intended to facilitate analysis and interpretation of the behaviour.

Prod location on the pectoral disc was measured as the angle between lines connecting the snout to the centre of the disc and the location of contact of the prod on the pectoral fins. Relative to the substrate, the prod angle of attack was similar across all recordings (119 ± 0.678 (mean ± SE)). Prod speed was measured as the speed of the prod upon contact with the animal. This was determined using three-dimensional

Pythagoras theorem from the displacement of the X (dorsal view camera), Y (dorsal view camera), and Z (horizontal view camera) coordinates and the duration of successive frames. Translational and yawing speeds of the centre of the disc were measured using the two-dimensional displacement of the centre point and its

33 relationship to the snout during each digitized frame. Translational and yawing accelerations were measured as the change in translation and rotational speeds, respectively, between successive frames.

Yaw displacement was measured as the total angle rotated throughout the entire escape response (beginning to end of phase 1 and phase 2). Turning radius typically describes the radius of an arc followed by the path of the body of the fish throughout a response (Blake et al., 1995; Walker et al., 2000). However, this definition can be misleading for a fish that rotates around a fixed point (i.e., zero turning radius) while the body translates linearly along the substrate, which would misleadingly equate to an arc of infinite radius. To provide evidence that P. motoro were in fact translating in a straight line while achieving yaw with a very small turning radius, we divided the linear displacement of the body disc by the distance of the path it actually travelled throughout the entire response. Ratios near 1 indicate the fish is translating in a straight line while the body rotates (yaws about a fixed point with low turning radius), rather than following the arc of a circle to turn. In comparison, if the path of the fish is following the arc of a circle, then the ratio would be near 0.64. Fast-start trajectory (FST) was measured as the angle between lines connecting the snout to the centre of the disc at rest and the location of the centre of the disc upon the termination of phase 1; FST revealed the ability of the fish to accelerate in different directions relative to the orientation of the snout at rest; an FST of 0°, 90°, and 180° define movement directly forwards, sideways, and backwards from the initial orientation of the P. motoro, respectively. Head orientation was measured as the angle between lines connecting the centre of the disc

34 at rest and the centre of the disc to the snout at the end of phase 1; head orientation revealed whether the body of the fish followed the orientation of the snout at the end of phase 1, as it does in BCF swimming fishes (see Introduction). For example, 180° indicates that the head is leading the centre of disc at the end of phase 1 (forward swimming), 90° indicates that the head is perpendicular to the motion of the centre of the disc (sideways swimming), and 0° indicates that the head is trailing the centre of the disc (backward swimming).

Statistical analysis

All statistical analyses were performed using R software (R Development Core

Team 2008). For some analyses, FST was categorized as forwards (0° to <45°), sideways (45° to <135°), and backwards (135° to 180°); paired t tests did not show differences between left and right sideways responses for the impacts of prod speed on escape speed and acceleration, so data from left and right sides were grouped together for analysis. For other analyses, FST was treated as a continuous variable. A total of seven forward, nine sideways, and eight backward responses were measured. For two of the sideways responses, the prod contacted the substrate rather than the pectoral disc. These trials were not included in analyses that considered the impact of prod speed on escape kinematics. However, they were included when considering the impact of prod location on FST and yaw displacement because the fish still showed a robust escape response consistent with those with similar prod locations, and there were not differences between statistical results when these data points were included versus not included. Linear regressions were employed to test the effects of prod

35 location on FST, FST on head orientation, and prod location on yaw displacement.

Linear regressions, accounting for repeated measures, were employed to test the effects of prod speed on translational acceleration and speed, as well as on yaw acceleration and speed. Escape performance (e.g., translational acceleration and speed) was then normalized by prod speed and a one-way ANOVA was used to test for differences between forward, backward, and sideways escapes. A p-value of 0.05 was the threshold for significance.

Results

Pectoral fin movement

As the prod contacted the animal, the initial finbeat (i.e., phase 1) was characterized as a rolling motion of the pectoral fins, up and over towards the dorsal midline of the body, effectively retracting P. motoro away from the threat (Fig. 1). Phase

1 in forward and backward escapes were always powered by two fins (left and right sides of the body simultaneously), whereas sideways escape was mostly powered with one fin contralateral to the prod location (77.8% of responses). Location of the onset and offset of the finbeat around the periphery of the disc was dependent on the prod location such that finbeats were observed to propagate (i.e., finbeat direction) from anterior to posterior, posterior to anterior, or the entire fin rolled up and over at once

(Fig. 1). During forward and backward responses, finbeat direction was always from anterior to posterior and posterior to anterior, respectively. During sideways escape, for

Fig. 1. Disc starts for forward response (upper row), sideways response (middle), and backward response (lower) of the ocellate river stingray (Potamotrygon motoro). Movement of 11 points around the periphery of the pectoral fin (white dots) were analysed relative to the centre of the disc to determine the direction of the resultant finbeat powering escape. During escape responses, fins rolled up and over towards the midline (black and grey shades indicate the dorsal and ventral pectoral fin surfaces, respectively). Phase 1 (broken arrow) corresponds to the first finbeat of the escape response. Phase 2 (solid arrow) was defined as the time from the termination of phase 1 to the time that P. motoro finished rotating or yawing and reached its preferred swimming orientation away from the prod.

72.7% of the total number of finbeats, the entire fin rolled up and over at once, whereas

18.2% occurred from anterior to posterior and 9.09% occurred from posterior to anterior.

Omnidirectional fast starts

From rest (phase 1), P. motoro displayed the capacity for omnidirectional movement across the substrate. However, the direction of movement was not dependent on the orientation of the head, but rather on the location of the prod. FST

(the trajectory of acceleration of the centre of the disc relative to the resting orientation of the fish, which describes the direction of movement of the body relative to the initial orientation of the head) decreased as prod location moved from the snout towards the tail (Fig. 2). Hence, P. motoro tended to escape more directly backwards (i.e., FST of

180°) as prod location shifted towards the rostrum, more directly sideways (i.e., FST of

90°) as prod location became more lateral, and more directly forwards (i.e., FST of 0°) as prod location shifted towards the base of the tail. Furthermore, head orientation (the angle between the centre of the disc at rest and the centre of the disc to the snout at the end of phase 1, which describes the orientation of the head at the end of phase 1 relative to movement of the body) changed from leading the centre of the disc when escaping forwards to a perpendicular orientation when escaping sideways and to trailing the centre of the disc when escaping backwards. Consequently, as P. motoro accelerated away from the prod, the body was not restricted to following the initial or final orientation of the head (Fig. 2).

Fig. 2. The ocellate river stingray (Potamotrygon motoro) can fast start away from a prod in all directions across the substrate; its body is not restricted to following the orientation of the head. A) The impacts of prod location on the fast-start trajectory (FST) measured

2 upon completion of phase 1 (linear regression, r = 0.942, pslope < 0.001, y = –1.13x + 180).

Prod location of 0° is the head and 180° is the tail. An FST of 0° indicates that P. motoro moves in the direction that its head was initially oriented, 90° indicates directly sideways (left or right) to the initial head orientation, and 180° indicates directly backwards. The broken line represents a line of unity. B) Circular plot showing the range of FST exercised by P. motoro from rest. Zero degrees indicates a fast start directly forwards, 90° indicates directly sideways, and 180° indicates directly backwards. Solid lines indicate escape that included leftward trajectory, whereas broken lines indicate rightward trajectory. C) The relationship between FST and head orientation (HO) of the escape response (r2 = 0.936, pslope < 0.001, y = –0.886x + 176). During phase 1, the body is not restricted to following the orientation of the head, whereby fish that move in the direction of the snout at the start of phase 1 (0° FST) are still moving in the direction of the snout at the end of phase 1 (HO near 180°), fish that move laterally relative to the orientation of the snout at the start of phase 1 (90° FST) are still moving laterally relative to the snout at the end of phase 1 (HO near 90°), and fish that move backwards at the start of phase 1 (180° FST) are still moving backwards at the end of phase 1 (HO near 0°). If the fish first rotated to follow the head during phase 1 when escaping away from the prod, then HO would remain near 180° throughout the full range of FST, which it does not. The broken line represents a line of unity.

Translational performance

An increase in prod speed evoked a faster phase 1 response such that for each category of escape direction, an increase in prod speed was associated with increased maximum translational speed and acceleration of the disc (Fig. 3). There were no differences between categories of escape direction for translational speed or acceleration during phase 1. An increase in prod speed was also associated with reduced latency between the time of contact of the prod and the initiation of movement

2 of the fins (pslope = 0.024, r = 0.209).

Phase 2 was defined as starting at the termination of phase 1 to the time that the animal finished rotating (i.e., yaw rotation) away from the prod (Fig. 1). During phase 2, the animal recovered from the initial finbeat by rolling the pectoral fins outward along the body to their original position. This recovery stroke was then followed either by an additional finbeat (57.1% of the responses for forward escape, 55.6% for sideways escape, and 87.5% for backward escape) or by passive gliding, as the animal transitioned into a forward swimming orientation away from the prod. During phase 2,

P. motoro did not achieve maximum translation accelerations that were greater than measured in phase 1; however, the maximum translational speed of phase 2 was usually greater than that obtained during phase 1 for sideways escape (77.8% of responses) and backward escape (62.5% of responses), but not for forward escape

(42.9% of responses), although the mean across measurements were all faster for phase 2 (Fig. 3). Hence, throughout the entire response (for each category of escape

Fig. 3. The impacts of prod speed on maximum translational performance of the centre of the disc during forward, backward and sideways escapes of the ocellate river stingray (Potamotrygon motoro). Triangles and dotted lines represent forward escape (0° to <45° fast-start trajectory (FST)), squares and broken lines represent sideways escape (45° to

<135° FST), and circles and solid lines represent backward escape (135° to 180° FST). (A) An increase in prod speed was associated with increased maximum translational acceleration of the centre of the disc measured during phase 1 in all escape directions

2 2 (forwards: r = 0.695, pslope = 0.0123, y = 34.9x – 1190; sideways: r = 0.854, pslope = 0.00182,

2 y = 28.4x – 408; backwards: r = 0.916, pslope < 0.001, y = 18.7x + 110). There were no differences in acceleration between the slopes of these three escape directions (p = 0.446). (B) An increase in prod speed was associated with increased maximum translational speed of the centre of the disc measured during phase 1 in all directions (forwards: r2 = 0.902,

2 pslope < 0.001, y = 0.952x + 12.8; sideways: r = 0.782, pslope = 0.00513, y = 0.971x – 0.626;

2 backwards: r = 0.937, pslope < 0.001, y = 0.813x + 23.7). There were no differences in speed between the slopes of these three escape directions (p = 0.113). (C) An increase in prod speed was associated with increased maximum translational speed of the centre of the disc measured during phase 2 of the escape response in all directions (forwards: r2 = 0.604,

2 pslope = 0.0244, y = 1.33x – 5.22; sideways: r = 0.567, pslope = 0.0309, y = 0.930x + 23.1;

2 backwards: r = 0.933, pslope < 0.001, y = 0.775x + 27.7). There were no differences in speed between the slopes of these three escape directions (p = 0.316).

43 direction), maximum translational acceleration always occurred during the first finbeat of phase 1, whereas maximum translational speed occurred during phase 1 or phase 2 depending on the escape direction and varying between individuals. There were no differences between categories of escape direction for maximum translational speed attained.

Yaw rotation

As P. motoro translocated away from the prod, they also yawed with essentially no turning radius (i.e., were able to spin without having to swim in an arc), seemingly in an effort to achieve a forward swimming orientation away from the prod while simultaneously moving directly away from the prod. This was demonstrated by the ratio of the direct-line displacement of the centre of the disc from start to end of the escape being nearly equivalent to the measured distance it travelled (0.953 ± 0.00038 (mean ±

SD)), where a value of 1 would indicate travel in a straight line compared with a value near 0.64 if the fish were following an arc of a circle while yawing. Initiation of yaw rotation tended to begin during the latter part of phase 1 (83% of total trials) or during phase 2 (8.3% of total trials); however, in a few instances, initiation of yaw rotation occurred simultaneously with the initiation of translational movement at the onset of phase 1 (8.3%). Throughout the entire escape response (i.e., phases 1 and 2), total yaw displacement increased as prod location became oriented more towards the head

(Fig. 4), where fish tended to rotate so that the head was eventually oriented away from the prod location by the end of phase 2 (Fig. 1). Although prod speed did not influence the yawing speed or acceleration in forward and sideways escapes, both increased with

44 prod speed in backward escape (Fig. 4). Maximum yawing speed tended to occur during phase 2 of the response (86.4% of observations), consistent with observations that initiation of yaw often occurred either late in phase 1 or during phase 2.

Discussion

Disc-start behaviours

We define fast-start escapes in stingrays with rounded pectoral discs as disc starts, functionally distinct from the more commonly described C- and S-starts employed by BCF swimming fishes, whereby the body is not restricted to following the orientation of the head, allowing the fish to escape in all directions across the substrate. A disc start is thus a unique type of fast start, with rapid, brief, and unsteady swimming used as a response to avoid or escape from a threat. Although the performances that we observed in the present study were not necessarily maximal, at least in all cases, they are notably greater than routine swimming and thus can be considered to be fast-start behaviour. It is not clear if disc starts involve Mauthner cell mediated activation, as are commonly associated with escape behaviour in many fishes (Eaton et al., 1977; Eaton et al., 2001). This distinction between a maximal, Mauthner cell mediated response and an unsteady swimming maneuver to rapidly retreat from a threat may have implications for the kinetics and kinematics of swimming during the response, and hence, interpretation of how different types of fast starts are accomplished.

Fig. 4. Yaw rotational performance of the ocellate river stingray (Potamotrygon motoro) during escape. Triangles and dotted lines represent forward escape (0° to <45° fast-start trajectory (FST)), squares and broken lines represent sideways escape (45° to <135° FST), and circles and solid lines represent backward escape (135° to 180° FST). (A) The impact

46 of prod location on the magnitude of yaw rotation measured from the start of phase 1 to

2 the end of phase 2 (r = 0.583, pslope < 0.001, y = –0.661x + 132). As prod location became more anterior (i.e., 0° snout, 180° tail), P. motoro rotated a greater amount toward a forward swimming orientation away from the prod, but P. motoro did not always swim directly away from the prod as evidenced by the scatter around the regression line. The thin broken line represents a line of unity. (B) An increase in prod speed was associated with increased maximum yaw speed, measured throughout the entire response, during backward escape, whereas there was no significant effect of prod speed on yaw speed in forward and

2 2 sideways escapes (forwards: r = –0.139, pslope = 0.627, y = 1.032x + 255; sideways: r =

2 0.0752, pslope = 0.277, y = 3.46x + 260; backwards: r = 0.681, pslope = 0.00716, y = 3.70x + 307). (C) An increase in prod speed was associated with increased maximum yaw acceleration achieved throughout the entire response in backward escape, whereas there was no significant effect of prod speed on yaw acceleration in forward and sideways

2 2 escapes (forwards: r = –0.200, pslope = 0.968, y = –4.05x + 804; sideways: r = 0.371, pslope =

2 0.0864, y = 95.9x + 3850; backwards: r = 0.823, pslope = 0.00116, y = 193x – 875).

Potamotrygon motoro is a dorsoventrally flattened fish that uses high drag and flexible pectoral fins to power MPF swimming. This mode of swimming, along with a relatively rigid body axis, have been proposed to lead to relatively poor maneuverability performance compared with most BCF swimmers (Parson et al., 2011). However, the swimming and maneuvering abilities of these fish have not been studied under circumstances where the fish might be expected to exert greater efforts than during routine swimming, and thus, if they might exceed these expectations of modest performance. It was found that the morphology of P. motoro promotes omnidirectional fast-start performance (i.e., acceleration and speed) along the benthos (i.e., disc starts), including the ability to move in all available freedoms of motion across the substrate: forwards–backwards, side-to-side, and yaw (Figs. 2 and 4). P. motoro demonstrated similar translational performance regardless of the orientation of the head relative to the direction of the prod (Fig. 3). Furthermore, as these fish retracted away from the prod, they simultaneously rotated in the yaw (i.e., left or right) freedom with next to no turning radius, quickly achieving a forward swimming orientation away from the prod, with a maximum yaw rate that was up to 27-fold greater than what was previously described for benthic stingrays during routine swimming maneuvers (Parson et al., 2011) (Fig. 4).

We thus conclude that the morphology of stingrays affords a unique tactic of fast-start swimming that allows impressive flexibility in escape behaviour in a benthic environment in spite of the apparent constraints of a rigid body axis.

Benthic stingrays can accelerate along the forward–backward, side-to-side, and yaw freedoms such that the body is not restricted to following the orientation of the

48 head; this range of maneuverability appears to be unique to disc starts (Figs. 1 and 2).

Some MPF swimming fishes can swim backwards, but it is unclear if they are capable of fast-start performance in this direction, particularly with the capacity to simultaneously rotate within a low turning radius (Lannoo and Lannoo, 1993; Webb and Fairchild,

2001). Most fishes use BCF swimming to power fast starts, performing C-starts and S- starts for escape (Domenici, 2010a; Domenici, 2010b; Domenici and Blake, 1997;

Wakeling, 2001). During these fast starts, the body is restricted to following the orientation of the head. This would appear to restrict BCF swimming fishes in the extent of omnidirectional movement that they can attain, at least over a small distance of travel. Although most BCF swimming fishes are likely not capable of fast starting in the sideways or backward freedoms, it is possible that omnidirectional maneuverability

(across the substrate) could still be achieved via forward swimming and high yaw displacement and rates within low turning radii. This type of rotational performance may be promoted by extremely flexible bodies (e.g., eel-like fishes) performing C-start behaviours. However, C-starts and S-starts tend to be powered by a preparatory stroke followed by a main propulsive stroke so that BCF swimming fishes would seemingly require two BCF motions to displace their entire body away from the initial orientation. In contrast, stingrays can readily retract the entire pectoral disc directly away from a prod via the first beat of the pectoral fins. Some BCF swimming fishes, with elongated and flexible bodies (e.g., eel-like fishes) living in structurally complex environments, can retract their head towards the body in one motion, often pulling their head or tail into a tunnel-like feature, defined as a withdrawal response (Bierman et al., 2004; Ward and

Azizi, 2004). Nevertheless, apparently these withdrawal responses often do not involve a propulsive component (i.e. acceleration) of the body.

A flattened and rounded pectoral disc likely promotes the capacity for omnidirectional disc starts across the substrate. The large and flexible pectoral fins can create propulsive forces in any direction across the substrate, whereas the dorsoventrally flattened shape of stingrays will reduce the mass of water that must be displaced as the fish accelerates along the anteroposterior and lateral axes (Webb,

1988). The latter should decrease the drag forces retarding acceleration, providing stingrays with the ability to effectively slice through the water and obtain comparable fast start performance regardless of direction across the substrate. P. motoro were never observed to escape upwards. Although they were never prodded from below or directly lateral on the edge of the disc to intentionally invoke an upward escape, the lack of upward escape is more likely due to the high profile drag that would be experienced by the fish during acceleration in the vertical axis (relative to its resting position), as well as the relatively substantial mass of water that must be displaced to move upwards due to the large surface area of the dorsal surface of the flattened pectoral disc.

Furthermore, even when a threat is moving along the substrate, stingrays maintain escape along the substrate and not up (S.G. Seamone, personal observation).

Pleuronectiformes (i.e., flatfish such as sole and flounder) are also flattened in the same plane as the substrate; these fishes are laterally flattened and flipped on their side

(Friedman, 2008). Yet rather than performing disc starts, flatfish perform BCF escape responses, undulating their longitudinal body axis and power C-starts perpendicular to

50 the ground (Brainerd et al., 1997; Webb, 1981). Whether these fishes have the capacity to disc start backwards and sideways is not clear; however, it is highly unlikely due to the asymmetrical shape from anterior to posterior and the reliance on BCF propulsion.

Hence, it is not only the flattened profile in stingrays that is responsible for the range in maneuverability across the substrate, but the flexible high-drag pectoral fins that wrap around the head and body that appear key in displacing water in different directions to promote disc-start behaviours.

Drag-based propulsion

The flexible and low aspect ratio (i.e. rounded) pectoral fins roll up and over during the disc start, forming a large propulsive surface that is dragged through the water along the anteroposterior and lateral axes (Fig. 1). This presents a large surface to exert drag-based forces on water and displace water for powering movement. In sideways escape, the entire fin on one side of the fish tends to roll up and over towards the dorsal midline at once. During forward and backward escapes, rolling of the fin is initiated along the anterior or posterior portion of the pectoral fins, respectively, and is propagated along the anteroposterior axis. These fin movements would generate momentum in the water, in different directions around the body depending on the direction of fin movement, to effectively accelerate the fish across the substrate in the counter direction to the finbeat (Webb, 1984a). The flexible and rounded fins of benthic stingrays are important for promoting drag-based propulsion and would appear to restrict disc starts to benthic stingrays, and possibly skates, that possess this rounded morphology. As such, although it has not yet been explored, it is unlikely that pelagic

51 stingrays can effectively perform disc starts with their high aspect ratio, triangular- shaped fins that tend to produce lift rather than generate drag (Fontanella et al., 2013;

Franklin et al., 2014; Rosenberger, 2001).

Modulation of escape performance and direction

Given that escape appears to be an energetically expensive event (Jayne and

Lauder, 1993; Westneat et al., 1998), it is not entirely surprising that fishes do not always employ maximum performance in response to a threat (Seamone et al., 2014;

Ydenberg and Dill, 1986). In the present study, contacting P. motoro with a greater prod speed evoked a greater magnitude of escape acceleration and speed in all translational directions away from the threat (Fig. 3), increased yaw acceleration and speed during backward escapes (to achieve a forward swimming orientation away from the threat) (Fig. 4), and reduced latency between the time of contact of the prod and the initiation of fin movement (see Results). Although maximum rates of yaw rotation in forward and sideways escapes were also associated with higher prod speeds, the relationships were not significant, which might be based in the relatively small amount of yaw displacement required to achieve a forward swimming orientation away from the prod (i.e., during forward escape, the fish is already oriented away from the prod).

Nevertheless, P. motoro clearly have the capacity to modulate the level of performance relative to the intensity of tactile stimulation, which may be an effective behaviour in a benthic environment. Routinely, benthic stingrays likely contact a range of different objects that are not threatening, especially in turbid environments such as the South

American river basins where P. motoro resides. Always employing maximum

52 performance and escaping away from all tactile interactions would be a waste of energy, as well as the potential for missing feeding and mating opportunities (Ydenberg and Dill, 1986).

Predators of stingrays likely include various types of ambush predators, in addition to some large animals that are very effective at swimming through the water column at high cruising and burst speeds, such as sharks and marine mammals (Cliff and Dudley, 1991; Strong et al., 1990; Visser, 1999). The ability of stingrays to escape downwards from such predators is restricted due to the substrate, and escaping upwards is likely not an appropriate behaviour in the face of faster predators with the capacity to attack at high speeds, particularly given the aforementioned limitations on moving a disc-shaped body in a direction other than parallel with the substrate.

However, larger predators tend to have lesser rotational maneuverability (Domenici,

2001). Thus, the ability of stingrays to readily escape in any direction across the substrate, and to quickly yaw to any direction while simultaneously accelerating away from an attack, promotes movement along the benthic plane that least exposes the stingray to the aptitudes of these predators, and perhaps forces the predator to reorient itself towards the escaping stingray, a time-consuming process. Therefore, disc-start behaviours across the substrate would seem to exploit the limitations of a large swimming predator. Remaining close to the substrate, stingrays could then bury when the opportunity arises.

Maneuverability

Disc-start behaviours in stingrays challenge the concept of maneuverability used to describe fast starts in other fishes. Our current understanding of maneuverability in fishes stems from those that undulate their body axis and power C-starts and S-starts

(Domenici, 2010a; Domenici, 2010b; Domenici and Blake, 1997; Wakeling, 2001), where the body follows the orientation of the head during the fast starts and maneuverability is achieved via forward thrust in conjunction with angular rotation

(predominantly yaw). Thus, maneuverability in fishes is most often discussed in the context of turning (i.e., yaw rotation) angle, turning radius, and turning rate, whereas forward acceleration and speed are often discussed separately as propulsive performance (Domenici, 2010b; Domenici and Blake, 1997; Domenici et al., 2004).

However, a maneuver defines a change in position (i.e., acceleration) of an object.

Thus, fishes suspended in a three-dimensional environment have the potential to accelerate (i.e., maneuver) and escape predators in six degrees of freedom of motion: three translational freedoms (forwards–backwards, sideways, up–down) and three rotational freedoms (pitch, yaw, and roll). Although translational acceleration (forwards, backwards, and sideways) in P. motoro is low relative to maximum forward acceleration in BCF swimmers, where some species can accelerate over 100 m/s2 (Domenici and

Blake 1997), BCF swimmers are not known to accelerate backwards or sideways to nearly the same extent that stingrays can. Furthermore, although yawing rates of P. motoro were relatively low (Fig. 4) compared with the maximum measured in BCF swimmers, whereby numerous species can rotate over 2000°/s (Domenici 2001), yawing in BCF swimmers typically occurs first as part of a C-start, followed by forward

54 translation, whereas P. motoro can yaw and translate simultaneously. Thus, although the relatively high yaw rates of BCF swimmers contribute to a shorter prelude to forward movement, they do not in themselves necessarily contribute to rapid translation towards or away from a prey or predator. A stingray that employs slower yaw rates, but which can be accomplished with simultaneous translation, may actually have a faster effective response. Thus, high yaw rates in BCF predators may not be as effective for pursuing an animal using disc starts as they might first appear. Also, stingrays have the capacity to yaw with next to no turning radius (He and Zhang 2013), perhaps zero (spin on the spot), that BCF swimmers cannot accomplish. Thus, interpreting maneuverability in fishes may be more complex than simply the ability to yaw at a high rate. We propose that although stingrays may not exhibit rotational and translational accelerations as high as some BCF swimmers numerically, they should still be considered to be highly maneuverable.

Chapter 3

The ocellate river stingray (Potamotrygon motoro) can generate suction

to station hold along the substrate

Abstract

This study tested whether the flattened and rounded pectoral disc of the ocellate river stingray (Potamotrygon motoro) has the capacity to generate suction along the substrate, as might be beneficial for station holding against high current or predator dislodgement. Animals were recorded using videography as lifting forces were applied to the tail of the stingray and quantified using a scale, while the suction pressure underneath the stingray was measured using a u-shaped manometer. There was a strong and positive relationship between the maximum lifting force exerted on the tail and the maximum suction pressure. During a suction event, the stingray thrusted its body downwards and passed waves of bending along the pectoral fins. Body and finbeat movements had a positive and synchronous relationship, but neither the frequency or number of body and finbeat movements had an impact on suction pressure. Dye injected beneath the stingray indicated that pectoral fin bending and spiracle pumping expelled water from underneath the disc. Hence, in response to a lifting force, stingrays create suction via the generation and maintenance of a vacuum and possibly Stefan adhesion beneath the pectoral disc.

Keywords: stingray, benthic, biomechanics, suction pressure, station holding

Introduction

Station holding is a behaviour by which an animal counters a displacement inducing disturbance, such as high current or predator dislodgement, to maintain position in a habitat (Carlson and Lauder, 2010; Gerstner, 2007; Webb, 1989; Webb et al., 1996). Fishes demonstrate a range of mechanisms to promote station holding, including rheotaxis, avoiding currents by finding refuge behind structures or in the benthic boundary-layer, in addition to morphological features that enable the fish to reduce drag and lift, and generate friction and suction forces along the substrate

(Arnold and Weihs, 1978; Beckert et al., 2015; Carlson and Lauder, 2010; Carlson and

Lauder, 2011; Ditsche et al., 2014; Gerstner, 2007; Gerstner and Webb, 1998; Gibson,

1969; Liao, 2003; Liao et al., 2003; Sewell, 1925; Taft et al., 2008; Wainwright et al.,

2013; Webb, 1989; Webb et al., 1996; Wilga and Lauder, 2001). Given that station holding has been proposed to be important for animals to potentially reduce energetic costs, to remain close to food sources, mates and desirable habitat, and in predator- prey behaviours (Betz and Kölsch 2004; Isabelle 1995; Liao et al. 2003; Nilsson et al.

2010; Tierney et al. 2017), exploring station holding behaviours in animals is expected to provide insight into relationships between shape, behaviour and ecology.

Fishes that are flattened in the same plane as the substrate, such as the laterally compressed flatfish (order Pleuronectiforms) and dorsoventrally depressed skates

(order Rajiformes) and rays (suborder Myliobatoidei), are more common to benthic than pelagic ecosystems in marine and freshwater environments, indicating a flattened shape might be particularly effective for living in a benthic environment. These fishes

57 demonstrate a range of locomotor mechanisms along the benthos including punting, rajiform undulation, disc-start escapes, ground effects, and burying behaviours (Blevins and Lauder, 2012; Corn et al., 2018; Rosenberger, 2001; Rosenberger and Westneat,

1999; Seamone et al., 2019; Webb, 1981; see Chapter 2, 4, and 5). Furthermore, these fishes are often exposed to disturbances including high currents in river systems and in surf and tidal zones, and predators such as sharks and marine mammals (Allen and

Pondella 2006; Goes de Araújo et al. 2004; Maxwell et al. 2009; Van Der Veer et al.

1994; Vaudo and Lowe 2006), suggesting that these fish may exercise mechanisms to maintain position along the benthos. Of note, the body of these fishes is similar to a foil, such that the flattened shape has an extremely high surface area to volume ratio, and therefore, parallel to the flow promotes relatively low drag and high lift coefficients

(Arnold and Weihs, 1978; Webb, 1989). Thus, station holding in these fishes is anticipated to reveal intriguing mechanisms to counter displacing vertical lifting forces along the substrate, while maintaining the capacity to perform the other aforementioned locomotor behaviours (i.e. swimming and burying). In response to current, both flatfish and rays demonstrate rheotactic behaviours, and as the fish start to slip along the benthos, fin-beating, body arching, body clamping along the substratum might be observed (Arnold and Weihs, 1978; Webb, 1989). Passing waves of bending along their propulsive surfaces has been suggested to increase the flow underneath the fish relative to that of the upper surface, and consequently, reduce the pressure gradient, and hence, the lifting force (Webb, 1989). The highly flattened ventral surface of these fishes has also been suggested to potentially act like a suction disc, enabling fishes to

58 adhere to the benthos (Arnold and Weihs, 1978; Fontanella et al., 2013), although this has not yet been explored.

The ocellate river stingray (Potamotrygon motoro) is a dorsoventrally flattened fish common to river systems of South America, that has rounded pectoral fins with high surface area, that wrap around the head and body, and are collectively referred to as the pectoral disc. These fins are used to power swimming via flexible movements, whereas the tail is long and thin with minimal caudal fin, and is used primarily for defence with a barbed stinger (Blevins and Lauder 2012; Campbell 1951; Fontanella et al. 2013; Rosenberger 2001; Rosenberger and Westneat 1999; Seamone et al. 2019).

This study aimed to test the prediction that the high surface area of the flattened ventral surface of the pectoral disc can generate suction pressures to hold station along the benthos in response to a displacing lift force, and to describe the mechanism that might enable them to achieve these pressures. Lifting forces were applied to the tail of the stingray and the suction pressure underneath the pectoral disc was measured. Body movements, and water movements under and around the stingray were also observed via videography and dye to provide insight into the mechanism by which stingrays may be generating suction pressures to station hold along the substrate.

Materials and methods

Animal husbandry

All procedures were approved by the University of Calgary animal care committee, following Canadian Council on Animal Care guidelines. Eight P. motoro were purchased from a licenced supplier and used in experiments (pectoral disc width

16.1 cm ± 0.509 S.E.M.; mass 216 g ± 21.5 S.E.M.). These fish were housed at the

University of Calgary in a cylindrical tank (180 cm diameter x 70 cm height, approximately 1400 L) with flow-through freshwater at 27° C, pH of 6.5, bubbled with air, 12:12 hour photoperiod. The stingrays were provided 2.5 cm3 of frozen bloodworms per stingray once a day, and were monitored to ensure feeding and health. Water chemistry was measured daily, and the fish had regular veterinary oversight. Stingrays were acclimated for two months prior to experiments.

Experimental apparatus

A separate rectangular tank was used for experimentation, with dimensions 45.5 cm length x 24 cm width x 20.5 cm height, made of polysulfone with a smooth bottom

(Ancare, New York, USA). The tank was placed 12 cm above the floor using a stand that secured the base of each end of the tank. A hole was drilled into the bottom of the tank, 14 cm from the lengthwise end of the tank, and a thick-walled Tygon tube with an inner diameter of 5 mm was affixed in the hole so that the open end was flush with the bottom of the tank. The tube was filled with water, and the opposite end was connected to a u-shaped manometer, filled with water and open to atmospheric air, used to

60 measure the pressure beneath the ventral surface of the stingray relative to the surrounding water in the tank. Maximum pressure-change measurements of the manometer were 21cm with a resolution of 0.1 cm. The manometer was secured to a stand and placed beside the tank, so the reference level was at the surface of the water in the tank (i.e. zero pressure when no external forces were applied). The tank was filled with water from the holding tank to a height of 16 cm from the bottom, and refilled with fresh water before each trial.

Experimental procedure

To measure whether stingrays could achieve suction pressures beneath the ventral surface of the disc, stingrays were first transferred individually to the experimental tank using a rubber mesh net. A soft lace, 1 cm in diameter, was wrapped around the tail at the location of the barbed stinger, and was secured in place using a small, spring locking, plastic lace-clamp. The lace was connected to a digital spring scale, with a resolution of 5 grams (Next-Shine, Guangdong, China), calibrated with weights of known mass. As the scale was lifted vertically, force applied to the tail (as indicated on the spring scale) and the pressure change in the manometer were recorded using a Panasonic Lumix DMC-FZ1000 digital camera filming at 60 fps

(Panasonic Corporation, Osaka, Japan). Force was applied to the tail for a maximum of

5.5 seconds (average 2.80 seconds ± 0.524 S.E.M.), long enough to obtain a steady lifting force and associated manometer recording. All 8 stingrays were initially subjected to these measurements of suction pressure. A second round of measurements was then made on the fish, whereby an additional suction event was measured for 3 of the 8

61 fish (pectoral disc width 16.79 cm ± 0.326 S.E.M.; mass 226.76 g ± 60.7 S.E.M.), for a total of 11 pressure measures, while the remaining 5 fish were used for observations of water flow during the suction event, as discussed below.

To explore whether stingrays were flushing water from beneath the pectoral disc during suction events, a 1.0 ml syringe filled with green food dye (Club House, Ontario,

Canada) was connected to the hole in the bottom of the tank. Just prior to lifting the tail of the ray to induce a suction response, approximately 0.5 ml of dye was injected through the hole in the bottom of the tank, beneath the ventral surface of the ray.

Subsequent movement of the dye by water flow generated by the ray was recorded using two GoPro Hero 6 cameras (GoPro Inc, California, USA), one filming from the dorsal and one filming from the lateral view at 240 fps at 1080p. Force applied to the tail and manometer pressures were not recorded during these trials, but similar forces were applied as during the other trials, as confirmed visually with the spring scale. This procedure was completed once each for five stingrays (pectoral disc width 16.4 cm ±

0.326 S.E.M.; mass 210 g ± 17.07 S.E.M.).

Data analysis

Video recordings of the scale and manometer were analysed using ImageJ

(bundled with 64-bit Java 1.8.0_112, Wayne Rasband Developers 1997). The duration of the burying event was defined as the initiation of the pressure change as the lifting force was applied to when the pressure change plateaued, and maximum pressure was

62 achieved. At the time the maximum pressure change (centimeters of H20) was noted from the manometer, the force recorded by the spring scale was noted. ImageJ was also used to assess the behaviours of the rays during the suction event (e.g. downward body thrusts, pectoral finbeats), whereas the movement of the dye around the pectoral disc from the dorsal and lateral view videos were visualized in Premier Pro CC 2017

(Adobe Systems, 2003). The number of finbeats was defined as the number of times the fins passed vertical waves of bending along the lateral edge and then returned to a flattened position. The number of body thrusts was defined as the number of times the body thrusted downwards. A position on the dorsal pelvis, defined by skin pigments, was tracked throughout the suction event, and thrusting velocity and acceleration were measured as the change between distance and speed, respectively, of that positions over the period between subsequent frames. The average frequency of the finbeats and the body thrusts was calculated as the number of finbeats or body thrusts divided by the duration of the suction event.

Statistical analysis

Using R software (R Development Core Team, 2008), linear regressions were employed to test the relationship between the independent variables (lift force applied to the stingray’s tail, duration of the suction event, the number and frequency of finbeats, the number and frequency of body thrusts, and the velocity and acceleration of body thrusts), and the dependent variable, suction pressure, measured by the manometer beneath the ventral surface of the pectoral disc. In addition, linear regressions were

63 used to test the relationships between finbeat and body thrust number, and finbeat and body thrust frequency. A p-value of 0.05 was used to test for significance.

Results

Suction pressure, lifting force and kinematics

As the lifting force on the tail increased, suction pressure beneath the ventral surface of the pectoral disc increased, and as the duration of the suction event increased, the suction pressure tended to increase, although this relationship was marginally not significant (Fig. 1). Furthermore, as the rate of change in the lifting force increased, the rate of change in the suction pressure increased (Fig. 1). Video observations revealed that the anterior portion of the pectoral disc remained in contact with the substrate, while

P. motoro passed waves of bending along the posterior half of the disc and thrusted the body downwards. As the number and frequency of body thrusts increased, the number and frequency of finbeats increased, and body thrusts and finbeats tended to occur in synchrony such that nearly each body thrust was accompanied by a finbeat (Fig. 2).

The number of finbeats and their frequency did not have an impact on suction pressure

(Fig. 3), and neither did the number of body thrusts or frequency (Fig. 3). However, as the maximum downwards acceleration of the body thrusting increased, suction pressure increased, and while marginally not significant, a similar relationship appeared to exist for maximum downward velocity of the body thrusting (Fig. 4).

Fig. 1. A lifting (vertical) force was exerted on the ocellate river stingray’s (Potamotrygon motoro) tail for a 1-5.5 second duration and the pressure beneath the body of the animal was measured simultaneously. A) The relationship between the lifting force that was applied to the stingray’s tail and the maximum suction pressure generated beneath the

2 ventral surface of the pectoral disc (pslope < 0.001, r = 0.80, y = 1.3x - 0.33). B) The relationship between the duration of the suction event and the suction pressure generated

2 beneath the ventral surface of the pectoral disc (pslope = 0.093, r = 0.20, y = 1.76x – 1.28). C) The relationship between the rate of change in the lifting force and the rate of change

2 in suction pressure (pslope < 0.001, r = 0.70, y = 0.895 + 0.481).

Fig. 2. A) The relationship between the number of finbeat and body thrusting motions

2 exercised by the ocellate river stingray (Potamotrygon motoro) (pslope < 0.001, r = 0.866, y = 0.88x + 0.65). B) The relationship between the average frequency of the finbeat and body

2 thrust motions (pslope < 0.001, r = 0.77, y = 1.01x + 0.14)

Fig. 3. A) The relationship between the number of times the body thrusted downwards towards the bottom of the tank and the suction pressure that was measured beneath the ventral surface of the pectoral disc of the ocellate river stingray (Potamotrygon motoro)

2 (pslope = 0.22, r = 0.07, y = 0.66x + 3.11). B) The relationship between the average frequency of the downward body thrusts and the suction pressure that was measured beneath the

2 ventral surface of the pectoral disc (pslope = 0.67, r = 0.02, y = -0.79 + 7.59). C) The relationship between the number of finbeats and the suction pressure generated beneath

2 the ventral surface of the pectoral disc (pslope = 0.135, r = 0.145, y = 0.84 x + 2.17). D) The

2 relationship between average finbeat frequency and suction pressure (pslope = 0.65, r = 0.00, y = -0.71x + 7.58).

Fig. 4. A) The relationship between the maximum velocity of the downward body thrusting and the suction pressure generated beneath the pectoral disc of the ocellate river stingray

2 (Potamotrygon motoro) (pslope = 0.053, r = 0.28, y = 0.57x + 2.48). B) The relationship between the maximum acceleration of the downward body thrusting and the suction

2 pressure generated beneath the pectoral disc (pslope = 0.006, r = 0.540, y = 0.02x + 1.84).

Dye observations

Movement of the dye that was injected beneath the body disc at the start of a suction event revealed that the bending fin movements flushed water from beneath the ventral surface of P. motoro, exiting along the lateral edge of the posterior half of the disc (Fig. 5 & 6). The entire lateral edge of the pectoral disc was then pressed down onto the substrate immediately following a passing wave, and dye was not expelled from under the pectoral fins until another wave passed (Fig. 5). In three P. motoro, dye was also observed to be jetted out of the spiracles while the animals were generating suction onto the substrate, indicating that water was being withdrawn from under the body disc through the gill slits, passed across the gills and ejected out the spiracles

(Fig. 6).

Fig. 5. A lateral view of the ocellate river stingray (Potamotrygon motoro) exercising a finbeat to flush dyed water from beneath the pectoral disc, followed by pressing the fin along the substrate to create a seal.

Fig. 6. A dorsal view of the ocellate river stingray (Potamotrygon motoro) exercising a finbeat (A1-4) and a branchial jet (B1-4) to flush water from beneath the pectoral disc.

Discussion

Fishes may resist upward lifting forces to maintain position in response disturbances such as high currents in rivers and surf zones, or predator dislodgement from the benthos (Arnold and Weihs, 1978; Webb, 1989). Given that the foil-shaped body in flattened fishes (see Introduction), such as stingrays, skates and flatfish, promotes relatively low drag but high lift coefficients, foil-shaped fishes are expected to be equipped with mechanisms to station hold in response to lifting forces (Arnold and

Weihs, 1978; Fontanella et al., 2013; Webb, 1989). This study demonstrates that the pectoral disc of stingrays can maintain a range of suction pressures exerted onto the substrate in response to upwards displacement forces exerted on the animal, via lifting the tail vertically in this case (Fig. 1). Observation of dye movement from beneath the body disc during suction events (Fig. 5 & 6), suggested P. motoro flushed water out from beneath the ventral surface of the disc via synchronous movements of the body and pectoral fins, movements that were closely related in number and frequency (Fig.

2), and sometimes supplemented by branchial flow. Therefore, it appears that stingrays most likely generate suction pressures beneath the pectoral disc via a vacuum that is formed and maintained beneath the pectoral disc by the motions of the body as the animal is pulled away from the substrate, and possibly due to Stefan adhesion as the body and fins are lifted out of contact with the substrate. Stefan adhesion refers to the normal stress acting between two surfaces when their separation is attempted in a fluid, whereby viscous forces resist the flow of the fluid into the forming space and consequently resists separation of the two surfaces (Brainerd et al., 1997). Thus, this phenomenon is particularly relevant to animals with a high and flattened surface area

72 living along the substrate in an aquatic environment (Brainerd et al., 1997; Wainwright et al., 2013), such as the ventral surface of the pectoral disc in stingrays.

As stingrays locomote along the benthos, such as in punting behaviours, they glide over a thin layer of water between the pectoral disc and the ground (Macesic and

Kajiura, 2010; Macesic et al., 2013). Furthermore, these fish draw water in beneath the rostrum and the pectoral disc to feed (Wilga et al., 2012). However, for suction to be achieved and effective, a seal between the substrate and the pectoral fins must be formed and maintained (Sewell, 1925; Wainwright et al., 2013). Motions of the body and fins were often executed as the lifting force was applied to the fish, which would help create the seal, and seemed to be repeated as the fish started to lose the seal as it began to slip from the substrate. As the body accelerated downwards, water located between the ventral surface of the disc and the substrate was forced towards the lateral edge of the fins, and was then flushed from beneath the disc via rolling motions along the posterior half of the pectoral fins, a similar bending motion as to when these fish exercise fast-start escape and bury into the substrate (Fig. 5) (Seamone et al., 2019; see Chapter 2, 4 and 5). Increasing the maximum acceleration and velocity of the body thrusts appeared to flush more water from beneath the disc, leading to a greater capacity to resist the lifting force that was applied at the tail (Fig. 4). In contrast, the number and frequency (the number of motions/duration of the suction event) of fin and body motions did not have a significant impact on suction pressure, and hence, these parameters were probably more closely associated with continued maintenance of the

73 seal as the ray started to slip from the substrate, which would allow P. motoro to resist lifting forces for longer durations (Fig. 1).

The pectoral disc of P. motoro permits a relatively large suction disc relative to other fishes (Beckert et al., 2015; Ditsche et al., 2014; Gerstner, 2007; Wainwright et al.,

2013). The highest recorded suction pressure of 16.7 cmH2O (about 1640 Pascal), was associated with 13 N of lifting force applied to the tail. This force would yield 1640

Pascal if exerted over an area of about 79 cm2, equivalent to a 10 cm diameter disc.

For the ray from which this recording was obtained, the disc was 16 cm in diameter.

Assuming the suction disc to be a symmetrical circle, this suggests that approximately two thirds of the disc area (10 cm/16 cm) is effectively generating suction against the substrate, and one third of the disc around the perimeter was involved in the maintenance of the seal. Furthermore, this ray was about 180 g, and hence, it resisted a force equivalent to about 7 times its weight in air. Although we did not aim to explore the maximum suction pressures that the stingrays could maintain, so as to avoid harming the fish, stingrays are probably limited in the capacity to generate suction relative to fishes with organs that are used primarily for adhesion (Ditsche et al., 2014;

Gibson, 1969; Wainwright et al., 2013). For example, with suction forces between 80-

230 times the body mass have been measured for the northern clingfish (Gobiesox maeandricus) (Ditsche et al., 2014; Wainwright et al., 2013). It is possible that selective pressures on the pectoral fins of stingrays to be employed for other behaviours, such as routine and fast-start swimming, and burying (Blevins and Lauder,

2012; Rosenberger and Westneat, 1999; Seamone et al., 2019; see Chapter 4 & 5), might limit the suction capabilities in stingrays.

The stingrays used in this study were creating suction against a smooth, plastic surface, which would usually differ from the particulate and sometimes relatively coarse substrate in their environment, and would likely reduce the ability to form a seal and allow some leaking of water beneath the animal when a suction is formed. That said, the northern clingfish, Gobiesox maeandricus, is able to adhere to irregular surfaces

(Wainwright et al., 2013), and the ability of the pectoral fins of stingrays to expel water from beneath the body disc, and furthermore, potentially mold their ventral surface to an irregular surface due to their anatomical flexibility, might allow the animals to maintain some suction even on relatively coarse substrates. Furthermore, given that stingrays tend to inhabit environments with sandy and mostly fine particulate sediments such as sand (Corcoran et al., 2013; Harada and Tamaki, 2003; Semeniuk and Dill, 2005;

Stokes and Holland, 1992), where movement of water through the sediment would be limited, they likely can achieve suction against the substrate in their natural environment. While this might limit the force and duration of a suction event in such environments, it still may afford a mechanism that flattened fishes can use to resist lifting displacement, as might occur in response to displacement from high currents in rivers and tidal and surf zones (Allen and Pondella, 2006; Arnold and Weihs, 1978;

Fish and Hoffman, 2015; Goes de Araújo et al., 2004; Webb, 1989), and when predators such as marine mammals and sharks might dislodge flattened fishes from the substrate (Cliff, 1995; Myers et al., 2007; Pádraig, 2000). Hence, understanding the

75 limitations of suction performance along different substrate surfaces should be explored further, because it holds promise to provide insight into the relationship between the biomechanics and ecology of these fishes.

Chapter 4

The ocellate river stingray (Potamotrygon motoro) exploits vortices of sediment

to bury into the substrate

Abstract

Particle image velocimetry and video analysis were employed to describe the mechanism used by the stingray Potamotrygon motoro to bury into sand. P. motoro repeatedly and rapidly pumped the body up and down while folding the posterior portion of the pectoral fin disc up and over, drawing water in and suspending sediment beneath the pectoral disc. As the fins folded up and over, vortices of sediment travelled along the ventral surface of the fins toward the fin tips, and were then directed onto the dorsal surface of the fins and towards the dorsal midline of the ray, where they dissipated and the sediment settled. As displacement and speed of the body pumping and finbeat motions increased, the speed of the sediment translating across the dorsal surface increased; accordingly, sediment coverage of the dorsal surface increased. Mean burying depth and coverage were ~2 cm and 83%, respectively, and the head and tail buried less than the pectoral fins and body. However, in the most vigorous burying events, vortices of sediment shed from each fin collided at the midline and annihilated, reorienting the flow and sending jets of sediment towards the head and the tail to effect complete burial. By directing vortices of fluidized sand, nearly full coverage was accomplished with little body displacement into the substrate, a mechanism of burying that appears to circumvent the constraints of high skin-drag that would otherwise limit the ability of a flattened body with high surface-area to dig through sediment.

Keywords: Burying, stingray, benthic, vortices, particle image velocimetry, video

Introduction

Burying and burrowing are locomotor behaviours exercised by animals that involve the displacement of sediment to move beneath the substrate. While these terms are often used interchangeably in the literature, burying has been defined as being encased by sediment, whereas burrowing has been defined as forming a cavity or tunnel that is created in the substratum (Atkinson and Taylor, 1988; Bellwood, 2002;

Garstang, 1897; Warner, 1977). Both burying and burrowing promote the potential for high drag forces between the body and the surrounding particles due to frictional interactions between grains in dry substrates or viscous interactions in saturated environments (i.e. skin drag). Hence, these behaviours have been considered more energetically expensive than other forms of locomotion (Dorgan et al., 2011; Hunter and

Elder, 1989; Trevor, 1978), although burrowing through materials similar to elastic-muds is not a substantial component of the total metabolic energy budget of clam worm

(Nereis virens) (Dorgan et al., 2011), and a wide range of terrestrial and aquatic animals bury and burrow (e.g., Arnold, 1995; Brown et al., 1972; Corn et al., 2018; Couffer and

Benseman, 2015; Deacon, 2006; Dorgan et al., 2007; Gidmark et al., 2011; Herrel et al.,

2011a; Jung et al., 2011; Ken’Ichi Kanazawa, 1992; Norris and Kavanau, 1966; Nye,

1974; Rodrigues et al., 2010; Sandbak and Murison, 1996; Schepper et al., 2007;

Winter et al., 2012). Considering that animals have been proposed to bury and burrow to potentially hide from predators, hide from prey, store food, engage in subsurface migration, station hold, avoid unfavourable conditions, and reproduce (Daniels, 1989;

Deacon, 2006; Hosoi and Goldman, 2015; Kuhlmann, 1992; Maladen et al., 2011;

McGaw, 2005; Rebach, 1974; Simon, 1991; Stein and Magnuson, 1976; Tatom-

Naecker and Westneat, 2018), exploring burying and burrowing behaviours are expected to provide valuable insight into the relationship between shape, movement, sediment dynamics, and ecology.

There appear to be at least three broad approaches used by animals to penetrate and move into sediment, and animals might employ a combination of these tactics. Crack propagation refers to the use of a wedge-like feature that focuses stress at the tip of a crack in the substrate, causing a fracture to propagate through the sediment, enabling the animal to penetrate through the sediment (Dorgan et al., 2005;

Dorgan et al., 2008; Dorgan et al., 2011; Jung et al., 2011; Winter et al., 2012). Crack propagation has been described as a mechanically efficient mechanism for invertebrate worms (e.g. clam worms) to burrow in fine grained sediments with cohesive and elastic properties, such as mud (Dorgan et al., 2005; Dorgan et al., 2011). In contrast, excavation (i.e. digging) refers to an animal using their bodies (e.g. undulations and oscillations) and appendages (e.g. fins and limbs) to push and pull on sediment generating action-reaction forces to break apart and displace sediment. Excavation is a common approach to burying and burrowing in a range of animals as such as mammals, amphibians, squamates, fish and invertebrates, and in a range of sediment types (e.g. mud, sand, gravel, wood chips) in both aquatic and terrestrial environments

(Arnold, 1995; Brown et al., 1972; Daniels, 1989; Deacon, 2006; Gidmark et al., 2011;

Herrel et al., 2011b; Herrel et al., 2011a; Ken’Ichi Kanazawa, 1992; Maladen et al.,

2011; McGaw, 2005; Mosauer, 1932; Nye, 1974; Rebach, 1974; Sandbak and Murison,

1996; Schepper et al., 2007; Sharpe et al., 2015a; Tatom-Naecker and Westneat,

2018). Finally, fluidization is a process by which energy from a flowing fluid is imparted to a bed of particles, thereby converting sediment from a solid-like state to a fluid-like state (Hosoi and Goldman, 2015). Of note, it takes more force to penetrate a wet particulate medium than a dry particulate medium of the same grain size (Sharpe et al.,

2015b). Fluidizing substrate decreases its viscosity and yield stress (Hosoi and

Goldman, 2015), which correspondingly decreases skin drag and the resistance experienced by the animal moving through the surrounding particles (Jung et al., 2011;

Winter et al., 2012). Furthermore, larger particles require higher water velocities for fluidization. Thus, fluidizing sediment to bury and burrow appears to be most common in fine grained sands, is employed by a range of aquatic animals such as fish, squid, and razor clams (Corn et al., 2018; Jung et al., 2011; McKee et al., 2016; Rodrigues et al., 2010; Winter et al., 2012), and may be particularly well suited for animals with a high surface area to avoid the drag associated with moving into the substrate.

Given that resistance to burrowing and burying is dominated by skin drag, which scales proportionally to the surface area of a body (Webb, 1974), body shape is anticipated to have an impact on burying and burrowing behaviours. For example, an elongated and cylindrical shape, common to many invertebrates, fish, lizards and snakes, decreases the surface area to volume ratio, and consequently, decreases skin drag; accordingly, these animals commonly migrate relatively high distances in sediment (Arnold, 1995; Gidmark et al., 2011; Herrel et al., 2011a; Herrel et al., 2011c;

Maladen et al., 2011; Schepper et al., 2007; Sharpe et al., 2015a; Tatom-Naecker and

Westneat, 2018; Wiens and Slingluff, 2001; Wiens et al., 2006). In contrast, a flattened

80 shape would seem to limit these behaviours due to a high surface area to volume ratio, and hence, high skin drag. Nevertheless, many animals that are flattened in the same plane as the substrate have been described to bury into sands, or mixed mud and sand, including species of lizards, crabs, flattened urchins, flatfish (Arnold, 1995; Bellwood,

2002; Corn et al., 2018; McKee et al., 2016). Rather than penetrating deep into sediment, flattened animals appear to exercise mechanisms for coverage of the body involving limited body displacement into the substrate, which tends to involve lifting sediment up and over onto the exposed surface. Lizards, crabs and urchins may excavate and wedge themselves just beneath the surface of the sediment at a slight angle to the substrate plane, and they may transport sediment onto the exposed surface using their extremities (Arnold, 1995; Bellwood, 2002; Ken’Ichi Kanazawa, 1992).

Flatfish (i.e. pleuronectiformes) are fishes that are laterally flattened and flipped on their side as they rest on the substrate and swim through the water column, and these fishes swim via undulations of the body and caudal fin that are oriented vertically to the substrate (Webb, 2002). To bury, flatfish exercise similar vertical waves of bending along the length of the body, which fluidizes the sediment underneath the animal and moves it into suspension and up and over onto the exposed surface of the fish (Corn et al., 2018; McKee et al., 2016). Our objective was to explore the mechanisms of burying in animals with flattened shapes that use a very different mechanism of body movement for locomotion than flatfish and other flattened animals, the stingray.

Stingrays (Suborder Myliobatoidei) are a group of fishes that are flattened in the same plane as the substrate, are common to benthic habitats with fine grained

81 sediments such as sand in both freshwater and marine environments (Allen and

Pondella, 2006; Aschliman et al., 2012; Goes de Araújo et al., 2004; Vaudo and Lowe,

2006), and are known to bury into the substrate. These fishes are dorsoventrally flattened with a stiffened body axis and enlarged and rounded pectoral fins that wrap around the head and body forming the pectoral disc (Blevins and Lauder, 2012;

Fontanella et al., 2013; Franklin et al., 2014; Parson et al., 2011; Schaefer and

Summers, 2005). Rather than axial bending as in flatfish, stingrays employ flexible movements of the pectoral fins to power routine and fast-start swimming (Blevins and

Lauder, 2012; Rosenberger, 2001; Rosenberger and Westneat, 1999; Seamone et al.,

2019). The aim of this study was to describe the mechanism and effectiveness of the ocellate river stingray (Potamotrygon motoro) to bury in sand. We hypothesized that to bury, rather than crack propagation or excavation, rapid movement of the body and fins generate flow beneath the pectoral disc to fluidize and suspend sediment onto the dorsal and exposed surface. Furthermore, we predicted that an increase in aspects of the kinematics of the body and fins would increase the speed of the sediment flows, move more sediment and ultimately increase the extent of burying. Videography was used to explore the kinematics and evaluate the effectiveness by which P. motoro buries, and particle image velocimetry was used to describe the dynamics of the fluidized sediment, and accordingly, reveal the mechanism of burying. This study advances our understanding of a flattened disc shape in relation to life in a benthic environment, describing a novel approach to burying in a flattened fish with flexible pectoral fins and a rigid body axis, limitations and benefits, and it is relevant to

82 underwater robotic designs for crypsis and station holding in a benthic environment, inspired by biomimicry (Koller-Hodac et al., 2010).

Materials and methods

Animals and housing

All procedures were approved by the University of Calgary animal care committee, following Canadian Council on Animal Care guidelines. Four ocellate river stingrays (Potamotrygon motoro) (pectoral disc width 14.9 ± 0.410 cm S.E.M.; mass

155 ± 10.2 g S.E.M.) were purchased from a licenced supplier and transported to facilities at the University of Calgary. P. motoro were housed in a cylindrical holding tank (180 cm diameter x 70 cm height, approximately 1400 L) with flow-through freshwater at 27° C, pH of 6.5, bubbled with air. Room lighting was provided with a

12:12h on:off cycle. The stingrays were provided 5 cm3 of frozen bloodworms per ray once a day, and were monitored to ensure feeding. The animals were acclimated for two months prior to experiments, water chemistry was measured daily, and the fish had regular veterinary oversight.

Data collection

A separate rectangular tank with dimensions 75 cm length x 32 cm width x 46 cm height was used for experimentation; water chemistry and temperature were the same as the holding tank. Aquarium substrate, with a maximum grain diameter of 1 mm

(Crystal River Sediment, CaribSea, Fort Pierce, Florida, USA), was evenly spread across the bottom of the tank, at a depth of 7 cm. An individual ray was transferred to the experimental tank via a rubber mesh net. Burying behaviours were captured using three cameras filming at 120 fps at 720p (Hero3 Black, GoPro Inc, San Mateo,

California, USA), one filming from the dorsal view and two filming across the substrate from each end of the tank. A measuring stick was placed on the bottom of the tank to calibrate the camera from the dorsal view, whereas measuring sticks were placed vertically at four different locations along the length of the tank and a linear regression relating distance from the camera to the filming location was used to calibrate cameras filming along the substrate of the tank. A pair of light emitting diodes attached to a plastic pipe submerged in the tank was used to synchronize the cameras immediately after the burying event was performed. A total of 20 responses were collected; each fish performed 5 burying events. The four stingrays were rotated into the experimental tank throughout the study, such that only a single burying event was recorded from an animal before it was placed back into the holding tank and another animal moved into the experimental tank.

Data analysis

Sediment flow patterns and speed

Particle image velocimetry via MATLAB PIVlab 2.01 (Garcia, 2011; Thielicke,

2014; Thielicke and Stamhuis, 2014) was used to describe the dynamics of sediment movement during burying. From the anterior view, 10 burying events, where the head of the ray was oriented toward the camera, were analysed to explore the pattern of the

84 displacement of the sediment granules from underneath the disc onto the dorsal surface. As the sediment moved across the dorsal surface, maximum speed and direction of the sediment flow was determined frame-by-frame in two-dimensions from the dorsal view for all 20 burying events, starting when the sediment moved from the pectoral fins towards the midline, and ending when the sediment reached the midline or when the ray came to a rest (i.e. if the sediment did not reach the midline). The 10 most rapid measures of sediment speed for each burying event were averaged and this average was used as sediment speed for statistical analysis. For both dorsal and anterior view analysis, the exposed body of the ray was masked for each frame such that PIV was measuring only sediment flow and not movement of the body (Fig. 1 & 2).

Furthermore, any air bubbles present in the video frames were masked so they did not interfere with PIV. Three passes were conducted on the videos: one with 64x64 pixel windows, one with 32x32, and one with 16x16. Increasing the interrogation area with a fourth pass of 128x128 pixel windows did not yield different results in 5 videos analysed, and decreasing the interrogation area with a fourth pass of 8x8 pixel windows was too small relative to the size and displacement of the sediment particle, yielding substantial noise in the data.

Kinematics of the body and fins

Digital video images were analysed using ImageJ. Using images from the camera filming an anterior view of the ray along the substrate, the location of an eye was tracked for successive frames to investigate the kinematics of the down-up pumping motions (i.e. oscillations) exerted by the body along the vertical axis relative to

85 the substrate. The number of body pumps was defined as the number of down-up oscillations during the burying event. The oscillations of the body were not symmetrical, and therefore, the downward motions were defined as a push (i.e. toward the substrate), whereas the upward motions were defined as a pull (i.e. away from the substrate), and displacement and speed for pushing and pulling motions were calculated separately. Displacement was defined as the cumulative change in the vertical position of the eye for a given push or pull, and the maximum and average displacement of the body pumping motions throughout burying were measured. Speed was calculated using the two-dimensional change in the vertical position of the eye over time between successive frames, and the average speed of the body pumping motions were measured in addition to the maximum speed. The onset of the body pumping motions (always a push) was defined as the point at which push speed was 10% of the maximum of the initial push speed. Average body-pump frequency was defined as the total number of oscillations divided by the duration (in seconds) of these oscillations.

Body pumping often began before motion of the pectoral fins, and therefore, the body pump oscillations that were performed during the finbeats were selected to evaluate the relationship between body pump frequency and finbeat frequency, defined below.

We anticipated that the lateral motion of the fins towards the midline would most impact the two-dimensional speed of sediment as it moved across the dorsal surface.

Hence, analysis of fin motions was simplified to only two dimensional movements as measured from the dorsal view, whereby the length of a line transecting the centre of the disc from the lateral edge of the dextral fin to the lateral edge of the sinistral fin was

86 measured for successive frames and used as a proxy for the displacement and speed of fin movement. Finbeats were assessed until the fins either came to rest or the lateral edges of the fins were covered by substrate. The number of finbeats was defined as the number of oscillations of the fins folding over towards the sagittal midline, and then recovering to the original position, during a burying event. The average finbeat frequency was defined as the number of oscillations divided by the duration of these oscillations. Finbeat displacement was defined as the cumulative two-dimensional change in length of the line from fin tip-to-tip, and the maximum and average displacements of the finbeats throughout the burying event were measured. Finbeat speed was calculated using the two-dimensional change in length of the line from tip-to- tip between successive frames, and the maximum and average speed of the finbeat motions towards the midline were measured. Displacement and speed values were divided by two, to provide values that were associated with one fin, and the onset of finbeat motions was defined as the point where finbeat speed was 10% of the maximum for the initial finbeat speed.

Burying performance

Burying duration was defined as the onset of the body pumping motions to when the body pumping came to rest or the eyes were covered by substrate. Burying depth was measured as the displacement of the eye from its resting position immediately before burying commenced, to its position after the burying event occurred or when the eye was covered by substrate. Sediment coverage of the ray, from the dorsal view, was measured to compare the extent of burying of the head, body, fins and tail, and the

87 effects of body kinematics on extent of coverage. From the dorsal view, a grid was created using ImageJ over the animal at rest before burying motions began, to define the surface area covered by sediment of different locations on the disc (Fig. 3). The outline of the ray was traced, and then points were marked around the lateral edge, separated by 30 degrees relative to the centre of the disc, starting from the rostrum, which was defined as 0 degrees. Then, anteroposterior and lateral transects were created to connect the points directly across from one another, creating a grid over the dorsal surface. Furthermore, the tail was divided into proximal (i.e. tail base proximal to the stinger) and distal (i.e. tail tip from the stinger outward) locations. This grid defined

18 different locations on the dorsal surface of the pectoral disc and tail (Fig. 3). The grid was then placed over the image of the animal once the burying event was completed, using the location of the eyes and the tail as a reference, which remained recognizable even when the rest of the ray was fully buried. For each location in the grid, the surface area covered by sediment was measured after burying. The percent of sediment coverage for each location was defined as the surface area of the location covered divided by the total surface area of that location, multiplied by 100. Additionally, the total sediment coverage of the ray (i.e. pectoral disc and tail) was defined as the sum of the amount of surface area covered for each location divided by the total surface area of the pectoral disc, multiplied by 100.

Statistical analysis

All statistical analyses were performed using R software and accounted for repeated measures (R Development Core Team, 2008). Data of sediment coverage

88 was transformed using the arcsine of the square root of the percentage prior to statistical analysis. Linear regressions were used to test the relationship between the frequency of the body pump and finbeat motions, and the impact of the following measures on sediment speed: number of finbeats, number of body pumps, average finbeat frequency, average body-pump frequency, maximum and average displacement and speed of the finbeats and the push and pull of the body pumping. Linear regressions were used to test the relationship of burying depth, duration, and sediment speed on sediment coverage. A one-way ANOVA and Tukey HSD Post-Hoc Test was used to test for differences in sediment coverage of the different individuals, in addition to the different locations of the pectoral disc. A mean value of sediment coverage for each location was calculated from the 5 different trials for each individual, such that the n for this statistical test was 4.

Results

Sediment flow patterns

To bury, P. motoro repeatedly pumped the body up and down as the pectoral fins folded up and over (Fig. 1). The body pumping appeared to draw water beneath the rostrum and into the space between the body and substrate as the anterior portion of the pectoral fins pressed into the substrate, indicated by movement of sediment being sucked towards the rostrum (Fig. 2). The flow and pumping motions suspended sediment beneath the pectoral disc, and the movement of the fins generated and directed vortices of suspended sediment along the ventral surface of the fins as they

Fig. 1. Particle image velocimetry analysis of a burying event exercised by the ocellate river stingray (Potamotrygon motoro), viewed from the anterior, showing the direction and relative speed of sediment granules. Granules are lifted as vortices, move along the ventral surface of the fins and are then shed towards the dorsal midline, where they collide and annihilate as indicated by the upwards jets. (A-B) From a resting position, the fish pressed its body downwards. (C-D) The rostrum lifted upwards, followed by the body

90 pulling upwards, as the fins folded up and over. (E) The rostrum subsequently pressed onto the substrate as the body pushed downwards and the fins recovered towards the original position by rolling outwards. (F-G) Again, the rostrum and then the body pulled upwards, as the fins folded up and over suspending and directing vortices of fluidized sediment along the ventral surface of the fins towards the dorsal midline. (H-I) The vortices of sediment moved towards the midline, as the body presses downwards. (K-M) The vortices collided along the midline, as another set of vortices were formed by subsequent finbeat and directed towards the midline. (N) The second set of vortices collided over the midline. (O) The sediment dissipated and settled onto the fish. In all images, every second vector is shown for clarity. Orange vectors indicate data interpolated from neighboring vectors.

Fig. 2. Particle image velocimetry analysis of a burying event exercised by the ocellate river stingray (Potamotrygon motoro), viewed from the dorsal surface, showing the measured speed and direction of sediment moving across the dorsal surface of the fish

92 towards the sagittal midline, where the sediment from each fin collides and generates anteroposterior jets of sediment flow that reorient the sediment towards the head and tail. (A-C) The fins folded up and over, suspending and directing sediment towards the midline. (D-F) The sediment moved towards the midline and collided, changing direction towards the head and tail base, as another surge of sediment was suspended and directed towards the midline for another collision by a subsequent finbeat. (G-H) The sediment settled over the dorsal surface of the fish. In all images, every second vector is shown for clarity. Orange vectors indicate data interpolated from neighboring vectors. Color bar scale at the bottom of the image indicates particle speed.

93 were lifted. The vortices moved toward the fin tips, and then over and onto the dorsal surface of the body as the fins were depressed again, shedding the vortices of sediment over the dorsal surface of the body toward the midline where the sediment then settled.

These patterns of sediment flow were resolved using video images and particle image velocimetry analysis from both anterior and dorsal views.

From the anterior view, P. motoro lifted the sediment as vortices along the ventral surface of the pectoral fins, such that a circular, rotating mass of sediment particles translated in the same direction as the movement of fins, while the particles rotated in a direction counter to the dorso-medial movement of the tips of the fins (Fig. 1). The vortices then rolled over the tips of the fins as the fins were depressed, and were shed across the dorsal surface where they dissipated. The fins subsequently recovered into the original position by rolling the tips away from the body midline, laterally and downwards along the dorsal surface of the fish, so that the fin tips were maintained in close proximity to the dorsal surface and thus moved underneath the suspended sediment during the recovery stroke. In 5 of the trials, from the anterior view, the sediment vortices that were shed from the two fins collided over the dorsal midline of the fish, where they annihilated and upward vertical jets of sediment were produced as a result, before the sediment settled onto the dorsal surface (Fig. 1). From the dorsal view, the sediment was shed primarily along the region of the pectoral fins approximately posterior to the eye and anterior to the tail, in a lateral direction towards the longitudinal midline (Fig. 2). The velocity profiles suggest the sediment settled first onto the fins, and then onto the body (Fig. 3). In the 5 trials where the sediment shed

94 from each fin collided along the midline, immediately following collision jets of high sediment velocity were created in the anteroposterior direction, resulting in sediment moving toward the head and the tail base, well anterior and posterior to the initial vortices of sediment that were shed over the dorsal surface by the fins (Fig. 2).

Kinematics, sediment speed and burying performance

The mean sediment coverage of all burying events was 82.5% ± 3.0 S.E.M, and ranged from 60.4 to 98.1 % (Fig. 4). The pectoral fins (location 1, 4, 5, 8, 9, 12, 13 and

16) buried significantly more than the head (location 2 and 3) and the tail stinger

(location 18), and the body and the tail base buried significantly more than the tail stinger; otherwise, sediment coverage between the different locations was not significant (Fig. 3). Furthermore, sediment coverage was significantly greater in ray 1 and 2 when compared to ray 4; otherwise, sediment coverage was not statistically different between individuals (Fig. 4).

Body pumping motions were often initiated before finbeat motions (65% of trials).

As the frequency of the body pumping increased, frequency of the finbeats also increased (Fig. 5), and the slope of the relationship between these parameters was not different than 1, indicating these motions occurred in synchrony and suggesting coordination between sediment fluidization beneath the body and the movement of that sediment by the fins onto the dorsal surface. The body pump and finbeat motions were coordinated such that maximum finbeat speed and maximum body-pull speed were

Fig. 3. Sediment coverage of the different locations on the dorsal surface of the ocellate river stingray (Potamotrygon motoro). Mean + SEM, n = 4 animals, with the 5 observations from each animal averaged to form a single observation at each location for each animal. See table 1 for statistical comparisons.

Table 1. P-value statistics for ANOVA comparisons of sediment coverage of the different locations on the dorsal surface of the ocellate river stingray (Potamotrygon motoro) (see Fig. 3 for locations). Row 2 shows the average sediment coverage for each location, with the standard error of the mean in row 3; the observations from 5 burying events on each animal were averaged at each location, and these averages used as the values for that animal.

Location 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Mean % 85.6 45.4 46.5 88.7 99.1 80.4 84.0 99.3 99.1 84.9 82.7 99.8 99.2 98.3 97.1 99.2 74.5 75.4

± SEM 3.80 8.81 8.41 4.41 0.57 5.54 4.10 0.49 0.55 5.02 5.81 0.23 0.78 1.21 1.46 0.59 6.41 1.52

1 0.24 0.30 1 0.81 1 1 0.79 0.81 1 1 0.64 0.73 0.90 0.95 0.74 1 <0.001

2 1 0.10 <0.001 0.24 0.16 <0.001 <0.001 0.10 0.13 <0.001 0.0011 0.0021 <0.001 <0.001 0.679 0.178

3 0.13 <0.001 0.30 0.20 <0.001 <0.001 0.135 0.17 <0.001 <0.001 0.0016 0.003 <0.001 0.751 0.139

4 0.96 1 1 0.95 0.96 1 1 0.87 0.92 0.99 1 0.93 1 <0.001

5 0.82 0.91 1 1 0.96 1 1 0.87 0.92 0.99 1 1 <0.001

6 1 0.79 0.81 1 1 0.64 0.73 0.90 0.95 0.74 1 <0.001

7 0.89 0.91 1 1 0.78 0.85 0.96 0.99 0.85 1 <0.001

8 1 0.95 0.92 1 1 1 1 1 0.33 <0.001

9 0.96 0.93 1 1 1 1 1 0.36 <0.001

10 1 0.87 0.92 0.98 1.0 0.92 1 <0.001

11 0.82 0.88 0.97 0.99 0.89 1 <0.001

12 1 1 1 1 0.21 <0.001

13 1 1 1 0.28 <0.001

14 1 1 0.47 <0.001

15 1 0.60 <0.001

16 0.28 <0.001

17 <0.001

Fig 4. Average total sediment coverage from 5 burying events of the four individuals of ocellate river stingray (Potamotrygon motoro). Vertical bars indicate the standard error of the mean, and horizontal dashed lines represent the ranges of coverage observed across the 5 burying events for each animal (1 vs 2, p = 0.216; 1 vs 3, p = 0.1; 1 vs 4, p < 0.001; 2 vs 3, p = 0.969; 2 vs 4, p = 0.0284; 3 vs 4, p = 0.0678).

98 closely aligned, as determined from the measured time course of these movements, whereas maximum body-push speed occurred during the recovery stroke of the fins.

It was anticipated that the vigor of the body and fin motions should relate to the extent of sediment movement, and so relationships between displacement, speed and frequency of body and fin movements and sediment movement were explored. The number of body pumps and finbeats, and the average body pump frequency and finbeat frequency, did not significantly impact the speed of the sediment as it moved towards the midline of the fish (Fig. 5). Conversely, as the maximum and average displacement of the body push, body pull, and maximum and average speed of body push and body pull increased, sediment speed increased (Fig. 6). Repeated measures of individuals influenced the positive relationship between the average pull displacement and sediment speed; ray 2 and 3 individually showed a positive relationship between the two parameters, whereas ray 1 and 4 individually showed no significance. As the maximum and average speed of finbeats and finbeat displacement increased, sediment speed increased (Fig. 7). Furthermore, as sediment speed increased, sediment coverage of the dorsal surface increased (Fig. 8). Moreover, burying duration did not have an impact on sediment coverage, and as burying depth increased the extent of sediment coverage increased (Fig. 8).

Fig. 5. A) The relationship between body pump frequency and finbeat frequency during

2 burying in the ocellate river stingray (Potamotrygon motoro) (pslope < 0.001, r = 0.86, y = 1.0 x + 0.26). B) The relationship between the number of finbeats or body pumps and sediment speed during a burying event. The black circles and solid line represent finbeat

100

2 number (pslope = 0.67, r = -0.045, y = - 0.65x + 25) and the white triangles and dashed line

2 represent body pump number (pslope = 0.64, r = -0.043, y = -0.46x + 25). C) The relationship between average finbeat or body pump frequency and sediment speed. The black circles

2 and solid line represent finbeat frequency (pslope , r = 0.12, y = -3.0x + 34) and the white

2 triangles and dashed line represent body pump frequency (pslope = 0.072, r = 0.13, y = 2.4x + 13). Data are results of 5 observations on each of 4 animals for all measures.

101

Fig. 6. A) The relationship between body-push displacement and sediment speed during burying in the ocellate river stingray (Potamotrygon motoro). The black circles and solid

2 line represent maximum push displacement (pslope < 0.001, r = 0.73, y = 10x + 11), and the

2 white triangles and dashed line represent average push displacement (pslope < 0.001, r = 0.77, y = 17x + 9.3). B) The relationship between body-pull displacement and sediment speed. The black circles and solid line represent maximum pull displacement (pslope < 0.001, r2 = 0.68, y = 18x + 12), and the white triangles and dashed line represent average

2 pull displacement (pslope < 0.001, r = 0.61, y = 24x + 12). C) The relationship between body- push speed and sediment speed. The black circles and solid line represent maximum push

2 speed (pslope < 0.001, r = 0.83, y = 0.56x + 13), and the white triangles and dashed line

2 represent average push speed (pslope < 0.001, r = 0.77, y = 2.0x + 11). D) The relationship between body-pull speed and sediment speed. The black circles and solid line represent

2 maximum pull speed (pslope < 0.001, r = 0.62, y = 1.1x + 13), and the white triangles and

2 dashed line represent average push speed (pslope < 0.001, r = 0.77, y = 3.2x + 11). Data are results of 5 observations on each of 4 animals for all measures.

102

Fig. 7. A) The relationship between fin displacement and sediment speed during burying in the ocellate river stingray (Potamotrygon motoro). The black circles and solid line

2 represent maximum fin displacement (pslope < 0.001, r = 0.52, y = 5.2x + 5.8 ) and the white

2 triangles and the dashed lines represent the average fin displacement (pslope <0.001, r = 0.68, y = 10x - 1.3). B) The relationship between fin speed and sediment speed. The black

2 circles and solid line represent maximum fin speed (pslope < 0.001, r = 0.74, y = 0.20x + 12),

2 and the white triangles and dashed lines represent average fin speed (pslope < 0.001, r = 0.64, y = 0.70x + 7.6). Data are results of 5 observations on each of 4 animals for all measures.

103

Fig. 8. A) The relationship between sediment speed and sediment coverage (pslope < 0.001, r2 = 0.83, y = 2.8x + 19). B) The relationship between sediment coverage and duration of

2 burying (pslope = 0.57, r = - 0.037, y = -2.4x + 85). C) The relationship between sediment

2 coverage and burying depth (pslope < 0.001, r = 0.68, y = 16x + 51). Data are results of 5 observations on each of 4 animals for all measures.

104

Discussion

Our study revealed that the mechanism of burying employed by P. motoro utilized effective control of sediment vortices, flows and collisions to direct sediment onto the entire dorsal surface of the animal, with the potential to modulate the extent of burying. The downwards-pushing and upwards-pulling movement of the body, coupled with movement of the pectoral fins and the rostrum that form and release a seal along the substrate, functioned like a piston with valves to manipulate the flow of water and generate changes in pressure that fluidized sand beneath the pectoral disc. As the fins folded up and over, vortices of sand were lifted upwards and along the ventral surface of the fins, transferred across the lateral edge of the fins toward the dorsal surface, and then were shed medially where they dissipated over the dorsal side of the fish, as the fins rolled outwards under the vortices of sediment and recovered to the original position

(Fig. 1). In the most vigorous burying events, the vortices of sediment travelled toward the dorsal midline where they collided and were annihilated, resulting in jets of sediment being directed upwards along the vertical axis, and forwards and backwards along the anteroposterior axis of the animal (Figs. 1 & 2). This led to rapid and nearly complete coverage of the exposed surface of the animal, even when burying in only a few centimeters of sediment, including the head, fins, body and most of the tail except for the stinger, despite using only the posterior two-thirds of the pectoral fins to lift sediment from the substrate. The following is a description of how the body and fin movements appear to generate these patterns of sediment dynamics and impact the extent of burying, why fluidizing sediment is an effective approach to burying in sandy sediments, and relationships of shape and burying behaviours in animals.

105

Fluidization of sediment occurs through the formation of a pressure gradient in a fluid that generates flow to dislodge and support the weight of the sediment particles.

The kinematics of body movements and the resulting sediment dynamics when burying suggest that these stingrays rely on changes in pressure underneath the pectoral disc that generate flows to fluidize sediment beneath the disc and move it along the ventral surface of the fins. Water from the surrounding environment was drawn into the cavity beneath the animal through suction pressure created as the body pulled upwards, flowing in via a tunnel created by raising the rostrum off the substrate while the anterior portion of the pectoral fins maintained contact with the substrate, similar to the mechanisms employed by stingrays to draw suction when feeding (Wilga et al., 2012).

In support of this mechanism, granules of sediment near the rostrum were often observed to be sucked in under the disc as the rostrum and body lifted upwards (Fig. 2).

Then, an increase in pressure beneath the pectoral disc most likely occurred as the body subsequently pushed towards the substrate while the rostrum and pectoral fins were pressed against the substrate, expelling the fluidized sediment toward the lateral edges of the pectoral disc. Furthermore, subsequent motion of the fins, lifting upwards and folding towards the dorsal midline, generated suction along the ventral side of the fins, drawing vortices of the fluidized sediment from underneath the body that were directed up and over onto the dorsal surface. The bending motions of the fin were similar to the bending motions exhibited during escape responses in these fish

(Seamone et al., 2019), and they transferred the sediment along the ventral and presumably low pressure surface of the fins as vortices, similar to the appearance of hydrodynamic vortices formed and shed along the trailing surface of the caudal fin

106 during stage 1 of the C-start in fishes (Borazjani et al., 2012; Tytell and Lauder, 2008)

(Fig. 1). While the formation of vortices is well established as being important in the generation of thrust for swimming fish (Borazjani and Sotiropoulos, 2009; Borazjani et al., 2012; Drucker and Lauder, 1999; Drucker and Lauder, 2002; Lauder, 2015; Lauder et al., 2003; Liu et al., 2017; Standen and Lauder, 2007; Tytell, 2004; Tytell, 2006; Tytell and Lauder, 2008; Wilga and Lauder, 2002), this appears to be the first evidence that fish use vortices to displace sediment to bury.

Individuals demonstrated different levels of sediment coverage when compared to one another, whereby one of the rays tended to only bury its pectoral fins leaving most of the body and the head unburied for all burying events, and another ray never buried its head fully (Fig. 4). Hence, this suggests different thresholds or motivations to bury, or perhaps a limitation in physical capability of individuals might exist in stingrays.

While it is not clear what lead to the variation in the extent of sediment coverage between individuals, variation in the overall relationships between several aspects of the kinematics, the speed of sediment movement, and sediment coverage, suggest that the mechanism employed by stingrays to bury provides the potential to modulate the speed of sediment flow and ultimately the extent of burying. When the speed and displacement of body pumping and finbeat motions were increased, this increased the speed of the sediment movement (Fig. 6 & 7) and fluidized more sediment to increase sediment coverage of the dorsal surface (Fig. 8). An increase in speed and displacement as the body pulled upwards away from the substrate most likely generated a greater magnitude of suction beneath the disc, and hence, a greater rate of

107 water flow in through the tunnel beneath the rostrum, similar to what is observed when stingrays increase suction in feeding (Wilga et al., 2012). This would likely fluidize more sediment beneath the body. Furthermore, an increase in the displacement and the speed of the body subsequently pushing downwards would likely increase the pressure beneath the disc, to loosen and expel more of the fluidized sediment toward the periphery. The capacity to expel more sediment would also allow the fish to bury deeper in the substrate (Fig. 8). In contrast, the frequency and number of body and fin movements was not related to burying (Fig. 5), nor was the duration of the burying event (Fig. 8), and the average frequency of the body pumping and finbeat motions remained within a narrow range in all individuals as they buried. This is opposed to observations in burying flatfish, that increase sediment coverage by increasing the frequency of body undulations (Corn et al., 2018; McKee et al., 2016). In P. motoro, increasing the number of body and finbeat movements would not necessarily increase the speed and thus distance that the sediment travels, and therefore, would have little impact on extent of burying. Furthermore, changes in the frequency of movements was more a reflection of changes in the duration of pauses between movements rather than the rate of movement itself, which again would have relatively little impact on the speed and distance that sediment would move. Therefore, it appears burying is not necessarily related to how long the ray partakes in the behaviour (duration of burying even, number of finbeats), rather it is dependent on the vigour of the burying event.

There was also evidence that a mechanism of burying via a flattened disc that utilizes sediment fluidization and translation of sediment vortices can selectively control

108 what specific regions of the dorsal surface are covered. The rays did not always cover the entire dorsal surface with sediment (Fig. 3), and the pattern of how the sediment moved up and over onto the dorsal side of the pectoral disc impacted the sediment coverage of different locations on the fish. Sediment initially moved onto the pectoral disc mostly along the portion of the fins that were involved in finbeats, which was predominantly the fins posterior to the eye (Fig. 2). The fins anterior to the eye tended to remain close to the substrate, contributing to forming the tunnel through which water was drawn under the body as the rostrum lifted upwards, but not contributing greatly to the finbeat motions that directed sediment onto the dorsal surface (Fig. 1 & 2). The rostrum pressed down onto the substrate, but it did not tend to flick sediment onto the surface as it lifted away from the substrate. As such, little substrate moved onto the pectoral disc across the rostrum and anterior portion of the pectoral fins. Likewise, the tail region was not involved in moving sediment onto the dorsal surface of the disc.

When the speed and displacement of the body and fin kinematics were relatively low, the vortices of sediment did not reach the midline before they dissipated, and consequently the sediment settled mostly on the pectoral fins, and only the region of the fins involved in lifting the sediment vortices over the fish (i.e. the anteroposterior middle portion). As the speed and displacement of the body pumping and finbeats became faster, sediment moved across the posterior fins and towards the sagittal midline, and settled on the fins and much of the body, but again primarily the posterior portion.

When the speed and displacement of the finbeats were greatest, the vortices of sediment from opposite sides of the body collided above the body midline, and the pathway of the sediment flow was redirected from a lateral orientation to an

109 anteroposterior and vertical orientation, producing jets of sediment towards the head and the tail, as well as upwards (Fig. 1 & 2). This accomplished covering regions of the body adjacent to the region of the fins involved in moving sediment. The anteroposterior and upwards jets of sediment suggest that stingrays might be exploiting the collision and annihilation of vortices to overcome the constraints of covering the head and most of the tail, where energy in the colliding and collapsing vortices is redirected radially and axially (Kudela and Kosior, 2014; Lim and Nickels, 1992).

Fluidizing sediment appears to be an effective approach for burrowing and burying in a range of aquatic animals, particularly in fine grained sediments that are loosely packed, such as sand (Corn et al., 2018; Jung et al., 2011; McKee et al., 2016;

Rodrigues et al., 2010; Winter et al., 2012). Studies of penetration and drag forces on smooth cylinders reveal that wet sand is more resistive than dry sand (Sharpe et al.,

2015b), and so fluidizing sediment can decrease the energy required to penetrate into saturated sediments (Winter et al., 2012). Furthermore, water is denser and more viscous than air, making it a more effective medium for fluidizing sediment and supporting the particles in suspension at lower flow velocities (Hosoi and Goldman,

2015). Whether terrestrial animals can generate air flow speeds that are great enough to fluidize sediment such as sand to bury is not clear, but we are not aware of any terrestrial animals that employ this mechanism. Moreover, when the weight of the sediment particle is considered, such as the diameter of the particle size, a greater rate of flow is required to generate a force sufficient to dislodge and suspend larger particles.

Flatfish, which also exercise motions of the body and fins to fluidize sediment from

110 beneath the fish to bury, showed decreased sediment coverage of the exposed surface as particle size of the sediment increased (Corn et al., 2018). Yet sediments with grain sizes smaller than sand may generate large cohesive forces, which can consequently lead to more force required to dislodge particles from the substrate than loose sands

(Hosoi and Goldman, 2015). Hence, the mechanism employed by stingrays to bury is likely to be most effective in loose and fine-grained sediments, such as sand, and these tend to be the type of benthic environments that stingrays inhabit (Corcoran et al., 2013;

Harada and Tamaki, 2003; Semeniuk and Dill, 2005; Stokes and Holland, 1992; Vaudo and Lowe, 2006).

There is mounting evidence that shape has a strong impact on burying behaviours in animals. Animals that bury or burrow into the substrate headfirst often have relatively reduced eyes, conical skull shapes and strengthened cranial bones, whereas in animals that burrow using the tail the post-cranial skeleton is highly reduced, fortified and pointed, such as in garden eels (e.g. Heterocongrinae); these modifications have been proposed to be advantageous in piercing through the sediment (Arnold,

1995; Herrel et al., 2011a; Sharpe et al., 2015a; Tatom-Naecker and Westneat, 2018).

Moreover, animals with high surface-area to volume ratios, including flattened urchins, flatfish, and stingrays, appear to reduce the potential constraints of high skin drag due to viscous and frictional forces between the body and the sediment by transporting sediment up and over onto the exposed surface, and burying or burrowing only a small distance into the substrate (Fig. 8) (Arnold, 1995; Bellwood, 2002; Corn et al., 2018;

Ken’Ichi Kanazawa, 1992; McKee et al., 2016). In contrast, animals with lower surface-

111 area to volume ratios, including worms, globular urchins, snakes, and lizards, accrue lower magnitudes of skin drag, and these animals tend to move much deeper and further through the substrate, which appears to be supplemented by reduced or absent extremities (Dorgan et al., 2005; Dorgan et al., 2011; Jung et al., 2011; Ken’Ichi

Kanazawa, 1992; Maladen et al., 2011; Sharpe et al., 2015a; Winter et al., 2012).

Furthermore, the mechanisms that animals exercise to displace sediment appear to be closely tied to the mechanisms used for other locomotor behaviours. Mice, frogs, crabs and lizards may use their limbs to move across the substrate surface, and to excavate sediment (Bellwood, 2002; Brown et al., 1972; Deacon, 2006; Kuhlmann, 1992; McGaw,

2005; Sandbak and Murison, 1996). Fish, lizards and snakes may use lateral undulations to swim through water or to move on land, and to fluidize or excavate sediment (Corn et al., 2018; Gidmark et al., 2011; Sharpe et al., 2015b; Sharpe et al.,

2015a). Squid generate jets of water via the mantle funnel to propel themselves through water and also to fluidize sand for burying (Rodrigues et al., 2010). Stingrays rely on the flexibility of the pectoral fins for a range of behaviours, such as routine swimming (Blevins and Lauder, 2012; Breder, 1926; Macesic et al., 2013; Rosenberger,

2001; Rosenberger and Westneat, 1999), escape responses (Seamone et al., 2019), feeding (Wilga et al., 2012), and as demonstrated in this present study, burying behaviours. The selective pressures of burying likely also impact shape in a manner that promotes or constrains performance and energetics of other locomotor behaviours.

For example, the low aspect ratio (span2/area) of the pectoral fins in stingrays promotes drag-based rajiform swimming(Blevins and Lauder, 2012; Rosenberger, 2001;

Rosenberger and Westneat, 1999), and this also makes them well-suited to create

112 suction and fluidize and displace vortices of sediment when the animal is stationary on the benthos. Conversely, pelagic stingrays tend to have triangular shaped fins with higher aspect ratios that generate greater magnitudes of lift relative to drag and promote mobuliform swimming, and thus are not suited for sediment displacement and burying.

Further exploration of these relationships may provide some insight into the clear transition from high to low aspect ratio fins in pelagic and benthic stingrays, respectively.

113

Chapter 5

Sediment selection of the ocellate river stingray (Potamotrygon motoro)

is associated with constraints in burying

Abstract

This study addresses how sediment size impacts burying performance of the ocellate river stingray (Potamotrygon motoro), and whether these fish demonstrate preference for sediment size in an environment where they are periodically threatened by a predator model. Time-lapse photography and high-speed video of burying events were used to measure preference for inhabiting and burying in sediment, as well as parameters related to burying performance, across four grain sizes of aragonite substrate (0.25-1.00 mm, 1.0-2.0 mm, 2.5-5.5 mm, 7.0-25.0 mm). Most of the stingrays spent more time inhabiting and buried more often in the finest grained sediment when provided the choice of all four sediments, whereas time spent and burying count in the three coarser sediments were not different from one another. To bury, P. motoro uses its pectoral fins to fluidize and suspend sediment from beneath the pectoral disc onto the exposed dorsal surface. As sediment size increased, sediment coverage of the exposed body surface decreased, time required to bury increased, rate of sediment coverage decreased, and burying effectiveness (i.e. rate of coverage divided by fin speed while burying) decreased. Thus, for a given finbeat speed, P. motoro could fully cover themselves faster with finer grained sediments, and these fish became restricted in the ability to bury, regardless of effort or time spent burying, as grain size increased. Therefore, we suggest that constraints imposed on fluidizing larger grain sizes are associated with the preference for very fine-grained substrates in benthic stingrays.

Keywords: sediment selection, behaviour, burying, performance, stingrays, benthic

114

Introduction

An extensive range of abiotic and biotic factors impact the selection of habitat in animals, such as temperature, salinity and humidity, ontogeny and developmental plasticity, intraspecific and interspecific competition, life history and reproduction, topographic characteristics such as substrate type, foraging, predation risk and locomotor performance (Aubret and Shine, 2008; Bartolino et al., 2011; Brown et al.,

1972; Calsbeek and Irschick, 2007; Cardona, 2007; Casterlin and Reynolds, 1977;

Freitas et al., 2016; Gilliam et al., 2016; Lecchini et al., 2005; Levin et al., 1997;

Lindberg et al., 2016; McConnaughey and Smith, 2000; McIvor and Odum, 1988;

Rodway and Regehr, 1999; Rosenfeld et al., 2005; Stokes and Holland, 1992). In particular, exploring the relationship between habitat selection and the impacts of habitat on the performance of locomotor behaviours such as burying, might provide insight into the relationship between shape, behaviour and ecology (Calsbeek and

Irschick, 2007), and may provide inspiration to the design of robotics for specific environments, via biomimicry. Burying is a behaviour whereby an animal displaces substrate and covers themselves with sediment, possibly to separate themselves from conditions such as high current, to enhance prey capture, and to hide from predators

(Arnold and Weihs, 1978; Bellwood, 2002; Daniels, 1989; Deacon, 2006; Gerstner and

Webb, 1998; McGaw, 2005; Simon, 1991). Many aquatic animals, including fish, squid, and razor clams fluidize sediment to bury. Fluidizing sediment entails creating a flowing fluid that imparts energy to a bed of particles, thereby converting sediment from a solid- like state to a fluid-like state (Hosoi and Goldman, 2015). Of note, larger particles require higher water velocities for fluidization, and evidence suggests that fluidizing

115 sediment may be more effective in sandy substrates compared to more coarse-grained sediments (Corn et al., 2018). Further, whether animals that fluidize sediment demonstrate preference for finer grained substrates in the presence of a threat is not understood.

Stingrays (suborder Myliobatoidei) and flatfish (order Pleuronectiformes) have convergently evolved a shape that is flattened like a foil in the same plane as the substrate (i.e. foil-shaped fishes), and tend to inhabit benthic environments composed of fine-grained sediments such as sand or mixed sand and mud (McConnaughey and

Smith, 2000; Vaudo and Lowe, 2006). These fishes bury by exercising motions of the body and fins to fluidize and suspend sediment underneath the fish, and move it up and over onto the exposed surface, and both fishes increase aspects of their burying kinematics to increase the speed of the flow, to fluidize more sediment and increase the extent of burying (see Chapter 4; Corn et al., 2018; McKee et al., 2016). Because the size of sediment particles that water can carry is limited by the speed of flow, larger particles would presumably require greater kinematics of the body and fins to be fluidized, suggesting that burying performance may become constrained by grain size in fishes that fluidize sediment to bury (Corn et al., 2018). Accordingly, sediment coverage increased in burying flatfish as particle size of the sediment decreased (Corn et al.,

2018). Furthermore, flatfish preferred finer grained substrates when provided the choice to inhabit sediments that differed in grain sizes (Gibson and Robb, 2000; Moles and Norcross, 1995). However, sediment selection and the constraints of sediment size on burying have not been measured in stingrays, which employed oscillations of their

116 pectoral fins to bury rather than lateral undulations of the body as in flatfish, providing the opportunity to explore how the specific mechanism of burying employed by stingrays might impact this relationship, and whether enhanced burying performance and preference for inhabiting and burying in finer grained substrates is a common theme to foil-shaped fishes that fluidize sediment to bury in a benthic environment.

Stingrays are dorsoventrally flattened with enlarged pectoral fins that wrap around the head and body, which collectively are referred to as the pectoral disc. To bury, stingrays fold their fins up and over to direct fluidized sediment along the ventral surface of the fins onto the dorsal surface (see Chapter 4). These fishes are believed to bury to camouflage and hide from predators, such as sharks and marine mammals

(Pádraig, 2000; Strong et al., 1990). This study aims to understand whether there is a preference for substrate size, and if there is an association between effectiveness of burying and preference for sediment size in the ocellate river stingray (Potamotrygon motoro), specifically when the animals is threatened, and hence, when increasing the extent of sediment coverage in a shorter period of time might be of great consequence.

We hypothesized that because P. motoro is dependent on fluidizing sediment when burying, and because larger particle sizes require greater rates of flow to be fluidized, these fish will become be constrained in burying performance as sediments become more coarse, such that even if stingrays can exercise motions of the fins that are great enough to generate sufficient flow that can fluidize coarser sediment, for a given finbeat speed, in finer sediments P. motoro can displace more sediment onto the exposed surface in a shorter time. Accordingly, we anticipate that P. motoro will demonstrate

117 preference for finer grained substrates when in an environment where they are periodically threatened, similar to what might be expected in the wild. Time-lapse photography was used to track and measure the time spent by P. motoro on each of four aragonite substrates that differed in grain size when exposed to all four substrates simultaneously. A model predator was thrust towards the fish at regular intervals throughout the trials to evoke burying behaviours, and to maintain a perception that there is potential danger in the environment. Subsequently, P. motoro was videoed burying in each sediment type from the dorsal view, and aspects of burying performance were measured, including sediment coverage of the body, burying duration, coverage rate, fin speed while burying and burying effectiveness.

Methods

Animal husbandry

All procedures were approved by the University of Calgary animal care committee, following Canadian Council on Animal Care guidelines. A total of 11 individuals of P. motoro were purchased from a licenced supplier and used in experiments (pectoral disc width 16.054 cm ± 0.464 S.E.M.; mass 198 g ± 20.4 S.E.M.).

P. motoro were housed at the University of Calgary in a cylindrical tank (180 cm diameter x 70 cm height, approximately 1400 L) with no sediment along the bottom, flow-through freshwater at 27° C, pH of 6.5, bubbled with air, and a 12:12 hour photoperiod. The fish were provided 2.5 cm3 of frozen bloodworms per stingray once a day, and were monitored to ensure feeding and health. Water chemistry was measured

118 daily, and the fish had regular veterinary oversight. Individuals were acclimated for six months prior to experiments.

Data collection

A rectangular Plexiglass tank with dimensions 213 cm length x 91 cm width x 91 cm height was used for experiments. To explore whether P. motoro demonstrated preference for occupying sediment of a particular grain size, the bottom of the tank was divided into quadrants, each with the same dimensions of 53 cm length x 23 cm width x

16 cm height. Each quadrant was filled with a different grain size of aragonite substrate

(Sediment 1, bahamas oolite, 0.25-1.00 mm grain size, 1530 kg/m3 average density;

Sediment 2, special grade reef 1.0-2.0 mm, 1360 kg/m3; Sediment 3, Florida crushed coral 2.5 – 5.5 mm, 1150 kg/m3; Sediment 4, Aruba puka shell 7.0-25.0 mm, 1120 kg/m3; CaribSea, Fort Pierce, Florida, USA). Sediment was spread evenly across the quadrant using a plastic rake. Water chemistry and temperature were the same as the holding tank, and water was filled to the surface of the tank and refreshed daily.

Temperature was maintained using seven 200 W heaters placed immediatley below the water surface, well above the substrate, and placed symmetrically around the tank: one above the junction of all four quadrants at the centre of the tank, one at each opposite end of the long axis of the tank attached to the tank wall, and one above each quadrant attached to the tank wall. Furthermore, three air stones were provided, placed just beneath the water surface so they would not interefere with the stingrays on the substrate and evenly spaced across the middle of the tank. A 12:12 hour photoperiod

119 was maintained, and no additional lighting beyond ambient room lights was used during this experiment.

An individual ray was transferred to the experimental tank via a rubber mesh net, and was provided an acclimation period of 36 hours (12 hours darkness, 12 hours light,

12 hours darkness). The fish was fed 2.5 cm3 of frozen bloodworms at the beginning of the ensuing 12-hour light period, and one hour after feeding experimental trials were initiated. The location of the ray was tracked for 20 hours over the span of two days

(i.e. 10 hours each day), during light periods, using time-lapse photography set to 5 second intervals with four GoPro cameras (GoPro Inc, San Mateo, California, USA) oriented towards the dorsal view over the center of each quadrant at a height of 91 cm from the surface of the sediment. Every 2 hours from the onset of these recording periods, a black plastic cylinder (18 cm diameter) mounted to a plastic pipe (length 100 cm x 2 cm diameter), intended to replicate a small shark predator, was thrust downwards into the sediment about 2 disc widths away from the ray, at a location towards the corner of the quadrant that the stingray was located in at the time, to induce a threat that might evoke burying behaviours. This location provided the ray with the potential to access each quadrant without interference of the predator model, including the quadrant it was currently in. The number of threats evoked for each fish was 12 (i.e.

6 each day).

120

A separate set of experiments was subsequently employed to explore burying performance across the four different substrates. Four separate containers in the shape of conical frustums (height 32 cm, upper radius 52 cm, and lower radius 45 cm) were placed on the bottom of the experimental tank in a zigzag pattern, such that each container was greater than 20 cm from the wall of the experimental tank. Each container was filled to the surface of the container with one of the four sediment types.

Plastic mesh with a 1x1 cm grid was wrapped around the perimeter of each container to a height 91 cm, to prevent P. motoro from swimming off the sediment and the container.

A GoPro camera filming a dorsal view at 240 fps and 1080 p was mounted directly above the sediment at a height of 91 cm from the surface of the sediment. Room lighting was turned off, and three LED lights (12 watts, Philips Electronics Ltd.

Markham, Ontario, Canada) were used to illuminate the sediment and filming area. A checkerboard was placed in the filming areas and was used to calibrate the camera.

Stingrays were transferred from the holding tank to the experimental tank using a rubber-mesh net and placed into one of the containers of sediment. Stingrays were provided a maximum of 30 minutes to bury, and if a burying event was not completed within 30 minutes the stingray was removed from the container and placed back in the holding tank. Each of the 11 rays were subjected to two trials for each sediment type, and hence, 22 trials for each sediment type were collected. Order of burying trials were in sequence from the smallest to the largest sediment, and then from the largest to the smallest sediment sizes. The number of burying events collected for each sediment were 22 for sediment 1, 22 for sediment 2, 21 for sediment 3, and 18 for sediment 4.

121

Data analysis

Time-lapse images were analyzed using Adobe Premier Pro (Adobe Systems,

Mountain View, California, USA, 2003). The percentage of time P. motoro spent occupying a given sediment was determined as the sum of the number of images that the ray was in a given quadrant, divided by the total number of images collected for each trial. The percentage of threats P. motoro experienced in a given sediment was determined as the sum of the threatening events that occurred in a given quadrant (i.e. sediment), divided by the total number of threats for each trial. The burying percentage that P. motoro excercised in a given sediment was determined as the sum of the number of burying events in that sediment type plus the number of times the stingray remained motionless and did not flee from the threat when the fish was already buried and covered by that sediment type, divided by the total number of threats for each trial.

For analysis of burying performance in each sediment type, digital videos were imported into Adobe Premier Pro and converted from HEVC to AVI format, and then analysed using ImageJ (Wayne Rasband Developers 1997). Sediment coverage was defined as the surface area of the dorsal surface covered by sediment divided by the total surface area of the dorsal surface before burying commenced, as viewed from above, multiplied by 100. Duration of the burying event was defined as the time from the initiation to the termination of the fin motions associated with burying, where initiation and termination were defined as a speed of fin movement that was at least

10% of the maximum fin speed measured during the burying event; alternatively, when the fins were not visible toward the latter stages of some burying events due to

122 suspended sediment, termination was defined as the point when sediment coverage of the pectoral disc was within 10% of the maximum sediment coverage of that burying event. The averate rate of coverage [%/s] was measured as the percent sediment coverage divided by the duration of the burying event. Finbeat speed while burying was measured using the two-dimensional change in length of the line from tip-to-tip of the left and right pectoral fins between successive frames, providing an estimate of the rate at which the fin tips moved toward the anteroposterior midline. The average speed of these finbeat motions was then normalized by the disc width of the fish [%/s]. Finally, effectivness of burying was defined as the unitless ratio of the average rate of sediment coverage divided by the average fin speed

Statistical analysis

The percentage of time spent in a given sediment, threats in a given sediment, and burying in a given sediment was transformed using arcsine of the square root prior to statistical analysis. A one-way ANOVA and Tukey HSD Post-Hoc Test was used to test for differences in the time spent, the threat count, the bury count, and escape count between the sediment types. All burying performance data was transformed by log10 for statistical testing (Corn et al., 2018). A one-way ANOVA accounting for repeated measures and Tukey HSD Post-Hoc Test accounting for repeated measures were used to test for differences in sediment coverage, duration, coverage rate, average fin speed, and effectivness of burying in each sediment type. Average values of the two trials were measured for each afromentioned parameter for each ray, such that an n of 11 was used for the statistical tests. A p-value of 0.05 was used to test for significance.

123

Results

Sediment selection

P. motoro spent a greater amount of time occupying the finest grained sediment

1 when compared to sediment 4 (Fig. 1). Otherwise, there were no differernces in the time spent in the different sediment sizes. Of note, differences between sediment 1 when compared to sediment 2 (p = 0.0594) and sediment 3 (p = 0.103) were marginally not significant. The total number of times P. motoro was threatened by the model in each sediment was 64, 31, 28 and 9, for sediment 1, sediment 2, sediment 3, and sediment 4, respectively, and the percentage of threats was greater in sediment 1 when compared to sediment 4, but otherwise there were no differences between the different sediment sizes (Fig. 2). The total burying count was 67, 26, 22 and 3, for sediment 1, sediment 2, sediment 3 and sediment 4, respectively, and the percentage of burying events was greater in sediment 1 when compared to sediment 4. Otherwise burying percentage was not different between the different sediments; (Fig.3). Of note, differences between sediment 1 when compared to sediment 2 (p = 0.0541) and sediment 3 (p = 0.0590) were marginally not significant.

124

Fig. 1. The percentage of time, over a 20 hour period, that 11 individuals of the ocellate river stingray (Potamotrygon motoro) spent in four different sediments differing in grain size. Vertical bars indicate the standard error of the mean, and horizontal dashed lines represent the maximum and minimum values observed (sediment 1 versus 2, p = 0.0594; sediment 1 versus 3, p = 0.103; sediment 1 versus 4, p = 0.00469; sediment 2 versus 3, p = 0.994; sediment 2 versus 4, p = 0.756; sediment 3 versus 4, p = 0.605).

125

Fig. 2. The percentage of threats that occurred in each of the four different sediments differing in grain size. 11 individuals of the ocellate river stingray (Potamotrygon motoro) each experienced 12 threats, for a total of 132 threats. Vertical bars indicate the standard error of the mean, and horizontal dashed lines represent the maximum and minimum values observed (sediment 1 versus 2, p = 0.311; sediment 1 versus 3, p = 0.238; sediment 1 versus 4, p = 0.015; sediment 2 versus 3, p = 0.998; sediment 2 versus 4, p = 0.504; sediment 3 versus 4, p = 0.603) .

126

Fig. 3. The percentage of burying events that occurred in four different sediments differing in grain size for 11 individuals of the ocellate river stingray (Potamotrygon motoro). Vertical bars indicate the standard error of the mean, and horizontal dashed lines represent the maximum and minimum values observed (sediment 1 versus 2, p = 0.0541; sediment 1 versus 3, p = 0.0590; sediment 1 versus 4, p < 0.001; sediment 2 versus 3, p = 1.00; sediment 2 versus 4, p = 0.422; sediment 3 versus 4, p = 0.400) .

127

Burying performance

During burying, video observation revealed that P. motoro fold their fins up and over, and suspend sediment along the ventral surface of the fins and then flick the sediment towards the midline onto the dorsal and exposed surface of the stingray. As sediment grain-size decreased, all measures of burying performance tended to be enhanced. Sediment coverage was greater in sediment 1 versus 4, sediment 2 versus

4, and sediment 3 versus 4, but otherwise sediment coverage was not different between sediments (Fig. 4). Burying duration was less in sediment 1 versus 2, sediment 1 versus 3, sediment 1 versus 4, and sediment 2 versus 4, but not different in sediment 2 versus 3, and sediment 3 versus 4 (Fig. 5). Coverage rate was greater in sediment 1 versus 3, sediment 1 versus 4, sediment 2 versus 4, and sediment 3 versus 4, but not different in sediment 1 versus 2 and sediment 2 versus 3 (Fig. 6). The average speed of the fins while burying was less in sediment 1 versus 3, but not greater in sediment 1 versus sediment 2, sediment 1 versus 4, sediment 2 versus 3, sediment 2 versus 4, and sediment 3 versus 4 (Fig. 7). Burying effectivness, defined as the coverage rate normalized by fin speed, was greater in sediment 1 versus 2, sediment 1 versus 3, sediment 1 versus 4, sediment 2 versus 3, sediment 2 versus 4,and sediment 3 versus

4 (Fig. 8). There was significant variation between individuals for all performance measurments (sediment coverage, p < 0.001; duration, p < 0.001; coverage rate, p <

0.001; fin speed, p < 0.001; effectivness, p < 0.001), whereby some individuals appeared to be better burying performers than other individuals, and this was consistent across sediment sizes.

128

Fig. 4. Sediment coverage of the dorsal surface in four different sediments differing in grain size for 11 individuals of the ocellate river stingray (Potamotrygon motoro). Vertical bars indicate the standard error of the mean, and horizontal dashed lines represent the maximum and minimum values observed (sediment 1 versus 2, p = 0.980; sediment 1 versus 3, p = 0.179; sediment 1 versus 4, p < 0.001; sediment 2 versus 3, p = 0.343; sediment 2 versus 4, p < 0.001; sediment 3 versus 4, p < 0.001).

129

Fig. 5. The duration of burying events in four different sediments differing in grain size for 11 individuals of the ocellate river stingray (Potamotrygon motoro). Vertical bars indicate the standard error of the mean, and horizontal dashed lines represent the maximum and minimum values observed (sediment 1 versus 2, p = 0.00165; sediment 1 versus 3, p < 0.001; sediment 1 versus 4, p < 0.001; sediment 2 versus 3, p = 0.330; sediment 2 versus 4, p = 0.0016; sediment 3 versus 4, p = 0.107).

130

Fig. 6. The rate of sediment covering the dorsal surface in four different sediments differing in grain size, for 11 individuals of the ocellate river stingray (Potamotrygon motoro). Vertical bars indicate the standard error of the mean, and horizontal dashed lines represent the maximum and minimum values observed; maximum value for sediment 1 was 220 %/s (sediment 1 versus 2, p = 0.316; sediment 1 versus 3, p = 0.00189; sediment 1 versus 4, p < 0.001; sediment 2 versus 3, p = 0.145; sediment 2 versus 4, p < 0.001; sediment 3 versus 4, p < 0.001).

131

Fig. 7. The average speed of the fins (percent disc width per second) while burying in four different sediments differing in grain size, for 11 individuals of the ocellate river stingray (Potamotrygon motoro). Vertical bars indicate the standard error of the mean, and horizontal dashed lines represent the maximum and minimum values observed (sediment 1 versus 2, p = 0.125; sediment 1 versus 3, p = 0.025; sediment 1 versus 4, p = 0.690; sediment 2 versus 3, p = 0.889; sediment 2 versus 4, p = 0.724; sediment 3 versus 4, p = 0.325).

132

Fig. 8. The effectiveness of burying in four different sediments differing in grain size, for 11 individuals of the ocellate river stingray (Potamotrygon motoro). Effectiveness was measured as the ratio of coverage rate to fin speed (see Methods for details). Vertical bars indicate the standard error of the mean, and horizontal dashed lines represent the maximum and minimum values observed (sediment 1 versus 2, p = 0.0490; sediment 1 versus 3, p < 0.001; sediment 1 versus 4, p < 0.001; sediment 2 versus 3, p = 0.0477; sediment 2 versus 4, p < 0.001; sediment 3 versus 4, p < 0.001).

133

Discussion

Burying is a cryptic behaviour that stingrays use to hide from predators, such as sharks and marine mammals (Pádraig, 2000; Strong et al., 1990). Stingrays commonly inhabit benthic environments with fine-grained sediments such as sand while feeding

(Corcoran et al., 2013; Harada and Tamaki, 2003; Semeniuk and Dill, 2005; Stokes and

Holland, 1992; Vaudo and Lowe, 2006), but whether an appearance to favor this sediment might be associated with their burying habits is not known. This study demonstrates that there is a preference to inhabit and bury in finer sediment compared to coarser sediment when Potamotrygon motoro are being threatened (Fig. 1 & 3), and a relationship between finer sediment grain size and enhanced burying performance

(Fig. 4-8). To bury, P. motoro exercise motion of the body and fins to suspend and fluidize sediment from beneath the pectoral disc, moving it up and over onto the exposed dorsal surface, and this mechanism became significantly restricted in its ability to effect burying as grain size of the sediment increased. Accordingly, P. motoro was more effective at burying in smaller grain sizes (Fig. 8), including covering on average over 90% of the body versus less than 30% (Fig. 4), requiring only half the time to bury

(Fig. 5), and rates of coverage about 10-fold faster (Fig. 6), in the finest grain sediment compared to performance in more coarse sediment, despite similar rates of finbeat movement (Fig. 7). This improved burying performance is related to the rate that the animals can move sediment off of the substrate, where for a given finbeat speed, the rate of coverage of the exposed surface is greater in finer sediment, and hence, the fish can suspend and transport larger quantities of sediment from beneath the disc onto the dorsal surface faster (Fig. 8). Thus, P. motoro has the capacity to achieve greater

134 coverage of the exposed surface in a shorter amount of time as sediment grain-size decreases, both of which are assumed to be pivotal assets when exercising burying as a defence mechanism against predators.

When threatened by the predatory model, P. motoro had the tendency to first exercise a disc-start escape to gain separation away from the model (Seamone et al.,

2019), and then buried into the surrounding sediment. Furthermore, when covered by sediment prior to being threatened by the model, P. motoro tended to remain buried and motionless, and did not flee from the threat. Hence, stingrays use burying behaviours as a mechanism of defence in response to a threat, and burying effectiveness appeared to have an impact on sediment selection, in the presence of a threat. While differences between sediment 1 versus 2 and 3 were marginally insignificant, P. motoro was threatened and buried more often in the finest sediment compared with the coarsest sediments (Fig. 2 & 3). Moreover, the stingrays spent over 50% of their time occupying the finest sediment, and only 20% or less on each of the coarser sediment types (Fig.

1). This suggests that stingrays might select sediment based on their ability to bury, likely in anticipation of responding to threats.

Despite the stingrays preferring the finest sediment on average, and having the best burying performances in the finest sediment on average, there was substantial variability in preference and burying performance across the specimens and sediment types (see maximum and minimum ranges in Figures). Surprisingly, there were two

135 individuals, one of which preferred to remain buried in sediment 2 and one in sediment

3, throughout nearly the entire duration of the experimental trials; the basis for the preference of these individuals for burying and resting in sediment 2 and 3 as opposed to sediment 1 is not clear, particularly given the clear pattern of improved burying capacity in finer sediments. Of note, burying performance of these individuals was above average for sediment 2 and 3, but not as great as their performance in sediment

1. Furthermore, P. motoro did not spend notably more time on average in sediment 2 when compared to sediment 3 and 4 (Fig. 1), despite having substantially better burying performance in sediment 2 compared to 3 and 4 in most aspects that were measured

(Fig. 4, 5, 8). This is likely due to the preference for sediment 1 rather than lack of preference for sediment 2 over 3 and 4, where most animals simply spent the more of their time in sediment 1 and relatively less in any of the larger grain sizes. We anticipate that there would be a gradient in preference that follows grain size and burying performance, where the preference for sediment would shift from sediment 1 to 2 if sediment 1 was not available, but this is not known.

Given that P. motoro exhibited limited success in burying in the coarsest sediment (Fig. 4, 5, 8), the observation that they rarely buried in this sediment when reacting to a threat (Fig. 3) suggests a strong preference for grain sizes that promote adequate burying performance as a defense mechanism. Of note, the limited effectiveness for burying in the coarser-grained sediments appears to be a physical limitation. It would be expected that the rate of movement of the fins would be a critical element in effectively lifting sediment off the substrate and moving it over the animals,

136 and that increased rates of movement might thus be employed by animals in coarser sediment to achieve adequate burial. However, while there were statistically greater fin speeds employed while burying in Sediment 2 and 3 relative to 1, the differences were quite small in magnitude, and the stingrays appeared to reduce their effort while burying in the coarsest Sediment 4 (Fig. 7). These results would suggest that the burial mechanics used by these stingrays are relatively inflexible, are already employed to near maximal even when burying in the finest sediment, and that they are thus physically incapable of increasing the effort to bury in coarser sediment. This would again result in a strong preference for finer sediments by stingrays if burial is an important defense mechanism.

Provided that burying performance is also constrained in flatfish as sediment grain-size increases (Corn et al., 2018), we suggest that burying mechanisms that suspend sediment via movements of a foil-shaped body are limited to being more effective at burying in finer grained substrates (Fig. 9). Larger particles require faster flows to be suspended and transported, and they are faster to fall out of suspension as flow speed slows, and these animals do not appear capable of achieving adequate flows to effect burying in larger particles of sediment. Furthermore, both stingrays and flatfish have the tendency to prefer sediments of finer substrates in behavioural experiments (Gibson and Robb, 2000; Moles and Norcross, 1995), and given the relationship between burying performance and sediment selection revealed in this present study, we suggest that the constraints imposed on burying in sediments with larger grain sizes probably makes these fishes more vulnerable to predation in these

137 types of benthic environments. Hence, better burying in finer grained substrates most likely contributes to the common observation of these fishes in sandy benthic environments (Corcoran et al., 2013; Harada and Tamaki, 2003; Semeniuk and Dill,

2005; Stokes and Holland, 1992; Vaudo and Lowe, 2006). Whether stingrays would more commonly exploit other benthic environments when relieved of predator stress, such as due to the reduction of shark populations due to human impacts (Luiz and

Edwards, 2011; Robbins et al., 2006), is unclear. Other factors, such as feeding and reproduction, likely also impact sediment selection.

138

Chapter 6

A rounded disc-shaped foil can be an effective shape for life in a benthic

environment

Summary

Fish show tremendous diversity in shape, swimming behaviour and habitat, which provides substantial opportunity to advance our understanding of locomotion in water, in a manner that can be relevant to biologists and engineers designing underwater robotics. Our understanding of how shape impacts swimming is dominated by studies of fishes that swim via axial bending along the length of the body and caudal fin (Blake, 2004; Domenici and Blake, 1997; Domenici and Hale, 2019; Sfakiotakis et al., 1999). In contrast, stingrays are dorsoventrally flattened fishes, shaped like a foil, with enlarged pectoral fins that wrap around the head and body, creating a broad surface referred to as the pectoral disc (Fontanella et al., 2013; Franklin et al., 2014).

The pectoral fins are used to power swimming, whereas the tail is long and thin and appears poorly suited for swimming due to a low surface area, and the body axis is relatively rigid anterior to the pelvic girdle (Parson et al., 2011) which prevents lateral undulation of the body and BCF propulsion. While pelagic stingrays tend to have wing shaped fins with high aspect ratios (span2/area) that power lift-based mobuliform propulsion, benthic stingrays tend to have more rounded pectoral fins with low aspect ratios that generate drag-based rajiform propulsion. Our knowledge of how both pelagic and benthic stingrays swim is predominately restricted to routine swimming (Blevins and

Lauder, 2012; Macesic and Kajiura, 2010; Macesic et al., 2013; Rosenberger, 2001;

139

Rosenberger and Westneat, 1999). This dissertation aimed to describe additional mechanisms of locomotor behaviours that are anticipated to employ substantially different mechanisms than those of BCF swimmers, including fast-start escape, station holding, and burying, using the ocellate river stingray, Potamotrygon motoro, as a model benthic stingray.

Disc-start escape

Fishes commonly use fast start swimming during predator-prey interactions, and these behaviours can provide valuable insight into the limitations of extreme swimming performance imposed by shape (Domenici and Blake, 1997; Domenici and Hale, 2019;

Wakeling, 2001). Whereas fast starts have been extensively studied in BCF swimming fishes for close to 50 years (Webb, 1976)(Webb, 1977), this study is the first to explore escape behaviour in stingrays, a fish where the rigid body axis and the low surface area of the caudal region might be anticipated to markedly impact the approaches used to power fast starts relative to BCF swimming fishes. During fast-start escape in P. motoro, the enlarged pectoral fins that are placed peripherally around the body functioned independently from one another to generate thrust through waves of bending that included anterior, posterior and lateral components, resulting in an exceptional flexibility that promoted the rapid generation of thrust in any direction around the center of mass. This allowed these fishes to fast start in all the freedoms of motion along the substrate (i.e. yaw, forward-backward, side-side), and accordingly to not only rotate away from the threat, but to immediately retract in any direction away from the threat along the benthic plane, regardless of their initial orientation.

140

In BCF swimming fishes, asymmetry in shape along the anteroposterior axis, in addition to propulsive waves of bending via axial bending that generates thrust posterior to the center of mass along the caudal fin, restricts movement to the forwards and yaw freedoms of motion when escaping along the substrate, and furthermore, BCF swimming fishes cannot turn without forward thrust (Webb, 1984a). In contrast, while P. motoro might not match the maximum forward acceleration and speed or the yawing rates of some BCF swimming fishes (Domenici and Blake, 1997), the body is not restricted to following the orientation of the head and these fishes can therefore more readily utilize more freedoms of motion than BCF swimming fishes. Not only can P. motoro achieve a full range of maneuverability along the substrate, the symmetrical disc-shape of P. motoro promoted similar translational accelerations and speeds, regardless of the direction of escape. Furthermore, they can achieve fast-start rotation around the vertical axis (i.e. yaw), and rotation can occur independently from translation, such that while escaping away from the threat they can rapidly reorient themselves into a forward swimming orientation with minimal turning radius. In result, despite the limited longitudinal flexibility of the body and a reduced caudal fin, the rounded pectoral disc and drag-based propulsion promoted a distinctive approach to escape along the substrate, which I have termed disc starts, that results in unusually unrestricted abilities to rapidly maneuver across the benthic environment, and that challenges the concept of maneuverability typically used for fishes.

141

Suction and station holding

Fishes employ station holding behaviours to resist disturbing forces such as current or possibly predation (Arnold and Weihs, 1978; Carlson and Lauder, 2010;

Webb et al., 1996). It has been proposed by prior studies that the ventral surface of fishes that are flattened in the same plane as the substrate might be able to generate suction to adhere to the substrate in station holding (Arnold and Weihs, 1978;

Fontanella et al., 2013). By measuring pressures underneath the pectoral disc of P. motoro in response to an upward lifting force applied to the tail, and video and dye to describe the motions of the body and pectoral fins, and the movement of water around the pectoral disc, I provide evidence that P. motoro can create a seal against the benthos that forms suction when lifting forces are applied to the fish, and provide insight into to the mechanism that generate suction. As a lifting force was applied to the tail, P. motoro pressed its body and fins into the substrate and also initiated motion of the body and fins, whereby the body thrust downwards in synchrony with waves of bending that were passed along the posterior portion of the pectoral fins, similar to rajiform undulation (Blevins and Lauder, 2012; Rosenberger, 2001; Rosenberger and Westneat,

1999). As the magnitude of the lifting force increased, suction pressure increased as the fish held position along the substrate. The relationship between the lifting force and suction pressure was highly correlated, suggesting that stingrays relied on mostly a passive mechanism of Stefan adhesion to generate suction. If stingrays relied on an active mechanism to generate suction, a weaker relationship between lifting force and suction pressure would be expected, because the fish would not be dependent on the lifting force to achieve suction. A passive mechanism was further supported by the

142 observation that a fish was able to resist about 8 N of lifting force and generate a suction pressure of about 7 cmH20 without exercising any finbeat or body thrust motions. That said, for Stefan adhesion to be effective, P. motoro must achieve direct contact between the ventral surface of the pectoral disc and the substrate. Video and dye observations demonstrated that stingrays can expel water from beneath the pectoral disc via bending of the pectoral fins and branchial pumping. They tended to execute these motions as the lifting force was applied and as the fish began to lose grip and slip from the substrate as it was pulled upwards (i.e. as the amount of surface area contact between the pectoral disc and the substrate appeared to decrease). Thus, this study concluded that the rounded disc common to benthic stingrays permits mostly a passive mechanism to achieve suction pressure to adhere to the benthos, but the flexible pectoral fins can actively create and maintain a seal along the substrate, probably to resist greater lifting forces for longer durations.

Burying mechanics

Stingrays are commonly observed to bury into the substrate and cover themselves with sediment, in aquariums and in their natural environment. Many BCF swimming fishes tend to employ lateral undulations that appear to push sediment aside and essentially dig into the substrate to bury (Gidmark et al., 2011; Tatom-Naecker and

Westneat, 2018), however, stingrays cannot bend laterally as they possess a rigid body axis, but rather use flexible pectoral fins in a mechanism that has not yet been described. Using video to analyze the kinematics of the body and fins, and particle

143 image velocimetry to describe the flows of sediment around the fish, this is the first study to describe the mechanism of burying in stingrays.

The body and fins exercised synchronous motions to generate water currents that fluidized sediment beneath the pectoral disc, creating vortices of suspended sediment that flowed along the ventral surface of the fins, then moved up and over onto the dorsal fin surface, and were subsequently shed onto exposed dorsal body surface where the sediment settled. To generate the flow patterns that fluidized sediment beneath the pectoral disc, the fins were repeatedly pumped up and down, folding up and over toward the ventral midline, drawing water in beneath the pectoral disc through a tent-like shape formed as the rostrum was lifted off the substrate while the anterior portion of the pectoral fins remained pressed onto the substrate. The fish then pressed the body downwards as the posterior portion of the pectoral fins folded up and over, generating vortices of suspended sediment that were forced from beneath the body and transported along the ventral surface of the pectoral fins as the fins lifted upwards. The vortices were then directed toward the dorsal surface of the fish as the fins folded over toward the midline, and shed and dissipated so that the sediment settled over the exposed surface. Increased speed and acceleration of the body pumping and finbeats lead to an increase in the speed of the sediment as it was transported across the dorsal surface, which lead to an increase in sediment coverage. Furthermore, this mechanism resulted in differences in the extent of burying of different locations of the dorsal surface, such that the fins buried more than the body, head and tail, and the body buried more than the head and tail. In the most vigorous burying events, the vortices of

144 sediment approaching from opposites sides collided above the midline of the fish, where they annihilated, redirecting sediment upwards and towards the head and the tail, increasing the extent of burying. Thus, P. motoro can achieve full coverage of the dorsal surface despite using only the posterior portion of the fins to transport sediment, and exploit vortices of sediment for effective control of burying performance.

Sediment preference and performance in burying

Stingrays commonly inhabit benthic environments with sandy sediments that are fine grained, which has been associated with their foraging strategies (Corcoran et al.,

2013; Harada and Tamaki, 2003; Stokes and Holland, 1992), but whether these specific environments are associated with enhanced locomotor performance in behaviours such as burying has not been explored. Furthermore, burying is believed to be a defence mechanism used by stingrays to hide from predators, but there is no direct evidence to support this, or how it may impact sediment selection. To better understand these behaviours, the sediment preference of P. motoro, when exposed to sediment of four different grain sizes and threatened by a predator model, and the limitations to burying performance in these sediments, was studied using video and time-lapse photography.

As P. motoro was threatened with the model, the fish tended to rapidly bury into the surrounding sediment that it was situated on at the time of being threatened. Hence, we might predict that under the circumstances of anticipating a threat, P. motoro would preferentially inhabit sediments that maximize burying performance. P. motoro spent substantially more time and buried more often in the finest grain size versus the coarser sediment sizes, despite this resulting in them being threatened more often in this

145 sediment. Accordingly, burying performance was greatest in the finest grain sediment, and decreased as sediment grains became coarser. As described above, to bury P. motoro folded the fins up and over towards the dorsal midline, drawing sediment from beneath the pectoral disc up and over onto the exposed surface. The rate that P. motoro displaced sediment from beneath the pectoral disc onto the exposed dorsal surface was significantly greater in finer sediments, enabling these fish to consistently achieve greater than 90% sediment coverage of the exposed surface in just under a second, whereas sediment coverage decreased, the rate of coverage decreased, and duration of the burying event increased as sediments became coarser. Furthermore, the speed of the fins required to bury was significantly less in the finest grain sediment compared to coarser sediments, and P. motoro was more effective at burying (rate of coverage relative to fin movement) as sediment size decreased. Hence, P. motoro actively chose to inhabit fine grain sediments that was associated with enhanced performance in burying behaviours.

Conclusion and future directions

While most species of fish rely on a range of surfaces to interact with water and to move, the studies in this dissertation demonstrate that P. motoro can accomplish a diverse range of locomotor behaviours using primarily a single broad surface of the pectoral disc. The pectoral fins of stingrays, which wrap around the head and body forming almost a 360 degree propulsive surface, are extremely flexible in both anatomy and in their apparent ability to generate thrust in any direction, promoting a range of behaviours, including the ability to use rajiform undulations of the pectoral fins to swim

146 very close to the substrate and perhaps benefit from ground effect, exercise slow precision maneuvering when feeding, glide through the water column, execute fast- start escape in all directions across the substrate from rest, achieve suction pressures along the substrate for station holding and feeding, and exploit vortices of sediment to achieve almost complete burial of the body in less than a second. Hence, rather than being constrained in swimming due to a rigid body axis and low surface area of the caudal region, the shape of P. motoro permits alternative mechanism to accomplish a diversity of behaviours that in some respects exceeds those described in BCF swimming fishes. These abilities lead me to conclude that the rounded disc-shaped foil of many benthic stingrays is highly effective for living in a benthic environment, and should not be considered a liability but rather an asset in substrates with fine grained sediments.

While the studies described in this dissertation have informed several aspects of our understanding of the implications of a rounded disc-shape of stingrays for locomotion in water, that understanding is still far from complete. Field studies might provide insight into whether the unusual performance abilities of benthic stingrays in terms of disc-start escape might enable them to exploit the comparatively large turning radii and low turning rates of their larger BCF swimming predators and thus defeat the attack. Furthermore, although P. motoro can achieve similar escape performance in all directions along the benthic plane, I did not test for maximum performance. Thus, modelling the amount of thrust required to achieve a given magnitude of acceleration and speed in forwards, backwards and sideways escape, might provide insight as to

147 whether backwards or sideways escape might actually be less effective as opposed to forward escape, and hence, maximum performance might be limited in these directions as opposed to escaping forwards. Moreover, particle image velocimetry of disc start behaviours would be informative regarding how the pectoral fins interact with the water to promote drag-based escape. Furthermore, a possible role of the tail in locomotion should be explored, and whether it can generate enough drag to contribute to swimming, or function as rudder, or as an inertial mass to promote rapid turning. In addition, how the transition from the rounded, low aspect-ratio fins in benthic stingrays to the high aspect-ratio fins in pelagic stingrays impacts fast-start performance, and whether the shape of pelagic rays permits different fast-start maneuverability than BCF swimming fishes or the disc starts of benthic stingrays, will advance our understanding of the relationship between shape, performance and environment in stingrays.

Furthermore, whether pectoral fins with high aspect ratios are penalized in burying and suction performance might provide insight into the limitations of these fins in a benthic environment. Moreover, modelling the force and circulation imparted on water by motion of the pectoral fins during burying behaviours will provide insight into the limitations sediment grain size imposes on the ability to fluidize and suspended sediment in vortices. Also, whether stingrays achieve enhanced energetics and performance from benthic locomotion due to ground effect, such as when these fish glide near the substrate in punting and walking behaviours, is not fully understood.

Hence, there is still much to learn, and broadening these studies to other shapes and species of foil fishes, and other forms of locomotion that they employ, holds great

148 promise to expand our knowledge of the dynamics and effectiveness of swimming in these flattened and negatively buoyant fishes.

149

References

Allen, L. G. and Pondella, D. J. (2006). Surf Zone, Coastal Pelagic Zone, and Harbors. In The Ecology of Marine Fishes California and Adjacent Waters (ed. Allen, L. G.), Pondella, D. J.), and Horn, M. H.), pp. 149–166. Berkeley: University of California Press.

Altringham, J. and Ellerby, D. (1999). Fish swimming: patterns in muscle function. J. Exp. Biol. 202, 3397–3403.

Archer, S. D. and Johnston, I. A. (1989). Kinematics of labriform and subcarangiform swimming in the Antarctic fish. J. Exp. Biol. 210, 195–210.

Arnold, E. N. (1995). Identifying the effects of history on adaptation: origins of different sand‐diving techniques in lizards. J. Zool. 235, 351–388.

Arnold, G. P. and Weihs, D. (1978). The hydrodynamics of rheotaxis in the plaice (Pleuronectes platessa L.). J. Exp. Biol. 75, 147–169.

Arnold, G. P., Webb, P. W. and Holford, B. H. (1991). The role of the pectoral fins in station-holding of Atlantic salmon parr (Salmo salar L.). J. Exp. Biol. 156, 625–629.

Aschliman, N. C., Nishida, M., Miya, M., Inoue, J. G., Rosana, K. M. and Naylor, G. J. P. (2012). Body plan convergence in the evolution of skates and rays (Chondrichthyes: Batoidea). Mol. Phylogenet. Evol. 63, 28–42.

Atkinson, A. R. and Taylor, A. C. (1988). Physiological ecology of burrowing decapods. Zool. Soc. London, Symp. Proc. 59, 201–226.

Aubret, F. and Shine, R. (2008). Early experience influences both habitat choice and locomotor performance in tiger snakes. Am. Nat. 171, 524–531.

Bartolino, V., Ciannelli, L., Bacheler, N. M. and Chan, K. (2011). Ontogenetic and sex-specific differences in density-dependent habitat selection of a marine fish population. Ecol. Soc. Am. 92, 189–200.

Beckert, M., Flammang, B. E. and Nadler, J. H. (2015). Remora fish suction pad attachment is enhanced by spinule friction. J. Exp. Biol. 218, 3551–3558.

150

Bellwood, O. (2002). The occurrence, mechanics and significance of burying behaviour in crabs (Crustacea: Brachyura). J. Nat. Hist. 36, 1223–1238.

Betz, O. and Kölsch, G. (2004). The role of adhesion in prey capture and predator defence in arthropods. Arthropod Struct. Dev. 33, 3–30.

Bierman, H. S., Schriefer, J. E., Zottoli, S. J. and Hale, M. E. (2004). The effects of head and tail stimulation on the withdrawal startle response of the rope fish (Erpetoichthys calabaricus). J. Exp. Biol. 207, 3985–3997.

Bilecenoglu, M. and Ekstrom, L. J. (2013). Pelvic fin walking and punting behaviour of Raja radula Delaroche, 1809 observed in the Sea of Marmara. Mediterr. Mar. Sci. 14, 158–161.

Blake, R. W. (1979). The energetics of hovering in the mandarin fish. J. Exp. Bio. 82, 25–33.

Blake, R. W. (1983). Mechanics of gliding in birds with special reference to the influence of the ground effect. J. Biomech. 16, 649–654.

Blake, R. W. (2004). Fish functional design and swimming performance. J. Fish Biol. 65, 1193–1222.

Blake, R. W. and Chan, K. H. S. (2011). Biomechanics of swimming in the pufferfish Diodon holocanthus: propulsive momentum enhancement is an adaptation for thrust production in an undulatory median and paired-fin swimmer. J. Fish Biol. 79, 1774–1794.

Blake, R. W., Chatters, L. M. and Domenici, P. (1995). Turning radius of yellowfin tuna (Thunnus albacares) in unsteady swimming manoeuvres. J. Fish Biol. 46, 536–538.

Blevins, E. L. and Lauder, G. V. (2012). Rajiform locomotion: three-dimensional kinematics of the pectoral fin surface during swimming in the freshwater stingray Potamotrygon orbignyi. J. Exp. Biol. 215, 3231–3241.

Blevins, E. L. and Lauder, G. V. (2013). Swimming near the substrate: A simple

151

robotic model of stingray locomotion. Bioinspiration and Biomimetics 8, 1–12.

Bone, Q. (1966). On the function of the two types of myotomal muscle fibre in elasmobranch fish. J. Mar. Biol. Assoc. United Kingdom 46, 321–349.

Borazjani, I. and Sotiropoulos, F. (2009). Numerical investigation of the hydrodynamics of anguilliform swimming in the transitional and inertial flow regimes. J. Exp. Biol. 212, 576–592.

Borazjani, I., Sotiropoulos, F., Tytell, E. D. and Lauder, G. V. (2012). Hydrodynamics of the bluegill sunfish C-start escape response: three-dimensional simulations and comparison with experimental data. J. Exp. Biol. 215, 671–684.

Brainerd, E. L., Page, B. N. and Fish, F. E. (1997). Opercular jetting during fast-starts by flatfishes. J. Exp. Biol. 200, 1179–1188.

Braun, C. D., Skomal, G. B., Thorrold, S. R. and Berumen, M. L. (2014). Diving behavior of the reef manta ray links coral reefs with adjacent deep pelagic habitats. PLoS One 9, 1–8.

Breder, C. M. (1926). The locomotion of fishes. Zoologica 4, 159–256.

Brown, L. E., Jackson, H. O. and R, B. J. (1972). Burrowing behavior of the chorus frog, Pseudacris streckeri. Herpetologica 28, 325–328.

Calsbeek, R. and Irschick, D. J. (2007). The quick and the dead: Correlational selection on morphology, performance, and habitat use in island lizards. Evolution. 61, 2493–2503.

Cardona, L. (2007). Habitat selection by grey mullets (Osteichthyes: Mugilidae) in Mediterranean estuaries: the role of salinity. Sci. Mar. 70, 443–455.

Carlson, R. L. and Lauder, G. V. (2010). Living on the bottom: kinematics of benthic station-holding in darter fishes (percidae: Etheostomatinae). J. Morphol. 271, 25– 35.

Carlson, R. L. and Lauder, G. V. (2011). Escaping the flow: boundary layer use by the darter Etheostoma tetrazonum (Percidae) during benthic station holding. J. Exp.

152

Biol. 214, 1181–1193.

Casterlin, M. E. and Reynolds, W. W. (1977). Aspects of habitat selection in the musquitofish Gamusia affinis. Hydrobiologia 55, 125–127.

Cliff, G. (1995). Sharks caught in the protective gill nets off KwaZulu-Natal, South Africa. 8. the great hammerhead shark Sphyrna mokarran (Rüppell). South African J. Mar. Sci. 15, 105–114.

Cliff, G. and Dudley, S. F. (1991). Sharks caught in the protective gill nets off Natal, South Africa. 4. the bull shark Carcharhinus leucas Valenciennes. South African J. Mar. Sci. 10, 253–270.

Corcoran, M. J., Wetherbee, B. M., Shivji, M. S., Potenski, M. D., Chapman, D. D. and Harvey, G. M. (2013). Supplemental feeding for ecotourism reverses diel activity and alters movement patterns and spatial distribution of the southern stingray, Dasyatis americana. PLoS One 8,.

Corn, K. A., Farina, S. C., Summers, A. P. and Gibb, A. C. (2018). Effects of organism and substrate size on burial mechanics of English sole, Parophrys vetulus. J. Exp. Biol. 221, 1–6.

Couffer, M. C. and Benseman, S. A. (2015). A young-of-the-year giant sea bass, Stereolepsis gigas, buries itself in sandy bottom: a possible predator avoidance mechanisms. Bull. South. Calif. Acad. Sci 114, 54–57.

Daniels, R. A. (1989). Sigificance of burying in Ammorcrypta pellucida. Copeia 29–34.

Deacon, R. M. J. (2006). Digging and marble burying in mice: Simple methods for in vivo identification of biological impacts. Nat. Protoc. 1, 122–124.

Ditsche, P., Wainwright, D. K. and Summers, A. P. (2014). Attachment to challenging substrates - fouling, roughness and limits of adhesion in the northern clingfish (Gobiesox maeandricus). J. Exp. Biol. 217, 2548–2554.

Domenici, P. (2001). The scaling of locomotor performance in predator-prey encounters: From fish to killer whales. Comp. Biochem. Physiol. - A Mol. Integr.

153

Physiol. 131, 169–182.

Domenici, P. (2010a). Context-dependent variability in the components of fish escape response: Integrating locomotor performance and behavior. J. Exp. Zool. Part A Ecol. Genet. Physiol. 313 A, 59–79.

Domenici, P. (2010b). Escape responses in fish: kinematics, performance and behavior. In Fish Locomotion: An Eco-Ethological Perspective (ed. Domenici, P.) and Kapoor, B. G.), pp. 123–170.

Domenici, P. and Blake, R. W. (1993a). The effect of size on the kinematics and performance of angelfish (Pterophyllum eimekei) escape responses. Can. J. Zool. 71, 2319–2326.

Domenici, P. and Blake, R. W. (1993b). Escape trajectories in angelfish (Pterophyllum Eimekei). J. Exp. Biol. 177, 253–272.

Domenici, P. and Blake, R. W. (1997). The kinematics and performance of fish fast- start swimming. J. Exp. Biol. 200, 1165–1178.

Domenici, P. and Hale, M. E. (2019). Escape responses of fish: a review of the diversity in motor control, kinematics and behaviour. J. Exp. Biol. 222, 1–15.

Domenici, P., Standen, E. M. and Levine, R. P. (2004). Escape manoeuvres in the spiny dogfish (Squalus acanthias). J. Exp. Biol. 207, 2339–2349.

Donley, J. M. (2003). Steady swimming muscle dynamics in the leopard shark Triakis semifasciata. J. Exp. Biol. 206, 1117–1126.

Dorgan, K. M., Jumars, P. A., Johnson, B., Boudreau, B. P. and Landis, E. (2005). Burrow extension by crack propagation. Nature 433, 475.

Dorgan, K. M., Arwade, S. R. and Jumars, P. A. (2007). Burrowing in marine muds by crack propagation: Kinematics and forces. J. Exp. Biol. 210, 4198–4212.

Dorgan, K. M., Arwade, S. and Jumars, P. A. (2008). Worms as wedges: Effects of sediment mechanics on burrowing behavior. J. Mar. Res. 66, 219–254.

Dorgan, K. M., Lefebvre, S., Stillman, J. H. and Koehl, M. A. R. (2011). Energetics of

154

burrowing by the cirratulid polychaete Cirriformia moorei. J. Exp. Biol. 214, 2202– 2214.

Drucker, E. G. (2003). Function of pectoral fins in rainbow trout: behavioral repertoire and hydrodynamic forces. J. Exp. Biol. 206, 813–826.

Drucker, E. G. and Lauder, G. V. (1999). Locomotor forces on a swimming fish: three- dimensional vortex wake dynamics quantified using digital particle image velocimetry. J. Exp. Biol. 202, 2393–2412.

Drucker, E. G. and Lauder, G. V. (2002). Experimental hydrodynamics of fish locomotion: functional insights from wake visualization. Integr. Comp. Biol. 42, 243– 257.

Dudley, R. (2002). Mechanisms and implications of animal flight maneuverability. Integr. Comp. Biol. 42, 135–140.

Eaton, R. C., Bombardieri, R. A. and Meyer, D. L. (1977). The Mauthner-initiated startle response in teleost fish. J. Exp. Biol. 66, 65–81.

Eaton, R. C., Lee, R. K. K. and Foreman, M. B. (2001). The Mauthner cell and other identified neurons of the brainstem escape network of fish. Prog. Neurobiol. 63, 467–485.

Ekstrom, L. J. and Kajiura, S. M. (2014). Pelvic girdle shape predicts locomotion and phylogeny in batoids. J. Morphol. 275, 100–110.

Fish, F. E. (1996). Transitions in drag-based to lift-based propulsion in mammalian swimming. Am. Zool. 641, 628–641.

Fish, F. E. and Hoffman, J. L. (2015). Stability design and response to waves by batoids. Integr. Comp. Biol. 55, 648–661.

Fish, F. E., Schreiber, C. M., Moored, K. W., Liu, G., Dong, H. and Bart-Smith, H. (2016). Hydrodynamic performance of aquatic flapping: Efficiency of underwater flight in the manta. Aerospace 3, 1–24.

Fontanella, J. E., Fish, F. E., Barchi, E. I., Campbell-Malone, R., Nichols, R. H.,

155

DiNenno, N. K. and Beneski, J. T. (2013). Two- and three-dimensional geometries of batoids in relation to locomotor mode. J. Exp. Mar. Bio. Ecol. 446, 273–281.

Fox, C. H., Gibb, A. C., Summers, A. P. and Bemis, W. E. (2018). Benthic walking, bounding, and maneuvering in flatfishes (Pleuronectiformes: Pleuronectidae): New vertebrate gaits. Zoology 130, 19–29.

Franklin, O., Palmer, C. and Dyke, G. (2014). Pectoral fin morphology of batoid fishes (Chondrichthyes: Batoidea): Explaining phylogenetic variation with geometric morphometrics. J. Morphol. 275, 1173–1186.

Freitas, C., Olsen, E. M., Knutsen, H., Albretsen, J. and Moland, E. (2016). Temperature-associated habitat selection in a cold-water marine fish. J. Anim. Ecol. 85, 628–637.

Friedman, M. (2008). The evolutionary origin of flatfish asymmetry. Nature 454, 209– 212.

Garcia, D. (2011). A fast all-in-one method for automated post-processing of PIV data. Exp. Fluids 50, 1247–1259.

Garstang, W. (1897). Contributions to marine bionomics: II. The function of antero- lateral denticulations of the carapace in sand-burrowing crabs. J. Mar. Biol. Assoc. UK 4, 396–401.

Gerstner, C. L. (2007). Effect of oral suction and other friction-enhancing behaviors on the station-holding performance of suckermouth catfish (Hypostomus spp.). Can. J. Zool. 85, 133–140.

Gerstner, C. L. and Webb, P. W. (1998). The station-holding performance of the plaice Pleuronectes platessa on artificial substratum ripples. Can. J. Zool. 76, 260–268.

Gibson, R. N. (1969). Powers of adhesion in Liparis montagui (Donovan) and other shore fish. J. Exp. Mar. Bio. Ecol. 3, 179–190.

Gibson, R. N. and Robb, L. (1992). The relationship between body size, sediment

156

grain size and the burying ability of juvenile plaice, Pleuronectes platessa L. J. Fish Biol. 40, 771–778.

Gibson, R. N. and Robb, L. (2000). Sediment selection in juvenile plaice and its behavioural basis. J. Fish Biol. 56, 1258–1275.

Gidmark, N. J., Strother, J. A., Horton, J. M., Summers, A. P. and Brainerd, E. L. (2011). Locomotory transition from water to sand and its effects on undulatory kinematics in sand lances (Ammodytidae). J. Exp. Biol. 214, 657–664.

Gilliam, J. F., Fraser, D. F. and Fraser, D. F. (2016). Habitat selection under predation hazard: test of a model with foraging minnows stable. Ecology. 68, 1856–1862.

Goes de Araújo, M., Charvet-Almeida, P., Almeida, M. P. and Pereira, H. (2004). Freshwater stingrays (Potamotrygonidae): status, conservation and management challenges. Inf. Doc. AC 20, 1–6.

Goto, T., Nishida, K. and Nakaya, K. (1999). Internal morphology and function of paired fins in the epaulette shark, Hemiscyllium ocellatum. Ichthyol. Res. 46, 281– 287.

Gray, J. (1933). The relationship between the waves of muscular contraction and the propulsive mechanism of the eel. J. Exp. Biol. 10, 386–390.

Hale, M. E. (2002). S- and C-start escape responses of the muskellunge (Esox masquinongy) require alternative neuromotor mechanisms. J. Exp. Biol. 205, 2005– 16.

Harada, K. and Tamaki, A. (2003). Assessment ot the predation impact of the stingray Dasyatis akajei on the population of the ghost shrimp Nihonotrypaea harmandi on an intertidal sandflat (preliminary report). Proc. Symp. ‘Ecology large bioturbators tidal flats shallow sublittoral sediments—from Individ. Behav. to their role as Ecosyst. Eng. 81–85.

Harper, D. G. and Blake, R. W. (1991). Prey capture and the fast-start performance of northern pike Esox lucius. J. Exp. Biol. 155, 175–192.

157

He, J. and Zhang, Y. (2013). Experimental investigation to the kinematics of a blue spotted ray like underwater propulsor. Res. J. Appl. Sci. Eng. Technol. 6, 2799– 2806.

Henningsson, P., Hedenström, A. and Bomphrey, R. J. (2014). Efficiency of lift production in flapping and gliding flight of swifts. PLoS One 9,.

Herrel, A., Choi, H. F., Dumont, E., Schepper, N. De, Vanhooydonck, B., Aerts, P. and Adriaens, D. (2011a). Burrowing and subsurface locomotion in anguilliform fish: Behavioral specializations and mechanical constraints. J. Exp. Biol. 214, 1379–1385.

Herrel, A., Choi, H. F., Dumont, E., De Schepper, N., Vanhooydonck, B., Aerts, P. and Adriaens, D. (2011b). Burrowing and subsurface locomotion in anguilliform fish: behavioral specializations and mechanical constraints. J. Exp. Biol. 214, 1379–1385.

Herrel, A., Choi, H. F., De Schepper, N., Aerts, P. and Adriaens, D. (2011c). Kinematics of swimming in two burrowing anguilliform fishes. Zoology 114, 78–84.

Higham, T. E., Rogers, S. M., Langerhans, R. B., Jamniczky, H. A., Lauder, G. V., Stewart, W. J., Martin, C. H. and Reznick, D. N. (2016). Speciation through the lens of biomechanics: locomotion, prey capture, and reproductive isolation. Proc. R. Soc. B.

Higham, T. E., Clark, R. W., Collins, C. E., Whitford, M. D. and Freymiller, G. A. (2017). Rattlesnakes are extremely fast and variable when striking at kangaroo rats in nature: three-dimensional high-speed kinematics at night. Sci. Rep. 7, 40412.

Holst, R. J. and Bone, Q. (1993). On bipedalism in skates and rays. Philos. Trans. R. Soc. B 339, 105–108.

Hosoi, A. E. and Goldman, D. I. (2015). Beneath our feet: strategies for locomotion in granular media. Annu. Rev. Fluid Mech. 47, 431–453.

Hunter, D. R. and Elder, H. (1989). Burrowing dynamics and energy cost of transport in the soft‐bodied marine invertebrates Polyphysia crassa and Priapulus caudatus. J.

158

Zool. 218, 209–222.

Isabelle, M. C. (1995). Effects of predatory crab effluent on byssus production mussels. J. Exp. Mar. Bio. Ecol. 188, 233–241.

Jagnandan, K. and Sanford, C. P. (2013). Kinematics of ribbon-fin locomotion in the bowfin, Amia calva. J. Exp. Zool. Part A Ecol. Genet. Physiol. 319, 569–583.

Jayne, B. C. and Lauder, G. V. (1993). Red and white muscle activity and kinematics of the escape response of the bluegill sunfish during swimming. J. Comp. Physiol. A 173, 495–508.

Jung, S., Winter, A. G. and Hosoi, A. E. (2011). Dynamics of digging in wet soil. Int. J. Non. Linear. Mech. 46, 602–606.

Kasapi, M. A., Domenici, P., Blake, R. W. and Harper, D. (1993). The kinematics and performance of escape responses of the knifefish Xenomystus nigri. Can. J. Zool. 71, 189–195.

Kawabe, R., Naito, Y., Sato, K., Miyashita, K. and Yamashita, N. (2004). Direct measurement of the swimming speed, tailbeat, and body angle of Japanese flounder (Paralichthys olivaceus). ICES J. Mar. Sci. 61, 1080–1087.

Ken’Ichi Kanazawa (1992). Adaptation of test shape for burrowing and locomotion in spatangoid echinoids. Palaeontology 35, 733–750.

Koester, D. M. and Spirito, C. P. (2003). Punting: an unusual mode of locomotion in the little skate, Leucoraja erinacea (Chonrichthyes: Rajidae). Copeia.

Koller-Hodac, A., Germann, D. P., Gilgen, A., Dietrich, K., Hadorn, M., Schatz, W. and Eggenberger Hotz, P. (2010). Actuated bivalve robot study of the burrowing locomotion in sediment. Proc. - IEEE Int. Conf. Robot. Autom. 1209–1214.

Korsmeyer, K. E., Steffensen, J. F. and Herskin, J. (2002). Energetics of median and paired fin swimming, body and caudal fin swimming, and gait transition in parrotfish (Scarus schlegeli) and triggerfish (Rhinecanthus aculeatus). J. Exp. Biol. 205, 1253–1263.

159

Kuhlmann, M. L. (1992). Behavioral avoidance of predation in an intertidal hermit crab. J. Exp. Mar. Bio. Ecol. 157, 143–158.

Lannoo, M. J. and Lannoo, S. J. (1993). Why do electric fishes swim backwards ? An hypothesis based on gymnotiform foraging behavior interpreted through sensory constraints. Environ. Biol. Fishes 36, 157–165.

Lauder, G. V. (2015). Fish locomotion: recent advances and new directions. Ann. Rev. Mar. Sci. 7, 521–545.

Lauder, G. V., Drucker, E. G., Nauen, J. C. and Wilga, C. D. (2003). Experimental hydrodynamics and evolution: caudal fin locomotion in fishes. Vertebr. Biomech. Evol. 117–135.

Lecchini, D., Shima, J., Banaigs, B. and Galzin, R. (2005). Larval sensory abilities and mechanisms of habitat selection of a coral reef fish during settlement. Oecologia 143, 326–334.

Levin, P., Petrik, R. and Malone, J. (1997). Interactive effects of habitat selection, food supply and predation on recruitment of an estuarine fish. Oecologia 112, 55–63.

Liao, J. C. (2003). The Karman gait: novel body kinematics of rainbow trout swimming in a vortex street. J. Exp. Biol. 206, 1059–1073.

Liao, J., Beal, D., Lauder, G. V. and Triantafyllou, M. (2003). Fish exploiting vortices decrease muscle activity. Science 302, 1566–1569.

Lighthill, M. J. (1969). Hydromechanics of aquatic animal propulsion. Annu. Rev. Fluid Mech. 1, 413–446.

Lighthill, M. J. (1970). Aquatic animal propulsion of high hydromechanical efficiency. J. Fluid Mech. 44, 265–301.

Lindberg, W. J., Frazer, T. K., Portier, K. M., Vose, F., Murie, D. J., Mason, D. M., Nagy, B., Hart, M. K., Lindberg, W. J., Frazer, T. K., et al. (2006). Density- dependent habitat selection and performance by a large mobile reef fish stable. Ecol. Appl. 16, 731–746.

160

Liu, G., Ren, Y., Zhu, J., Bart-Smith, H. and Dong, H. (2015). Thrust producing mechanisms in ray-inspired underwater vehicle propulsion. Theor. Appl. Mech. Lett. 5, 54–57.

Liu, G., Ren, Y., Dong, H., Akanyeti, O., Liao, J. C. and Lauder, G. V. (2017). Computational analysis of vortex dynamics and performance enhancement due to body-fin and fin-fin interactions in fish-like locomotion. J. Fluid Mech. 829, 65–88.

Lucifora, L. O. and Vassallo, A. I. (2002). Walking in skates (Chondrichthyes, Rajidae): Anatomy, behaviour and analogies to tetrapod locomotion. Biol. J. Linn. Soc. 77, 35–41.

Luiz, O. J. and Edwards, A. J. (2011). Extinction of a shark population in the Archipelago of Saint Paul’s Rocks (equatorial Atlantic) inferred from the historical record. Biol. Conserv. 144, 2873–2881.

Macesic, L. J. and Kajiura, S. M. (2010). Comparative punting kinematics and pelvic fin musculature of benthic batoids. J. Morphol. 271, 1219–1228.

Macesic, L. J., Mulvaney, D. and Blevins, E. L. (2013). Synchronized swimming: Coordination of pelvic and pectoral fins during augmented punting by the freshwater stingray Potamotrygon orbignyi. Zoology 116, 144–150.

Maia, A. and Wilga, C. D. (2013). Function of dorsal fins in bamboo shark during steady swimming. Zoology 116, 224–231.

Maladen, R. D., Ding, Y., Umbanhowar, P. B., Kamor, A. and Goldman, D. I. (2011). Mechanical models of sandfish locomotion reveal principles of high performance subsurface sand-swimming. J. R. Soc. Interface 8, 1332–1345.

Martin, R. A., Hammerschlag, N., Collier, R. S. and Fallows, C. (2005). Predatory behaviour of white sharks (Carcharodon carcharias) at Seal Island, South Africa. J. Mar. Biol. Assoc. United Kingdom 85, 1121–1135.

Maxwell, D. L., Stelzenmüller, V., Eastwood, P. D. and Rogers, S. I. (2009). Modelling the spatial distribution of plaice (Pleuronectes platessa), sole (Solea solea) and thornback ray (Raja clavata) in UK waters for marine management and

161

planning. J. Sea Res. 61, 258–267.

McConnaughey, R. A. and Smith, K. R. (2000). Associations between flatfish abundance and surficial sediments in the eastern Bering Sea. Can. J. Fish. Aquat. Sci. 57, 2410–2419.

McGaw, I. J. (2005). Burying behaviour of two sympatric crab species: Cancer magister and Cancer productus. Sci. Mar. 69, 375–381.

McIvor, C. C. and Odum, W. E. (1988). Food, predation risk, and microhabitat selection in a marsh fish assemblage. Ecology 69, 1341–1351.

McKee, A., MacDonald, I., Farina, S. C. and Summers, A. P. (2016). Undulation frequency affects burial performance in living and model flatfishes. Zoology 119, 75–80.

Moles, A. and Norcross, B. L. (1995). Sediment preference in juvenile pacific flatfishes. Netherlands J. Sea Res. 34, 177–182.

Mosauer, W. (1932). Adaptive convergence in the sand reptiles of the Sahara and of California: A study in structure and behaviour. Copeia 72–78.

Myers, R. A., Baum, J. K., Shepherd, T. D., Powers, S. P. and Peterson, C. H. (2007). Cascading effects of the loss of apex predatory sharks from a coastal ocean. Science 1846–1850.

Nifong, J. C. and Lowers, R. H. (2017). Reciprocal intraguild predation between Alligator mississippiensis (American alligator) and elasmobranchii in the Southeastern United States. Southeast. Nat. 16, 383–396.

Nilsson, J., Kristiansen, T. S., Fosseidengen, J. E., Stien, L. H., Fernö, A. and van den Bos, R. (2010). Learning and anticipatory behaviour in a “sit-and-wait” predator: the Atlantic halibut. Behav. Processes 83, 257–266.

Norris, K. S. and Kavanau, J. L. (1966). The burrowing of the western shovel-nosed snake, Chionactis occipitalis hallowell, and the undersand environment. Copeia 1966, 650–664.

162

Nye, P. A. (1974). Burrowing and burying by the crab Macrophthalmus hirtipes. New Zeal. J. Mar. Freshw. Res. 8, 243–254.

Olla, B. L., Samet, C. E. and Studholme, A. L. (1972). Activity and feeding behavior of the summer flounder (Paralichthys dentatus) under controlled laboratrory conditions. Fish. Bull. 70, 1127–1136.

Pádraig, J. (2000). Stingray spines: A potencial cause of killer whale mortality in New Zealand. Aquat. Mamm. 26, 143–147.

Parson, J. M., Fish, F. E. and Nicastro, A. J. (2011). Turning performance of batoids: limitations of a rigid body. J. Exp. Mar. Bio. Ecol. 402, 12–18.

Rebach, S. (1974). Burying behavior in relation to substrate and temperature in the Hermit Crab, Pagurus longicarpus. Ecology 55, 195–198.

Robbins, W. D., Hisano, M., Connolly, S. R. and Choat, J. H. (2006). Ongoing collapse of coral-reef shark populations. Curr. Biol. 16, 2314–2319.

Rodrigues, M., Garcí, M. E., Troncoso, J. S. and Guerra, A. (2010). Burying behaviour in the bobtail squid Sepiola atlantica (Cephalopoda: Sepiolidae). Ital. J. Zool. 77, 247–251.

Rodway, M. S. and Regehr, H. M. (1999). Habitat selection and reproductive performance of food-stressed herring gulls. Condor. 101, 566–576.

Rosenberger, L. (2001). Pectoral fin locomotion in batoid fishes: undulation versus oscillation. J. Exp. Biol. 204, 379–394.

Rosenberger, L. and Westneat, M. W. (1999). Functional morphology of undulatory pectoral fin locomotion in the stingray Taeniura lymma (Chondrichthyes: dasyatidae). J. Exp. Biol. 202, 3523–39.

Rosenblum, H. G., Long, J. H. and Porter, M. E. (2011). Sink and swim: kinematic evidence for lifting-body mechanisms in negatively buoyant electric rays Narcine brasiliensis. J. Exp. Biol. 214, 2935–2948.

Rosenfeld, J. S., Leiter, T., Lindner, G. and Rothman, L. (2005). Food abundance

163

and fish density alters habitat selection, growth, and habitat suitability curves for juvenile coho salmon (Oncorhynchus kisutch). Can. J. Fish. Aquat. Sci. 62, 1691– 1701.

Rozhdestvensky, K. V. (2006). Wing-in-ground effect vehicles. Prog. Aerosp. Sci. 42, 211–283.

Ruiz-Torres, R., Curet, O. M., Lauder, G. V. and MacIver, M. A. (2014). Kinematics of the ribbon fin in hovering and swimming of the electric ghost knifefish. J. Exp. Biol. 217, 3765–3766.

Sandbak, T. and Murison, R. (1996). Voluntary alcohol consumption in rats: Relationships to defensive burying and stress gastric erosions. Physiol. Behav. 59, 983–989.

Schaefer, J. T. and Summers, A. P. (2005). Batoid wing skeletal structure: Novel morphologies, mechanical implications, and phylogenetic patterns. J. Morphol. 264, 298–313.

Schepper, N. De, Kegel, B. D. and Adriaens, D. (2007). Morphological specializations in Heterocongrinae (Anguilliformes: Congridae) related to burrowing and feeding. J. Morphol. 356, 1042–1154.

Schriefer, J. E. and Hale, M. E. (2004). Strikes and startles of northern pike (Esox lucius): a comparison of muscle activity and kinematics between S-start behaviors. J. Exp. Biol. 207, 535–544.

Seamone, S. G., Blaine, T. and Higham, T. E. (2014). Sharks modulate their escape behavior in response to predator size, speed and approach orientation. Zoology 117, 377–382.

Seamone, S. G., McCaffrey, T. M. and Syme, D. A. (2019). Disc starts: the pectoral disc of stingrays promotes omnidirectional fast starts across the substrate. Can. J. Zool. 605, 597–605.

Semeniuk, C. A. D. and Dill, L. M. (2005). Cost/benefit analysis of group and solitary resting in the cowtail stingray, Pastinachus sephen. Behav. Ecol. 16, 417–426.

164

Semeniuk, C. A. D. and Dill, L. M. (2006). Anti-predator benefits of mixed-species groups of cowtail stingrays (Pastinachus sephen) and whiprays (Himantura uarnak) at rest. Ethology 112, 33–43.

Sewell, R. B. S. (1925). The adhesive apparatus of the “Sucking-fish”. Nature 115, 48– 49.

Sfakiotakis, M., Lane, D. M. and Davies, J. B. C. (1999). Review of fish swimming modes for aquatic locomotion. IEEE J. Ocean. Eng. 24, 237–252.

Sharp, N. C. C. (1997). Timed running speed of a cheetah (Acinonyx jubatus). J. Zool. 241, 493–494.

Sharpe, S. S., Koehler, S. A., Kuckuk, R. M., Serrano, M., Vela, P. A., Mendelson, J. and Goldman, D. I. (2015a). Locomotor benefits of being a slender and slick sand- swimmer. J. Exp. Biol. 218, 440–450.

Sharpe, S. S., Kuckuk, R. and Goldman, D. I. (2015b). Controlled preparation of wet granular media reveals limits to lizard burial ability. Phys. Biol. 12, 1–16.

Simon, T. P. (1991). Startle response and causes of burying behavior in captive eastern sand darters, Ammocrypta pellucida (Putnam). Proc. Indiana Acad. Sci. 100, 155–160.

Standen, E. M. (2008). Pelvic fin locomotor function in fishes: three-dimensional kinematics in rainbow trout (Oncorhynchus mykiss). J. Exp. Biol. 211, 2931–2942.

Standen, E. M. and Lauder, G. V. (2007). Hydrodynamic function of dorsal and anal fins in brook trout (Salvelinus fontinalis). J. Exp. Biol. 210, 325–339.

Stein, R. and Magnuson, J. (1976). Behavioral response of crayfish to a fish predator. Ecology 57, 751–761.

Stickney, R. R., White, D. B. and Miller, D. (1973). Observations of fin use in relation to feeding and resting behavior in flatfishes (Pleuronectiformes). Copeia 1973, 154–156.

Stokes, M. D. and Holland, N. D. (1992). Southern stingray (Dasyatis americana)

165

feeding on lancelets (Branchiostoma floridae). J. Fish Biol. 41, 1043–1044.

Strong, W. R., Snelson, F. F. and Gruber, S. H. J. (1990). Hammerhead shark predation on stingrays: an observation of prey handling by Sphyrna mokarran. Copeia 3, 836–840.

Taft, N. K., Lauder, G. V. and Madden, P. G. A. (2008). Functional regionalization of the pectoral fin of the benthic longhorn sculpin during station holding and swimming. J. Zool. 276, 159–167.

Takagi, T., Kawabe, R., Yoshino, H. and Naito, Y. (2010). Functional morphology of the flounder allows stable and efficient gliding: An integrated analysis of swimming behaviour. Aquat. Biol. 9, 149–153.

Tatom-Naecker, T. M. and Westneat, M. W. (2018). Burrowing fishes: kinematics, morphology and phylogeny of sand-diving wrasses (Labridae). J. Fish Biol. 860– 873.

Thielicke, W. (2014). The flapping flight of birds - analysis and application.

Thielicke, W. and Stamhuis, E. J. (2014). PIVlab – towards user-friendly, affordable and accurate digital particle image velocimetry in MATLAB. J. Open Res. Softw. 2,.

Tierney, K., Watkinson, D. A., Ruppert, J. L. W., Poesch, M. and Veillard, M. F. (2017). Comparative swimming and station-holding ability of the threatened Rocky Mountain Sculpin (Cottus sp.) from four hydrologically distinct rivers. Conserv. Physiol. 5, 1–12.

Trevor, J. H. (1978). The dynamics and mechanical energy expenditure of the polychaetes Nephtys cirrosa, Nereis diversicolor and Arenicola marina during burrowing. Estuar. Coast. Mar. Sci. 6, 605–619.

Tucker, V. a (1988). Gliding birds: descending flight of the white-backed vulture, Gyps africanus. J. Exp. Biol. 344, 325–344.

Tucker, V. (1998). Gliding flight: speed and acceleration of ideal falcons during diving and pull out. J. Exp. Biol. 201, 403–414.

166

Tucker, V. A. and Parrott, G. C. (1970). Aerodynamics of gliding flight in a falcon and other birds. J. Exp. Biol. 52, 345–367.

Tytell, E. D. (2004). The hydrodynamics of eel swimming II. Effect of swimming speed. J. Exp. Biol. 207, 3265–3279.

Tytell, E. D. (2006). Median fin function in bluegill sunfish Lepomis macrochirus: streamwise vortex structure during steady swimming. J. Exp. Biol. 209, 1516–1534.

Tytell, E. D. and Lauder, G. V. (2008). Hydrodynamics of the escape response in bluegill sunfish, Lepomis macrochirus. J. Exp. Biol. 211, 3359–3369.

Tytell, E. D., Standen, E. M. and Lauder, G. V. (2008). Escaping Flatland: three- dimensional kinematics and hydrodynamics of median fins in fishes. J. Exp. Biol. 211, 187–195.

Umbanhowar, P. B., Komsuoglu, H., Koditschek, D. E., Goldman, D. I., Umbanhowar, P. B., Goldman, D. I., Matin, A. K. M. A., Gibb, A. C., Gibb, A. C., Goldman, D. I., et al. (2016). Phototactic guidance of a tissue-egineered soft- robotic ray. Science 353, 158-162.

Van Der Veer, H. W., Aliaume, C., Miller, J. M., Adriaans, E. J., Witte, J. I. J. and Zerbi, A. (1994). Ecological observations on juvenile flatfish in a tropical coastal system, Puerto Rico. Netherlands J. Sea Res. 32, 453–460.

Vaudo, J. J. and Lowe, C. G. (2006). Movement patterns of the round stingray Urobatis halleri (Cooper) near a thermal outfall. J. Fish Biol. 68, 1756–1766.

Vernes, K. (2001). Gliding performance of the northern flying squirrel (Glaucomys Sabrinus) in mature mixed forest of Eastern Canada. J. Mammal. 82, 1026–1033.

Visser, I. N. (1999). Benthic foarging on stingrays by killer whals (Orcinus orca) in New Zealand waters. Mar. Mammal Sci. 15, 220–227.

Wainwright, D. K., Kleinteich, T., Kleinteich, A., Gorb, S. N. and Summers, A. P. (2013). Stick tight: Suction adhesion on irregular surfaces in the northern clingfish. Biol. Lett. 9, 1–5.

167

Wakeling, J. M. (2001). Biomechanics of fast-start swimming in fish. Comp. Biochem. Physiol. - A Mol. Integr. Physiol. 131, 31–40.

Walker, J. A. and Westneat, M. W. (2002). Performance limits of labriform propulsion and correlates with fin shape and motion. J. Exp. Biol. 205, 177–187.

Walker, J. A., Breder, C. M., Domenici, P., Blake, R. W., Gerstner, C. L., Howland, H. C., Norberg, U., Rayner, J. M. V., Randall, J. E., Schrank, A. J., et al. (2000). Does a rigid body limit maneuverability? J. Exp. Biol. 203, 3391–6.

Ward, A. B. and Azizi, E. (2004). Convergent evolution of the head retraction escape response in elongate fishes and amphibians. Zoology 107, 205–217.

Warner, G. F. (1977). The biology of crabs. 202. London: Elek Science.

Webb, P. W. (1974). Hydrodynamics and energetics of fish propulsion. Bull. Fish. Res. Board Can. 190, 1–159.

Webb, P. W. (1976). The effect of size on the fast-start performance of rainbow trout Salmo cairdneri, and a consideration of piscivorous predator-prey interactions. J. Exp. Biol. 65, 157–177.

Webb, P. W. (1977). Effects of median-fin amputation on fast-start performance of rainbow trout (Salmo Gairdneri). J. Exp. Biol. 68, 123–135.

Webb, P. W. (1981). The effect of the bottom on the fast start of flatfish Citharichthys stigmaeus. Fish. Bull. 79, 271–276.

Webb, P. W. (1984a). Body form, locomotion and foraging in aquatic vertebrates. Am. Zool. 24, 107–120.

Webb, P. W. (1984b). Form and function in fish swimming. Sci. Am. 251, 72–82.

Webb, P. W. (1988). Simple physical principles and vertebrate aquatic locomotion. Am. Zool. 28, 709–725.

Webb, P. W. (1989). Station-holding by three species of benthic fishes. J. Exp. Biol. 320, 303–320.

168

Webb, P. W. (2002). Kinematics of plaice, Pleuronectes platessa, and cod, Gadus morhua, swimming near the bottom. J. Exp. Biol. 205, 2125–2134.

Webb, P. W. (2004). Maneuverability - General issues. IEEE J. Ocean. Eng. 29, 547– 555.

Webb, P. W. and Fairchild, A. G. (2001). Performance and maneuverability of three species of teleostean fishes. Can. J. Zool. 79, 1866–1877.

Webb, P. W. and Keyes, R. S. (1982). Swimming kinematics of sharks. Fish. Bull. 80, 803–812.

Webb, P. W. and Skadsen, J. (1980). Strike tactics of Esox. Can. J. Zool. 58, 1462– 1469.

Webb, P. W., Gerstner, C. L. and Minton, S. T. (1996). Station-holding by the mottled sculpin, Cottus bairdi (Teleosteir: Cotteidae), and other fishes. Copeia 2, 488–493.

Weihs, D. (1973). Mechanically efficient swimming technique for fish with negative buoyancy. J. Mar. Res. 194–209.

Westneat, M. W., Hale, M. E., Mchenry, M. and Long, J. (1998). Mechanics of the fast-start: muscle function and the role of intramuscular pressure in the escape behavior of Amia calva and Polypterus palmas. J. Exp. Biol. 201, 3041–55.

Wiens, J. J. and Slingluff, J. L. (2001). How lizards turn into snakes: A phylogenetic analysis of body-form evolution in anguid lizards. Evolution 55, 2303–2318.

Wiens, J. J., Brandley, M. C. and Reeder, T. W. (2006). Why does a trait evolve multiple times within a clade? Repeated evolution of snakeline body form in squamate reptiles. Evolution 60, 123–141.

Wilga, C. D. and Lauder, G. V. (2001). Functional morphology of the pectoral fins in bamboo sharks, Chiloscyllium plagiosum: benthic vs. pelagic station-holding. J. Morphol. 249, 195–209.

Wilga, C. D. and Lauder, G. V. (2002). Function of the heterocercal tail in sharks: quantitative wake dynamics during steady horizontal swimming and vertical

169

maneuvering. J. Exp. Biol. 205, 2365–2374.

Wilga, C. D., Maia, A., Nauwelaerts, S. and Lauder, G. V. (2012). Prey handling using whole-body fluid dynamics in batoids. Zoology 115, 47–57.

Winter, A. G., Deits, R. L. H. and Hosoi, A. E. (2012). Localized fluidization burrowing mechanics of Ensis directus. J. Exp. Biol. 215, 2072–2080.

Ydenberg, R. C. and Dill, L. M. (1986). The economics of fleeing from predators. Adv. Study Behav. 16, 229–249.

170